Chapter  111:  Wound-Healing Technologies: Low-Level Laser and Vacuum-Assisted Closure

Types of Skin Wounds/Ulcers

Skin wounds are a heterogeneous and complex group of disorders with a wide variety of causes (Table 1) (Pierce, 2001). Approximately 70 percent are classified as pressure ulcers, diabetic ulcers, or vascular ulcers (Stadelman, Digenis, and Tobin, 1998a; Valencia, Falabella, Kirsner, et al., 2001). Vascular ulcers are further classified as due to arterial or venous insufficiency. Other less-frequent causes include inflammatory conditions, malignancies, burns, and radiation injuries (Valencia, Falabella, Kirsner, et al., 2001). Often the causes of wounds are multifactorial, such as in the diabetic patient who has both arterial insufficiency and peripheral neuropathy (Valencia, Falabella, Kirsner, et al., 2001). Each wound type has distinct physiologic characteristics, and exists in a unique host environment with varied clinical and psychosocial factors (Valencia, Falabella, Kirsner, et al., 2001).

Wounds are often classified as acute or chronic. Acute wounds are generally less than 8 weeks in duration and have not yet completed the natural healing cycle. Chronic wounds are defined as wounds that have failed to proceed through an orderly and timely process that produces anatomic and functional integrity (Lazarus, Cooper, Knighton, et al., 1994). Chronic wounds either require a prolonged time to heal, do not heal completely, or recur frequently.

Phases of Wound Healing

There are three phases of wound healing: (1) inflammatory, (2) proliferative, and (3) remodeling (Steed, 2003b; Harding, Morris, and Patel, 2002). These phases are distinct, but overlap in time during the healing process.

During the inflammatory phase, neutrophils and macrophages enter the wound site. Neutrophils act primarily to prevent and respond to infection; macrophages release inflammatory mediators such as cytokines and growth factors (Henry and Garner, 2003), which clear the wound of devitalized tissue and set the stage for cellular regeneration. The proliferative phase begins after two or three days and is marked by a predominance of fibroblasts and endothelial cells. Fibroblasts secrete growth factors and extracellular matrix components that lead to tissue regeneration (Henry and Garner, 2003). Endothelial cells form the new blood vessels that are also necessary for tissue regeneration. The final phase is the remodeling phase, in which intact skin replaces scar tissue. This phase is characterized by continued cycles of new cellular component formation and degradation of the scar by proteases (Eming, Smola, and Krieg, 2002; Henry and Garner, 2003).

Wounds that heal properly progress through these phases in an orderly fashion within about 8 weeks. Nonhealing wounds are often “stuck” in one of these stages, usually continued inflammation or proliferation (Douglass, 2003; Henry and Garner, 2003). A large number of factors can impede wound healing (Figure 1) and may predispose a patient to the development of chronic wound(s) (Williams and Barbul, 2003; Steed, 2003b). These include both systemic factors (e.g., poor nutrition, metabolic derangements, and drugs) and local factors (e.g., tissue hypoxia, infection, dry wound bed) (Stadelman, Digenis, and Tobin, 1998b).

While the above paradigm is widely accepted in conceptualizing wound care, there are alternative frameworks that have been proposed. Mustoe (2004) addresses limitations of current wound-healing science by addressing unifying aspects of chronic wounds, rather than their differences. He proposes that most, if not all, chronic wounds share common features of (1) the cellular and systemic effects of aging, (2) repeated ischemia-reperfusion injury, and (3) bacterial colonization with the accompanying inflammatory response. In this model, treatment approaches logically address all three aspects of chronic wounds.

Conventional Treatment of Chronic Skin Wounds

Optimal management of wounds starts with prompt recognition and accurate diagnosis in order to properly treat wounds at the earliest stage possible. Early recognition depends on identification of patients at risk, education for appropriate patients, and proper surveillance (U.S. Food and Drug Administration, 2000). An accurate diagnosis can be made from the appearance of the wound and the patient's risk factors in up to 75 percent of skin wounds (de Araujo, Valencia, Federman, et al., 2003). In some cases, specialized testing such as blood flow measurement is necessary (de Araujo, Valencia, Federman, et al., 2003; Valencia, Falabella, Kirsner, et al., 2001).

Conventional treatment for established wounds incorporates common principles that apply to the management of all wounds, including debridement of necrotic tissue, maintenance of a moist wound bed, and control of infection (Table 2). These common elements are combined with treatment modalities targeted to each wound type and the clinical characteristics of the patient (Lionelli and Lawrence, 2003; Steed, 1998a; U.S. Food and Drug Administration, 2000) (Table 2). Optimal treatment also entails consideration of the appropriate intensity of treatment (Ratliff, Bryant, Dutcher, et al., 2002). For example, depending on the context, dressing changes may be performed once a day or several times per day. Nutritional support can entail a wide variety of approaches that may differ considerably. Treatment regimens that are considered intensive often involve a multidisciplinary team of clinicians, nurses, therapists, and other ancillary staff. Unfortunately, there are no widely accepted, standardized protocols that define optimal standard treatment or the appropriate intensity of treatment delivery.

For wounds that do not heal with conservative therapy, surgical intervention may be considered. Surgical restoration of adequate blood flow is the goal for wounds caused by vascular insufficiency. Arterial revascularization procedures are often curative; however, venous surgery is of uncertain benefit for this purpose (de Araujo, Valencia, Federman, et al., 2003; Valencia, Falabella, Kirsner, et al., 2001).

Skin grafting can be performed for chronic, nonhealing skin wounds that are not amenable to surgical revascularization. Skin grafts are usually taken from a portion of intact skin of the same individual (autograft), but may also be taken from a cadaveric source (allograft). The specific indications for skin grafting are not well standardized (Valencia, Falabella, Kirsner, et al., 2001). Also, skin grafting is not always successful, as the donor skin may not “take” at the graft site in up to 25 percent of cases (Valencia, Falabella, Kirsner, et al., 2001). In addition, the procedure is associated with a substantial amount of morbidity, such as pain and wound infections (Jones and Nelson, 2000). A recent Cochrane review of skin grafting for venous ulcers found the available efficacy studies to be of poor methodologic quality and concluded that there was limited evidence on whether skin grafting improves the rate of healing (Jones and Nelson, 2003). Finally, amputation may be the treatment of last resort for wounds that fail all other methods, if the benefit of healing the wound outweighs the adverse effects of amputation.

The setting in which wounds are treated varies widely, from home treatment to specialized wound treatment centers. This may influence the specific treatment modalities used and/or the intensity of treatment provided. In clinical practice, there is a high degree of variability in wound treatment, and there is evidence that standard wound care deviates substantially from optimal treatment (ECRI, 2000). Thus, patients who present with nonhealing wounds may not have received similar prior care. It is possible that many of these “nonhealing” wounds may actually heal with an adequate trial of optimal care. The variability in prior care is also a concern for clinical trials, since this variability contributes to the heterogeneity of the study sample.

