88. Effectiveness of Antimicrobial Adjuncts to Scaling and Root Planing Therapy for Periodontitis

Objectives

Periodontal diseases are bacterial infections that occur at or below the gum line. In contrast to gingivitis, which affects only the gums, periodontitis (severe periodontal disease) may involve the soft tissue and bone supporting the teeth. An estimated 70 percent of the adult U.S. population is affected by these infections. This includes 20–30 percent who have periodontitis that threatens the loss of teeth. Approximately $5 billion is spent on treatment of periodontal diseases each year. This report deals with the treatment of chronic periodontitis in adults.

The key question is whether scaling and root planing (SRP) accompanied by an antimicrobial agent, as a supplemental or adjunct treatment, results in improved outcomes that persist over time in adults with chronic periodontitis when compared with SRP alone (or SRP and placebo) . The primary outcomes of interest in this report are reductions in probing depth (PD) and gains in clinical attachment level (CAL). Of secondary interest are reductions in selected disease-causing bacteria, particularly reduction in the percentage of spirochetes present in dental plaque or in fluid from the gingival crevice.

Methodology

Findings

The authors conducted separate analyses of the following agents as adjuncts to SRP: tetracycline, minocycline, metronidazole, the combination of metronidazole and amoxicillin, and chlorhexidine. For tetracycline, minocycline, and metronidazole, they did separate analyses for systemically and locally delivered adjunct treatments. Local treatment delivery methods included irrigants, gels, ointments, microcapsules, and impregnated strips, chips, and fibers.

The authors also analyzed agents that appeared in the literature as part of only one or two identified eligible studies. These were grouped together, either as other antibiotics (doxycycline, azithromycin, spiramycin, and ofloxacin), or as other antimicrobials (fluorides, hydrogen peroxide, povidone iodine, triclosan, and tetrapotassium peroxydiphosphate).

Tetracycline. For systemic tetracycline (five studies), there was a greater reduction in PD with adjunct treatment than using SRP alone, but no individual difference reached statistical significance. The meta-analysis produced an estimated overall difference of 0.15 mm in PD reductions, favoring the use of SRP with systemic tetracycline over SRP alone, but this difference also did not reach statistical significance. One of the four studies that measured CAL gain produced a statistically significant reduction of 0.31 mm, favoring the use of the adjunct with SRP over SRP alone.

The weight of the available evidence supports the effectiveness of locally applied tetracycline as an adjunctive therapy. Of the 16 studies of locally applied tetracycline preparations, four demonstrated statistically significant PD reductions ranging from 0.41 mm to 0.93 mm, favoring the experimental group. The overall estimated PD reduction—0.47 mm—was statistically significant, favoring the adjunct treatment. Only two studies in this group showed a statistically significant gain in CAL, 0.15 mm and 0.48 mm, respectively; the overall effect size from the meta-analysis was a statistically significant 0.24 mm CAL gain.

Minocycline. Neither of the two studies of systemic minocycline used as an adjunct to SRP provided any statistically significant evidence for its use in reducing PD or increasing gains in CAL.

The eight studies of locally applied minocycline are more supportive of its use as an adjunct to SRP. Four studies reported statistically significant reductions in PD. These ranged from 0.30 mm to 1.10 mm, with this latter amount reported for persons whose initial probing depth was 7 mm or greater. The mean effect size from the meta-analysis was a statistically significant 0.49 mm reduction in PD, favoring use of local minocycline. A very similar result was reported for CAL gain, with three studies showing statistically significant gains in CAL of 0.39 mm to 0.80 mm. The mean effect size from the meta-analysis was a statistically significant 0.46 mm gain in CAL and favored the use of the adjunct.

Metronidazole. Only two of the seven studies of systemic metronidazole used as an adjunct to SRP reported statistically significant reductions in PD over SRP alone. They ranged from 0.47 mm to 1.64 mm and represented subpopulations with initial probing depths of 4 mm to 6 mm and more than 6 mm, respectively. Two studies also reported statistically significant gains in CAL with the adjunctive use of systemic metronidazole, ranging from 0.47 mm to 1.19 mm, again in persons with relatively deep initial PD.

Four of the 11 studies of SRP plus locally delivered metronidazole yielded statistically significant reductions in PD ranging from 0.18 mm to 0.80 mm. The overall effect size estimated from the meta-analysis was 0.32 mm favoring local metronidazole as an adjunct to SRP; this effect was found to be statistically significant. Two studies reported statistically significant CAL gains of 0.40 mm and 0.66 mm, again favoring the adjunctive use of local metronidazole. The mean effect size estimated from the meta-analysis was only 0.12 mm, favoring adjunctive local metronidazole, but it is statistically significant.

Metronidazole and Amoxicillin Combination. Only one of the four studies of this systemically administered drug combination plus SRP reported a statistically significant greater PD reduction than SRP alone (0.7 mm). One of the four studies of CAL gain reported a statistically significant improvement over SRP alone, but the exact amount of the difference was not reported.

Chlorhexidine. Of the 17 studies of locally administered chlorhexidine included in the review, most had small numbers of subjects but larger numbers of sites or pockets as the unit of analysis. Despite this, only two of these trials (both using chlorhexidine chips) produced statistically significant PD reductions. The reductions favoring the use of chlorhexidine as an adjunct to SRP ranged from 0.26 mm to 0.33 mm. Conversely, one large trial showed a statistically significant net PD reduction of 0.46 mm favoring the control group. The statistically significant overall effect size from the meta-analysis was 0.24 mm, reflecting the moderating effect of the contrary results from the larger trial reporting greater PD reduction with SRP alone.

Gains in CAL with the use of chlorhexidine as an adjunct were generally lower than were the reductions in PD. Three studies had statistically significant results ranging from 0.16 mm to 0.28 mm, favoring chlorhexidine use. The statistically significant mean effect size estimated from the meta-analysis was 0.16 mm.

Other Antibiotics. The seven trials in the group of other systemic antibiotics (doxycycline, spiramycin, the combination of spiramycin and metronidazole, azithromycin, amoxicillin and clavulanic acid, and amoxicillin plus chlorhexidine) were quite varied in size, duration, and other variables. The authors were not able to combine these trials into a meta-analysis. Three of the studies reported statistically significant results for PD reduction, ranging from 0.47 mm (for spiramycin) to 0.87 mm (for azithromycin, among patients with initial PD levels of 6 mm or greater). Two studies reported statistically significant results for CAL gains; only one gave specific data, a gain of 1.3 mm with doxycycline. Given the diversity of these therapeutic agents, means of therapy, and overall study designs, the authors believe that caution is warranted in interpreting these studies as convincing evidence of effectiveness, especially in the light of the generally negative results for other, more commonly studied systemic antibiotics.

Only two trials dealt with other local antibiotics (doxycycline gel and ofloxacin inserts), and only the one with doxycycline provided data showing a 0.44 mm PD reduction and a 0.37 mm CAL gain, both statistically significant. These results are promising, as they come from a relatively large trial, but the strength of the evidence should be interpreted conservatively when compared to that represented by the multiple studies of the more commonly used local adjunct therapies.

Other Antimicrobials. It is not possible to say much about the group of five studies (one with two experimental arms) grouped together as other antimicrobials (amine fluoride gel, stannous fluoride gel, triclosan gel and dentifrice, hydrogen peroxide, povidone­–iodine, and tetrapotassium perioxydiphosophate), all of which are locally delivered. As regards PD reduction, one of the six trials reported a statistically significant 0.8 mm net reduction at 52 weeks, favoring hydrogen peroxide used as an adjunct to SRP; however, for CAL gains, no study had statistically significant improvements favoring the treatment group. In light of the level of improvements from adjunct use of some locally administered antibiotics, the PD findings for hydrogen peroxide may seem promising, but they are from only a single, small study.

Conclusions

Although the findings differ for each antimicrobial and mode of delivery, the authors make some important overall observations relating to the key question. First, relative to the PD reductions achieved from the baseline measurement to the study end-point measurement, the difference in measurements between the treatment and control groups typically favored the treatment group, but was relatively modest. With respect to CAL gains, the picture was similar, but the effects are smaller and statistical significance was less common.

Of the antimicrobials investigated, studies of locally applied tetracycline and minocycline—and locally delivered chlorhexidine—have fairly consistent results in moderately large studies that often reach statistical significance; improvements observed in these studies typically average in the neighborhood of 0.3 mm to 0.6 mm. The other agents and delivery modes produced less consistent outcomes and fewer outcomes that reached statistical significance; the majority of studies showed small, statistically nonsignificant PD improvements. CAL outcomes were not as positive as those for PD. The question remains, the authors note, whether such improvements are clinically meaningful.

Periodontal Disease and Periodontitis

Periodontal diseases are bacterial infections that occur at or below the gum line; they include both gingivitis and periodontitis. The former affects only the gingival tissue, while the latter not only affects the gingivae but also the bone supporting the teeth. This systematic review concerns chronic periodontitis, which itself is typically described as mild, moderate, or advanced on the basis of gingival inflammation, pocket formation, loss of gingival attachment, bone resorption, and number of teeth involved.^{1, 2} According to the Surgeon General’s report on oral health, most adults are affected by these infections; a decade ago, nearly 36 million persons ages 35 through 79 had some form of periodontitis.^{1, 3} Of persons ages 45 to 54 (one of the two most affected age groups), 14 percent have severe periodontitis.^{1, 4} Prevalence rates and severity of periodontitis are higher among males than females and among blacks and Mexican Americans than whites.¹

Expenditures on dental services were estimated in 1998 to be almost $54 billion; they were expected to exceed $60 billion in 2000.⁴ Of this dental bill, perhaps nearly $5 billion is now spent on periodontal services (in 1999, an estimated $4.4 billion was spent on periodontal procedures alone).¹ As documented in the Surgeon General’s report, periodontal diseases can be associated with a variety of other serious health conditions (e.g., diabetes, cardiovascular disease, stroke); the diseases themselves and the need to seek dental or periodontal care can have impacts on numerous quality-of-life indicators (e.g., social interaction, limitations in usual daily activities, psychological status and sleep, and diet and nutrition) that reflect patient-oriented concerns.

For the past 100 years, many investigations have attempted to define the etiologic agents of these diseases.⁵ The microbiology of periodontal infections is quite complicated, and numerous bacterial agents have been implicated in their etiology. Perhaps as much as 50 percent of the subgingival flora of chronic periodontitis has not yet been characterized (Gary Armitage, DDS, Personal Communication, May 7, 2003). Nonetheless, small groups of specific bacterial species are now considered to be important in the initiation or progression (or both) of periodontitis;⁶ often mentioned are Bacteroides forsythus, Porphyromonas gingivalis, Treponema denticola, and Actinobacillus actinomycetemcomitans.^{7- 9}

Scaling and root planning (SRP) is generally the first treatment employed for periodontitis. It is considered a nonsurgical procedure for which local anesthesia is often given to numb the infected gingiva (gums) around the teeth to be subgingivally scaled and planed. Scaling may be performed with hand instruments alone or with the aid of an ultrasonic scaler. It is done to clean teeth thoroughly below the gum line, removing bacterial plaque, calculus (tartar), debris, necrotic tissue, and pus from pockets that form around infected teeth. Root planing involves cleaning and smoothing the root surface of an infected tooth after scaling so that the gingival tissue can heal close to the root, shrinking the tissue and reducing the depth of the pocket that had formed. SRP is intended to reduce the bacterial load, shrink swollen and inflamed gingiva, and recondition the subgingival ecology, making it biologically compatible with optimal healing and reattachment of epithelium to the root surface.

Two commonly used clinical measures of periodontal disease progression and restoration of oral health are probing depth (PD, sometimes referred to as probing pocket depth) and clinical attachment level (CAL). These measures are made with specially marked periodontal probes held parallel to the tooth and inserted under the free gingival margin and gently “walked” to the base of the sulcus (i.e., pocket). The probes are typically marked with rings or bands that measure distance in millimeters. The PD is generally measured as the distance from the base of the sulcus to the top of the free gingival margin. The CAL is often measured as the difference between the PD and the distance from the free gingival margin and a natural fixed anatomical marker on the tooth called the cemento-enamel junction (CEJ). As the name implies, the CEJ is where the cementum and enamel on the tooth are joined. Specially fabricated stents are also used as an alternative to a fixed anatomical landmark to measure changes in CAL.

Figure 1. Diagrammatic Representation of Mechanical (more...)

   Figure 1. Diagrammatic Representation of Mechanical and/or Antimicrobial Therapy on Both the Microbial Load and Disease Risk

Figure 1

Figure 1

Figure 1

Specifically, we examine the evidence, for adults with chronic periodontitis, as to whether SRP accompanied by an adjunctive antimicrobial agent, compared to SRP alone, improves outcomes that persist over time. The primary outcomes of interest in this report are PD reductions, CAL gains, and secondarily, reductions in selected pathogens. The clinical rationale for this question relates to developments in controlling periodontal infections.

Controlling Periodontal Infections

The juxtaposition of three developments makes concerted efforts to control periodontal infections more realistic today than in the past. The first major advance was the development of molecular diagnostic tools that can rapidly and inexpensively examine large numbers of plaque samples and identify specific microbial species. A second advance was the recognition that specific microbial complexes occur together in plaque. The third was the development of new tools to reduce the supra- and subgingival bacteria, such as chlorhexidine mouthwash, triclosan dentifrice, electronic toothbrushes, and systemic and local drug delivery systems. From data derived from these three scientific avenues, three strategies have been proposed for reducing the risk of periodontal diseases. Each attempts to intercept the disease process at critical points in its development.

1. Reduce supragingival plaque. Supragingival plaque reduction by home care and professional cleaning is the most universally practiced periodontal treatment available; it is considered essential in the treatment of periodontal diseases.

2. Control pathogen transmission. Introduction of an antibacterial mouthwash and toothpaste may insulate sites from infected pathogen reservoirs elsewhere in the mouth. Hujoel et al. tested rinsing once per week and observed a 45 percent reduction in tooth loss after 1 year.¹⁰ Quirynen et al.¹¹ and De Soete et al.,¹² examining one-stage, full-mouth disinfection, observed a parallel significant reduction in periodontal pathogens and improvement in clinical health following chlorhexidine rinses.

3. Disinfect pathogen reservoirs. Many investigators have recognized disease reservoirs as seeding sources for intraoral spread of disease and as an important consideration in determining therapeutic outcome.¹³ Of the infection sources in the oral cavity, untreated sites elsewhere in the mouth represent the most obvious potential source of re-infection. At least three mechanisms are used to address this threat: SRP; local drug delivery; and systemic antibiotics.

SRP. SRP has been used effectively in periodontal therapy for more than 1,000 years.^{14, 15} The concept that eliminating periodontal pockets that support pathogen growth decreases the risk of periodontal disease is generally accepted. Some clinical studies indicate that most patients with periodontal disease can be maintained by regular SRP alone.^{16- 19} By itself, SRP produces a very modest transient reduction in bacteria; they can return to pretreatment levels within 2 weeks.^{20, 21} When total bacterial load changes, a subset of bacteria associated with periodontitis is also depressed, including P. gingivalis, B. forsythus, and T. denticola. ²² The microbial impact, however, appears to be short lived with sites re-infecting after 3 months. This suggests that a more effective initial therapy might reduce the labor of continued maintenance and also further reduce the disease risk.

Local drug delivery. Several local antibacterial agents have been tested for intra-pocket delivery. These include doxycycline gel, metronidazole gel, chlorhexidine chips, minocycline microspheres, and tetracycline fibers. All these agents, either alone or in combination with SRP, appear to reduce pocket depth²³ and may also alter oral bacteria.

Systemic antibiotics. A host of systemic antibacterial agents has been tested: amoxicillin, metronidazole, metronidazole plus amoxicillin, azithromycin, clindamycin, and ciprofloxacin. The studies, some considered in recent systematic reviews examining these antibacterial agents, either alone or in conjunction with mechanical therapy, suggest that antibacterial therapy alone may, in some cases, be as effective as SRP therapy.^{15, 24}

Origins of this Evidence Report

This is the fourth in a series of systematic reviews of dental topics prepared for the Agency on Healthcare Research and Quality (AHRQ) with the support and collaboration of the National Institute of Dental and Craniofacial Research (NIDCR) (for Acknowledgments, see Appendix A). The first was a review of dental caries diagnosis and management that cut across the entire population and stages of life.²⁵ The second report, rather than focusing on a specific dental disease condition or a particular treatment approach for the general population, dealt with several aspects of the treatment of a special population subgroup – persons infected with human immunodeficiency virus (HIV) and those living with acquired immune deficiency syndrome (AIDS).²⁶ The third report reviewed the cardiovascular effects of the use of epinephrine – an ingredient of dental anesthesia and retraction cord inserted around teeth to reduce bleeding – in hypertensive dental patients.²⁷

The specific clinical question to be addressed in this evidence report emerged from a working group meeting that NIDCR convened on the National Institutes of Health campus on April 2, 2001. This invited group of eight experts from the field of periodontics represented the range of experience, activity, and perspective within the discipline, including academia, public health, clinical practice, and research. In selecting the area of periodontal disease, NIDCR sought to include in its array of evidence reports the next largest domain of the profession’s clinical activities after dental caries, for which we had already prepared a report. In bringing together the working group, NIDCR was opening the topic selection process to the needs and sensitivities of the field.

A set of six questions emerged from the working group meeting. Two involved diagnostic issues; one related to distinguishing aggressive from chronic periodontitis, and the other to assessing the validity of methods to predict periodontal destruction. Four questions addressed disease management or treatment issues. These included assessments of the effectiveness of SRP compared to other treatments, the nature of professional maintenance needed after periodontal therapy, how risk factors modify the outcomes of periodontal therapy, and whether predictable therapies exist for regenerating supporting tissue lost to periodontal disease. For a variety of reasons, the group rated the question of the effectiveness of SRP therapy for chronic periodontitis as the top issue for the NIDCR to consider in the evidence report.

The project team subsequently refined and clarified the question through discussions with the NIDCR staff and later through communication with the Technical Expert Advisory Group (TEAG) assembled for this particular topic area. The original question on effectiveness of SRP was too broad to be covered in a single evidence report, as it would have involved too much literature and required more time and resources than were available. The consensus decision, therefore, was to focus on the primary comparison of interest: In adults with chronic periodontitis, does SRP therapy in conjunction with the use of chemical antimicrobial agents, when compared to SRP alone, improve clinical outcomes that persist over time?

Among many possible indicators of improved clinical outcomes, we chose to use reported measures of PD reduction, gain in CAL, and pathogen reduction, with no accompanying increase in adverse events. To answer the key question, the project team systematically identified, critically appraised, and synthesized the evidence emanating from published primary human clinical trials research that produced data to allow examination of this question.

The question reflects two fairly common concerns in dentistry. First, practicing dentists may not be aware of the available research with respect to the effectiveness over time of therapies adjunctive to SRP for persons with chronic periodontitis. Second, the research may not be as comprehensive or definitive as it should be. Thus, all judged this particular focus to be of clinical, research, and practical significance.

Technical Expert Advisory Group

AHRQ guidelines require identification of a technical expert advisory group (TEAG) for evidence reports, in this case in the specialized area of managing periodontal diseases. Our TEAGs advance AHRQ’s broader goals of (a) creating and maintaining science partnerships and public-private partnerships and (b) meeting the needs of an array of potential consumers and users of its products. Thus, a TEAG is both an additional resource and a sounding board throughout the project.

Table 1. Technical Expert Advisory Group for the Evidence (more...)

