NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Jonas D, Van Scoyoc E, Gerrald K, et al. Drug Class Review: Newer Diabetes Medications, TZDs, and Combinations: Final Original Report [Internet]. Portland (OR): Oregon Health & Science University; 2011 Feb.

Cover of Drug Class Review: Newer Diabetes Medications, TZDs, and Combinations

Drug Class Review: Newer Diabetes Medications, TZDs, and Combinations: Final Original Report [Internet].

Show details


Inclusion Criteria

All citations were reviewed for inclusion using the criteria described in Table 2. Studies meeting these criteria and comparing at least one of the drugs of interest with an eligible comparator were included. Eligible drugs and comparators are listed in Table 3.

Table 2. Study inclusion and exclusion criteria.

Table 2

Study inclusion and exclusion criteria.

Table 3. Eligible drugs and comparators.

Table 3

Eligible drugs and comparators.

Literature Search

To identify articles relevant to each key question we searched MEDLINE®, Embase, the Cochrane Library, and the International Pharmaceutical Abstracts. Initially, we conducted 5 separate searches to ensure overlap and consistency with the 3 reports that were being updated and to capture additional references relevant to the new inclusion criteria. We used the generic and brand names of included drugs, and study designs as search terms. We combined the results of all the searches and removed duplicate references. The full search strategies are presented in Appendix C. Update searches were conducted on July 28, 2010 to ensure that recent publications were captured.

We attempted to identify additional studies through hand searches of reference lists of included studies and reviews. In addition, we requested dossiers of published and unpublished information from the relevant pharmaceutical companies for this review. All received dossiers were screened for studies or data not found through other searches. All citations were imported into an electronic database (Endnote®X.0.2, Thomson Reuters).

Study Selection

Selection of included studies was based on the inclusion criteria created by the Drug Effectiveness Review Project participants, as described above. Two reviewers independently assessed titles and abstracts of citations identified through literature searches for inclusion using the criteria above. Full-text articles of potentially relevant citations were retrieved and again were assessed for inclusion by both reviewers. Disagreements were resolved by consensus. Results published only in abstract form were not included because inadequate details were available for quality assessment.

Data Abstraction

The following data were abstracted from included trials: study design; population characteristics, including sex, age, and ethnicity; eligibility and exclusion criteria; interventions; comparisons; numbers randomized or treated, and the numbers analyzed; and results for each outcome. We recorded intention-to-treat results when reported. If true intention-to-treat results were not reported, but loss to follow-up was very small, we recorded these results and noted that they were modified intention-to-treat results. In cases where only per protocol results were reported, we recorded these results and noted that they were per protocol results. We considered whether results were intention-to-treat, modified intention-to-treat, or per protocol when assessing the internal validity of studies (as described below). Data abstraction was performed by one reviewer and independently checked by a second reviewer and differences were resolved by consensus. When studies reported duration in number of months, we converted this to number of weeks by multiplying months by 4.33 and rounding up or down. Number of weeks is presented in the tables of study characteristics throughout the report. When recording data on lipids, we converted mmol/L to mg/dL. To convert total cholesterol and HDL and LDL cholesterol, we used the following formula: divide mmol/L by 0.0259 to get mg/dL. To convert triglycerides, we used the following formula: divide mmol/L by 0.0113 to get mg/dL.

Validity Assessment

Two reviewers independently assessed each study and differences were resolved by consensus. We assessed the internal validity (quality) of trials based on the predefined criteria (see These criteria are based on those developed by the US Preventive Services Task Force and the National Health Service Centre for Reviews and Dissemination (United Kingdom).9, 10 We rated the internal validity of each trial based on the methods used for randomization, allocation concealment, and blinding; the similarity of compared groups at baseline; maintenance of comparable groups; adequate reporting of dropouts, and attrition; loss to follow-up; and the use of intention-to-treat analysis. Trials that had a fatal flaw were rated poor quality; trials that met all criteria were rated good quality; the remainder were rated fair quality. As the fair-quality category is broad, studies with this rating vary in their strengths and weaknesses: The results of some fair-quality studies are likely to be valid, while others are only possibly valid. A poor-quality trial is not valid; the results are at least as likely to reflect flaws in the study design as a true difference between the compared drugs. A fatal flaw may be reflected by one aspect introducing a high risk of bias or by failure to meet combinations of items of the quality assessment checklist. We did not include poor quality studies in our analysis. A particular randomized trial might receive 2 different ratings, for different outcomes.

Observational studies included for the assessment of adverse events were also rated for quality. The criteria used reflect aspects of the study design that are particularly important for assessing adverse event rates.

Included systematic reviews were also rated for quality. We rated the internal validity based a clear statement of the questions(s); reporting of inclusion criteria; methods used for identifying literature (the search strategy), validity assessment, and synthesis of evidence; and details provided about included studies. These studies were categorized as good when all criteria were met.