Emerging Treatments for Skin Wounds

Vacuum-assisted closure and low-level laser therapy are two potential alternatives for treating skin wounds. Low-level laser-assisted wound healing uses a low-energy, low-power, low-level laser, also known as a “cold” laser. It is hypothesized that delivery of low-energy laser therapy in this way may stimulate the physiologic process of wound healing, thus facilitating and/or accelerating the healing process. This physiologic rationale is supported by in vitro studies, and some animal models (Basford, Hallman, Sheffield, et al., 1986; Kana, Hutchenreiter, Haina, et al., 1981; Robinson, Garden, Taute, et al., 1987).

Lasers used in wound-healing applications include the gallium-aluminum (GaAl), gallium-arsenide (GaAs), and helium-neon (HeNe) laser. The power of these lasers ranges from 0.001 watts (1 mW) to 0.05 watts (50 mW), producing minimal heating of tissue. These low-energy lasers do not damage tissue directly, as do high-energy lasers that are used to ablate or vaporize tissue. Critical aspects of laser treatment delivery include the wavelength, power density (mW/cm²), and energy density (Joules/cm²). Variability in these parameters may lead to variation in tissue response to treatment (Eells, Henry, Summerfelt, et al., 2003).

Vacuum-assisted closure uses negative pressure to assist wound healing. Negative pressure drains fluid from the wound, thus removing the substrate for growth of microorganisms. Negative pressure may also accelerate granulation tissue formation and promote angiogenesis (Lionelli and Lawrence, 2003). The mechanical stimulation of cells by tensile forces may also play a role, by increasing cellular proliferation and protein synthesis (Morykwas and Argenta, 1997).

The technique involves application of a sterile, open-pore foam dressing directly on the wound. The wound is then sealed with an occlusive drape in order to create a closed, controlled environment. A fenestrated vacuum tube is attached to the wound dressing, and connected to a collection device. Negative pressure is applied at 50–125 mm/Hg, resulting in a decrease in the local interstitial pressure, and effluent from the wound is drawn out into the collection device (Lionelli and Lawrence, 2003; Morykwas and Argenta, 1997). Initially, the vacuum pressure is applied continuously. As the amount of drainage decreases, the vacuum may subsequently be applied on an intermittent basis (Morykwas and Argenta, 1997). The vacuum dressing is usually changed at approximately 48-hour intervals (Morykwas and Argenta, 1997).

Both the lasers used in wound healing and vacuum-assisted closure devices have been cleared for marketing by the U.S. Food and Drug Administration's (FDA's) 510(k) process, a regulatory mechanism that does not require submission of data from controlled efficacy trials.

There are a variety of other emerging treatments for skin wounds that are in various stages of development and FDA approval/clearance (Bennett, Griffiths, Schor, et al., 2003; Cross and Mustoe, 2003; Eming, Smola, and Krieg, 2002; Lionelli and Lawrence, 2003; Petrie, Yao, and Eriksson, 2003). These include: topical growth factors; bioengineered skin products; electrical stimulation; therapeutic ultrasound; novel dressings (e.g., hydrocolloids, alginates); hyperbaric oxygen; and gene therapy. However, discussion of these technologies is outside the scope of this evidence report.

Confounding Factors in Healing and Treatment

As summarized in Figure 1, many factors influence wound healing (Harding, Morris, and Patel, 2002). Local factors include severity of wound (size/depth), viability of surrounding tissue, or the presence of infection or a foreign body. Systemic factors include age, functional status, nutritional status, and comorbid illnesses, such as diabetes and/or renal disease. The large number of factors that contribute to wound healing, and the high degree of variability in wound characteristics, patient characteristics, and treatment delivery result in many potential confounding factors when attempting to measure treatment effect.

Since treatment varies widely in clinical practice, it is difficult to determine whether a patient has actually received an adequate course of treatment, and whether a nonhealing wound should truly be called “refractory.” In randomized, controlled trials, a relatively large proportion of “refractory” wounds heal with standard treatment (i.e., control arm). In two recent randomized, controlled trials of bioengineered skin substitute versus standard care (Falanga, Margolis, Alvarez, et al., 1998; Veves, Falanga, Armstrong, et al., 2001), 38 percent and 49 percent of “refractory” wounds, respectively, healed completely in the standard-care arm. Even in wounds present for at least 1 year (Falanga and Sabolinski, 1999), a substantial minority (19 percent) healed with standard treatment.

As a result of these multiple confounding factors, it is difficult to interpret outcomes from single-arm trials that lack a control group, since improvement may be due to factors other than the specific intervention being tested. A concurrent control group is necessary to permit measurement of a treatment effect above that related to optimization of standard treatment or due to the natural history of wound healing. Randomized assignment to treatment group is essential in maximizing the likelihood that confounding factors are equally distributed across treatment groups.

Design of Randomized, Controlled Trials for Wound-Healing Treatments

The FDA has prepared a draft guidance document that outlines difficulties in conducting trials to assess the effectiveness of interventions for wound healing. This document offers guidance in optimal design of such trials (U.S. Food and Drug Administration, 2000). The principals set forth by the FDA are summarized here and have been adapted in the development of the protocol for this systematic review.

Patient selection. The study population should ideally consist of patients with one particular type of skin wound, because of the different pathophysiology of each wound type and potential differences in response to therapy. A standardized definition of an adequate course of optimal care should be used in order to enroll a clinically refractory population. Alternatively, a run-in treatment period in which all patients receive an adequate course of optimal care may be utilized in order to exclude patients who heal with optimal standard care. This ensures that patients who enter the study are truly refractory to standard care.

Specific enrollment criteria that exclude conditions known to impede healing, such as very deep wounds or immunosuppression, may be helpful in reducing variability in measured outcomes. This may aid in determining the specific effect of treatment, but may also lead to reduced generalizability.

Patient assessment. Thorough assessment prior to treatment is important in accurately characterizing the features of the wound and in measuring potential confounders of outcome. Accurate recording of wound size, depth, and duration are important, since these are major predictors of healing (de Araujo, Valencia, Federman, et al., 2003).

Wounds can be measured by a variety of objective means, including direct tracing, planimetry, and stereophotogrammetry (ECRI, 2000). Wound imaging by photographic methods may also aid in the objective measurement of wounds. Other objective measures such as transcutaneous oxygen tension (t_cpO₂), ankle/brachial index, and microfilament testing may also be helpful in assessing the baseline wound characteristics.

A thorough assessment and measurement of other potential confounding variables should be performed at baseline and at followup time points. These include patients' clinical and demographic characteristics, comorbid medical conditions, and prior treatment received.