Table 1. Technical Expert Advisory Group for the Evidence Report on Effectiveness of Adjuncts to Scaling and Root Planing Therapy for Periodontitis

Name	Affiliation
Gary Armitage, DDS (ADA Representative)	Professor and Chair Division of Periodontology UCSF School of Dentistry San Francisco, California
Jack Caton, DDS (AAP Representative)	Professor and Chief, Division of Periodontics, University of Rochester Eastman Dental Center Rochester, New York
Sara G. Grossi, DDS, MS	Senior Research Scientist Department of Oral Biology SUNY at Buffalo Buffalo, New York
Daniel Fine, DMD	Professor of Oral Biology Department of Oral Biology University of Medicine and Dentistry of New Jersey Newark, New Jersey
Marjorie Jeffcoat, DMD	Professor and Chair Department of Periodontics School of Dentistry University of Alabama - Birmingham Birmingham, Alabama
Anthony Neely, DDS, PhD	Lecturer on Epidemiology of Periodontal Disease Department of Periodontology University of Detroit-Mercy School of Dentistry Detroit, Michigan

To ensure scientifically robust work, we asked the TEAG to provide reactions to work in progress and to advise on substantive issues or possibly overlooked areas of research. TEAG members participated in e-mail communications

to discuss the key clinical questions, initial drafts of causal pathways, and proposed inclusion and exclusion criteria for research articles;

to provide comments concerning the article abstraction forms, the content proposed for inclusion in the evidence tables, and the final versions of the key clinical question; and

to discuss the proposed content of the evidence tables and the completeness of the search.

Because of their extensive knowledge of the literature and ongoing research in this specialized area of dentistry dealing with treatment of chronic periodontitis, as well as their active involvement in the associated professional societies, we also asked TEAG members to participate in the peer review process by commenting on the draft evidence report, and four did so.

Organization of this Report

The remainder of this evidence report is organized in the following manner. Chapter 2 provides details about our literature search and review methodology. Specifically included are the analytical framework for our key clinical question and our approach to conducting the systematic review, applying the inclusion/exclusion criteria, abstracting data from articles, maintaining quality control, and similar details. Chapter 3 presents the results of our analyses. Chapter 4 gives our concluding discussion, and Chapter 5 notes weaknesses and gaps we found in the research and offers recommendations for a research agenda related to the question addressed on the added effectiveness of therapies adjunctive to SRP in treating chronic periodontitis. Chapter 6 provides the references cited in the body of the evidence report.

Chapter 7 contains the evidence tables and supporting information. Finally, the complete bibliography of literature considered and used in developing the evidence report (including all articles reviewed in the literature search and all references cited in Chapters 1 to 5) appears in Chapter 8. The three appendices provide acknowledgments (Appendix A), our data abstraction form (Appendix B), and the quality review checklist (Appendix C).

Overview

This chapter documents the procedures that the RTI-University of North Carolina at Chapel Hill Evidence-based Practice Center (RTI-UNC EPC) used to develop a comprehensive evidence report on the effectiveness over time of adjunctive therapies used in addition to scaling and root planing (SRP) in treating adults with chronic periodontitis. To set the framework for the review, we discuss first the key question that we address and the related underlying causal pathway for this topic. A detailed description of the literature search process follows; it includes descriptions of the Medical Subject Headings (MeSH terms) used in the principal search, other search sources, the inclusion and exclusion criteria, and the application of these criteria to the results of the searches. We note steps for reviewing studies that met the inclusion and exclusion criteria, abstracting data onto data abstraction forms, creating evidence tables, writing a draft report for external peer review, and revising the draft following peer review. We also discuss quality issues, in particular, the RTI-UNC EPC’s quality control procedures for determining eligibility for inclusion, carrying out data abstraction, checking evidence and text tables against articles, and grading the quality of individual studies.

Key Question and Casual Pathway

Figure 2. Added Effectiveness of Therapies Adjunctive (more...)

   Figure 2. Added Effectiveness of Therapies Adjunctive to Scaling and Root Planing for Treatment of Chronic Periodontitis: Causal Pathway

Figure 2

We constrained the review to chronic periodontitis among adults and excluded studies pertaining solely to more aggressive forms of periodontal disease. Thus, we did not examine studies relating to treatment of periodontitis that is described as refractory, localized, juvenile, aggressive, related to human immunodeficiency virus or acquired immune deficiency syndrome (HIV/AIDS), or related to diabetes; neither did we consider studies that included a mix of periodontal diseases but that did not present separate analyses of the chronic form of the disease.

Ultimately, we focused on the following adjunctive therapies, which in some cases include both systemic and local delivery modalities:

• Tetracycline

• Minocycline

• Metronidazole

• Metronidazole in combination with amoxicillin

• Chlorhexidine

• Other antibiotics (e.g., spiramycin, doxycycline, azithromycin, Augmentin), and

• Other antimicrobials (e.g., povidone iodine, hydrogen peroxide, fluoride).

The key question defines the population of interest as adults receiving SRP for chronic periodontitis. The principal audience for this review comprises dental practitioners and researchers, especially in periodontology; however, we believe the issues may be of interest to physicians, and so we did not, in the searches, exclude studies done in medical care settings.

Of all the potential outcome measures in this literature, we had to constrain those for this systematic review to a very small number. As discussed in Chapter 1, after discussion with clinical experts asssisting the team and the Technical Expert Advisory Group, we focused our analyses on two clinical measures: probing depth (PD, measured in millimeters of reduction) and clinical attachment level (CAL, measured in millimeters of gain). Even with this narrow focus, the nature of these measures can vary significantly across studies (i.e., across dental examiners) on the basis of several factors: the level of gingival inflammation at the time of measurement, diameter of the probe tip, probing force, and angulation of the probe,²⁸ often these factors reflect the nature of the original or graduate training of the examiners, but in any case they raise the level of incomparability across studies to some unknown degree. We also targeted one microbiological measure ─ presence or percentage reduction in spirochetes ─ for the qualitative analyses, but we did not try to use this for the meta-analyses, as it was even less often reported than either PD or CAL.

We reported the longest period of follow-up for entering information into evidence tables; when data were provided only by subgroups (e.g., subgroups defined by different levels of PD at baseline), we attempted to retain all that information in evidence tables or, if the volume of information was too large, to highlight data from the subgroups with the worst baseline periodontal disease severity. In some cases, articles otherwise of interest did not report on these measures in any direct way but may have given other empirical evidence, and when such information showed a statistically significant net difference between the treatment and control groups (i.e., a difference of the differences between baseline and the end point), we tried to reflect that in discussing results.

Literature Search and Analysis Strategy

Literature Searches

Table 2. Strategy and Results of Original and Follow-up (more...)

Table 2. Strategy and Results of Original and Follow-up MEDLINE Searches

Step	Search Terms and Strategy	Number of Articles
1	Explode periodontitis	10,670
2	Limit 1 to randomized controlled trial	417
3	Explode single-blind method/ or explode double-blind method/or explode random allocation/ or explode epidemiologic study characteristics/ or explode comparative study	164,891
4	1 and 3	3,120
5	2 and 4	3,189
6	Limit 5 to human	2,987
7	Explode dental scaling/ or dental scaling.mp	1,757
8	Explode root planing/ or root planing.mp	1,087
9	7 or 8	2,047
10	Limit 9 to randomized controlled trial	260
11	9 and 3	841
12	10 or 11	896
13	Limit 12 to human	836
14	6 and 13	517
15	Explode metronidazole	7,150
16	Explode chlorhexidine	3,106
17	Explode doxycycline	3,270
18	Explode tetracycline	12,460
19	Explode minocycline	1,644
20	Explode amoxicillin	5,034
21	Explode amoxicillin-potassium clavulanate combination/ or augmentin.mp	1,067
22	Povidone iodine	1,359
23	15 or 16 or 17 or 18 or 19 or 10 or 21 or 22	31,725
24	Explode antibiotics	347,992
25	23 or 24	356,095
26	14 and 25	203
27	From 26 keep 1-200	200
28	From 26 keep 201-203	3
29	November 2002 rerun of the above search for new (unduplicated) articles	41
30	December 2002 search beginning with periodontal disease and steps 2-29	2

Table 3. Strategy and Results of EMBASE Search

Table 3. Strategy and Results of EMBASE Search

Step	Search Terms and Strategy	Number of Articles
1	Periodontitis	2,099
2	Randomized controlled trial	69,265
3	Single blind procedure	3,883
4	Double blind procedure	45,593
5	Randomization	5,018
6	Comparative study	38,078
7	2 or 3 or 4 or 5 or 6	129,259
8	1 and 7*	68

*68 records yielded 36 new (unduplicated with MEDLINE) records (articles).

To comprehend fully the scope of chronic periodontitis research, we had to design and implement several different search approaches. We searched for articles on clinical trial research on periodontitis that gave particular attention to SRP. We also searched for specific antibiotics and other chemical antimicrobials to see if those terms produced new articles or only ones we had already found. To find relevant articles on all aspects of periodontitis, we searched MEDLINE for papers published from 1966 through the second week of February 2002. This search is fully documented with the terms (MeSH and key words) and the counts of articles identified at each step of the search process in Table 2. Table 3 provides additional specifications of our inclusion and exclusion criteria.

Our initial MEDLINE search identified 10,670 articles (Table 3). After limiting to trials and human studies, 2,987 articles remained (Step 6). We further searched for “dental scaling” and “root planing” trials, yielding 836 articles, then searched for the specific drugs listed in Table 2. We had a total of 517 articles for consideration (Step 14); the 203 drug citations (Step 26) are a subset of the 517. This search was updated in November 2002, yielding 41 additional articles (step 29 of Table 2). In addition, we ran a somewhat different MEDLINE search in December 2002 starting with the term periodontal disease (instead of periodontitis); this search (line 30 of Table 2) yielded several dozen ostensibly new articles, but because periodontal disease includes gingivitis, virtually all proved to be irrelevant to our topic, and we added only two new items. The supplemental EMBASE search (Table 3) identified 68 articles, of which 36 were new. We identified three additional studies through references. Finally, external peer reviewers brought three additional studies (from the international literature) to our attention. In all, we examined a total of 599 articles (cited in the Bibliography [Chapter 8]).

Inclusion/Exclusion Criteria

Table 4. Inclusion/Exclusion Criteria For Studies of (more...)

Table 4. Inclusion/Exclusion Criteria For Studies of Added Effectiveness of Periodontitis Therapies Adjunctive to Scaling and Root Planing

Include studies if:	Exclude studies if:
They compare scaling and root planing (SRP) to SRP with some (single/combination agent) adjunctive therapy (chemical and antimicrobial).	Either they have no comparison/control group or the comparison/control group is historical.
They use concurrent comparison/control and experimental/test groups.	They compare therapies but do not include SRP alone (or with placebo).
They compare multiple adjunctive therapies to SRP alone (or with placebo) but report outcomes separately for each therapy.	They do not perform SRP in the experimental/test group in addition to the adjunctive therapy.
They measure outcomes at one or more specified intervals.	They use bundled adjunctive therapies.
	They do not report therapy outcomes separately for each therapy.
They do not specify time intervals for outcomes reported or do not report outcomes separately for each interval.
They are published in English.	Periodontitis treated is not chronic or adult type (i.e. is aggressive, juvenile, early onset, refractory, etc.).
They involve a human population.	The population under study has a behavioral factor or comorbidity that complicates the therapy or alters prognosis (smokers, HIV/AIDS, diabetes, etc.).

Table 4 specifies the major criteria that we applied to research studies for deciding on their inclusion or exclusion. We applied some in our searches, but opted to cast the net broadly and mainly exclude articles through review of titles and abstracts or the articles themselves. From an earlier preliminary search, we found a considerable number of trials in MEDLINE that appeared to be applicable to our question. For that reason, we searched only for reports of primary research described as trials and those with common characteristics of trials such as blinding and randomization. Although we included review articles that reported on similar types of research (to check references for additional articles), we did not incorporate them directly in this report. We excluded letters, commentaries, editorials, clinical case reports, and practice or treatment guidelines from our searches. Finally, we included only human populations in our search and did not review animal studies.

In our manual review of the articles identified and obtained, one of the senior investigators made an initial decision about inclusion for abstraction and another senior investigator reviewed that decision. In addition, the criteria for inclusion/exclusion were printed on the abstraction form so that abstractors could, if necessary, call for another senior review of eligibility. We determined whether the adjunctive therapy was antimicrobial and chemical at this point; we dropped articles about use of lasers (used presumably to destroy microbes but not a chemical or antimicrobial agent), and those about anti-inflammatory agents (clearly a chemical agent but not antimicrobial). Very late in the process, we also excluded articles reporting on sub-antimicrobial doses of doxycycline, clearly an antibiotic but, at such doses, not intended as an antimicrobial agent.

Many of the studies we identified in our search were not investigations directly relevant to our key question. Many were studying whether a chemical antimicrobial was as effective as SRP in treating periodontitis. However, several studies had a combined (SRP and antimicrobial) treatment group in addition to an SRP-only or SRP-with-placebo treatment group, an antimicrobial-only treatment group, and a no-treatment control group. If the investigators reported comparisons and outcomes of the combined treatment and SRP-only groups for measures we focused on for this report, we included the study. Further exclusions involved studies in which (a) the SRP provided to the group receiving the adjunctive therapy was not the same as that given to the SRP-only group, (b) the periodontitis being treated was not of the chronic (or adult) type but rather a more aggressive variety (e.g., juvenile, early onset, refractory), and (c) the population with the periodontitis had a complicating comorbidity (e.g., HIV/AIDS or diabetes) or a risk factor (e.g., smoking).

We included articles in languages other than English in our searches because often the titles and abstracts are translated into English; this allowed us to assess whether we were missing potentially important areas of work. If the full articles were not in English, however, we did not include them in the review. This approach may have caused us to omit some materials in other languages, but our previous systematic reviews have shown that relevant studies done outside of English-speaking nations that would otherwise have met our inclusion criteria would likely have been published in an English-language journal. Thus, we do not think that restricting the full review on this key question to English-language documents introduced any serious bias.

Meta-Analysis

The studies in this evidence report were exclusively clinical trials, albeit many were small and underpowered as individual studies. Investigators often reported the two clinical outcome measures of interest ─ probing depth (PD) and clinical attachment level (CAL) ─ in ways that would allow us to express effect sizes in the same way (as millimeters of change). These conditions permitted us to consider performing a series of meta-analyses for studies of at least some of the adjunctive therapies, so that we could quantitatively summarize the work and calculate an overall effect size measure.

We had a total of 70 studies that were candidates for inclusion in one of the meta-analyses. Of those, 29 studies, five with multiple arms, met the criteria that we established for inclusion in the meta-analysis:

• The study had to provide a measure of the treatment effect at 6 months post-baseline, although we allowed for a 3-month window on either side of the 6-month point. Thus, we included studies reporting results from 3 to 9 months (i.e., 12 to 39 weeks). For some articles, the true endpoint of the study fell within this range and they were included; for studies lasting longer than 9 months, the authors had to have reported results from some point within this range; and studies shorter than 3 months were automatically excluded.

• Included studies had to indicate treatment effect by either PD change (i.e., reduction) CAL change (i.e., gain), or both.

• Included studies had to indicate the between-group difference in means (the treatment effect), the standard error or the 95 percent confidence interval of the treatment effect, and the sample size for each study group. Studies could also be included if they provided enough information to allow us to calculate these numbers, such as the within-group differences and their standard deviations or standard errors. Some studies gave the difference between the mean differences of the two study groups (i.e., the effect size of the adjunctive treatment), but often we had to calculate the overall effect size from the mean differences between the baseline to follow-up means for the experimental treatment group and the control group. Similarly, we often had to estimate the standard error of the difference of the mean differences.

• To proceed with a meta-analysis of a given therapy, we required that at least three studies related to one of the clinical outcomes meet the above criteria. The studies of systemic minocycline, systemic metronidazole, and combined metronidazole and amoxicillin failed to meet this requirement.

Analysis

Analysis was based on the general inverse variance method of estimation available in RevMan. This method calculates a pooled, or overall, effect for each outcome, a test of significance for the treatment effect across all studies (a Z statistic), and a measure of heterogeneity. The heterogeneity statistic is a rough indicator of whether all included studies are indeed comparable (null hypothesis is that treatment effect does not differ among trials). The heterogeneity statistic is also used to calculate an I² statistic, which indicates, approximately, the proportion of total variation in the study estimates that can be attributed to heterogeneity rather than sampling error.²⁹ The method of estimation can handle situations in which effect is specified as fixed or random; we modeled both but have reported only fixed effects for our analysis, as the results were quite similar.

Eight studies did not provide required information on certain aspects of their results. Typically this meant that, to include them, we had to calculate or estimate standard errors ourselves.^{30- 37} Rather than risk dramatically overestimating the “real” standard error, possibly by as much as 100 percent or more, and unfairly reduce the likelihood of finding statistical significance, we used a set of statistical procedures incorporated in SAS macros³⁸ to estimate the standard errors of the difference of differences to include in the meta-analysis.

Quality Rating of Articles

To grade the quality of individual studies (articles), as is expected for AHRQ systematic reviews,³⁹ we developed a quality rating checklist for articles that dealt with internal validity, external validity, and analytic dimensions (Appendix C). The 13 items in the checklist relate mostly to study design elements. We customized items on the checklist to fit the question and literature, but many of the component items were taken directly from, or represent slight modifications of, existing rating scales used by the RTI-UNC EPC, reflecting suggestions from work done by this EPC. ^{39, 40} CONSORT criteria also figured prominently in our thinking because the studies were trials of various kinds.⁴¹

We pretested draft forms on several articles and eliminated or reworded some items. We gave scores to articles by summing the number of items on the quality rating form checked as “yes” and dividing by 13, the number of items. The EPC Administrator and Project Director independently assigned quality grades to all studies using the final form (Appendix C). We entered both quality scores (essentially two percentages) into the evidence tables. Although not formally validated, our rating scheme adopts the basic strategy of quality grading and provides a relative basis by which we and others can assess the overall strength of the research available to address our key question. We do not employ the quality score as a way of reviewing articles for inclusion in the evidence report or the meta-analysis. This approach is in accord with what recent research has found: no reliable relationship between overall quality rating measures and estimates of treatment effects in trials.⁴²

Development of the Evidence Report

Following completion of evidence and/or text tables, we sent them to main authors of the report, together with a general outline of the results chapter of the report. The authors had previously agreed to present mainly qualitative syntheses of the information in the tables, with primary attention to PD and CAL findings and data on spirochetes when available; they would call out information on specific articles only when those studies offered clinically significant findings or insights into the key question. Authors writing from either text or evidence tables were also asked to check data in the tables against articles whenever any table entry was unclear or inconsistent across tables. The authors returned their sections to the project director and EPC Co-Director, who developed an overall synthesis of the results and the discussion chapter. In addition, the project director and other members of the project team developed the research agenda chapter, drawing on the limitations and gaps in the existing literature and on promising leads from the studies reviewed.

We submitted the draft evidence report for external peer review in mid-April 2003 (see Appendix A). Upon receipt of reviews, the EPC staff compiled them into a peer reviewer matrix, discussed many issues with AHRQ and NIDCR staff, and then revised the report as appropriate, documenting in detail the disposition of all significant clinical, analytic, or policy-relevant changes. We expanded the meta-analyses somewhat to include more studies where we were able to use the SAS macros to produce more estimates of standard errors than we had been able to do before peer review. As noted, we also added several recent articles mentioned by peer reviewers that met our inclusion criteria but had not appeared in any MEDLINE or EMBASE searches. The revised version of the report was submitted to AHRQ and NIDCR for further review before it was put into absolute final form.