Grading the Strength of Evidence

We graded strength of evidence based on the guidance established for the Evidence-based Practice Center Program of the Agency for Healthcare Research and Quality.11 Developed to grade the overall strength of a body of evidence, this approach incorporates 4 key domains: risk of bias (includes study design and aggregate quality), consistency, directness, and precision of the evidence. It also considers other optional domains that may be relevant for some scenarios, such as a dose-response association, plausible confounding that would decrease the observed effect, strength of association (magnitude of effect), and publication bias. We considered all placebo-controlled evidence to be indirect (not directly comparing medications). We considered all evidence from intermediate outcomes (e.g. HbA1c) to be indirect (not directly reporting health outcomes).

Table 4 describes the grades of evidence that can be assigned. Grades reflect the strength of the body of evidence to answer key questions on the comparative effectiveness, efficacy, and harms of the drugs included in this review. Grades do not refer to the general efficacy or effectiveness of pharmaceuticals. Two reviewers assessed each domain for each outcome and differences were resolved by consensus.

Table 4. Definitions of the grades of overall strength of evidence.

Table 4

Definitions of the grades of overall strength of evidence.

We graded the strength of evidence for the outcomes deemed to be of greatest importance to decision makers and those most commonly reported in the literature. For example, these included HbA1c and weight changes, among others. Because of time and resource constraints we did not grade the strength of evidence for every possible outcome reported everywhere in the included literature.

Data Synthesis

We constructed evidence tables showing the study characteristics, quality ratings, and results for all included studies. We reviewed studies using a hierarchy of evidence approach, where the best evidence is the focus of our synthesis for each question, population, intervention, and outcome addressed. Studies that evaluated one included drug of interest against another provided direct evidence of comparative effectiveness and adverse event rates. Where possible, these data are the primary focus. Direct comparisons (i.e., head-to-head comparisons of included medications) were preferred over indirect comparisons; similarly, effectiveness and long-term safety outcomes were preferred to efficacy and short-term tolerability outcomes.

In theory, trials that compare included drugs of interest with other drug classes (i.e., active controls) or with placebos can also provide evidence about effectiveness. This is known as an indirect comparison and can be difficult to interpret for a number of reasons, primarily heterogeneity of trial populations, interventions, and outcomes assessment. Data from indirect comparisons are used to support direct comparisons, where they exist, and are used as the primary comparison where no direct comparisons exist. Indirect comparisons should be interpreted with caution.

Quantitative analyses were conducted using meta-analyses of outcomes reported by a sufficient number of studies that were homogeneous enough that combining their results could be justified. In order to determine whether meta-analysis could be meaningfully performed, we considered the quality of the studies and the heterogeneity among studies in design, patient population, interventions, and outcomes. When meta-analysis could not be performed, the data were summarized qualitatively. Since a large number of recent good quality meta-analyses were available for comparisons of the TZDs with other medications of interest to this review, we did not conduct our own meta-analyses for those comparisons (except for the head-to-head comparison of pioglitazone and rosiglitazone for HbA1c).

Random-effects models were used to estimate pooled effects.12 Forest plots graphically summarize results of individual studies and of the pooled analysis.13

The Chi-squared statistic and the I2 statistic (the proportion of variation in study estimates due to heterogeneity) were calculated to assess heterogeneity in effects between studies.14, 15 An I2 from 0 to 40% might not be important, 30% to 60% may represent moderate heterogeneity, 50% to 90% may represent substantial heterogeneity, and ≥75% represents considerable heterogeneity.16 The importance of the observed value of I2 depends on the magnitude and direction of effects and on the strength of evidence for heterogeneity (e.g. P value from the chi-squared test, or a confidence interval for I2). Whenever including a meta-analysis with considerable statistical heterogeneity in this report, we provide an explanation for doing so, considering the magnitude and direction of effects. .16 Potential sources of heterogeneity were examined by analysis of subgroups of study design, study quality, patient population, and variation in interventions. Quantitative analyses were conducted using Stata version 11.1 and Comprehensive Meta Analysis version 2.2.055.

When describing conclusions and key findings in this report, we sometimes refer to “no difference” between two treatments. We use this wording to indicate that the available evidence did not support a statistically or clinically significant difference between the two treatments.

Peer Review

We requested and received peer review of the report from 3 content or methodology experts. Their comments were reviewed and, where possible, incorporated into the final document. All comments and the authors’ proposed actions were reviewed by representatives of the participating organizations of the Drug Effectiveness Review Project before finalization of the report. Names of peer reviewers for the Drug Effectiveness Review Project are listed at

Public Comment

This report was posted to the Drug Effectiveness Review Project website for public comment. We received comments from 6 persons, 6 representing pharmaceutical companies, 0 representing professional or advocacy organizations, and 0 individuals with no reported affiliation.

Copyright © 2011, Oregon Health & Science University.
Bookshelf ID: NBK54210
PubReader format: click here to try


Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...