Treatment issues. Double-blinding of treatment is the optimal study design to minimize bias in treatment delivery and outcome assessment. A sham placebo should be considered in the control arm to allow for double-blinding. However, double-blinding may be difficult for some devices, especially for a therapy such as vacuum-assisted closure. The difficulties of double-blinding need to be balanced against the benefit in minimizing bias in interpreting the trial outcomes.

Researchers should ensure that high-quality standard treatment is delivered to the control group. “High-quality” treatment means that all of the main modalities of standard care from wound treatment guidelines are included (Table 2). No definitive standard treatment guidelines exist, but there are guidelines that incorporate modalities of standard care (e.g., from the Wound Ostomy and Continence Nurses Society [www.wocn.org]).

It is also important to ensure that standard treatment modalities are identical between groups in order to avoid performance bias. The experimental treatment arm should not include additional elements of standard care that are not delivered to the control group. The experimental treatment arm should not incorporate a greater intensity of standard care than the control arm. The importance of equal intensity of care was demonstrated in a prior multicenter trial of platelet-derived growth factor for chronic wounds (Cross and Mustoe, 2003). In this study, the rate of healing was significantly higher in centers that incorporated more frequent debridement (Cross and Mustoe, 2003).

An adequate followup period is required to demonstrate durability of response and adverse effects. It is recommended that patients remain enrolled in studies for at least 3 months following initiation of treatment (U.S. Food and Drug Administration, 2000). This is the minimum amount of time required to evaluate the number of healed wounds that recur. Some experts recommend an even longer minimum duration. For example, Steed (2003b) recommends that the minimum duration of a clinical trial include a run-in period of standard care, followed by 20 weeks of treatment, and an additional 12 weeks of followup.

Outcomes and outcome measurement. Outcome measurement should focus on outcomes that are quantitative and clinically meaningful (Jeffcoate and Harding, 2003; Steed, 2003b). The most important outcomes to be considered are: (1) the percent of patients with complete healing; and (2) time to complete healing.

Other outcomes that may also be clinically meaningful are: (1) facilitating surgical wound closure; (2) change in wound size; (3) improved cosmesis; (4) improved activities of daily living; (5) improved quality of life; (6) pain; (7) transcutaneous oxygen tension; (8) infections; and (9) need for debridement.

In some cases, particularly for vacuum-assisted closure, the treatment may not be expected to result in complete healing. Rather, the treatment may be intended to advance the wound to a stage where healing is possible, either by continued conventional treatment or by surgical closure. These goals represent intermediate treatment outcomes. If the overall treatment strategy is successful, the benefit of these intermediate outcomes will ultimately be reflected in improved rates of complete healing. The intermediate outcome states are more difficult to measure, but are likely partly represented by the secondary outcomes of wound size and facilitation of surgical closure.

Outcome assessment should also include measurement of adverse events that result from the treatment or from the natural history of the disorder. These include: (1) local adverse effects (pain, discharge, dermatitis); (2) immune reactions; (3) infections; (4) limb amputations; and (5) discontinuation from treatment, including assessment of whether discontinuation is a result of the treatment.

Ascertainment of outcomes should be ideally performed by an independent, blinded individual. This is especially important in situations where patients and/or treating physicians are not blinded to treatment.

Low-Level Laser Treatment

In the treatment of chronic, nonhealing wounds, what are the outcomes of low-level laser therapy for specific indications and patient types:

1. as a substitute for standard therapy; or

2. as an adjunct to standard therapy, compared with standard therapy alone?

Vacuum-Assisted Closure

In the treatment of various wounds, what are the outcomes of vacuum-assisted closure for specific indications and patient types:

1. as a substitute for standard dressings; and

2. as an adjunct to standard therapy, compared with standard therapy alone?

Patient Populations

Low-level laser treatment. With respect to low-level laser treatment, chronic wounds may be classified in a variety of ways. The simplest way is to distinguish between cutaneous ulcers and burns. However, a more comprehensive classification system places chronic wounds into these categories:

• pressure ulcers

• metabolic disorders (e.g., diabetes mellitus)

• vascular insufficiency

• inflammatory disorders

• malignancies

• infections

• miscellaneous (e.g., burns)

Vacuum-assisted closure. The review for vacuum-assisted closure addressed:

• chronic wounds (as above)

• acute wounds

• traumatic wounds

• subacute wounds

• dehisced wounds

• partial thickness burns

• diabetic ulcers

• pressure ulcers

• flaps

• grafts

Study populations with both acute and chronic wounds will be examined carefully with respect to the duration of the wound and the types of interventions that have been performed prior to treatment with low-level laser therapy or vacuum-assisted closure. Some wounds may be described as refractory; that term should be defined as specifically as possible in terms of the types and duration of previous treatments. Similarly, the term “chronic” should be defined in as much detail as possible.

Practice Settings

Low-level laser treatment and vacuum-assisted closure may be used in the following settings:

• Surgical centers

• Hospitals

• Specialized wound care centers

• Nursing or rehabilitation facilities

• Physicians' offices

• Physical therapy offices

• Homes

Interventions/Technologies of Interest

Standard care. Standard wound care is multifactorial. Among its components are:

• Debridement

• Dressings

• Topical or systemic medications

• Compression

• Skin grafting

• Skin equivalents

• Improved nutrition

• Convalescence

• Physical therapy

• Treatment of underlying disorder

Low-level laser treatment. This review will focus on lasers that have been described as low-energy, low-power, low-level or “cold” lasers. The power of these lasers ranges from 0.001 watts (1 mW) to 0.05 watts (50 mW), producing minimal heating of tissue. Lasers used in wound healing applications include the gallium-aluminum (GaAl), gallium-arsenide (GaAs) and helium-neon (He-Ne) laser. Characteristics of laser treatment that would be of interest include laser type, intensity (measured in Joules per square centimeter of wound surface [Joules/cm²]), duration of each session, frequency of sessions, and overall duration of treatment. Other prior and concurrent treatments will be examined in detail.

Vacuum-assisted closure. The vacuum-assisted closure technique involves application of a sterile, open-pore foam dressing directly on the wound, which is then sealed with an adhesive drape, thus converting an open wound to a closed, controlled wound. An evacuation tube, embedded in the dressing, feeds into a collection canister. When subatmospheric pressure is applied, effluent from the wound is drawn out. Attention will be paid to the degree of negative pressure applied, frequency of dressing changes, and duration of use of the vacuum device. Other prior and concurrent treatments will be examined in detail.

Outcomes of Interest

In general, outcomes should be standard, valid, reliable and clinically meaningful. A 2000 draft guidance document produced by the FDA (U.S. Food and Drug Administration, 2000) stated that wound healing outcomes should focus on the probability or speed of achieving complete wound closure. Intermediate outcomes such as wound size are problematic because of uncertainty about the validity of measurement techniques and clinical meaningfulness.