Overview

This chapter presents the findings for the key question in this evidence report concerning the added effectiveness of therapy adjunctive to scaling and root planing (SRP) in the treatment of chronic periodontitis over time. The key outcomes are reductions in probing depth (PD), gains in clinical attachment level (CAL), and secondarily, microbial changes, chiefly reductions in the percentage of spirochetes present in crevicular fluid or plaque samples.

We present results in separate sections according to the specific agents used adjunctively and to the mode of delivery (either systemic or local) and when more than just one or two articles addressed a particular agent or combination of agents. The agents we report on in separate sections include tetracycline, minocycline, metronidazole, the combination of metronidazole and amoxicillin, and chlorhexidine. We grouped studies of azithromycin, spiramycin, amoxicillin clavenate, and doxycycline, which did not have enough studies to treat separately, into a sixth section called other antibiotics. Finally, we present a seventh section for other kinds of antimicrobial agents such as povidone iodine, hydrogen peroxide, and fluorides. Chapter 7 has the full evidence tables for each main category of treatment; brief summary tables appear at the end of this chapter.

Tetracycline

Systemic Tetracycline

Qualitative and Descriptive Results

Table 5. Main Features and Findings on Probing Depth (more...)

Table 5. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Systemic

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Listgarten et al., 197843	R NP NR	25 weeks	N: T: C:	12 subjects 6 subjects 6 subjects	250 mg tetracycline 4xday, days 0 to 14 and days 42 to 56	+0.20, NS	Data NR, NR
Lindhe et al., 198344	R PL DB	50 weeks	N: T: C:	14 subjects 7 subjects 7 subjects	250 mg tetracycline 4xday, days 0 to 14 and 250 mg 1xday days 15 to 350	+0.8, NR	+0.3, NR
Al-Joburi et al., 198945	R PL DB	24 weeks	N: T: C	51 subjects 27 subjects 24 subjects	250 mg tetracycline 4xday, days 0 to 14	Initial PD ≤ 3 mm: +0.12, NS Initial PD 4-6 mm: -0.05, NS Initial PD ≥ 7 mm: +0.19, NS	Initial PD ≤ 3 mm: +0.49, NS Initial PD 4-6 mm: +0.04, NS Initial PD ≥ 7 mm: +0.45, NS
Haffajee et al., 199546	R PL DB	43 weeks (10 months)	N: T: C:	24 subjects 13 subjects 11 subjects	250 mg tetracycline 3xday, days 0 to 30, plus 0.12% chlorhexidine rinse for same 30 days	+0.29, NS	Data NR, NS
Ramberg et al., 200147	Non-R NP NR	52 weeks	N: T: C:	89 subjects 28 subjects 61 subjects	250 mg tetracycline 4xday, days 0 to 21 and 0.2% chlorhexidine rinse 2xday	+0.3, NS	+0.31, P < 0.001

R, randomized; Non-R, nonrandomized; PL, placebo-controlled; NP, no placebo; DB, double-blind; SB, single blind; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

Five clinical trials appearing between 1978 and 2001 of systemic tetracycline, a broad-spectrum antibiotic of long standing that is effective against a wide array of bacteria, met the inclusion criteria (Table 5 and Evidence Table 1a). Four trials were randomized;^{43- 46} the most recently reported trial was not.⁴⁷ Taken together, the studies comprised 190 subjects who completed the studies, 81 on tetracycline regimens with SRP and 109 receiving SRP alone or with placebo. Subjects in the trials differed in periodontal disease experience. Three studies included subjects with advanced or severe periodontitis, one included “relatively young individuals” with severe disease,⁴³ and the fifth included only subjects who had demonstrated active disease with attachment loss equal to or greater than 2.5 mm at one or more sites in a pretrial monitoring period.⁴⁶

The studies differed in the selection of sites used to characterize a subject’s response to therapy. One study included only two sites per subject, both interproximal sites with PDs of 7 mm or more.⁴⁵ A second study based analyses on patients with at least three pairs of contralateral teeth with PDs of 5 mm or greater.⁴³ A third study included six measurement sites for all teeth present except third molars;⁴⁶ a fourth included two teeth per subject with 6 mm or greater PDs;⁴⁴ and the fifth study did not report the basis for the selection of sites.⁴⁷

SRP procedures differed across these studies. In one trial, the investigators used modified Widman flap surgery for all sites determined to be active in the pre-intervention period and for all sites with PDs of 4 mm or greater;⁴⁶ SRP was repeated at 3, 6, and 9 months. Another study provided repeated SRP during the observation period at 15 and 22 weeks.⁴³ In the other studies, SRP was provided once, at baseline, and no other details were noted.

The experimental regimen varied across the five trials. The standard dose was 250 mg, with a frequency of either three or four times per day for different treatment periods lasting 14 days,⁴⁵ 21 days,⁴⁷ and 30 days.⁴⁶ One regimen repeated an initial 14-day dosing pattern at the beginning of the sixth week;⁴³ a fifth approach continued past 14 days to the end of the study (day 350) with a single 250 mg dose per day.⁴⁴ Finally, the trials varied in longest follow-up period: 24 weeks to 52 weeks. Three trials reported results from interim periods.

For four studies reporting overall PD measurements, experimental subjects had 0.8 mm,⁴⁴ 0.3 mm,⁴⁷ 0.29 mm,⁴⁶ and 0.2 mm⁴³ greater mean reductions in the experimental group; of these, three were not significant and one⁴⁴ was not tested. A fifth study reported PD reduction by original PD values, and the largest PD reduction was 0.19 (not significant) for those with initial PDs of 7 mm or greater.⁴⁵ No interim measures differed significantly between experimental and control groups.

Five studies examined CAL gains, but only three reported data on CAL gains. One trial found a 0.31 mm net gain at 52 weeks (P < 0.001) for the experimental group.⁴⁷ In the other two trials reporting CAL values, one noted a 0.3 mm net improvement⁴⁴ and the other reported gains from 0.04 mm to 0.49 mm (depending on initial PD value), but all were not significant.⁴⁵ No interim measure for these studies was reported as being significantly different between groups. Of the two studies that did not report data, one did state that difference in CAL gain between the groups was not significant.⁴⁶

Three trials analyzed the proportion of spirochetes in the oral microflora. One trial found that significant differences favoring the experimental group at 2 and 8 weeks had disappeared by week 24.⁴⁵ Another team reported a larger proportional change in the experimental group (48 percent to 0 percent) than in the control group (37 percent to 8 percent), but they did not test the difference statistically.⁴⁴ A third study found similar overall change in both groups (34.8 percent or 36.3 percent to 6.3 percent or 6.5 percent), with the experimental groups showing a larger decline to 0 percent (not tested) at an interim examination.⁴³ Finally, one trial examined oral microflora but did not separate results for the tetracycline groups from results for another experimental group assessing amoxicillin clavenate.⁴⁶ The percentage of sites colonized with several putative pathogens decreased significantly for both group combined.

Quantitative Analysis of Systemic Tetracycline Effects

Figure 3. Meta-analysis of Systemic Tetracycline and (more...)

   Figure 3. Meta-analysis of Systemic Tetracycline and SRP vs. SRP Alone: Probing Depth

*Only subjects with initial PD ≥ 4 mm to ≤ 6 mm.

Probing Depth. Of the five studies reporting on PD reduction reviewed above, we included three in the meta-analysis of PD effect size (Figure 3).^{43- 45} Both excluded studies had study periods greater than 9 months.^{46, 47}

The estimate of overall effect size for PD is 0.15 mm (95 percent CI, -0.29 mm to 0.58 mm). Effect sizes of the two studies favoring use of adjunctive local tetracycline range from 0.20 to 0.90. The third study demonstrated a greater reduction in PD for SRP alone (-0.05 mm).⁴⁵ Based on their 95 percent CIs, none of these PD effect sizes differed significantly from zero. Because of its appreciably smaller standard error of the difference and larger sample size, the Al-Joburi et al. study contributed more weight to the meta-analysis results.⁴⁵ Given the mixed direction of the differences and small study samples, the nonsignificant overall effect size in this meta-analysis is not surprising.

Clinical Attachment Level. Only two of the four studies examining gain in CAL reviewed above had data and ran for the appropriate length of time.^{44, 45} As explained earlier, we thus did no meta-analysis for CAL gain resulting from adjunctive use of systemic tetracycline.

Local Tetracycline

Qualitative and Descriptive Results

Table 6. Main Features and Findings on Probing Depth (more...)

Table 6. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Local Tetracycline

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Goodson et al., 198548	R NP SB	52 weeks	N: T: C:	10 subjects 9 teeth 9 teeth	25% tetracycline fibers for 10 days	Data NR, NR	Initial PD > 3 mm: +0.34, NS Initial PD > 6 mm: +0.33, NS
MacAlpine et al., 198531	Non-R PL NR	24 weeks	N: T: C:	11 subjects 16 sites 16 sites	50 mg/ml tetracycline irrigation every 2 weeks for 22 weeks	+0.4, NR	+0.3, NS
Nylund and Egelberg, 199051	Non-R PL NR	52 weeks	N: T: C:	20 subjects 20 subjects 20 subjects	50 mg/ml tetracycline irrigation once every 2 weeks for 3 months	Data NR, NS	Data NR, NS
Minabe et al., 199152	R NP NR	8 weeks	N: T: C:	16 subjects 10 sites 8 sites	Tetracycline immobilized collagen film, 4 consecutive weekly administrations	+0.7, NS	Data NR, NS
Unsal et al., 199459	R NP NR	12 weeks	N: T: C:	15 subjects 7 subjects 8 subjects	40% tetracycline in white petroleum, single application	-0.43, NS	-0.23, NS
Jeong et al., 199433	R NP NR	12 weeks	N: T1: T2: C:	16 subjects 16 subjects 16 subjects 16 subjects	T1: 5% tetracycline in gel, 1 application T2: 5% tetracycline with citric acid in gel, 1 application	T1: +0.27, NS T2: +0.93, P < 0.05	T1: +0.14, NS T2: +0.73, NS
Newman et al., 199457	R NP SB	26 weeks (6 months)	N: T: C:	105 subjects 105 subjects 105 subjects	12.1 cm of tetracycline fiber for 10 days	+0.73, P < 0.01	+0.48, P < 0.05
Shiloah and Patters 199454	R PL SB	4 weeks	N: T: C:	7 subjects 12 sites 12 sites	5% aqueous tetracycline, 1 application	Data NR, NS	Data NR, NS
Drisko et al., 199549	R NP SB	52 weeks	N: T: C:	116 subjects 116 subjects 116 subjects	17 mg tetracycline fibers for 10 days	Data NR, NS	Data NR, NS
Darhous et al., 199550	Non-R NP NR	8 weeks	N: T: C:	7 subjects 7 subjects 7 subjects	100 mg/ml tetracycline irrigation, 5 minutes, 1 application	+1.04, NR	+1.8, NR
Trombelli et al., 199655	R NP SB	8 weeks (60 days)	N: T1: T2: C:	12 subjects 20 sites 24 sites 19 sites	T1: 100 mg/ml tetracycline irrigation, 4 minutes, 1 application T2: 25% tetracycline fibers for 10 days	T1: +0.4, P = 0.011 T2: +0.5, P = 0.011	T1: -0.2, NS T2: +0.4, NS
Lie et al., 199834	R NP DB	26 weeks	N: T: C:	18 subjects 18 subjects 18 subjects	3% tetracycline ointment after each SRP session, 1 week apart	Defect sites:* +0.6, NS Nondefect sites:* -0.3, NS	Defect sites:* +1.0, NS Nondefect sites:* +0.2, NS
Tonetti et al., 199835	R NP SB	26 weeks	N: T: C:	123 subjects 63 subjects 60 subjects	25% tetracycline fibers for 10±3 days	+0.4, NS	-0.1, NS
Yalcin et al., 199956	R NP SB	7 weeks	N: T: C:	17 subjects 221 sites 191 sites	Slow-release tetracycline fibers for 10 days	+0.41, P = 0.047	+0.20, NS
Kinane and Radvar, 199953	R NP SB	26 weeks (6 months)	N: T: C:	39 subjects 19 subjects 20 subjects	25% tetracycline fibers for 10 days	+0.67, P = 0.008	+0.15, P < 0.05
Friesen et al., 200258	R NP SB	26 weeks (6 months)	N: T1: T2: C:	24 subjects 24 subjects 24 subjects 24 subjects	13.5 mg tetracycline strips for 7 to 10 days T1: single strip T2: multiple strips	T1: +0.43, NR T2: +0.87, NR	T1: +0.44, NS T2: +0.48, NS

R, randomized; Non-R, nonrandomized; PL, placebo-controlled; NP, no placebo; DB, double-blind; SB, single blind; NB, not blinded; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

*Defect and nondefect sites refer to, respectively, the mean of measurements from the buccal and lingual side versus measurements for all remaining parts of the tooth.

Sixteen clinical trials, published between 1985 and 2002, met the inclusion criteria (Table 6 and Evidence Table 1b). All but three used randomized designs. In 10 studies, the examiner(s) were not aware of treatment assignment, and in 13, placebos were not used. Most of the trials assessed the effects of the intervention on sites with at least 5 mm PDs; four studies either included patients with shallower PDs (as small as 3 mm)^{33, 48, 49} or did not report site selection criteria.⁵⁰ Three studies assessed only sites with class II furcation involvements.^{35, 51, 52} One study required demonstration of active disease immediately before inclusion.⁵³

Five trials reported site-based analyses,^{31, 52, 54- 56} one used teeth, and the rest used the subject as the unit of analysis, with one tooth or site per quadrant per subject being most commonly assessed.

The extensiveness and methods used for the SRP varied, ranging from ultrasonic scaling during one visit to multiple visits over 4 to 7 hours for hand scaling. Tetracycline intervention vehicles included fibers, irrigation, collagen film, ointment, gel, and strips. In one study arm, tetracycline was combined with citric acid;³³ in all others, tetracycline was used alone. In one irrigation study, tetracycline was administered every 2 weeks for 22 weeks.³¹ Effects were assessed over durations as short as 4 weeks and as long as 52 weeks. All but one of the studies that reported statistically significant differences had used tetracycline fibers.

Of these 16 studies, 12 reported PD data sufficient to determine the net reduction (i.e., the difference between the experimental and control groups’ baseline to follow-up differences). Of these 12 studies, four found statistically significantly greater PD reductions associated with the experimental group. The differences were 0.93 mm at 12 weeks (combined with citric acid gel, P < 0.05),³³ 0.73 mm at 6 months (P < 0.01),⁵⁷ 0.67 mm at 26 weeks (P = 0.008),⁵³ and 0.41 mm at 7 weeks (P = 0.047).⁵⁶ In eight studies in which differences were either not tested or not significant, reductions in the experimental group were greater than those in the control group; some of these were of a magnitude similar to the statistically significant differences: 1.04 mm;⁵⁰ 0.7 mm;⁵² 0.6 mm;³⁴ 0.4 mm and 0.5 mm with irrigation and fibers, respectively;⁵⁵ 0.43 mm and 0.87 mm with one and multiple strips, respectively;⁵⁸ 0.4 mm;³¹ 0.4 mm;³⁵ and 0.27 mm without citric acid.³³ One study found a nonsignificant 0.43 mm difference favoring the control group.⁵⁹ The four remaining studies reported nonsignificant differences but did not give the magnitude or direction of those differences.^{48, 49, 51, 54} Interim results also showed statistically significant greater reduction in PD associated with the experimental arm in two studies.^{35, 53} In one of these,³⁵ the significant difference was not maintained at the final assessment.

Sixteen studies reported results of some kind for CAL gain, some for two different patient groups or treatments. Of these, two reported significantly greater CAL gains associated with local tetracycline treatment compared to SRP alone: 0.48 mm at 26 weeks (P < 0.05)⁵⁷ and 0.15 mm at 26 weeks (P < 0.05).⁵³ Nine studies reported nonsignificant or untested differences favoring the experimental group: 1.8 mm,⁵⁰ 1.0 mm,³⁴ 0.73 mm (with citric acid),³³ 0.44 mm and 0.48 mm for a single and multiple strips, respectively,⁵⁸ 0.34 mm and 0.33 for groups with initial PD values of greater than 3 mm and greater than 6 mm, respectively,⁴⁸ 0.3 mm,³¹ 0.2 mm,⁵⁶ and 0.14 mm (gel).³³ Three studies reported nonsignificant differences favoring the control group: 0.23 mm,⁵⁹ 0.2 mm for tetracycline irrigation,⁵⁵ and 0.1 mm.³⁵ Finally, three studies reported nonsignificant differences but did not report the magnitude or direction of those differences.^{49, 51, 60} No reported interim results were significantly different between the experimental and control groups.

Finally, five studies examined microbiological outcomes,^{31, 33, 49, 52, 54} but none reported significant differences in these outcomes at final assessments. An interim (3-month) assessment in one study found a significantly greater reduction in the proportion of P. gingivalis in the experimental group.⁴⁹

In general, the studies that reported side effects noted some irritation associated with the application of the experimental therapy and, less frequently, candidiasis. These conditions resolved when therapy ended.

Quantitative Analysis of Local Tetracycline Effects

Figure 4. Meta-analysis of Local Tetracycline and SRP (more...)

   Figure 4. Meta-analysis of Local Tetracycline and SRP versus SRP Alone: Probing Depth

*Tetracycline gel with citric acid used as treatment.

Probing Depth. Of the 16 studies reporting on PD reduction reviewed above, we included six in the meta-analysis of PD effect size (Figure 4).^{31, 33, 34, 53, 57, 59} One study included two different adjunctive local tetracycline treatments – one without and the other with citric acid added to the tetracycline³³ – so the meta-analysis had seven entries. Of the 10 studies excluded, five had study periods of less than 3 months,^{50, 52, 54- 56} and four provided no data from which to calculate effect size measures or variances.^{48, 49, 51, 58} We excluded the final study because a test for heterogeneity showed that it was at too great variance with the other studies, suggesting that it represented a different intervention.³⁵

The estimate of overall effect size for PD is 0.47 mm (95 percent CI, 0.22 mm to 0.72 mm). Effect sizes of included studies favoring use of adjunctive local tetracycline range from 0.27 to 0.93. All but one of the study effect sizes represented results that demonstrate greater PD reduction using adjunctive local tetracycline with SRP than using SRP alone. One study demonstrated a greater reduction in PD for SRP alone (-0.43).⁵⁹ Based on the 95 percent CIs, the PD effect sizes of only two studies differed significantly from zero.^{33, 57} Two studies with appreciably smaller⁵⁷ or larger³¹ standard errors of difference between means contributed relatively more and less weight, respectively, to the meta-analysis results.

The statistically significant PD result supports the added effectiveness of locally applied tetracycline as an adjunct to SRP in the treatment of chronic periodontitis in adults. It does not address, however, whether a mean change of 0.47 mm is clinically meaningful. In contrast to this less than half a millimeter difference between using and not using some form of local tetracycline as an adjunct to SRP, the effects of SRP alone on reduction in PD in these studies ranged from 0.71 mm to 2.30 mm for the same periods of time.

Figure 5. Meta-analysis of Local Tetracycline and SRP (more...)

   Figure 5. Meta-analysis of Local Tetracycline and SRP versus SRP Alone: Clinical Attachment Level

*Tetracycline gel with citric acid used as treatment.