• Primary outcomes:

  • incidence of complete wound closure

  • time to complete closure

  • adverse events

• Secondary outcomes

  • facilitating surgical closure

  • need for debridement

  • infections

  • pain

  • activities of daily living

  • quality of life

  • improved cosmesis

Other secondary outcomes abstracted were change in wound size and transcutaneous oxygen tension (t_cpO₂); however, these were considered to be of less clinical importance.

Study Design Criteria

• Initial assembly of comparable groups: adequate randomization, including concealment and whether potential confounders (e.g., other concomitant care, patient characteristics) were distributed equally among groups

• Maintenance of comparable groups (includes attrition, crossovers, adherence, contamination)

• Important differential loss to followup or overall high loss to followup

• Measurements: equal, reliable, and valid (includes masking of outcome assessment)

• Clear definition of interventions

• All important outcomes considered

• Analysis: adjustment for potential confounders, intention-to-treat analysis

Definition of Quality Ratings

In applying the Harris rating system to the studies selected for this systematic review, several rules were followed. To conclude that a study achieved initial assembly of comparable groups, it had to use an adequate randomization method and had to have equal distribution of confounders. Adequate randomization was defined as either central randomization or use of opaque envelopes (concealment). For the purposes of this review, equal distribution of confounders was defined as a minimal difference (less than 20%) in mean values between groups on age, wound duration and wound size. Low loss to followup and maintenance of comparable groups was defined as loss less than 20% of the initial sample and no differential loss to followup between groups.

To consider measurements reliable, valid and equal, the article had to provide a clear description of wound measurement methods that appeared reproducible. Examples include use of photographic or digital transfer of wound tracings and/or use of computer software to calculate wound size. Liquid or plaster used to measure wound volume was also acceptable. Use of a blinded outcome assessor was also necessary to fully satisfy this quality dimension. Clear, detailed descriptions of both control and treatment interventions were sought. Analysis of results was considered appropriate if the investigators adjusted for confounders and analyzed by intention-to-treat, which was defined as analyzing all randomized patients or no more than 5% loss of the initial sample. See Table 3 for the quality criteria and ratings system applied to the evidence tables in Chapter 3.

Good. Meets all criteria: Comparable groups are assembled initially and maintained throughout the study (followup at least 80 percent); reliable and valid measurement instruments are used and applied equally to the groups; interventions are spelled out clearly; all important outcomes are considered; and appropriate attention to confounders in analysis. In addition, for RCTs, intention-to-treat analysis is used.

Fair. Studies will be graded “fair” if any or all of the following problems occur, without the fatal flaws noted in the “poor” category below: Generally comparable groups are assembled initially but some question remains whether some (although not major) differences occurred with followup; measurement instruments are acceptable (although not the best) and generally applied equally; some but not all important outcomes are considered; and some but not all potential confounders are accounted for. Intention-to-treat analysis is done for RCTs.

Poor. Studies will be graded “poor” if any of the following fatal flaws exists: Groups assembled initially are not close to being comparable or maintained throughout the study; unreliable or invalid measurement instruments are used or not applied at all equally among groups (including not masking outcome assessment); and key confounders are given little or no attention. For RCTs, intention-to-treat analysis is lacking.

Overview

The only previous systematic reviews available on the use of laser therapy for wound healing have been produced by a single group in the United Kingdom (Flemming and Cullum, 2003; Cullum, Nelson, Flemming et al. 2001; Flemming, Cullum and Nelson, 1999). These reviews found no supportive evidence for a benefit of low level laser therapy in healing of venous leg ulcers. All 4 studies abstracted by these reviews are included in the present review.

Among excluded studies, 3 were randomized controlled trials. Two were excluded because they did not select patients with chronic wounds (Lagan, Clements, McDonough, et al., 2001; Fernando, Hill, and Walker, 1993). The third study reported only on an outcome that was not of interest to this review, skin temperature (Schindl, Heinze, Schindl et al., 2002). Four comparative studies published in foreign languages were excluded. One German study was excluded because it did not select patients with chronic wounds (Zimmerman, 1990). Three Russian studies were examined by a Russian reader who determined that subjects were not assigned to groups randomly (Babadzhanov and Sultanov, 1998; Gostishchev, Vertianov, Novochenko, et al., 1987; Gostishchev, Vertianov, Shur, et al., 1985). No other nonrandom comparative studies published in English were found. All other excluded studies were case series or case reports.

Review of search results identified a total of 11 studies (n=413; (Bihari and Mester, 1989; Crous and Malherbe, 1988; Franek, Krol, and Kucharzewski, 2002; Iusim, Kimchy, Pillar, et al., 1992; Lagan, McKenna, Witherow, et al., 2002; Lucas, Coenen, and De Haan, 2000; Lucas, van Gemert, and de Haan, 2003; Lundeberg and Malm, 1991; Malm and Lundeberg, 1991; Nussbaum, Biemann, and Mustard, 1994; Santoianni, Monfrecola, Martellotta, et al., 1984) that met study selection criteria for low-level laser therapy (Evidence Table 1).

Details about studies meeting selection criteria are provided in Evidence Tables 2–5, each of which is divided into three subsets. The “A” subsets include seven studies with placebo controls, the “B” subsets contain three studies assessing the effects of low-level laser therapy plus standard treatment versus standard treatment alone, and the “C” subsets describe one study comparing the use of ultraviolet light and low-level laser therapy. Evidence Table 2 presents patient characteristics. Evidence Table 3 focuses on treatment details. Evidence Table 4 describes outcomes assessment. Evidence Table 5 provides results. Information shown in Evidence Tables 1–5 served as the basis for study quality ratings. Study quality ratings are included in Evidence Table 6 (See Table 3, Chapter 2, Methods, for study quality criteria and ratings).

Evidence Table 1. Summary of Low-Level Laser Therapy Studies

Evidence Table 2. Low-Level Laser Therapy, Patient Characteristics

Evidence Table 3. Low-Level Laser Therapy, Treatments

Evidence Table 4. Low-Level Laser Therapy, Outcome Assessment

Evidence Table 5. Low-Level Laser Therapy, Results

Evidence Table 6. Study Quality Ratings, Low-Level Laser Therapy

Placebo-controlled studies. Of the 11 studies meeting study selection criteria, seven studies enrolling 262 patients compared standard treatment plus sham low-level laser versus standard treatment plus active low-level laser (Bihari and Mester, 1989; Franek, Krol, and Kucharzewski, 2002; Iusim, Kimchy, Pillar, et al., 1992; Lagan, McKenna, Witherow, et al., 2002; Lundeberg and Malm, 1991; Malm and Lundeberg, 1991; Santoianni, Monfrecola, Martellotta, et al., 1984). Six of seven studies included patients with primarily lower extremity venous ulcers, while the seventh (Santoianni, Monfrecola, Martellotta, et al., 1984) selected a heterogeneous group of patients with diabetes or peripheral vascular disease.

Study quality. Overall study quality ratings were poor for all seven placebo-controlled studies (Evidence Table 6). None of these seven studies demonstrated that groups were comparable on all three critical baseline characteristics: age, wound size and wound duration (Evidence Table 2). Randomization methods were poorly described. High loss to followup was common. It was rare for studies to use both reliable and valid measurement techniques and blinded outcome assessors. Less than half of the studies reported on the main outcome, complete healing, and none made it clear that analysis was by intention-to-treat and controlled for confounders.

It is unclear whether any study assembled groups that were comparable on a sufficient range of key baseline patient characteristics. The study by Franek, Krol, and Kucharzewski (2002) assembled a control group receiving compressive and topical therapy plus sham low-level laser treatment that had a shorter mean wound duration (30 months) than the group receiving standard care plus active low-level laser therapy (41 months). The other six placebo-controlled studies did not provide comparisons between groups on wound duration at baseline. The studies by Franek, Krol, and Kucharzewski (2002), Malm and Lundeberg (1991), Lundeberg and Malm (1991), and Iusim, Kimchy, Pillar, et al. (1992) enrolled patient groups of comparable mean age, as well as comparable wound size.

In order for this review to describe the randomization method as adequate, the article had to state either that central randomization was performed or opaque (concealed) envelopes were employed. None of the seven placebo-controlled studies satisfied this quality component. Loss to followup exceeded 20 percent of the initial sample size in two studies: Malm and Lundeberg (1991) and Lundeberg and Malm (1991). Description of treatment and control procedures was clear in all cases.

Wound measurement was assumed to be reliable and valid if the article described use of photographic or digital transfer of wound tracings and/or use of computer software to calculate wound size. The studies by Franek, Krol, and Kucharzewski (2002) and Lagan, McKenna, Witherow, et al. (2002) satisfied this quality component. Data abstraction also sought information on whether a blinded observer performed outcome assessment. Blinded outcome assessors were used by Lagan, McKenna, Witherow, et al. (2002), Malm and Lundeberg (1991), Lundeberg and Malm (1991), and Bihari and Mester (1989). In only one study (Lagan, McKenna, Witherow, et al., 2002) was it clear that investigators used both blinded assessment and measurement that was reliable and valid.

None of these articles stated whether statistical analyses used adjustment procedures to deal with baseline confounding variables. Intention-to-treat analysis was performed by Lagan, McKenna, Witherow, et al. (2002), Bihari and Mester (1989), and Iusim, Kimchy, Pillar, et al. (1992). The primary endpoint, complete healing, was reported in three studies: Malm and Lundeberg (1991), Lundeberg and Malm (1991), and Bihari and Mester (1989). Analysis of results was considered appropriate if both adjustment for confounders and intention-to-treat analysis was carried out. None of these studies satisfied both components.

Complete healing. Three of seven studies reported data on the primary outcome specified for this systematic review, complete healing. Using the GaAs laser, Malm and Lundeberg (1991, n=42) performed a life-table analysis of time to complete healing by 12 weeks and found no significant difference between groups. Lundeberg and Malm (1991, n=46) used the HeNe laser and similarly found no differences between groups on life-table analysis at 12 weeks. Selecting three groups of 15 subjects each, Bihari and Mester (1989) compared a sham low-level laser group with one group treated with a hand-held HeNe laser and another group treated with a machine-scanned HeNe/pulsed infrared laser. Results favored the hand-held laser group over sham, but the difference was not statistically significant: the relative risk and 95 percent confidence interval (RR, 95 percent CI) of complete healing at 9 months was 2.0 (0.9–4.5). Results were significant in the machine-scanned laser group (RR=2.4, 95 percent CI: 1.1–5.1); however, two control patients were excluded from the analysis. If they had achieved complete healing, the results would not be statistically significant. None of the available studies provides clear evidence that use of laser treatment leads to a higher probability of complete healing, compared with sham treatment.

Change in wound area. Six of seven studies reported on this outcome. At 12 weeks' followup, Franek, Krol, and Kucharzewski (2002) found no significant differences between groups on either the mean rate of change in relative area or in the mean rate of change of relative suppurative area. Lagan, McKenna, Witherow, et al. (2002) found no difference between groups on percent change in wound area at 12 weeks. Malm and Lundeberg (1991) reported no difference between groups in rate of wound area change over 12 weeks. Lundeberg and Malm (1991) found that the average percent change at 12 weeks did not differ between groups. Santoianni, Monfrecola, Martellotta, et al. (1984) observed no differences between groups in the area epithelialized. Iusim, Kimchy, Pillar, et al. (1992) compared sham low-level laser with one group treated by red light laser and another treated with infrared light laser. By percent change in area at 20 days, the red light laser performed significantly better than sham, which did not differ from the infrared laser group. Only one of six studies reported a significant advantage favoring low-level laser treatment over sham.

Other outcomes. Franek, Krol, and Kucharzewski (2002) reported that the rate of change in relative defect volume was better in the standard treatment alone arm versus the low-level laser arm. Lagan, McKenna, Witherow, et al. (2002) found that visual analog scale pain did not differ between groups.

Studies without a placebo control. Three studies enrolling 151 patients compared patients receiving standard treatment alone with those undergoing standard treatment plus low-level laser (Lucas, van Gemert, and de Haan, 2003; Lucas, Coenen, and De Haan, 2000; Nussbaum, Biemann, and Mustard, 1994). All three selected patients with pressure ulcers.

Study quality. Both the Lucas, van Gemert, and de Haan (2003) and the Lucas, Coenen, and De Haan (2000) studies showed that groups were comparable on age, wound size and wound duration. The groups in the study by Nussbaum, Biemann, and Mustard (1994) were comparable on age and wound duration, but not on wound size. The Lucas, van Gemert, and de Haan (2003) study stated that central randomization was used, while the other two studies provided insufficient details about the randomization technique. Loss to followup exceeded 20 percent of the initial sample in the Nussbaum, Biemann, and Mustard (1994) study. Wound measurement was reliable and valid in all three studies and blind outcome assessors were used in all. Both control and treatment interventions were clearly described in all three studies. The Lucas, van Gemert, and de Haan (2003) and the Lucas, Coenen, and De Haan (2000) studies analyzed results by intention-to-treat, while Nussbaum, Biemann, and Mustard (1994) did not; neither did it mention use of adjustment for confounders.

The Lucas, van Gemert, and de Haan (2003) study was the only study that met all five quality dimensions and received an overall good rating. The earlier study by Lucas, Coenen, and De Haan (2000) met all but one of the quality dimensions, and received an overall fair rating. The main detail lacking from this study was an adequate description of the randomization procedure. The Nussbaum, Biemann, and Mustard (1994) study was given an overall poor quality rating. Randomization was poorly described, wound size differed between groups, loss to followup was high and appropriate analysis of results was not carried out.