^†Multiple tetracycline strips used as treatment.

Clinical Attachment Level. Of the 16 studies examining gain in CAL reviewed above, we were able to include nine in this meta-analysis (Figure 5).^{31, 33- 35, 48, 53, 57- 59} Two studies included two different adjunctive local tetracycline treatments: citric acid in one gel group and none in the other,³³and single and multiple tetracycline strips.⁵⁸ Thus, the final meta-analysis had 11 entries. Among the seven excluded studies, five had study periods of less than 3 months,^{50, 52, 54- 56} and two had no way for us to calculate CAL effect size measures or variances.^{49, 51}

The estimate of overall effect size for CAL is 0.24 mm (95 percent CI, 0.07 mm to 0.42 mm). Effect sizes of included studies favoring use of adjunctive local tetracycline ranged from 0.01 mm to 1.00 mm. All but two of the study effect sizes represented greater CAL gain using adjunctive local tetracycline with SRP than using SRP alone. Of these two, one demonstrated a greater reduction in CAL for SRP alone (-0.23 mm),⁵⁹ and one favored neither study group.³⁵

Based on the estimated 95 percent CIs, only two studies had CAL effect sizes that differed significantly from zero.^{33, 34} One study had an appreciably smaller³⁵ standard error of difference between means and another had a standard error that was appreciably larger³¹ than the others, so they contributed considerably more or less weight, respectively, to the meta-analysis results.

The statistically significant CAL result supports the added effectiveness of locally applied tetracycline as an adjunct to SRP in the treatment of chronic periodontitis in adults, but whether a mean gain of 0.24 mm is clinically meaningful remains unclear. In contrast to this 0.24 mm difference, the effects of SRP alone on gain in CAL in these studies ranged from -0.13 mm to 1.61 mm for the same periods of time.

Minocycline

Systemic Minocycline

Table 7. Main Features and Findings on Probing Depth (more...)

Table 7. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Systemic Minocycline

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Ciancio et al., 198261	R PL DB	10 weeks	N: T: C:	26 subjects 13 subjects 13 subjects	200 mg minocycline per day for 7 days	Data NR, NS	N/A
Atilla et al., 199662	Non-R NP NR	6 weeks	N: T1: C1: T2: C2:	21 subjects 5 (PD 4-5 mm) 6 (PD 4-5 mm) 5 (PD ≥ 6 mm) 5 (PD ≥ 6 mm)	100 mg minocycline per day for 14 days	Initial PD 4-5 mm: -0.06, NS Initial PD ≥ 6 mm: +0.49, NS	N/A

R, randomized; Non-R, nonrandomized; PL, placebo-controlled; NP, no placebo; DB, double-blind; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

Qualitative and Descriptive Results. Two studies of systemic minocycline met inclusion criteria (Table 7 and Evidence Table 2a).^{61, 62} The 1982 trial by Ciancio et al. was a randomized, placebo-controlled, double-blinded trial,⁶¹ whereas the 1996 Atilla et al. study did not report study blinding but noted that it was neither placebo controlled nor randomized. These were small trials of patients with moderate to severe periodontitis, comprising in all 47 subjects who completed the trials, 23 on minocycline regimens with SRP and 24 with SRP only. The studies differed in numerous ways: dosage and duration of minocycline (200 mg a day for 7 days⁶¹ versus 100 mg a day for 4 days⁶²); focus on probing depth (4 mm to 5 mm versus 6 mm or higher⁶²); nature and periodicity of SRP; and outcomes assessed (the only overlap being PD).

With respect to PD levels, Ciancio et al. assessed outcomes at baseline and at 7, 14, 35, 49, and 70 days but only reported “no significant changes in any study group during the experimental period.”⁶¹ The Atilla et al. study reported PD differences according to the initial PD.⁶² Among subjects with initial PD of 4 mm to 5 mm, those receiving systemic minocycline experienced 0.06 mm less reduction than those receiving SRP only (not significant). By contrast, among patients with initial PD of 6 mm or greater, those receiving adjunctive minocycline experienced 0.49 mm greater reduction than controls (not significant); nonetheless, the authors commented that systemic minocycline might be a useful adjunct to nonsurgical SRP “in the presence of deep pockets, especially for reinfected cases.”

Neither study reported CAL data. One reported that patients receiving both SRP and minocycline had notable, long-lasting changes in subgingival microbiologic findings (e.g., cell counts of spirochetes).⁶¹ Owing to the small size of these trials, their nonsigifncant findings on PD (and no data on CAL), and other differences, we did not do any meta-analysis.

Local Minocycline

Table 8. Main Features and Findings on Probing Depth (more...)

Table 8. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Local Minocycline

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
van Steenberghe et al., 199363	R PL DB	12 weeks	N: T: C:	81 subjects 42 subjects 39 subjects	2% minocycline ointment applied at baseline and weeks 2, 4, and 6	Initial PD ≥ 5 mm: +0.3, P = 0.0018 Initial PD ≥ 7 mm: +1.0, P = 0.0001	Initial PD ≥ 5 mm: 0.0, NS Initial PD ≥7 mm: +0.40, NS
Jones, et al., 199464	R PL DB	26 weeks (6 months)	N: T: C:	17 subjects 11 subjects 6 subjects	1 mg microencapsulated minocycline injected into ≥5 mm pockets, single application	Data NR, NS	Data NR, NR
Graca et al., 199732	R PL DB	12 weeks	N: T: C:	26 subjects 13 subjects 13 subjects	2% minocycline gel applied at baseline and weeks 2 and 4	+0.34, NS	+0.39, P < 0.05
Kinane and Radvar, 199953	R NP DB	26 weeks (6 months)	N: T: C:	39 subjects 20 subjects 19 subjects	2% minocycline gel applied at baseline and weeks 2 and 4	+0.39, NS	+0.36, NS
van Steenberghe et al., 199965	R PL DB	65 weeks (15 months)	N: T: C:	93 subjects 46 subjects 47 subjects	2% minocycline ointment applied at baseline week 2, and months 1, 3, 6, 9, and 12	Initial PD ≥ 5 mm: +0.77, P < 0.0001 Initial PD ≥7 mm: +1.10, P < 0.0001	Initial PD ≥5 mm: +0.49, P < 0.0001 Initial PD ≥7 mm: +0.43, P < 0.0001
Williams et al., 200166	R PL DB	39 weeks (9 months)	N: T: C:	467 subjects 237 subjects 230 subjects	2% minocycline microspheres in polymer gel applied at baseline and months 3 and 6	+0.32, P < 0.001	Data NR, NR
Henderson et al., 200267	R NP SB	26 weeks (6 months)	N: T: C:	15 subjects 15 subjects 15 subjects	1 mg microencapsulated minocycline, single application	+0.7, P ≤ 0.05	+0.8, P = 0.04
Van Dyke et al., 200268	R NP DB	26 weeks (6 months)	N: T: C:	22 subjects 12 subjects 10 subjects	1 mg microencapsulated minocycline, single application	+0.28, NS	+0.48, NS

R, randomized; Non-R, nonrandomized; PL, placebo-controlled; NP, no placebo; DB, double-blind; SB, single blind; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

Qualitative and Descriptive Results. We included eight studies of local applications of minocycline, all appearing between 1993 and 2002 (Table 8 and Evidence Table 2b).^{32, 53, 63- 68} All were randomized; five were placebo-controlled;^{32, 63- 66} and all but one⁶⁷ were reported to be double-blind studies. Taken together, the eight trials involved 760 subjects who completed the trials; of these 396 received adjunctive minocycline (237 in the Williams et al. trial alone⁶⁶) and 379 received only SRP (230 in the Williams et al. trial). The Williams et al. trial is described as an intention-to-treat design and is a multi-center trial.

Subjects were variously described as having moderate to severe chronic (or advanced) periodontitis. Specifics about site criteria for inclusion, such as minimum number of teeth or sites, minimum PD levels, or bleeding, differed across the studies.

Four trials reported SRP details: subjects were hand-scaled in both van Steenberghe et al. studies (for a maximum of 15 minutes per quadrant in the 1999 study),^{63, 65} and in the other studies, subjects had both hand and ultrasonic SRP for 90 minutes.^{67, 68} The trials differed in many respects (when details were reported); for example, assessment times and duration of measurement included seven times (baseline through 15 months⁶⁵), five times (baseline through 9 months⁶⁶), for two studies three times (baseline through 6 months^{53, 67} and baseline through 12 weeks³²), and for one study four times (baseline through 6 months).⁶⁸ Moreover, the experimental regimen also differed across the trials. Two studies used gels (2 percent at baseline and at 2 and 4 weeks^{32, 53}); two other studies used ointments (2 percent at baseline and at various weekly or monthly points thereafter;^{63, 65} one study used 1 mg of minocycline as a single application at baseline;⁶⁴ one used 1 mg of 2-percent minocycline microspheres in a 3-mg polymer gel (at baseline and months 3 and 6⁶⁶); and two used single applications of 1-mg microencapsulated minocycline.^{67, 68}

With respect to PD effects, four trials reported a greater effect of adjunctive local minocycline – i.e., more reduction among experimental than control groups that was statistically significant for at least some patients. Experimental subjects experienced 1.0 mm more reduction at 12 weeks (for those with 7 mm or greater initial PD, P = 0.0001),⁶³ 0.77 mm and 1.10 mm more reduction at 65 weeks (respectively, for subjects with 5 mm or greater and 7 mm or greater initial PD, P < 0.0001 in both cases),⁶⁵ 0.7 mm at 26 weeks (P ≤ 0.05),⁶⁷ 0.32 mm at 39 weeks (P < 0.001) and 0.3 mm at 12 weeks (for those with initial PDs of 5 mm or greater, P = 0.0018).⁶³ The remaining trials showed net PD reductions favoring the local minocycline groups for which significance was not reported or the test was not significant: 0.39 mm,⁵³ 0.34 mm,³² and 28 mm.⁶⁸ The last study did not report data but did indicate the difference between the groups was not significant.⁶⁴ In the largest trial lasting 9 months, the investigators reported that the percentage of sites with PD reductions of 1 mm or greater, or 2 mm or greater, was higher in the treatment group by, respectively, 10 percent and 11.6 percent; both values reflected a statistically significant difference (P < 0.001) from those for patients receiving SRP and placebo.⁶⁶

Of the six trials that reported actual CAL data, three reported significantly greater net gains in CAL for the minocycline treatment groups: 0.8 mm at 26 weeks (P = 0.04),⁶⁷ 0.49 mm at 65 weeks (patients with baseline pockets of 5 mm or greater, P < 0.0001), and 0.43 mm at 65 weeks (baseline pockets of 7 mm or greater, P < 0.0001),⁶⁵ and 0.39 mm at 12 weeks (P < 0.05).³² Others reported no or nonsignificant differences in gains favoring the minocycline groups: 0.48 mm,⁶⁸ 0.4 for patients with initial PDs of 7 mm or greater,⁶³ 0.36 mm,⁵³ and 0.0 mm for patients with initial PD of 5 mm or greater.⁶³ Of the two remaining trials that did not report data on CAL gain, one indicated that the between group difference was not significant⁶⁴ and the other did not say.⁶⁶

None of these trials reported on percentage changes in spirochetes. One trial reported that concentrations of P. gingivalis and P. intermedia were lower following a minocycline ointment treatment than following SRP with placebo ointment at 2, 4, 6, and 12 weeks post-treatment, as was the concentration of A. actinomycetemcomitans at weeks 6 and 12.⁶³ The van Steenberghe team then later reported significant differences in several microbiological outcomes (e.g., P. gingivalis, P. intermedia, C. rectus, T. denticola, E. corrodens, F. nucleatum, and A. actinomycetemcomitans) at various follow-up points from month 1 to month 15 between patients receiving adjunctive minocycline ointment and nonsurgical SRP treatment alone.⁶⁵ There were significantly greater reductions in the P. gingivalis, T. denticola, and C. rectus counts from baseline,⁶⁵ but for the other microbial outcomes the changes were not significant. The Jones team used DNA probes for microbiological assessments of A. actinomycetemcomitans, P. gingivalis, P. intermedia, E. corrodens, and C. rectus. ⁶⁴ The P. gingivalis prevalence was completely nondetectable at 1 month and had a 60 percent reduction at 6 months for the treatment group. The SRP-alone group reductions were never significant from baseline.

Attrition was not well reported in these studies. Authors reported a wide array of adverse effects (among both treatment and control groups), some relatively serious and some not. They included dental infection, abscesses, gingivitis, gingival edema, stomatitis, root sensitivity, tooth sensitivity, dental pain, local irritation, headache, diarrhea, and other “minor” clinical reactions (e.g., redness).

Figure 6. Meta-analysis of Local Minocycline and SRP (more...)

   Figure 6. Meta-analysis of Local Minocycline and SRP versus SRP Alone: Probing Depth

Quantitative Analysis of Local Minocycline Effects. Six of the eight studies examining the effect on PD of locally applied minocycline reviewed above were included in the meta-analysis (Figure 6).^{32, 53, 65- 68} Reasons for excluding the other two studies were that one had no data for us to calculate an effect size measure⁶⁴ and the other had too great variance with the rest of the studies according to a test of heterogeneity, suggesting that it represented a different intervention.⁶³

The estimate of overall effect size for PD is 0.49 mm (95 percent CI, 0.40 mm to 0.58 mm). Effect sizes of included studies favoring use of adjunctive local minocycline range from 0.28 to 0.70. Despite the fact that all of the study effect sizes represent greater PD reduction using adjunctive local minocycline with SRP than using SRP alone, based on the 95 percent CIs, the PD effect sizes of only two of those studies differed significantly from zero.^{65, 66} Those same two studies had appreciably smaller standard errors of difference between means and larger study samples, and consequently they contributed almost all of the weight to the meta-analysis results.

The statistically significant PD result supports the added effectiveness of locally applied minocycline as an adjunct to SRP in the treatment of chronic periodontitis in adults. It does not address, however, whether a mean change of 0.49 mm is clinically meaningful. One context for judging the import of this half-millimeter difference is that the effects of SRP alone on reduction in PD in these studies range from 0.71 mm to 2.30 mm for the same periods of time.

Figure 7. Meta-analysis of Local Minocycline and SRP (more...)

   Figure 7. Meta-analysis of Local Minocycline and SRP versus SRP Alone: Clinical Attachment Level

Of the eight studies examining the effect on CAL of locally applied minocycline reviewed above, we included five in the meta-analysis (Figure 7).^{32, 53, 65, 67, 68} Of the three excluded studies, two had no way for us to calculate CAL effect sizes,^{64, 66} and one had a test for heterogeneity of variances suggesting that that study represented a different intervention than that of the other studies.⁶³

The estimate of overall effect size for CAL is 0.46 mm (95 percent CI, 0.32 mm to 0.60 mm). Effect sizes of included studies favoring use of adjunctive local minocycline range from 0.04 mm to 0.80 mm. All these effect size estimates represent greater CAL gains using adjunctive local minocycline with SRP than using SRP alone, but based on the 95 percent CIs, the CAL effect sizes of only one of those studies differed significantly from zero.⁶⁵ That study had an appreciably smaller standard error of difference between means and a larger study sample and consequently contributed almost all of the weight to the meta-analysis results.

The statistically significant CAL result supports the added effectiveness of locally applied minocycline as an adjunct to SRP in the treatment of chronic periodontitis in adults. As with the PD results, which were of about the same magnitude, the clinical ramifications of a change of this size remains unclear in the context of CAL reductions in these studies for SRP alone of -0.13 mm to 1.61 mm.

Metronidazole

This section concerns the adjunctive use of metronidazole, an antiprotozoa agent with bacteriocidal effects on anaerobic species such as spirochetes, when used either systemically or locally in conjunction with SRP in patients with periodontitis. We present qualitative and descriptive results for both systemic and local metronidazole and a meta-analysis of eligible studies for local metronidazole. Metronidazole used in combination with amoxicillin is presented in the following section.

Systemic Metronidazole

Table 9. Main Features and Findings on Probing Depth (more...)

Table 9. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Systemic Metronidazole

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Loesche et al., 198469	R PL DB	30 weeks	N: T: C:	14 subjects 7 subjects 7 subjects	250 mg metronidazole 3xday for 7 days	Initial PD 4-6 mm: +0.14, NS Initial PD ≥ 7 mm: +1.64, P < 0.03	Initial PD 4-6 mm: +0.10, NS Initial PD ≥ 7 mm: +1.19, P = 0.05
Joyston-Bechal et al., 198671	R PL DB	156 weeks (3 years)	N: T: C:	28 subjects 15 subjects 13 subjects	1% chlorhexidine gel for first 10 weeks, then 200 mg metronidazole: 1 at evening of 3rd visit + 3xday for 5 days; repeated 4 weeks later	+0.41, NS	N/A
Soder et al., 199072	R PL DB	26 weeks (6 months)	N: T: C:	92 subjects 46 subjects 46 subjects	400 mg metronidazole 3xday for 7 days	+0.05, NR	N/A
Loesche et al., 199173	R PL DB	52 weeks	N: T: C:	39 subjects 18 subjects 21subjects	250 mg metronidazole 3xday for 7 days	Initial PD ≤ 3 mm: -0.07, NS Initial PD 4-6 mm: -0.06, NS Initial PD ≥ 7 mm: +0.41, NS	Initial PD ≤ 3 mm: +0.10, NS Initial PD 4-6 mm: +0.13, NS Initial PD ≥ 7 mm: +0.32, NS
Loesche et al., 199274	R PL DB	104 weeks (2 years)	N: T: C:	33 subjects 15 subjects 18 subjects	250 mg metronidazole 3xday for 7 days	Initial PD ≤ 3 mm: -0.15, P = 0.08 Initial PD 4-6 mm: +0.47, P < 0.01 Initial PD ≥ 7 mm: +1.05, P < 0.01	Initial PD ≤ 3 mm: +0.37, P = 0.07 Initial PD 4-6 mm: +0.47, P < 0.01 Initial PD ≥ 7 mm: +0.66, NS
Noyan et al., 199775	R NP NR	6 weeks	N: T: C:	10 subjects 5 subjects 5 subjects	250 mg metronidazole 3xday for 7 days, then SRP again	+0.60, NR	+0.41, NR
Palmer et al., 199876	R NP SB	24 weeks	N: T: C:	58 subjects 31 subjects 27 subjects	200 mg metronidazole 3xday for 7 days	-0.06, NS	+0.16, NS
Rooney et al., 200270	R PL DB	26 weeks	N: T: C:	31 subjects 16 subjects 15 subjects	200 mg metronidazole plus placebo 3xday for 7 days; plus 0.2% chlorhexidine irrigation	Initial PD ≥ 6 mm: Data NR, P < 0.05	Initial PD ≥ 6 mm: Data NR, NS

R, randomized; ; PL, placebo-controlled; NP, no placebo; DB, double-blind; SB, single blind; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

Qualitative and Descriptive Results. Eight studies, published between 1984 and 2002, used systemic metronidazole as the sole antibiotic in conjunction with SRP and otherwise met our inclusion criteria (Table 9 and Evidence Table 3a).^{69- 76} Six trials were randomized, placebo-controlled, and double-blind trials.^{69- 74} Two trials were randomized but not placebo-controlled,^{75, 76} one was blinded only for examiners,⁷⁶ and the other did not report blinding.⁷⁵

In all, these trials had 305 completers, of whom 153 were treatment subjects. Subjects had adult periodontal disease characterized as moderate to severe (or advanced). All control subjects received SRP with or without a placebo; experimental subjects received SRP with metronidazole alone.