Complete healing. All three studies reported on complete healing, but none provide support that use of laser therapy results in a higher probability of complete healing as compared to standard treatment alone. Lucas, van Gemert, and de Haan (2003) did not show a higher probability of complete healing at 6 weeks with laser. In the study by Nussbaum, Biemann, and Mustard (1994), patients receiving low-level laser treatment had a higher mean number of weeks to complete healing than standard treatment alone, by a followup as high as 20 weeks, but no statistical test result was provided. Lucas, Coenen, and De Haan (2000) observed that four of eight control patients and three of eight low-level laser patients achieved complete healing by 6 weeks.

Change in wound area. In Lucas, van Gemert, and de Haan (2003), no significant differences between groups were observed for absolute improvement, relative improvement or natural log-transformed improvement. Nussbaum, Biemann, and Mustard (1994) found that the rate of wound size change for laser was not better than that of control. Lucas, Coenen, and De Haan (2000) found no significant difference between groups in percent change of wound area.

Other outcomes. Lucas, Coenen, and De Haan (2000) assessed the Norton wound scale and found no differences between groups. These authors also found no difference between groups in the incidence of stage IV pressure ulcers.

Studies comparing ultraviolet light and low-level laser therapy. One study of only six patients with chronic leg venous ulcers compared use of medical treatment plus ultraviolet light with medical treatment plus laser therapy (Crous and Malherbe, 1988). The overall study quality rating was poor. It is unclear whether the two groups of three subjects were comparable on age, wound duration or wound size. The randomization procedure was inadequately described. Interventions were clearly described. Wound measurement was performed with photographs and a blinded observer. Adjustment for confounders was not mentioned. Results were given for all randomized patients. Complete healing was not reported.

The article reported on percent change in wound area and wound perimeter. The mean change in area in the ultraviolet light group was 34 percent, compared with 50 percent for low-level laser. Mean change in perimeter was 18 percent for ultraviolet light and 27 percent for low-level laser. No statistical test results were reported, but t-tests performed for this systematic review did not find statistically significant differences between groups.

Conclusions

Eleven studies (Bihari and Mester, 1989; Crous and Malherbe, 1988; Franek, Krol, and Kucharzewski, 2002; Iusim, Kimchy, Pillar, et al., 1992; Lagan, McKenna, Witherow, et al., 2002; Lucas, Coenen, and De Haan, 2000; Lucas, van Gemert, and de Haan, 2003; Lundeberg and Malm, 1991; Malm and Lundeberg, 1991; Nussbaum, Biemann, and Mustard, 1994; Santoianni, Monfrecola, Martellotta, et al., 1984) met the study selection criteria for Part I of this review, nine of which were rated poor in quality (Bihari and Mester, 1989; Crous and Malherbe, 1988; Franek, Krol, and Kucharzewski, 2002; Iusim, Kimchy, Pillar, et al., 1992; Lagan, McKenna, Witherow, et al., 2002; Lundeberg and Malm, 1991; Malm and Lundeberg, 1991; Nussbaum, Biemann, and Mustard, 1994; Santoianni, Monfrecola, Martellotta, et al., 1984), while one was rated good quality (Lucas, van Gemert, and de Haan, 2003) and one was rated fair (Lucas, Coenen, and De Haan, 2000).

Seven studies (n=262) compared standard care plus placebo with the combination of standard care and sham laser therapy (Bihari and Mester, 1989; Franek, Krol, and Kucharzewski, 2002; Lagan, McKenna, Witherow, et al., 2002; Iusim, Kimchy, Pillar, et al., 1992; Lundeberg and Malm, 1991; Malm and Lundeberg, 1991; Santoianni, Monfrecola, Martellotta, et al., 1984). Most of these patients had lower extremity venous ulcers. Of the three studies that reported on complete healing (Bihari and Mester, 1989; Lundeberg and Malm, 1991; Malm and Lundeberg, 1991), one provides weak evidence of a higher rate of healing for patients treated by machine-scanned laser versus those receiving sham laser (Bihari and Mester, 1989).

Standard treatment alone versus standard treatment plus laser was compared in three studies, which reported on a total of 151 patients with pressure ulcers (Lucas, Coenen, and De Haan, 2000; Lucas, van Gemert, and de Haan, 2003; Nussbaum, Biemann, and Mustard, 1994). All three studies reported on complete healing. One of these was rated as good in quality, and this higher quality study did not show a higher probability of complete healing at 6 weeks with the addition of laser treatment (Lucas, van Gemert, and de Haan, 2003), nor did it show benefit for any of the other reported outcomes. Use of medical treatment plus ultraviolet light with medical treatment plus low-level laser therapy was compared in one study of six patients with chronic venous ulcers (Crous and Malherbe, 1988). That study did not show a higher probability of complete healing at 6 weeks with the addition of laser treatment.

Overall, the quality of this body of evidence is poor, and does not permit definitive conclusions. However, the available data suggests that the addition of laser therapy does not improve wound healing, as the vast majority of comparisons in these studies do not report any group differences in the relevant outcomes. It is unlikely that the lack of significant differences is the result of a type II error, since there are no trends or patterns of outcomes that favor the laser group.

Overview

A single previous systematic review is available on the use of vacuum-assisted closure for treating chronic wounds (Evans and Land, 2003). The authors concluded that the 2 small trials (Joseph, Hamori, Bergman, et al., 2000; McCallon, Knight, Valiulus, et al., 2000) that met their selection criteria offer weak evidence that vacuum-assisted closure is more efficacious than moist dressings. They noted that small sample sizes and methodological limitations require that the results of these 2 studies be interpreted with extreme caution. While Evans and Land restricted themselves to chronic wounds, the present review is broader in focus. Both studies reviewed in that report are also included here.

Two randomized trials on the use of vacuum-assisted closure are excluded from the current review (Buttenschoen, Fleischmann, Haupt, et al., 2001; Genecov, Schneider, Morykwas, et al., 1998) because they provided data only on outcomes that were not of interest to this review. The former reported on immune response markers and the latter gave data on skin biopsies. Two comparative studies published in Chinese were reviewed by a Chinese reader and found to be nonrandomized (Huang, Yao, and Huang, 2003; Yao, Huang, and Ma, 2002). No other nonrandomized comparative studies published in English were found. All other excluded studies were case series or case reports.

Six studies using vacuum-assisted closure met study selection criteria, with a collective total of 135 patients (Eginton, Brown, Seabrook, et al., 2003; Ford, Reinhard, Yeh, et al., 2002; Moues, Vos, van den Bemd, et al., 2004; Joseph, Hamori, Bergman, et al., 2000; McCallon, Knight, Valiulus, et al., 2000; Wanner, Schwarzl, Strub, et al., 2003). Details about these studies are given in Evidence Tables 7–12; information in these tables served as the basis for study quality ratings, which may be viewed in Evidence Table 13. Evidence Table 7 summarizes the included studies. Evidence Table 8 presents patient inclusion and exclusion criteria. Evidence Table 9 shows patient characteristics. Evidence Table 10 gives details of treatment. Evidence Table 11 includes information on how outcomes were assessed. Evidence Table 12 depicts results.