The trials differed markedly in both therapeutic regimens and outcomes. As to dose, frequency, and duration of the antibiotic: three studies used 200 mg three times a day, one trial for 5 days⁷¹ and two for 7 days;^{70, 76} one team, in three trials, used 250 mg three times a day for 7 days;^{69, 73, 74} another group used the same dosage but repeated SRP;⁷⁵ finally, one study used 400 mg three times a day for 7 days.⁷² The studies lasted from 6 weeks to 156 weeks. The trial outcome measures also varied considerably: average mm PD reduction per subject or per site; average mm CAL gain per subject or per site; various microbial measures; the percentage of sites per patient gaining, losing, or not changing PD or CAL; the percentage without disease; the number of sites with a given level of disease; and the percentage of sites needing surgery.

With respect to PD differences, only one of the trials failed to report any PD data although it did report a significant difference between the groups favoring the metronidazole treatment.⁷⁰ Of the others, two studies from the same group reported statistically significant net PD reductions favoring the metronidazole group. For patients with initial PD levels of 4 mm to 6 mm, the earliest study reported 0.14 mm gain at 30 weeks (not significant) and the latest study, 0.47 mm at 104 weeks (P < 0.01); for subjects with initial PD values greater than 6 mm or 7 mm and higher, the figures were, respectively, 1.64 mm (P < 0.03) and 1.05 (P < 0.01).^{69, 74} Other studies reported net gains of 0.6 mm,⁷⁵ 0.41 mm,⁷¹ 0.41 mm for patients with initial PDs of 7 mm or higher,⁷³ and 0.05 mm,⁷² but either these values were not significant or significance was not reported. The 1991 Loesche et al. study reported a greater PD reduction of 0.06 mm in the control group at 52 weeks, though the results were not statistically significant.⁷³ One trial only presented percent change of original PD.⁷⁰

Six of the eight trials reported CAL results.^{69, 70, 73- 76} The studies from the Loesche et al. teams reported the following net gains favoring the metronidazole groups:^{69, 73, 74} for those with initial PDs of 4 mm to 6 mm, 0.47 (P < 0.01), 0.10 mm (not significant), and 0.13 (not significant), and for those with initial PDs of either greater than 6 mm or 7 mm and higher, 1.19 (P = 0.05), 0.32 (not significant) and 0.66 (not significant). Two other trials reported net gains of 0.41 mm (test not reported)⁷⁵ and 0.16 mm (not significant).⁷⁶ The sixth study did not report data on the gain in CAL but did indicate that it was not significant.⁷⁰

Five investigations reported on microbial results (specifically spirochetes).^{69, 72- 74, 76} For example, the proportions of spirochetes dropped from baseline to the final observation from 59.1 to 22.0 for the test group and from 60.0 to 34.8 for controls, with the difference between groups approaching significance (P = 0.06).⁷⁴ Similar findings were reported for the 1991 and 1984 studies from this research team.^{69, 73} Palmer et al. also reported reductions in percentages of spirochetes between baseline and 8-week and then 24-week follow-up for both experimental and control patients, but the 24-week findings (from 47.1 percent to 25.8 percent for the metronidazole group and from 47.2 percent to 25.6 percent in the SRP-only group) reflected no significant difference between the two study groups. Finally, Soder et al. reported that the total number of microorganisms counted at the follow-up visits did not differ significantly between the metronidazole and placebo groups.⁷²

Other results covered numerous heterogeneous outcomes. For example, at 6 weeks the percentage of teeth per patient needing surgery was lower for the treatment than the control group.⁷³ At 24 weeks, the percentages of sites with PDs of 4 mm or greater and the percentages of sites improved per patient were greater for experimentals than controls.⁷⁶ Two studies examining a total of 18 experimental subjects in terms of the percentage of deep sites suggested that metronidazole in conjunction with SRP is less effective than scaling alone at 156 weeks⁷¹ and 260 weeks.⁷²

Finally, three trials reported some adverse effects.^{69, 72, 74} They included severe diarrhea, gastric discomfort, and, less seriously, “metallic taste.”

Local Metronidazole

Table 10. Main Features and Findings on Probing Depth (more...)

Table 10. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Local Metronidazole

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Aziz-Gandour and Newman 198677	R PL DB	12 weeks	N: T: C:	16 subjects 210 surfaces 118 surfaces	0.05% metronidazole irrigation for 28 days	Data NR; P < 0.01	N/A
Moran et al., 199037	R NP DB	12 weeks	N: T: C:	33 pockets 15 pockets 18 pockets	Acrylic strips impregnated with metronidazole	+0.9, NS	+0.3, NS
Noyan et al., 199775	R NP NR	6 weeks	N: T: C:	10 subjects 5 subjects 5 subjects	25% metronidazole gel applied on days 0 and 7	+0.78, NS	+0.66, P < 0.01
Awartani and Zulqarnain 199828	R NP SB	14 weeks	N: T: C:	12 subjects 360 sites 378 sites	25% metronidazole gel applied on days 0 and 7	-0.12, NS	N/A
Lie et al., 199834	R NP DB	26 weeks (6 months)	N: T: C:	18 subjects 18 subjects 18 subjects	25% sustained release metronidazole gel after SRP sessions on days 0 and 7	Defect sites:* +0.5, NS Nondefect sites:* -0.1, NS	Defect sites:* +0.7, NS Nondefect sites:* 0.0, NS
Palmer et al, 199876	R NP SB	24 weeks	N: T: C:	53 subjects 26 subjects 27 subjects	25% metronidazole gel subgingival application on days 0 and 7	Data NR, NS	-0.04, NS
Kinane and Radvar, 199953	R NP SB	26 weeks (6 months)	N: T: C:	36 subjects 19 subjects 20 subjects	25% metronidazole gel 2xday applied on day 0 and 7	+0.22, NS	+0.004, NS
Riep et al., 199978	R NP SB	13 weeks (3 months)	N: T: C:	29 subjects 29 subjects 29 subjects	25% metronidazole gel applied 5 times over 10 days	0.0, NS	+0.17, NS
Al Mubarak et al., 200080	R NP SB	13 weeks (90 days)	N: T: C:	14 subjects 14 subjects 14 subjects	25% metronidazole gel applied on days 0 and 7	+0.8, P < 0.03	N/A
Griffiths et al., 200079	R NP SB	39 weeks (9 months)	N: T: C:	88 subjects 1,770 sites 1,780 sites	25% metronidazole gel 1xweek for 3 weeks	+0.5, P < 0.001	+0.4, P < 0.001
Stelzel and Flores-de-Jacoby, 200036	R NP SB	37 weeks (9 months)	N: T: C:	59 subjects 59 subjects 59 subjects	25% metronidazole gel applied 2xday on days 0 and 7	+0.18, P < 0.05	+0.07, NS

R, randomized; ; PL, placebo-controlled; NP, no placebo; DB, double-blind; SB, single blind; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

*Defect and nondefect sites refer to, respectively, the mean of measurements from the buccal and lingual side versus measurements for all remaining parts of the tooth.

Qualitative and Descriptive Results. In all, we identified 11 clinical trials appearing between 1986 and 2000 that used local metronidazole in conjunction with SRP and that met the inclusion criteria (Table 10 and Evidence Table 3b).^{28, 34, 36, 37, 53, 75- 80} One trial was a randomized, placebo-controlled, double-blind design;⁷⁷ most of the studies were randomized, not placebo-controlled, and single-blind designs.^{28, 36, 53, 76, 78- 80} The studies lasted over a range of 6 weeks to 39 weeks (9 months). Overall, the completing number of subjects (treatment and control groups) was 368; units of analyses included subjects, sites, and surfaces.

All subjects had adult or chronic periodontal disease ranging from “mild to moderate” to “moderate to severe,” but most authors did not comment on severity. Most research teams used a 25 percent gel of metronidazole as the intervention therapy, each with reapplications after 1 week or more often.^{28, 34, 36, 53, 75, 76, 78- 80} Other groups used a variety of dosages and modes, such as 0.05 percent solution used with jet irrigation subgingivally⁷⁷ and 20 percent ethylcellulose film.³⁷ Control subjects or sites received SRP with or without a placebo; the experimental subjects received SRP with metronidazole alone.

For the PD outcomes, all 11 studies gave some results. Three reported data showing a statistically significant difference in the net PD reduction that favored the treatment group: 0.8 mm at 13 weeks (P < 0.03),⁸⁰ 0.5 mm at 39 weeks (P < 0.001),⁷⁹ and 0.18 mm at 37 weeks (P < 0.05).³⁶ One additional study reported that the net difference between treatment and control groups was significant at 12 weeks (P < 0.01) but did not report the difference.⁷⁷ Four trials reported net PD reductions in favor of metronidazole that were not statistically significant: 0.9 mm,³⁷ 0.78 mm,⁷⁵ 0.22 mm,⁵³ for “defect sites” at 26 weeks,³⁴ and one study did not report the data.⁷⁶ Two studies reported nonsignificant net PD reductions favoring the control groups ─ 0.12 mm²⁸ and 0.1 mm for “nondefect sites”³⁴ ─ and one reported no net difference at all.⁷⁸

Eight of these 11 studies provided information on CAL gains. Only two studies reported significant net gains for the treatment groups: 0.66 mm at 6 weeks (P < 0.01)⁷⁵ and 0.4 mm at 39 weeks (P < 0.001).⁷⁹ Several other studies reported nonsignificant CAL gains favoring local metronidazole groups: 0.17 mm,⁷⁸ 0.07 mm,³⁶ 0.03 mm, ³⁷ 0.004 mm,⁵³ and 0.7 mm and 0.0 mm for defect and nondefect sites.³⁴ One study reported a nonsignificant greater net gain in the attachment level for the control group of 0.04 mm.⁷⁶

As to changes in the presence of spirochetes, one trial showed that treatment lessened the percentage of spirochetes in both treatment and control groups. At 8 weeks of follow-up, the change was significantly less in the locally delivered metronidazole group, but at 24 weeks the differences were not significant.⁷⁶

Generally, none of these studies reported adverse events. One group noted that about half of the patients receiving metronidazole as a gel reported a bitter taste.³⁶

Figure 8. Meta-analysis of Local Metronidazole and (more...)

   Figure 8. Meta-analysis of Local Metronidazole and SRP versus SRP Alone: Probing Depth

Quantitative Analysis of Local Metronidazole Results. Seven of the 11 studies examining the effect on PD of locally applied metronidazole reviewed above were included in the meta-analysis (Figure 8).^{34, 36, 37, 53, 78- 80} Two studies were not included because they lacked data allowing us to calculate PD effect size measures,^{76, 77}one was too short,⁷⁵ and the final one did not provide data on variation.²⁸

The estimate of overall effect size for PD is 0.32 mm (95 percent CI, 0.20 mm to 0.44 mm). Effect sizes of included studies favoring use of adjunctive local metronidazole range from 0.18 mm to 0.90 mm. All but one of the study effect sizes represented results that demonstrate greater PD reduction using adjunctive local metronidazole with SRP than using SRP alone. The one exception found a zero difference between the treatment and control groups.⁷⁸ Based on the 95 percent CIs, the PD effect sizes of four studies differed significantly from zero.^{37, 53, 79, 80} The two studies with appreciably smaller standard errors of difference between means contributed relatively more weight to the meta-analysis results.^{53, 79}

The statistically significant PD result supports the added effectiveness of locally applied metronidazole as an adjunct to SRP in the treatment of chronic periodontitis in adults. It does not address, however, whether a mean change of 0.32 mm is clinically meaningful. In contrast to this approximately one-third of a millimeter difference between using and not using some form of local metronidazole as an adjunct to SRP, the effects of SRP alone on reduction in PD in these studies ranged from 0.71 mm to 2.50 mm for the same periods of time.

Figure 9. Meta-analysis of Local Metronidazole and (more...)

   Figure 9. Meta-analysis of Local Metronidazole and SRP versus SRP Alone: Clinical Attachment Level

Of the eight studies examining CAL effects of local metronidazole, we used seven in the meta-analysis (Figure 9).^{34, 36, 37, 53, 76, 78, 79} The excluded study was too short.⁷⁵

The estimate of overall effect size for CAL is 0.12 mm (95 percent CI, 0.01 mm to 0.24 mm). Effect sizes of included studies favoring use of adjunctive local metronidazole range from 0.07 mm to 0.70 mm. Five of the study effect sizes suggest greater gain in CAL using adjunctive local metronidazole with SRP than using SRP alone. One of the exceptions found a zero difference between the treatment and control groups;⁵³ the other was an effect size favoring SRP alone.⁷⁶ Based on the 95 percent CIs, the CAL effect sizes of 0.4 mm and 0.7 mm in two studies differed significantly from zero.^{34, 79} The two studies with appreciably smaller standard errors of the difference between means again contributed relatively more weight to the meta-analysis results.^{53, 79}

As with PD effects of locally applied metronidazole, the statistically significant CAL result supports its added effectiveness as an adjunct to SRP, but the clinical importance of a mean change of 0.12 mm is open to question. By contrast, the effects of SRP alone on CAL gains in the same time periods ranged from 0.20 mm to 1.60 mm.

Metronidazole and Amoxicillin

Table 11. Main Features and Findings on Probing Depth (more...)

Table 11. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Systemic Metronidazole plus Amoxicillin

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Berglundh et al., 199881	R PL NR	104 weeks	N: T: C:	16 subjects 8 subjects 8 subjects	250 mg metronidazole 3xday plus 375 mg amoxicillin 2xday for 2 weeks	+0.5, NR	+0.3, NR
Flemmig, Milian, et al., 199882	R NP SB	52 weeks	N: T: C:	38 subjects 18 subjects 20 subjects	250 mg metronidazole plus 375 mg amoxicillin 3xday plus 0.06% chlorhexidine irrigation 1xday for 8 days	Data NR, NS	Data NR, NS
Winkel et al., 200183	R PL DB	24 weeks	N: T: C:	49 subjects 23 subjects 26 subjects	250 mg metronidazole plus 375 mg amoxicillin 3xday for 7 days	+0.7, P < 0.05	+0.4, NS
Rooney et al., 200270	R PL DB	26 weeks	N: T: C:	30 subjects 15 subjects 15 subjects	200 mg metronidazole plus 250 mg amoxicillin 3xday for 7 days; plus 0.2% chlorhexidine irrigation	Initial PD ≥ 6 mm: Data NR, P < 0.001	Initial PD ≥ 6 mm: Data NR, P < 0.05

R, randomized; PL, placebo-controlled; NP, no placebo; DB, double-blind; SB, single blind; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

The combined antibiotic regimens differed in dosage, frequency, and duration. Three studies used the combination of 250 mg of metronidazole and 375 mg amoxicillin three times a day for 7 days,⁸³ 8 days,⁸² and two times a day for 2 weeks;⁸¹ one used 200 mg metronidazole and 250 mg amoxicillin three times a day for 7 days.⁷⁰

Two trials gave data on mean PD outcomes. The net PD reduction favoring the treatment groups were 0.7 mm at 13 weeks (P < 0.05)⁸³ and 0.5 mm at 104 weeks (significance not reported).⁸¹ In the remaining trials, data were not reported, although one noted that the net difference for initial probing depths of 6 mm or greater was significant favoring the treatment (P ≤ 0.001).⁷⁰

Similarly, two studies reported data on CAL gains: a net gain of 0.3 mm at 104 weeks where significance was not reported⁸¹ and one of 0.4 mm at 24 weeks (not significant).⁸³ Of the two studies not giving CAL data, one said the net difference was significant (P ≤ 0.05) for probing depths greater than 6 mm⁷⁰ and one said it was not significant.⁸²

None of the studies examined spirochetes, but all four reported on other microbiological outcomes.^{70, 81- 83} One team found that the treatment eliminated A. actinomycetemcomitans, P. gingivalis and P. intermedia at 8 weeks (significance not reported).⁸¹ For subjects with A. actinomycetemcomitans microbes, another team found that a significantly higher incidence of CAL gain of 2 mm or more was achieved in the experimental group over control (P < 0.05), but the opposite was true for subjects with P. gingivalis, who had a loss of attachment (P < 0.05) at 52 weeks.⁸² A third team reported significant differences for the treatment versus control for microbiological outcomes (A. actinomycetemcomitans, P. gingivalis, P. intermedia and others) only at 1 month and not for the final assessment at 6 months.⁷⁰ Finally, the remaining team found significant differences between the experimental and placebo groups in the decreased number of subjects who were positive for other microbiological outcomes (P value not reported).⁸³

Some of these trials also reported positive findings for other measures. In one study, for example, at 12 weeks the percentage of deep sites per patient and mean deep and shallow site attachment levels were all improved for the treatment group compared to controls.⁸³ In another, at 26 weeks the percentage of sites for those with initial PD 6 mm or greater dropped from 15.9 percent to 1.3 percent for the treatment group; figures for the full placebo-control group were 19.3 percent and 12.4 percent, for a net difference favoring the dual antibiotic group of 7.7 percentage points (reported as statistically significant);⁷⁰ similar findings were reported for sites with 3 mm or less PD initially. CAL findings indicated that patients in the combination antibiotic group had significant improvements—i.e., decreased percentage of sites with high attachment losses and increased percentage of sites with low attachment losses—compared with levels for the placebo group. Finally, some research teams reported greater elimination or suppression of some periodontal pathogens (e.g., A. actinomycetemcomitans).^{81, 82}

Various adverse effects were reported in these trials. They fell mainly into the category of gastrointestinal problems (e.g., diarrhea),^{82, 83} with two cases of skin rash and one case of nausea after alcohol use.⁸³

Chlorhexidine

Table 12. Main Features and Findings on Probing Depth (more...)

Table 12. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Local Chlorhexidine

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Braatz et al., 198530	R PL NR	24 weeks	N: T: C:	14 subjects 54 sites 52 sites	Irrigation with 2% chlorhexidine solution daily for 24 weeks	+0.3, NS	+0.2, NS
MacAlpine 198531	R PL NR	24 weeks	N: T: C:	11 subjects 16 sites 16 sites	Irrigation with 2% chlorhexidine solution every 2 weeks for 24 weeks	+1.0, NS	+0.9, NS
Aziz-Gandour and Newman, 198677	R PL DB	12 weeks	N: T: C:	12 subjects 147 surfaces 118 surfaces	Irrigation with 0.02% chlorhexidine solution 1xday for 28 days	Data NR, NS	N/A
Watts and Newman, 198684	R PL DB	12 weeks	N: T: C:	11 subjects 128 sites 134 sites	Irrigation with 0.02% chlorhexidine solution 1xday for 28 days	Data NR, NS	N/A
Wennstrom et al., 198785	R PL SB	52 weeks	N: T: C:	10 subjects 28 subjects 24 subjects	Irrigation with 0.2% chlorhexidine solution 3xweek	+0.1, NR	-0.1, NS
Southard et al., 198986	R NP DB	15 weeks	N: T: C:	8 subjects 8 subjects 8 subjects	Irrigation with 2% chlorhexidine solution at day 0 then 1xweek for 3 weeks	-0.1, NS	+0.1, NS
Taggart et al., 199087	NR PL NR	10 weeks	N: T: C:	10 subjects 10 subjects 10 subjects	Irrigation with 0.02% chlorhexidine solution	+0.1, NS	+0.1, NS
Reynolds et al., 199288	R PL DB	4 weeks (28 days)	N: T: C:	60 subjects 90 sites 90 sites	Irrigation with 0.12% chlorhexidine solution, 1 time	Data NR, NR	N/A
Shiloah and Patters, 199454	R PL SB	4 weeks	N: T: C:	7 subjects 12 sites 12 sites	Irrigation with 0.12% chlorhexidine solution, 1 time	Data NR, NS	Data NR, NS
Oosterwaal et al., 199189	R PL DB	36 weeks	N: T: C:	10 subjects 10 subjects 10 subjects	0.2% chlorhexidine gel applied 3x within 10 minutes after SRP	Data NR, NS	N/A
Unsal et al., 199459	R NP NR	12 weeks	N: T: C:	15 subjects 7 subjects 8 subjects	1% chlorhexidine gel applied once	-0.25, NS	-0.34, NS
Soskolne et al., 199790	R NP DB	26 weeks (6 months)	N: T: C:	94 subjects 94 subjects 94 subjects	2.5 mg chlorhexidine chip inserted into pockets 5-8 mm at day 0 and 3 months	-0.46, P < 0.001	+0.16, P < 0.05
Jeffcoat et al., 199891	R PL DB	39 weeks (9 months)	N: T: C:	419 subjects 211 subjects 208 subjects	2.5 mg chlorhexidine chip inserted into pockets at baseline and months 3 and 6	+0.26, P < 0.00056	+0.20, P < 0.012
Heasman et al., 200192	R NP SB	26 weeks (6 months)	N: T: C:	24 subjects 24 subjects 24 subjects	2.5 mg controlled release chlorhexidine chip	+0.33, P = 0.05	+0.28, P = 0.048
Azmak et al., 200293	R NP SB	26 weeks (6 months)	N: T: C:	20 subjects 20 subjects 20 subjects	2.5 mg chlorhexidine chip in pockets	Data NR, NS	Data NR, NS
Grisi et al., 200294	R NP SB	39 weeks (9 months)	N: T: C:	19 subjects 10 subjects 9 subjects	2.5 mg chlorhexidine chip at day 0, 3 and 6 months	-0.2, NS	-0.4, NS
Quirynen et al., 200095	Non-R NP NR	35 weeks (8 months)	N: T: C:	24 subjects 12 subjects 12 subjects	Combination of 1% chlorhexidine gel for brushing and subgingival irrigation + 0.2% chlorhexidine rinse + spray within 24 hours, then 0.2% rinse and spray 2xday for 60 days	Single-root: -0.1, NS Multi-root: -0.5, NS	Initial PD ≥7 mm: Single-root: -0.3, NS Multi-root: -0.3, NS Initial PD < 7 mm Data NR, NR

R, randomized; Non-R, nonrandomized; PL, placebo-controlled; NP, no placebo; DB, double-blind; SB, single blind; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

Below we discuss the nine studies, published between 1985 and 1994, that included chlorhexidine applied solely as a rinse or through irrigation.^{30, 31, 54, 77, 84- 88} Those are followed by seven studies involving direct applications via chips or gels (without additional chlorhexidine rinse or irrigation)^{59, 89- 94}and then by one trial that involved complex one-stage “full-mouth disinfection” within 24 hours, comprising gel, rinse, spray, and/or irrigation in combination.⁹⁵

Of these 17 studies, 15 are described as randomized and 10 as placebo-controlled; seven are double-blind; an additional five are single-blind studies (chiefly of examiners). All patients had periodontitis described variously as mild, moderate, severe, or advanced (or severity was not reported). In all, the total number of patients completing these studies was 767, but many analyses were done on sites, surfaces, or pockets.

Two trials were large and multi-site investigations (one of 419 subjects [211 treatment and 208 control]⁹¹ and one of 94 subjects [401 treatment pockets and 412 control pockets]⁹⁰); one was medium-sized (60 subjects);⁸⁸ and the remainder were small (no more than 24 completers, but most were in the 10-subject range).

The approaches to experimental treatments varied considerably in terms of timing, frequency, and relationship to SRP; some trials included Bass brushing as part of the test approach. However, dosage of irrigated chlorhexidine was variable (ranging from 0.02 percent to 2.0 percent); some trials using chips identified it as Perio-Chip® (2.5 mg chlorhexidine); and gels tended to be either 0.2 percent or 1 percent chlorhexidine. Control subjects tended to receive SRP alone or with some form of saline or water irrigation.

Similarly, the trials measured a wide array of clinical and microbiological outcomes at quite heterogeneous follow-up periods. Of the trials that reported on PD changes, some reported data in terms of the percentage of sites with smaller depths after treatment, rather than actual depths in millimeters; few studies reported on CAL in millimeters; and virtually no studies reported on spirochetes. Only a small number of research teams looked for or commented on adverse events, but as chlorhexidine is considered easily tolerated, this may not be surprising.

Direct Gingival Applications

Chlorhexidine Chips. Five trials appearing between 1997 and 2002 tested a chip application of chlorhexidine as an adjunct to SRP against either SRP alone or SRP with a placebo chip (Table 12).^{90- 94} These were 2.5 mg chips inserted into pockets of varying depths at baseline and at various points after that; the trials lasted from 26 to 39 weeks.

With respect to PD, two trials reported statistically significant net improvements for the treatment group: 0.33 mm at 26 weeks (P = 0.05)⁹² and 0.26 mm at 39 weeks (P < 0.0056).⁹¹ Another study cited a 0.2 mm improvement in favor of the control group but did not report significance,⁹⁴ and a fourth gave no data but reported a nonsignificant difference.⁹³ Finally, the fifth study reported a 0.46 mm net PD reduction favoring the SRP-only group (P < 0.001).⁹⁰

Three of the chip trials cited net CAL gains favoring the chlorhexidine chip groups: 0.28 mm at 26 weeks (P = 0.048),⁹² 0.20 mm at 39 weeks (P < 0.012),⁹¹ and 0.16 mm at 26 weeks (P < 0.05).⁹⁰ Of the other two studies, one did not report all final data in millimeters but commented that the changes were not significant⁹³ and the other showed only a nonsignificant 0.4 mm net PD reduction favoring the control group.⁹⁴

No chip trial reported on percentages of spirochetes. One research team reported toothache, upper respiratory tract infection, and headache as adverse events but noted that such side effects generally happened with similar frequency for the treatment and control groups except for toothache (e.g., pain, tenderness, and sensitivity similar), which was significantly higher in the chlorhexidine group (P = 0.042).⁹¹

Chlorhexidine Gels. Two trials used chlorhexidine gel as an adjunct to SRP and tested this modality against SRP alone or SRP with placebo gel (Table 12).^{59, 89} One study did not give tabulated data on PD reductions but the net difference was noted as not significant;⁸⁹ in the other, the 0.25 mm net reduction (not significant) favored the control group.⁵⁹ One trial noted only a nonsignificant 0.34 mm CAL gain favoring the control group.⁵⁹ The one trial that addressed percentages of spirochetes showed a decrease over time for the treatment group, but no significant net benefit emerged for the treatment group at the conclusion of the study.⁸⁹ Neither study reported on adverse effects.

Chlorhexidine Combination Treatment. The one study that employed an all within-24 hours full mouth disinfection also employed multiple forms of chlorhexidine.⁹⁵ These included chlorhexidine gel for brushing, subgingival chlorhexidine irrigation, and chlorhexidine rinse and spray, the latter twice a day for 60 days. In this study, net PD reductions favoring the control group of 0.1 mm (single-root teeth) and 0.5 mm (multi-root teeth) were not significant. The same was true for CAL gains of 0.3 mm (single-root teeth) and 0.3 mm (multi-root teeth) that favored the control group.

Figure 10. Meta-analysis of Local Chlorhexidine and (more...)

   Figure 10. Meta-analysis of Local Chlorhexidine and SRP versus SRP Alone: Probing Depth

Quantitative Analysis of Local Chlorhexidine Effects. Of the 17 studies examining the effect on PD of locally applied chlorhexidine reviewed above, we retained eight in the meta-analysis (Figure 10).^{30, 31, 59, 90- 92, 94, 95} Of the nine excluded studies, one had a study period of less than 3 months,⁵⁴ two were longer than 9 months with no intermediate data reported,^{86, 87} and six did not report data that could be used to calculate a PD effect size.^{77, 84, 86, 88, 89, 93}

The estimate of overall effect size for PD is 0.24 mm (95 percent CI, 0.13 mm to 0.35 mm). Effect sizes favoring use of adjunctive local chlorhexidine range from 0.14 to 1.00 mm. Five of these effect sizes reflected greater PD reduction using adjunctive local chlorhexidine with SRP than using SRP alone; the other three studies had PD effect sizes of from -0.10 mm to -0.25 mm that favored SRP alone.^{59, 94, 95} Based on the 95 percent CIs, the PD effect sizes of only three studies differed significantly from zero; all favored SRP with local chlorhexidine.^{31, 90, 91} The two studies with appreciably smaller standard errors of difference between means contributed considerably more weight to the meta-analysis results.^{90, 91}

The statistically significant PD result supports the added effectiveness of locally applied chlorhexidine as an adjunct to SRP in the treatment of chronic periodontitis in adults. As with other adjunctive therapies with a mean changes in this range (0.24 mm), however, whether they are consequential clinically remains debatable, given that the effects of SRP alone on reduction in PD in these chlorhexidine studies ranged from 0.70 mm to 3.00 mm for the same periods of time.

Figure 11. Meta-analysis of Local Chlorhexidine and (more...)

   Figure 11. Meta-analysis of Local Chlorhexidine and SRP versus SRP Alone: Clinical Attachment Level

Of the 13 studies examining the CAL effects of locally applied chlorhexidine reviewed above, we included seven in the meta-analysis (Figure 11).^{30, 31, 59, 90- 92, 94} Among the six excluded studies, one had a study period of less than 3 months,⁵⁴ two had study periods longer than 9 months with no intermediate data reported,^{85, 87} and three did not report data that could be used to calculate a CAL effect size.^{86, 93, 95}

The estimate of overall effect size for gain in CAL is 0.16 mm (95 percent CI, 0.04 mm to 0.28 mm). Effect sizes of included studies favoring use of adjunctive local chlorhexidine range from 0.16 mm to 0.90 mm. Two of the study effect sizes, ranging from 0.34 to 0.40, represented results that demonstrate greater CAL gains using SRP alone.^{59, 94} Based on the 95 percent CIs, the CAL effect sizes differed significantly from zero for only two studies.^{31, 91} The two studies with appreciably smaller standard errors of differences between means and larger sample sizes contributed relatively more weight to the meta-analysis results.^{90, 91}

The statistically significant CAL result supports the added effectiveness of locally applied chlorhexidine as an adjunct to SRP, but the clinical significance of a mean change of 0.16 mm is also debatable. The effects of SRP alone on CAL gains in these chlorhexidine studies ranged from 0.31 mm to 1.40 mm for the same periods of time.

Other Antibiotics

Systemic Antibiotics

Table 13. Main Features and Findings on Probing Depth (more...)

Table 13. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Other Systemic Antibiotics

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Chin Quee et al., 198796	R PL DB	26 weeks (6 months)	N: T: C:	50 subjects 26 subjects 24 subjects	3 tablets (750,000 IU of spiramycin and 125 mg of metronidazole), 2xday for 14 days	Data NR, NS	Data NR, P < 0.05
Al-Joburi et al., 198945	R PL DB	24 weeks	N: T: C:	52 subjects 28 subjects 24 subjects	500 mg spiramycin 2xday for 14 days	Initial PD ≤ 3 mm: +0.42, NS Initial PD 4-6 mm: -0.40, NS Initial PD ≥ 7 mm: -0.28, NS	Initial PD ≤ 3 mm: +0.92, NS Initial PD 4-6 mm: -0.22, NS Initial PD ≥ 7 mm: -0.08, NS
Bain et al., 199497	R PL DB	24 weeks	N: T: C:	189 subjects 93 subjects 96 subjects	1,500,000 IU of spiramycin “500” capsules, 2xday for 14 days	+0.47, P < 0.0075	+0.29, NS
Haffajee et al., 199546	R PL DB	43 weeks (10 months)	N: T: C:	21 subjects 10 subjects 11 subjects	250 mg amoxicillin with clavulanic acid 3xday for 30 days	+0.29, NS	Data NR, NS
Ng and Bissada, 199898	R PL SB	24 weeks	N: T: C:	16 subjects 8 subjects 8 subjects	200 mg doxycycline on day 1 then 100 mg 1xday for 6 weeks	-0.6, NS	+1.3, P ≤ 0.05
Rooney et al., 200270	R PL DB	26 weeks (6 months)	N: T: C:	31 subjects 16 subjects 15 subjects	250 mg amoxicillin and placebo (calcium lactate tablets 3xday for 7 days, plus 0.2% chlorhexidine rinse	Initial PD ≥ 6 mm: Data NR, P < 0.05	Initial PD ≥ 6 mm: Data NR, NR
Smith et al., 200299	R PL DB	22 weeks	N: T: C:	44 subjects 23 subjects 21 subjects	500 mg azithromycin 1xday for 3 days at week 2	Initial PD 1-3 mm: +0.14, NS Initial PD 4-5 mm: +0.52, P < 0.01 Initial PD ≥ 6 mm: +0.87, P < 0.05	N/A

R, randomized; PL, placebo-controlled; DB, double-blind; SB, single blind; ; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

Seven clinical trials of other antibiotics, all given as systemic agents, met the inclusion criteria (Table 13 and Evidence Table 6a).^{45, 46, 70, 96- 99} Two studies used spiramycin alone as the adjunctive therapy.^{45, 97} The others used different agents for the adjunctive systemic therapies: a combination of spiramycin and metronidazole,⁹⁶ doxycycline (a synthetic derivative of tetracycline),⁹⁸ azithromycin (an antibiotic related to erythromycin),⁹⁹ amoxicillin with clavulanic acid,⁴⁶ and amoxicillin with a placebo and chlorhexidine rinse.⁷⁰ All were randomized, placebo-controlled trials; all but one was a double-blind trial.⁹⁸

Taken together, the trials comprised 359 completing subjects, 204 on antibiotic regimens with SRP (just under half from the Bain et al. trial⁹⁷), and 199 receiving SRP with placebo. Subjects in the trials differed in terms of their periodontal disease experience or severity from severe or advanced periodontitis to moderate. One study included subjects with at least two sextants with scores on the Community Periodontal Index of Treatment Needs (CPITN) of 4.⁹⁹ All studies used the subject as the unit of analyses.

All studies reported details of the SRP procedures. In two studies, SRP was started at baseline appointment and completed in two sessions of 3 hours each, performed within 1 week of each other by one operator.^{45, 96} In another, SRP was started at baseline and performed in 3 to 5 hours during the 2 weeks of drug therapy; this involved seven operators in different centers.⁹⁷ In another study, 152 teeth received one session of SRP performed under local anesthesia using ultrasonic and hand instruments.⁹⁸ Finally, in the most recent study, SRP was performed by one hygienist using hand instruments.⁹⁹

Experimental regimens differed widely across these trials. Two studies had the standard dose of 500 mg of spiramycin, twice a day for 14 days;^{45, 97} another had a combination dose of 750,000 IU of spiramycin and 125 mg of metronidazole twice a day for 14 days.⁹⁶ Other regimens included: 200 mg of doxycycline the first day followed by 100 mg per day for 6 weeks;⁹⁸ 500 mg azithromycin capsules once a day for 3 days at week 2 after the final session of SRP;⁹⁹ 250 mg amoxicillin with clavulanic acid three times per day for 30 days⁴⁶ and the other had 250 mg amoxicillin with a placebo three times a day for 7 days along with a chlorhexidine rinse.⁷⁰ Studies ranged in length from 22 to 43 weeks.

Only five studies reported PD data.^{45, 46, 97- 99} While the remaining two studies did not report data, they did report that the difference between the groups was not significant.^{70, 96} In the doxycycline trial, the experimental group was significantly different from the placebo group only at weeks 3 and 6; at 12 and 24 weeks, the two groups did not differ significantly, with the 0.6 mm nonsignificant difference at 24 weeks favoring the placebo group.⁹⁸ One spiramycin study found significant differences in PD reduction favoring the drug group at 2 weeks into the study; at 24 weeks, the net reduction favoring the group receiving spiramycin was 0.47 mm (P = 0.0075).⁹⁷ Another spiramycin study showed greater reductions in PD favoring the experimental group in only the least severe class (probing PD ≤ 3 mm) of 0.42 mm at 24 weeks but the difference was not significant. For the other severity groups, the results favored the placebo groups—0.40 mm in those with PD 4 mm to 6 mm and 0.28 mm in those with PD ≥ 7 mm—and neither was significant.⁴⁵ In the azithromycin study, the mean values of subjects’ average PDs differed between the azithromycin and the SRP-only groups at baseline; thus, the investigators used analysis of covariance to render the experimental and control groups equivalent at baseline. The study results showed greater reductions in PD favoring the experimental group in each of three severity classes (probing depths of 1 mm to 3 mm, 4 mm and 5 mm, and 6 mm or greater) at various measurement points. At 22 weeks, the net reductions in PD for patients with the most advanced periodontitis was 0.87 mm (P < 0.05) and for these in the intermediate group, 0.52 mm (P < 0.01).⁹⁹

Six trials measured CAL.^{45, 46, 70, 96- 98} The doxycycline trial showed a 1.3 mm gain in CAL at 24 weeks (P ≤ 0.05).⁹⁸ The three spiramycin studies had mixed results. In one (used with metronidazole), the experimental group exhibited a significantly (P < 0.05) greater gain in attachment level than did the control group but provided no data;⁹⁶ another reported a nonsignificant 0.29 mm gain in CAL for the experimental group.⁹⁷ In the third, at 24 weeks a gain of 0.92 mm favoring the treatment group was not significant (for those with least severity), and results favored control groups (although also not significant) for those with greater severity.⁴⁵ The amoxicillin and clavulanic acid trial reported only a nonsignificant difference at the end of the trial,⁴⁶ and in the other amoxicillin trial no data or significance test were given.⁷⁰

Two trials performed microbiological examinations on the studies’ subjects and found a significant decline in the spirochetes level. In one, the proportion of spirochetes among experimental subjects receiving spiramycin decreased from 28 percent at baseline to 3 percent at 24 weeks; the respective values for the placebo group were 30 percent and 11 percent (P < 0.05).⁴⁵ In the other, the proportion of spirochetes declined significantly among experimentals relative to controls at 14 days and thereafter; at 6 months, the proportions were 3 percent for experimental subjects and 15 percent for controls (P < 0.05).⁹⁶

Most of these studies did not report adverse events or effects of the experimental intervention (or reported that they did not observe any). Adverse reactions mentioned, which occurred infrequently, included nausea, diarrhea, gastrointestinal upset, and abdominal pain.

Local Antibiotics

Table 14. Main Features and Findings on Probing Depth (more...)

Table 14. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Other Local Antibiotics

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Kimura et al., 1991100	R PL NR	4 weeks	N: T: C:	27 subjects 27 subjects 27 subjects	Controlled-release ofloxacin insert applied 1x week for 2 weeks, then SRP again and inserts applied 1xweek for next 3 weeks	N/A	N/A
Eickholz et al., 2002101	R PL DB	26 weeks (6 months)	N: T: C:	108 subjects 108 subjects 108 subjects	15% doxycycline gel, 1 subgingival application	+0.44, P = 0.0066	+0.37, P = 0.038

R, randomized; ; PL, placebo-controlled; ; DB, double-blind; NR, not reported; N/A, not applicable because not one of the clinical measures; NS, not significant

Two clinical trials of local antibiotics as the adjunctive therapy met the inclusion criteria (Table 14; Evidence Table 6b). One trial used local ofloxacin (a broad-spectrum antibiotic);¹⁰⁰ the other used doxycycline gel.¹⁰¹ Both were randomized, placebo-controlled studies; one was a multi-center, double-blind trial.¹⁰¹

In all, 135 subjects completed the two trials. All subjects in both studies received SRP, and each subject had one site treated with the local antibiotic regimens with SRP and a different site treated with placebo.