Evidence Table 7. Summary of Vacuum-Assisted Closure Studies

Evidence Table 8. Vacuum-Assisted Closure, Patient Selection Criteria

Evidence Table 9. Vacuum-Assisted Closure, Patient Characteristics

Evidence Table 10. Vacuum-Assisted Closure, Treatments

Evidence Table 11. Vacuum-Assisted Closure, Outcome Assessment

Evidence Table 12. Vacuum-Assisted Closure, Results

Evidence Table 13. Vacuum-Assisted Closure, Study Quality Ratings

All studies used the V.A.C.® (Kinetic Concepts, Inc., KCI) device. Three studies included patients who primarily had pressure ulcers (Wanner, Schwarzl, Strub, et al., 2003; Ford, Reinhard, Yeh, et al., 2002; Joseph, Hamori, Bergman, et al., 2000). Joseph, Hamori, Bergman, et al. (2000) selected patients with chronic (i.e., ≥4 weeks' duration) nonhealing wounds of various etiologies, of which 78 percent were pressure ulcers. Pressure ulcers in the Ford, Reinhard, Yeh, et al. (2002) study were of 4 weeks duration or longer. Wanner, Schwarzl, Strub, et al. (2003) did not specify the duration of the pressure ulcers. Eginton, Brown, Seabrook, et al. (2003) included diabetic foot wounds not expected to heal within 1 month and McCallon, Knight, Valiulus, et al. (2000) selected diabetic foot wounds of more than 1 month duration.

The comparison of interventions was conventional/standard wound care (mainly, moist dressings changed at least once daily) versus vacuum-assisted closure in five studies (Moues, Vos, van den Bemd, et al., 2004; Wanner, Schwarzl, Strub, et al., 2003; Joseph, Hamori, Bergman, et al., 2000; Eginton, Brown, Seabrook, et al., 2003; McCallon, Knight, Valiulus, et al., 2000). One study (Ford, Reinhard, Yeh, et al., 2002) compared vacuum-assisted closure with the Healthpoint® system, consisting of gel/pad products (Iodosorb®, Iodoflex™, Panafil®).

Study quality. All six studies were rated poor in quality (Evidence Table 13). Only one study made it clear that an adequate randomization method was used (i.e., sealed envelopes in Moues, Vos, van den Bemd, et al., 2004). One study (McCallon, Knight, Valiulus, et al., 2000) used an allocation method that was probably inadequate to be considered true randomization: coin flip to assign the first patient, then alternating group assignment.

No study indicated that groups were comparable on all three key baseline characteristics (age, wound duration, and wound size, see Evidence Table 9). Age was comparable between groups in these four studies: Moues, Vos, van den Bemd, et al. (2004); Wanner, Schwarzl, Strub, et al. (2003); Joseph, Hamori, Bergman, et al. (2000); and McCallon, Knight, Valiulus, et al. (2000). Vacuum-assisted closure patients were younger than control patients in the study by Ford, Reinhard, Yeh, et al. (2002). Wanner, Schwarzl, Strub, et al. (2003) assembled groups that were comparable in wound size. Wounds were smaller in the vacuum-assisted closure group than the control group in the study by Joseph, Hamori, Bergman, et al. (2000). None of the seven studies provided information on the comparability of groups with respect to wound duration. All studies gave clear descriptions of interventions. All studies also described reliable and valid measurement methods and three used blinded observers (Joseph, Hamori, Bergman, et al., 2000; Ford, Reinhard, Yeh, et al., 2002; Eginton, Brown, Seabrook, et al., 2003). One study (Ford, Reinhard, Yeh, et al., 2002) reported on the primary endpoint, complete healing, while McCallon, Knight, Valiulus, et al. (2000) provided data on time to satisfactory healing. Only Wanner, Schwarzl, Strub, et al. (2003) performed adjustment for confounders in the data analysis. Joseph, Hamori, Bergman, et al. (2000) provided the only intention-to-treat analysis.

Complete healing. The proportion of patients who achieved complete healing was reported in only one study (Ford, Reinhard, Yeh, et al., 2002). In the control group receiving gel products for full-thickness pressure ulcers, two of 15 wounds had complete healing within 8–10 weeks, compared with two of 20 in the vacuum-assisted closure group. The relative risk of complete healing was 0.75 with a 95 percent confidence interval from 0.12 to 4.73. Although the numerical rate of complete healing was lower in the vacuum-assisted closure group, the 95 percent confidence interval is quite wide and overlaps with 1.0, indicating a lack of statistical difference between groups.

McCallon, Knight, Valiulus, et al. (2000) defined a related outcome, “satisfactory healing,” as achieving definitive closure either by reaching a stage suitable for surgical intervention such as skin grafting (delayed primary intention) or by complete healing without surgical intervention (secondary intention). The mean time to satisfactory healing was 42.8 days in the control group and 22.8 days in the vacuum-assisted closure group, a difference that was not statistically significant in this study of 10 patients with diabetic foot wounds. Most of the vacuum-assisted closure wounds (4 of 5) were healed by delayed primary intention, while most of the control wounds (3 of 5) healed by secondary intention.

Facilitation of surgical closure. One study (Moues, Vos, van den Bemd, et al., 2004) reported on the time to readiness for surgical closure, among patients with full-thickness wounds of various etiologies. Log-rank test analysis of Kaplan-Meier time to readiness did not show any statistically significant differences between groups. The median time to readiness for surgical closure was 6 days for vacuum-assisted closure patients and 7 days for conventionally treated patients (p=0.19).

Change in wound area. Two studies of diabetic foot wounds reported nonsignificant results favoring vacuum-assisted closure over moist dressings in percent change in wound area, and one study of full-thickness wounds reported a similar finding. Eginton, Brown, Seabrook, et al. (2003) used a crossover design in six patients who first had 2 weeks of either moist dressings or vacuum-assisted closure, then switched to the other for 2 weeks. The mean change in area was an increase of 5.9 percent for the control intervention and a decrease of 16.4 percent for vacuum-assisted closure. In the study by McCallon, Knight, Valiulus, et al. (2000, n=10), the mean change in wound area in the control group was a gain of 9.5 percent, compared with a mean decrease of 28.4 percent in the vacuum-assisted closure group. In a subset of only 52 percent of the original group of patients with full-thickness wounds, Moues, Vos, van den Bemd, et al. (2004) found a significantly higher daily percent change in wound area among vacuum-assisted closure patients (3.8), compared with conventionally treated patients (1.7, p<0.05).