These two trials differed in several dimensions. With respect to periodontal disease experience, one study included subjects with chronic periodontitis diagnosed by showing on affected teeth more than 30 percent bone loss by radiographs.¹⁰⁰ The other study included subjects suffering from moderate to severe periodontitis with at least 3 single-rooted teeth either with PD depth of 5 mm and bleeding on probing or with PD of 6 mm or more.¹⁰¹

Details of the experimental regimens and SRP procedures regimens also differed. In the doxycycline trial, investigators used a subgingival application of a newly developed, biodegradable 15-percent doxycycline gel (DOXI); subgingival SRP was performed with hand instruments at all test and respective neighboring teeth under local anesthesia, and then mechanical debridement of one test tooth was limited to 10 minutes.¹⁰¹ In the other study, investigators used a controlled-release, film-shaped insert of ofloxacin that was applied once a week from baseline through 35 days; supragingival SRP was done for the first 2 weeks and root planning and subgingival scaling for 4 weeks after that.¹⁰⁰

Only the doxycycline trial reported on clinical measures.¹⁰¹ At 26 weeks both net PD reduction (0.44 mm, P = 0.0066) and CAL gain (0.37 mm, P = 0.038) favored the experimental group over the control group.¹⁰¹ Interim measures were said to show improvement compared to baseline but statistical significance was not reported. The ofloxacin trial analyzed the proportion of spirochetes over a 4-week period.¹⁰⁰ Two weeks after supragingival scaling period, the difference between the experimental and the placebo groups favored the former (9.5 percent versus 20.2 percent, P < 0.05); however, as the SRP treatment changed to mechanical subgingival debridement, proportions of spirochetes dropped for all groups, and differences between treatment and placebo groups were no longer statistically significant through the remainder of the study. Only the doxycycline study reported any adverse effects (swelling at one treatment tooth).

Other Antimicrobials

Table 15. Main Features and Findings on Probing Depth (more...)

Table 15. Main Features and Findings on Probing Depth and Clinical Attachment Level: Trials of Other Antimicrobials

Author, Year	Randomization Placebo Blinded	Duration	Number Completing Treatment Control		Dosage and Mode	Treatment Group PD Reduction (mm) Level of Significance	Treatment Group CAL Gain (mm) Level of Significance
Wennstrom et al., 198785	R PL SB	52 weeks	N: T: C:	10 subjects 10 subjects 10 subjects	3% hydrogen peroxide irrigation 3xweek for 2 weeks	+0.8, P < 0.05	+0.1, NS
Listgarten et al., 1989103	R PL DB	8 weeks	N: T: C:	40 subjects 20 subjects 20 subjects	7% tetrapotassium peroxydiphosphate irrigation 2xday for 8 weeks	+0.65, NS	+0.25, NS
Oosterwaal et al., 199189	R PL DB	36 weeks	N: T: C:	10 subjects 10 subjects 10 subjects	T1:1.25% amine fluoride gel 3x within 10 minutes T2: 4% stannous fluoride gel 3x within 10 minutes	T1:Data NR, NS T2: Data NR, NS	T1: N/A T2: N/A
Furuichi et al., 1997102	R PL DB	2 weeks	N: T: C:	16 subjects 32 sites 32 sites	0.6% triclosan gel + 0.3% triclosan dentifrice 2xday for 2 weeks, repeated again after 1 week washout period	-0.1, NS	N/A
Rosling et al., 2001104	NR PL NR	676 weeks (13 years)	N: T: C:	150 subjects 58 subjects 92 subjects	0.1% povidone iodine solution	+0.6, NR	-0.59, NR

R, randomized; PL, placebo-controlled; DB, double-blind; SB, single blind; NR, not reported; N/A, not applicable because not one of the clinical measures used; NS, not significant.

Although experimental regimens were applied subgingivally in all studies, none of these studies used exactly the same antimicrobial agent; the treatment regimens, including the SRP approaches, varied considerably too and included the following:

irrigation with hydrogen peroxide 3 times a week, with SRP started at week 32 with six visits during a period of, on average, 6 weeks to complete;⁸⁵

application of 1.25 percent amine fluoride gel on one site (tooth) or application of 4 percent stannous fluoride gel on another site, with both treatments applied three times within 10 minutes and selected pockets subjected at baseline to mechanical debridement with hand or ultrasonic instruments;⁸⁹

application of supragingival gel and dentifrice, both containing triclosan, twice a day for two time periods, each lasting for 14 days with 1 week of wash-out period between them, as well as subgingival gel application at days 0 and 7 in each time period, with the assigned quadrant anaesthetized and teeth exposed to meticulous SRP until root surfaces were hard and smooth;¹⁰²

use of 0.1 percent of iodofor solution (a water solution of povidone-iodine) as a cooling liquid for an ultrasonic device to provide nonsurgical therapy, with nonsurgical supra- and subgingival SRP under local anesthesia using ultrasonic device for 1 hour for four to six sessions (intervals between sessions never exceeding 1 week);¹⁰⁴ and,

subgingival irrigation of a 7-percent solution of tetrapotassium peroxydiphosphate twice a day for 8 weeks before SRP was done; SRP was performed through thorough subgingival scaling by hand (which took place at week 32 of the study, so week 32 was considered the baseline of the reported clinical results).¹⁰³

Taken together, the trials included 236 subjects and 326 sites (148 sites on antimicrobial regimens with SRP, and 178 sites on control regimens). Subjects in the studies experienced a full spectrum of periodontitis: moderate periodontitis;¹⁰² moderate to severe;^{85, 103} only severe periodontitis;⁸⁹ and advanced destructive periodontitis.¹⁰⁴

The studies differed markedly in the selection of sites used to characterize a subject’s response to therapy, but generally all studies involved patients with PDs in the range of 5 mm to 10 mm. One study based analyses on two pairs of contralateral sites with at least 5 mm PD that bled on probing at baseline.¹⁰² A second study selected for each experimental group one interdental pocket of 7 mm to 9 mm that affected a single-rooted tooth.⁸⁹ A third study included subjects with a minimum of eight nonmolar teeth, at least two of which in each dentate quadrant had PDs of 6 mm or more;¹⁰⁴ another study selected two or three interproximal sites in each jaw quadrant on the basis of the presence of PDs of 6 mm or more.⁸⁵ Finally, another study included subjects with PDs of 5 mm to 10 mm.¹⁰³

All teams used net reduction in PD as a clinical outcome.^{85, 89, 102- 104} The hydrogen peroxide study reported net PD reduction of 0.8 mm at 52 weeks (P < 0.05);⁸⁵ two other trials reported nonsignificant reductions of 0.65 mm (tetrapotassium peroxydiphosphate solution¹⁰³) and 0.6 mm (povidine-iodine¹⁰⁴) after 8 weeks among persons at baseline with 4 mm to 7 mm probing depths. In the triclosan study, the 0.1 mm reduction favored the control groups;¹⁰² the fluorides study did not report data but only that the results were not significant but favored the treatment group.⁸⁹

Three trials reported results for CAL – two in terms of actual gains or losses in attachment^{85, 104} and one in terms of less of a loss of attachment.¹⁰³ The hydrogen-peroxide trial reported a CAL gain of 0.1 mm for the treatment group (not significant).⁸⁵ It also reported the percentage distribution of sites that demonstrated a gain of more than 1 mm of clinical attachment level at the end of the phase that included both SRP and hydrogen-peroxide treatment; for both hydrogen-peroxide and control groups, 40 percent of sites showed an attachment gain of that magnitude (for no difference between the groups). In the tetrapotassium peroxydiphosphate trial, the net difference in attachment level between test and control groups at the end of 8 weeks was 0.25 mm (a nonsignificant finding).¹⁰³ Finally, in the longest study (of povidine-iodine), the control group had a greater attachment gain after 13 years of 0.597 mm (significance not reported).¹⁰⁴

Three trials examined microbiological outcomes including spirochetes.^{89, 102, 103} None of these studies showed that the treatment groups differed significantly from the control or placebo groups at the end of their respective observation periods. In the triclosan trial,¹⁰² the mean percentages dropped by day 14 to about 13 percent for both groups (from 46 percent among test patients and 37 percent among controls). In the tetrapotassium peroxydiphosphate study, the drop in percentage of spirochetes for both treatment and control groups was reported to be statistically significant between baseline and 8 weeks, but the net difference between groups was noted as not significant.¹⁰³ Finally, for fluoride gel treatment, the percentages of spirochetes was reported to have dropped significantly for both treatment groups (amine fluoride and stannous fluoride) and placebo between baseline and 36 weeks, but the differences between groups at the end were not significant.⁸⁹

By and large, these trials did not report on adverse events or effects. The tetrapotassium peroxydiphosphate trial noted mucosal irritation.¹⁰³

Overview

The key question examined in this report was: How does the effectiveness of scaling and root planing therapy (SRP) by itself for the treatment of chronic periodontitis compare to SRP accompanied by adjunctive therapy at varying lengths of time? To focus the evaluation of effectiveness, we relied primarily upon two clinical measures – reduction in probing depth (PD) and gains in clinical attachment level (CAL) – and one specific microbial measure – percentage reduction of spirochetes.

A total of 70 studies (unduplicated) met our inclusion criteria for this evidence report. However, some of these studies had multiple arms involving different adjunctive treatments; therefore, we included some studies more than once in the synthesis (Chapter 3) and evidence tables (Chapter 7). We provide qualitative and descriptive information for all these studies in Chapter 3; in addition, we did a total of nine meta-analyses when we had three or more studies that provided appropriate data for a 6-month follow-up period (plus or minus 3 months).

The adjunctive therapies were all chemical antimicrobials, applied either locally or systemically, which we grouped into 11 separate categories: systemic and local tetracycline, systemic and local minocycline, systemic and local metronidazole, metronidazole in combination with amoxicillin (systemic), chlorhexidine (local), systemic and local other antibiotics, and other local antimicrobials. The populations, severity of periodontitis, types of teeth treated, number of teeth treated, and added supportive therapies used all differed from study to study. In addition, thoroughness of SRP also differed across studies; we accepted studies only when the investigators used the same SRP approach for both the treatment and control groups. Some of the studies performed modified Widman flaps as needed to gain better access for debridement of study teeth. The time periods covered for follow-up of the therapies also varied from only a couple of weeks to 13 years.

The studies demonstrated that, for virtually any length of follow-up period reported, SRP whether or not accompanied by an adjunctive therapy resulted in statistically and clinically significant PD reductions and CAL gains between the baseline and study endpoint measurements. Those data set a context within which to interpret our findings, but effects of SRP are generally well known and so that was not a question that we sought to examine.

Rather, as agreed by the sponsors (the National Institute for Dental and Craniofacial Research [NIDCR], the Agency for Healthcare Research and Quality [AHRQ]), our Technical Expert Advisory Group (see Chapter 1), and our clinical consultants and colleagues from the University of North Carolina and elsewhere, we set out to determine whether the available research findings on using chemical antimicrobial therapy with SRP made an added contribution beyond that of the SRP alone. In simple terms, this value is the result of subtracting the baseline-to-follow-up differences for the SRP-only (or SRP with placebo) groups from the baseline-to-follow-up differences for the experimental or treatment groups using SRP and adjunctive therapy, and we characterized this as the net PD reduction or net CAL gain in our discussion.

Summary of Results

Overall Comments

Table 16. Summary of Qualitative and Quantitative Results (more...)

Table 16. Summary of Qualitative and Quantitative Results for Probing Depth and Clinical Attachment Level

Number of Studies/ Number of Positive Studies*	Effect Sizes (in mm): Range and Meta-Analytic Estimate (95% CI) †	Number of Studies/ Number of Positive Studies	Effect Sizes (in mm): Range and Meta-Analytic Estimate (95% CI)
Tetracycline, systemic	5/0	Range: NA MA : 0.15 (-0.29 - 0.58)	5/1	Range: 0.31 MA: none done
Tetracycline, local	16/5	Range: 0.40-0.93 MA: 0.47 ( 0.22 - 0.72)	16/2	Range: 0.15 - 0.48 MA: 0.24 (0.07 - 0.42 )
Minocycline, systemic	2/0	Range: NA MA: none done	0/0	Range: NA MA: none done
Minocycline, local	8/4	Range: 0.30‡ – 1.10' MA: 0.49 (0.40 - 0.58)	8/3	Range: 0.39 – 0.80 MA: 0.46 (0.32 - 0.60)
Metronidazole, systemic	8/3**	Range: 0.47¶ – 1.64║ MA: none done	6/2	Range: 0.47¶ – 1.19# MA: none done
Metronidazole, local	11/4**	Range: 0.18 – 0.80 MA: 0.32 (0.20 - 0.44)	8/2	Range: 0.40 – 0.66 MA: 0.12 (0.01 – 0.24)
Metronidazole with amoxicillin, systemic	4/2**	Range: 0.7 MA: none done	4/1**	Range: NR MA: none done
Chlorhexidine, local	17/2	Range: 0.26 – 0.33 MA: 0.24 (0.13 - 0.35)	13/3	Range: 0.16 – 0.28 MA: 0.16 (0.04 – 0.28)
Other antibiotics, systemic	7/3**	Range: 0.47 – 0.87†† MA: none done	6/2**	Range: 1.30 MA: none done
Other antibiotics, local	1/1	Range: 0.44 MA: none done	1/1	Range: 0.37 MA: none done
Other antimicrobials, local	5/1	Range: 0.8 MA: none done	4/0	Range: NA MA: none done

*Positive studies are defined as those showing statistically significant effects in favor of the adjunctive therapy as contrasted with scaling and root planing alone. For details on studies, see the specific text tables in Chapter 3 or evidence tables in Chapter 7.

†

CI, confidence interval; MA, meta-analysis; mm, millimeters NA, Not applicable; NR, not reported.

‡ 0.30 mm PD reduction for baseline probing depths of 5 mm or greater.

'1.10 mm PD reduction for baseline probing depths of 7 mm or greater.

¶ 0.47 mm PD reduction and CAL gain for baseline probing depths of 4 mm to 6 mm.

1.64 mm PD reduction for baseline probing depths of more than 6 mm.

# 1.19 mm CAL gain for baseline probing depths of more than 6 mm.

** One of these studies did not report any specific data, only a significant difference.

† †0.87 mm PD reduction for baseline probing depths of 6 mm or greater.

Table 16 summarizes our main findings, focusing on just the two clinical measures (PD and CAL) because those measures were far more commonly reported than changes in microbial environments (even changes in the percentages of spirochetes). We further focus in that table on just the number of “positive” studies – i.e., those that showed a statistically significant net PD reduction or net CAL gain favoring the adjunctive therapy in question. We specify the total number of studies and the number of positive studies (columns 1 and 3) and the range of effect sizes in millimeters for just those positive studies (columns 2 and 4). Finally, we recap the meta-analytic results, showing the overall estimated mean effect size in millimeters and the 95 percent confidence interval (CI) around that estimate (columns 2 and 4).

Across these 11 groupings of adjunctive therapies, we had a total of 84 intervention arms for PD reductions (recalling that some studies examined multiple types of antimicrobials). The two most commonly studied adjunctive therapies were local tetracycline (16) and chlorhexidine (17), followed by local metronidazole (11); least studied were systemic minocycline and the other local antibiotics (two each). Generally, investigators reported less frequently on gains in CAL, so we had only a total of 70 intervention arms; again, the commonly studied therapies were local tetracycline (16) and chlorhexidine (13), followed by local metronidazole and local minocycline (eight each), and the least studied were also systemic minocycline and other local antibiotics (two each).

We found it interesting to compare aggregate results (numbers of positive studies and percentages of total studies) for systemic and local therapies. Of the five categories of systemic therapies, we included 25 studies that measured PD reductions; of these, 7 (28 percent) reported a statistically significant net result favoring the adjunctive therapy (in this case metronidazole, metronidazole with amoxycillin, and various other systemic antibiotics). By contrast, of the six categories of local therapies, we included 60 studies with PD data; of these 16 (27 percent) reported a significant net result for the adjunctive therapy (mainly tetracycline and metronidazole).

For systemic therapies with CAL data, we included 19 studies; of these 6 (32 percent) reported net CAL gains favoring the adjunctive agent (tetracycline, metronidazole, and other systemic antibiotics). For the local treatments, we included 50 studies; of these, 11 (22 percent) had significant net gains (spread across all categories except the group of other antimicrobials).

In short, taking systemic and local applications together, tetracycline, metronidazole, and chlorhexidine were the most frequently studied therapies. PD reductions were measured slightly more often than CAL gains, but both were far more commonly reported than microbial changes. Some, but by no means all, investigators reported data by subgroups defined largely by baseline PD levels, but across those trials, the categories differed somewhat. For neither the systemic therapy studies nor the local therapy studies, taken as two groups, were more than about one-quarter to one-third reflective of statistically significant PD reductions or CAL gains.

Therapy-Specific Findings

Concluding Comments

Several themes emerge from these findings. First, PD reductions seemed to be more frequently measured (and statistically significant) than CAL gains although the two are clearly related; whether this has any practical or clinical ramifications is debatable, however. Second, adjunctive local antibiotics appeared to have more impact than adunctive systemic antibiotics, measured in terms of net PD reductions or CAL gains relative to SRP alone. Third, judging from trials with statistically significant results based on either our qualitative synthesis or our meta-analyses, the major PD reductions were in the range of about one-quarter to one-half millimeter, and the major CAL gains in the range of about one-tenth to one-half millimeter. As noted earlier, we take no stand on what might be considered a clinically meaningful change, but note that if the dental community were to consider improvements in the neighborhood of 0.50 mm as clinically important, then some of the therapies studied here do fall into that domain.

Fourth, combining PD and CAL results suggests that local minocycline might be the most promising adjunctive therapy (meta-analysis estimates of 0.49 mm for PD reduction and 0.46 mm for CAL gain), followed by local tetracycline (estimates of 0.47 mm for PD reduction and 0.24 mm for CAL gain). Local metronidazole and chlorhexidine results are well below these levels.

Fifth, in the absence of statistical significance, or when the evidence base is very small in terms of overall numbers of studies, the question of whether the evidence for adjunctive treatment is meaningful in a clinical sense does not need to be addressed, except to say the available evidence does not support its use. On the basis of the literature reviewed in this report, some experts might reach this conclusion for systemic tetracycline, systemic minocycline, systemic metronidazole, metronidazole with amoxicillin, and the various other systemic or local antibiotics and antimicrobials; others might disagree.

Sixth, by and large, harms from these adjunctive therapies are relatively minor. We take note, however, of concerns about bacterial resistance from overuse of systemic antibiotics, and we would urge that the positive findings reported here be interpreted in terms of whether the PD or CAL improvements justify that risk (for the individual patient but, perhaps more importantly, over the population).

Seventh, other important factors – supportive and follow-up care as well as self-care – may well affect the long-term periodontal status of patients as much if not more than use of these adjuncts to SRP, perhaps especially for patients with relatively early or moderate periodontitis. We did not review any body of literature directly on this point, but some results of trials that we did review suggested that added effectiveness of adjunctive treatment was greater in circumstances of more severe periodontitis where supportive or self-care may be less well executed. These situations may include patients with refractory periodontitis or who have deep pockets, defects or furcation involvement, or circumstances in which modified Widman flap surgery is not done (which would enable proper debridement of otherwise hard-to-reach areas). Routine use of appropriate (i.e., efficacious) adjunctive therapies might arguably be reserved for patients such as these.

Eighth, we cannot say from the trials reviewed here now how long the added effects of adjunctive treatment last (regardless of whether we would conclude they are either statistically or clinically significant). The endpoints for these studies varied tremendously, and even trying to narrow the field for the meta-analysis to trials lasting 6 months required us to allow in results from trials lasting 3 months to 9 months. What seems to occur in these studies is that if the adjunctive treatment is to have a more positive clinical effect (e.g., reduce PD or increase CAL) than SRP alone, then that effect seems to appear within a few weeks (1 to 2 months). However, with time, the difference in effect between SRP alone and SRP with adjunctive therapy narrows. Nonetheless, at all time periods, the SRP with adjunctive therapy seems to be more effective than SRP only, even if the net differences are quite small and not statistically significant.

Finally, putting all these results into the context of the results of SRP alone is imperative. SRP alone seems to produce significant improvements in mean PD reductions or CAL gains in the range of 1.5 mm to 2 mm or more, clearly making it the standard for nonsurgical (and nonpharmacologic) treatment of chronic periodontitis. The improvements produced by adjunctive antimicrobials beyond those levels – i.e., approximately one-quarter to one-third of the impact of SRP alone – pose a difficult “value” question for clinicians and patients alike that goes quite beyond the question of what adjunctive antibiotics to use. For example, one can question whether these improvements justify the added effort on the part of periodontists and dentists (and their staffs) or of patients or the likely added costs (either to dental insurance plans or to patients facing out-of-pocket payments). Moreover, as discussed in the next chapter, this literature on adjunctive therapies has enough drawbacks and gaps that a substantial research agenda remains before many of these issues can be resolved.

Clinical Issues Concerning Antimicrobials

Our recommendations for further research on the role of antimicrobials as adjuncts in the treatment of chronic periodontitis reflect several concerns. First, some issues remain for the specific key question of this evidence report, which dealt with the added effectiveness of particular antimicrobials when they are used as an adjunct to scaling and root planing (SRP), including whether any antimicrobials warrant further investigation in this regard. Second is the design and analysis of any future studies of this question, as the limitations of the existing literature are not trivial. Some of those limitations may relate to the actual reporting of the trials or other studies reviewed here, not the underlying design and conduct of the investigations per se. Moreover, the research reviewed in this evidence report provides a fairly broad range of expected effects of adjunctive antimicrobial use, but those effects remain far smaller in magnitude than the benefits achieved by SRP alone. A third topic, therefore, centers on the issue of what size difference between SRP alone and SRP with an adjunctive antimicrobial has clinical significance or relates to outcomes of particular meaning to patients and their dentists or periodontists.

We are limiting our consideration of future research directions or priorities to which antimicrobials, if any, warrant further examination in the context of use as adjuncts to SRP. We did not review literature relating to, for instance, antimicrobial use as an alternative to SRP, and we did not include every possible antimicrobial, in every possible modality, in this evidence report. Therefore, we do not comment further on any potential for new research in those areas. We also do not comment on research focused solely on the effectiveness of SRP per se, which appears to be well grounded in robust evidence accumulated over the years, as that was not a key question for this systematic review.

Improving Study Design and Conduct

Chapter 3 noted many of the difficulties we encountered in identifying appropriate research articles that would meet our a priori inclusion criteria and then in reviewing the included material in any coherent and systematic way. We may thus have omitted some relevant literature from this report, but even more important is the likelihood that some of the research that we did include could not be fully used or was open to incorrect interpretations because of poor reporting practices, confusing study designs, underpowered studies, and poorly conducted investigations.

The reporting practices may be the easiest to correct in the future (even though analysis and reporting of the data from these types of studies seem to have improved in recent years). Several authoritative statements from international groups provide clear instructions on appropriate ways to report on systematic reviews (QUORUM¹⁰⁵), randomized controlled trials (CONSORT¹⁰⁶), and observational studies (MOOSE¹⁰⁷). Authors and journal editors alike should take heed of these guidelines as a critical step in improving this literature overall.

Other guidance can come from the growing movement to grade the quality of individual articles that are included in reviews such as this one to begin with. Among the critical work now available is a lengthy report on systems to grade the quality of studies (i.e., articles) and rate the strength of evidence from the RTI-University of North Carolina Evidence-based Practice Center³⁹ and related methods of the US Preventive Services Task Force.¹⁰⁸

Study design issues, such as randomization, allocation concealment, blinding, and similar elements, must be given more careful consideration. Randomization should be a standard for all trials in this area. Use of placebo controls, and not simply variation in treatment arms, will be another useful step for trials attempting to establish the efficacy of a given medication. Every effort should be made to blind (mask) all parties (subjects, treatment providers, and outcome examiners) to the group (treatment arm[s], control) to which the patients belong. This is necessary to reduce the possibility of bias, which past research suggests typically exaggerates the effect of the treatment over what is experienced in the control group. In keeping with the reporting standards noted above, investigators should report clearly on randomization, control groups, and level of blinding achieved in their studies.

In addition, the study sample should be large enough to have adequate power to detect a statistically significant and clinically desired difference. Many of the studies we reviewed had apparently reasonable effect sizes, but they were based on small samples with large variances and could not have reached statistical significance. The problem here may be two-fold: the size of the original samples and the possibility of attrition (especially for studies with very long follow-up periods) such that samples at completion of the study became too small to provide adequate power for the analyses. This latter issue may pose particular challenges for investigators who propose to carry out intention-to-treat analyses but have instead to rely on final data only on completers.

Study reports often were unclear as to the underlying denominators for results, sometimes reporting on persons enrolled but then presenting data on some other unit. Thus, researchers should make it clear what the unit of analysis is – persons, teeth, sites, or pockets – and on what basis their means and measures of variance have been calculated. Specifically, investigators should ensure that their reports specify the number of units on which the mean for each group has been calculated and the variance (either standard deviation or standard error). Without this or comparable information, they or others cannot easily include the results in meta-analyses.

In addition, when using split-mouth designs, analytic techniques that take into account the nesting of observations within subjects need to be used, and when tests of statistical significance between groups are performed on multiple groups, techniques that adjust for the true significance level need to be used and reported. It remains to be demonstrated whether split-mouth designs of local therapies can adequately control the contamination or spill-over effect to be able to measure the true difference between the test and control groups.

Researchers in this area need to establish what measures are most meaningful for reporting treatment effects. Reaching some consensus on core outcomes for studies would help immensely for future systematic reviews on these topics, because the sheer number of possible outcomes complicated our work. With the inputs from our technical expert panel and representatives of the sponsors of this review, we selected PD reduction and CAL gain as the targeted clinical outcomes for several reasons. They appeared fairly frequently and consistently in the literature over the period covered by our review. They are also meaningful measures for clinicians, who can take such measures themselves to monitor the effects of treatment on their own patients. In the more recent literature, however, we saw a move away from reporting outcomes in terms of these metrics to outcomes that are somewhat less easy to understand or to measure objectively and reliably. Among them are variables such as percentage changes in prognosis, shifts from one category of treatment to another (extraction or surgery to maintenance), and other measures involving time (e.g., period of noninfection, time to recurrence). Moreover, investigators would find that easily used and understood statistical techniques are more readily available for analysis of metric data than for the analysis of percentages. A consistent, agreed-upon set of “critical” outcome measures would foster better comparisons across research projects and with past research. If the field moves to some of these newer outcome variables, attention will need to be given to standardizing how they are defined and reported and developing ways to convey absolute results and variances.

Some observers have noted that this literature contains little about measurement error and how it might affect reported results. Among the concerns are ambiguities about the level of training of those doing the SRP, the extensiveness and thoroughness of the SRP, the level of training and standardization of persons collecting the clinical measures, inaccuracies in measuring PD or CAL (or level of pre-existing inflammation), reliability of measurements across multiple examiners, and similar factors subject to variability in assessment and reporting. Moreover, time devoted to SRP, which is now the best proxy for the thoroughness of SRP, is a relatively imprecise measure and does not, in any case, ensure that SRP treatment was comparable across studies, patients, teeth, surfaces, or sites. Among the suggestions for overcoming some of these problems, at least in research venues, is the use of fiber optic devices that permit visual inspection of root surfaces and determination of the thoroughness of subgingival calculus removal and the level of cleanliness and smoothness of the root surface. The idea is that teeth (or surfaces, etc.) would be considered eligible for entry into a trial only after they had met some basic standard of SRP success. Whether moving to such a direct measure of SRP performance in place of time spent on SRP would yield more reliable and valid results, given the presumed additional costs to the research project, is itself an empirical question.

Finally, investigators need to be clearer as to the underlying diagnoses for their subjects. This point concerns two sources of ambiguity for those involved with developing the evidence based on these questions. The first problem is the mix of terms different research teams used for what was apparently the same disease: sometimes periodontal disease, sometimes periodontitis, with several different adjectives (adult, chronic, severe, moderate, mild) used, sometimes alone and sometimes in combination. We made every effort to focus this review on chronic periodontitis in adults (and in particular to eliminate studies in which patients could have had refractory or aggressive periodontitis), but on occasion we needed to draw an inference as to whether chronic periodontitis was indeed the disorder in question. Greater standardization of disease descriptors and their definitions, at least for use by the research community, would be helpful.

The second issue concerned whether subjects were being treated for periodontitis for the first time or were being retreated for chronic periodontitis. This confusion reflected in part the unpredictable use of descriptors such as recurrent, persistent, or refractory; although refractory may have a generally well-understood meaning within the periodontal and dental research community, recurrent and persistent have less agreed-upon definitions or connotations. The current literature generally did not make clear whether persons receiving retreatment had unsuccessful earlier treatment or were simply being retreated after successful treatment at some time in the past (i.e., were on some form of maintenance schedule). If researchers are including “maintenance” patients in their trials, they should explain this decision. More generally, investigators need to be certain that they are including only the types of patients for whom positive results from the particular study would be applicable in everyday practice.

DENTAL DATA ABSTRACTION FORM

Abstraction Form

Question: HOW DOES THE EFFECTIVENESS OF SCALING AND ROOT PLANING ACCOMPANIED BY ADJUNCTIVE THERAPY FOR CHRONIC PERIODONTITIS COMPARE TO SCALING AND ROOT PLANING THERAPY BY ITSELF AT VARYING LENGTHS OF TIME POST-TREATMENT?

A. Administrative Information

1. Abstractor:___________________________________ 2.Abstract Date:________________

3. Abbreviated study citation

___________________________________________________________________________

first author journal abbr

___________________________________________________­­­________________________

year volume pages

4. Abbreviated article title: ______________________________________________________________________________

5. Included?

No

If no, excluded because: (check one)

SRP different in control/comparison group than in treatment group

No SRP-only control/comparison group/sites

Treatment groups/sites did not also receive SRP

Literature review or meta-analysis

Practice guideline or editorial

Special population/disease studied, described as: with HIV/AIDS or diabetes; localized disease; juvenile or early onset disease; in smokers; around implants;

Other (Specify): ___________________________________________________

Yes (go to B. Study Design Information)

B. Study Design Information

1. Type of design (Check One):

Randomized controlled trial(RCT)

Non-Randomized controlled trial

RCT with crossover

Other (Specify)___________________________________________________________

__________________________________________________________________________

2. Is it a split-mouth design?

Yes

No

3. Does study investigate (check one):

Single-rooted teeth

Teeth bifurcated (molars)

Both

Not reported

4. Does study characterize periodontal disease as (check all that apply):

Mild

Moderate

Severe

Not reported

5. Placebo used in control/comparison (SRP only) group/sites?

Yes

No

6. Number of eligible active treatment groups: ________

7. Study period (Specify begin and end calendar month/year in which conducted)

From __________________ To ____________________

8. Study location(s) (Specify city/nation) _____________________________________________

9. Study/treatment site(s):

Hospital dental clinic

Graduate program university clinic

Undergraduate dental clinic

Health department clinic

Private dental practice

Not mentioned

Other (Specify)_________________________________________________________

10. Duration of treatment:_______ days/ weeks/months

11. Specify eligible adjunctive therapy(ies):

Amoxicillin

Augmentin

Chlorhexidine

Doxycycline

Metronidazole

Minocycline

Spiramycin

Povodine iodine

Tetracycline

Azithromycin

Clindamycin

Other (Specify)_______________________________________________________

12. Dose, mode of delivery, and schedule of treatment group(s) (XX mgs, as a rinse (for local), Y times /day, week, month) and SRP only group (or with placebo if used):

Group 1(or only treatment group): __________________________________________

Group 2: _______________________________________________________________

Group 3:_______________________________________________________________

SRP only / or SRP with Placebo:________________________________________________

13. Duration of evaluation period:_____ days or _______ weeks or ______ months

14. Drug use compliance monitoring:

Yes

No

Not reported

Not applicable (e.g., professionally administered/applied)

15. Describe the SRP as they performed it: _____________________________________________ _____________________________________________________________________________

16. How much time on average was spent doing SRP? ___________________________________

17. Was it performed with

An ultrasonic/Cavitron

By hand

Both

Not reported

18. Describe supportive therapy provided to study participants and indicate frequency/intervals.

Oral instruction How often: _______________________

Plaque removal How often: _______________________

Repeated SRP How often: _______________________

Other _______________________________________________________________

Not reported

19. Was baseline assessment performed before or after SRP?

Before

After

Not Reported

20. Number of evaluation time points (e.g. 4 times not including baseline) ______ and regular intervals of __________ (e.g., every two weeks). If intervals vary, specify evaluation points: __________________________________________________________________________

21. Blinding to active drug:

Examiners blind

Patient blind

Patients and examiners blind (double blind)

No blinding

Not reported

22. Outcomes (parameters measured) and citations:

CLINICAL MEASUREMENT

Bleeding on Probing (BOP/SBI) Citation: _______________________________

Plaque Index (PlI) Citation: _______________________________

Probing Pocket Depth (PPD) Citation: _______________________________

Clinical Attachment Level (CAL) Citation: _______________________________

Gingival Index (GI) Citation: _______________________________

Gingival Recession Citation: _______________________________

Tooth Loss Citation: _______________________________

Tooth Mobility Citation: _______________________________

Other (Specify) Citation: _______________________________

RADIOGRAPHIC TECHNIQUES

Bone Loss/Regeneration Citation: _______________________________

Other (Specify) Citation: _______________________________

MICROBIOLOGICAL METHODS

Chromatography Citation: _______________________________

DNA/RNA probe Citation: _______________________________

Crevicular fluid microbiota Citation: _______________________________

Plaque microbiota Citation: _______________________________

Other (Specify) Citation: _______________________________

PATIENT SELF REPORT

Patient Satisfaction Citation: _______________________________

Other (Specify) Citation: _______________________________

C. Sample Information

1. Description of population sampled (recurrent disease, never treated, clinic patients, private care)___ ___________________________________________________________________________________

___________________________________________________________________________________

2. Initial sample size -- persons (teeth):

_____________= Total subjects (sites)

_____________= Group 1 or only active drug group subjects (sites)

_____________= Group 2 (second active drug) subjects (sites)

_____________= Group 3 (third active drug) subjects (sites)

_____________= SRP only/SRP with placebo subjects (sites)

3. Subject/Site allocation to treatment/control conditions (Check one):

Random

Systematic

Cluster

Convenience

Unknown

Other (Specify)________________________________________________

4. Specific inclusion and exclusion criteria. (Specify tooth or person level considerations.)

Pregnant/breast feeding excluded

Concurrent drug therapy excluded

Diabetes history excluded

Diagnosed systemic infection excluded

Serious medical illness excluded

Need for prophylactic antibiotic before dental treatment excluded

Use of study drug within ____ months excluded

Hypersensitivity/allergic to study drug excluded

At least _____ teeth in _____ quadrant/mouth (circle one) with ____mm pocket included

Age at least _______ included

Other: ___________________________________________________________________ _________________________________________________________________________ _________________________________________________________________________

D. Group Characteristics

Total (All Study Groups Combined)

1. Age: mean:____________ median: ___________ range: ____________________

2. Gender: number (%) male:_______________ number (%) female:______________

3. Race/Ethnicity: number (%) NH White: _______________

number (%) NH Black: _____________

number (%) Hispanic (independent of race): ______________

number (%) NH Asian/PacificIslander: __________

number (%) other (specify):________

Group 1

1. Age: mean:____________ median: ___________ range: ____________________

2. Gender: number (%) male:_______________ number (%) female:______________

3. Race/Ethnicity: number (%) NH White: _______________

number (%) NH Black: _____________

number (%) Hispanic (independent of race): ______________

number (%) NH Asian/PacificIslander: __________

number (%) Other(Specify):________

Group 2

1. Age: mean:_______ median: ______ range: ______________________

2. Gender: number (%) male:_______________ number (%) female:______________

3. Race/Ethnicity: number (%) NH White: _______________

number (%) NH Black: _____________

number (%) Hispanic (independent of race): ______________

number (%) NH Asian/PacificIslander: __________

number (%) Other(Specify):________

Group 3

1. Age: mean:_______ median: ______ range: ______________________

2. Gender: number (%) male:_______________ number (%) female:______________

3. Race/Ethnicity: number (%) NH White: _______________

number (%) NH Black: _____________

number (%) Hispanic (independent of race): ______________

number (%) NH Asian/PacificIslander: __________

number (%) Other(Specify):________

SRP only/SRP plus placebo Group

1. Age: mean:_______ median: ______ range: ______________________

2. Gender: number (%) male:_______________ number (%) female:______________

3. Race/Ethnicity: number (%) NH White: _______________

number (%) NH Black: _____________

number (%) Hispanic (independent of race): ______________

number (%) NH Asian/PacificIslander: __________

number (%) Other(Specify):________

E. Provider/Examiner information

1. Number of examiners:_________

2. Examiners received:

Instruction in use of measures (written, pictures)

Standardization (Practice on patients)

Calibration (Gold standard examiner)

None/ Not reported

Not applicable (No clinical examination/measures: only one examiner)

3. How are examiners assigned to patients at assessments?

Random at each assessment

Repeat at each assessment

Not reported

Only one examiner

F. Analysis Information

1. Analysis adjusted for clustered observations

Yes

No

Not reported

2. Type of analysis reported:

Intent to treat

All with any follow-up

Full participants only

3. Reasons/criteria for exclusion from analysis: _____________________________________________________________________________________ ______________________________________________________________________________________

G. Outcome Information

1. Did study look for adverse effects?

Yes, found none

Yes, but none reported

Yes, found adverse effects (number and type) by study group.

Specify: ______________________________________________________________________ ___________________________________________________________________________ ___________________________________________________________________________

Never mentioned looking

2. Frequency with which modified Wideman flap procedure was performed in study?

Always/Often (25% of time or more)

Occasionally (5-24 % of time)

Rarely or Never (0-4 % of time)

Frequency Not Reported

Comparison of Assessment 1 (Baseline) to Assessment 2

Comparison of Assessment 2 / Baseline to Assessment 3

Comparison of Assessment 3 / Baseline to Assessment 4

Comparison of Assessment 4 / Baseline to Assessment 5

Comparison of Assessment 5 / Baseline to Assessment 6

QUALITY REVIEW CHECKLIST