Change in wound volume. Four studies have reported on changes in wound volume: all three studies of pressure ulcers and one study on diabetic foot wounds. Wanner, Schwarzl, Strub, et al. (2003) included 22 patients with pelvic pressure ulcers. The endpoint was time to 50 percent decrease in wound volume. The mean time for traditional care was 28 days, compared with 27 days for vacuum-assisted closure. The unadjusted p value was 0.9, while adjustment for initial wound volume yielded a p value of 0.2. In the study by Ford, Reinhard, Yeh, et al. (2002) of 19 patients with full-thickness pressure ulcers, after 3–10 months, the mean percent change in wound volume was -42.1 percent in the group receiving gel products and -51.8 percent in the vacuum-assisted closure group (p=0.46). The group of 24 patients with nonhealing wounds (78 percent pressure ulcers) studied by Joseph, Hamori, Bergman, et al. (2000) were evaluated in two ways. First, the mean percent reduction in volume at 6 weeks was compared: 30 percent for standard wound care and 78 percent for vacuum-assisted closure (p=0.038). Second, a Cox proportional hazards model analysis found that use of vacuum-assisted closure and initial exposure of tendon or bone were significant predictors of time to greater than 95 percent reduction in volume. In the crossover study of six patients with diabetic foot wounds (Eginton, Brown, Seabrook, et al., 2003), the mean percent reduction in wound volume was 0.1 percent in the moist dressing phase and 59 percent in the vacuum-assisted closure phase (p<0.005).

Change in wound dimensions. Three studies report changes in length, width and depth of wounds, and two of the three studies report significant differences in favor of the vacuum-assisted closure group for one or more of these outcomes. In the Ford, Reinhard, Yeh, et al. (2002) study of pressure ulcers, the mean changes (cm) for the gel product group and vacuum-assisted closure group, respectively, were: -18.7 and -36.9 (length, p=0.10); -19.0 and -40.0 (width, p=0.11); and -31.0 and -33.6 (depth, p=0.90). Mean percent change in the Joseph, Hamori, Bergman, et al. (2000) study of nonhealing wounds, for standard care and vacuum-assisted closure, respectively, were: -38 and -46 (length, p=0.38); -35 and -63 (width, p=0.02); -20 and -66 (depth, p<0.001). In the crossover study by Eginton, Brown, Seabrook, et al. (2003) of diabetic foot wounds, the following comparisons of mean percent change values for moist dressings and vacuum-assisted closure, respectively, were observed: +6.7 and -4.3 (length, p>0.05); +2.4 and -12.9 (width, p>0.05); and -7.7 and -49 (depth, p<0.05).

Complications. Two studies reported data on complications during wound treatment. Ford, Reinhard, Yeh, et al. (2002) reported two deaths (group assignments not specified) and one vacuum-assisted closure patient with diabetes, hypertension, vascular insufficiency and sepsis who required distal lower extremity amputation. More cases of pre-existing osteomyelitis improved in the vacuum-assisted closure group (37.5 percent) than in the group receiving gel products (0 percent), but the difference was not statistically significant (p=0.25). In the Joseph, Hamori, Bergman, et al. (2000) study, eight of 18 wounds treated with standard care developed complications, compared with three of 18 vacuum-assisted closure wounds (p=0.0028). Complications included: fistulas; wound infection; osteomyelitis; and calcaneal fractures.

Biopsy results. The study comparing gel products and vacuum-assisted closure in 22 patients with full thickness pressure ulcers (Ford, Reinhard, Yeh, et al., 2002) reported quantitative biopsy results. The mean number of polymorphonuclear (PMN) leukocytes per high-powered field increased in the gel product group, but decreased in the vacuum-assisted closure group (p=0.13). Lymphocytes also increased in the gel product group and decreased in the vacuum-assisted closure group (p=0.41). The mean number of capillaries declined in both groups, but to a slightly lesser extent in the vacuum-assisted closure group (p=0.75).

Randomized trials in progress. KCI, the manufacturer of the V.A.C.® device, has shared protocol documents for 10 randomized trials in progress (Evidence Table 14). These protocols cover a wide variety of wound types, including burns, pressure ulcers, diabetic ulcers, traumatic and surgical wounds, venous stasis wounds, and diabetic wounds. Large sample sizes are planned, determined by power analyses. Sophisticated randomization techniques will be used in many trials. A wide range of outcomes will be assessed, often by a blinded observer. Plans to adjust for confounders in the analysis, if necessary, are common. Concerns remain about the criteria for allowable withdrawals, including: noncompliance; worsened condition; complications; and treatment difficulties/failures. If such withdrawals are excluded from analysis, it would constitute violation of the intention-to-treat principle.

Evidence Table 14. Vacuum-Assisted Closure, KCI Randomized Trials in Progress

KCI also furnished abstracts presented at the 2^nd World Union of Wound Healing Societies, which met in Paris, France, July 8–12, 2004. While these abstracts provide too little detail for meaningful analysis in this systematic review, they are summarized in Evidence Table 15 to document the progress of ongoing randomized trials.

Conclusions

This body of evidence is insufficient to support conclusions about the effectiveness of vacuum-assisted closure in the treatment of wounds. There are only six trials that met the inclusion criteria for this review and the included trials were of small size and poor quality. With the exception of one study of 54 patients with incomplete followup, all studies included fewer than 25 patients. The randomization method was clearly adequate in only one study. No study made it clear that groups were comparable on all three key baseline characteristics (age, wound duration, wound size). None provided group information about wound duration. A single study adjusted for confounders in the data analysis and another performed an intention-to-treat analysis.

Some outcomes in the available trials show a significant benefit for the vacuum-assisted closure group, while others do not. Only one study gave data on the probability of complete healing, showing no significant difference between groups. A study reporting time to satisfactory healing also found no significant difference between groups. One study found no difference between vacuum-assisted closure and control in time to readiness for surgical closure. Three studies reported on change in wound area; one of which found a difference between vacuum-assisted closure and control, while two did not. Among four studies addressing change in wound volume, two found a significant advantage for vacuum-assisted closure and two did not achieve statistical significance. One study found significant changes in wound width and depth for vacuum-assisted closure and another found it only for depth. It is possible that the lack of significant results in some or all of these trials result from a type II error. In most cases, the numerical results favor the vacuum-assisted closure group. Power calculations are lacking for these trials, but their small size raises the possibility that they are underpowered.

The randomized, controlled trial protocols provided by KCI outline much larger trials that are condition-specific and address many of the quality problems found in the published studies. If implemented and completed successfully as planned, these trials will provide substantial advances in the evidence base for vacuum-assisted closure therapy, and may allow more definitive conclusions on the efficacy of vacuum-assisted closure.

Evidence Table 15. Abstracts Presented at the 2nd World Union of Wound Healing Societies, Paris, France; July 8–12, 2004