• DOI:10.7326/M15-2650
• Corpus ID: 32016657

Diabetes Medications as Monotherapy or Metformin-Based Combination Therapy for Type 2 Diabetes

@article{Maruthur2016DiabetesMA,
  title={Diabetes Medications as Monotherapy or Metformin-Based Combination Therapy for Type 2 Diabetes},
  author={Nisa M Maruthur and Eva Tseng and Susan M. Hutfless and Lisa M Wilson and Catalina Suarez‐Cuervo and Zackary D Berger and Yue Chu and Emmanuel Iyoha and Jodi B Segal and Shari D. Bolen},
  journal={Annals of Internal Medicine},
  year={2016},
  volume={164},
  pages={740-751}
}

• N. Maruthur, Eva Tseng, +7 authors S. Bolen
• Published 19 April 2016
• Medicine
• Annals of Internal Medicine

Type 2 diabetes and its complications are a substantial public health burden, affecting 9.3% of the U.S. adult population (1, 2). Most patients with type 2 diabetes eventually require glucose-lowering pharmacologic therapy, with a goal of reducing long-term complications. More than 7 classes of diabetes medications, which differ in their effects on glucose-lowering, safety, and other important outcomes, are recommended as first- or second-line therapy (3, 4). The Agency for Healthcare Research and Quality (AHRQ) published 2 detailed reports comparing monotherapies and medication combinations for adults with type 2 diabetes (5, 6), but many new medications have been approved by the U.S. Food and Drug Administration (FDA) since then. These include medications in a new class, the sodiumglucose cotransporter 2 (SGLT-2) inhibitors, and several new dipeptidyl peptidase-4 (DPP-4) inhibitors and glucagon-like peptide-1 (GLP-1) receptor agonists. Evidence on the newer versus older diabetes medications continues to amass, and we performed an updated systematic review of the comparative effectiveness and safety of medications for type 2 diabetes in terms of intermediate, long-term, and safety outcomes. This review focused on monotherapy comparisons and combination comparisons that include metformin, with the aim of providing the range of stakeholders, including patients and clinicians, a synthesis of the current evidence on the most common monotherapies and combination therapies used to treat type 2 diabetes. Methods With input from a technical expert panel and representatives from AHRQ, we developed a protocol (available at www.effectivehealthcare.ahrq.gov). The full evidence report (7) has additional details on the methods and results, including search strategies and evidence tables. Data Sources and Searches We searched MEDLINE, EMBASE, and the Cochrane Central Register of Controlled Trials. The search strategy included terms for the diabetes medications of interest and terms for type 2 diabetes (Appendix Table 1 provides the MEDLINE search strategy). Appendix Table 1. MEDLINE Search String We ran the search developed for the prior review with the date restrictions of April 2009 through March 2015. We ran an additional search that included the Medical Subject Heading terms and text words for all of the new medications included in this update, without any date restrictions. After completion of the evidence report (7), we searched MEDLINE through December 2015, and updated our findings where the strength of evidence changed from low or insufficient to moderate or high. We hand-searched the reference lists of all newly included articles and relevant systematic reviews. In addition, we searched ClinicalTrials.gov to identify relevant registered trials and reviewed the FDA Web site for any unpublished additional studies relevant to the topic as part of our gray-literature search. Study Selection Two reviewers independently screened titles, abstracts, and full-text articles for inclusion and resolved differences through consensus. We included English-language studies of nonpregnant adults with type 2 diabetes that evaluated at least 3 months of use of a diabetes medication or drug combination of interest. We included head-to-head monotherapy comparisons of metformin, thiazolidinediones, sulfonylureas, DPP-4 inhibitors, SGLT-2 inhibitors, and GLP-1 receptor agonists; comparisons of metformin alone with a metformin-based combination; and comparisons of metformin-based combinations where the second medication was one of the monotherapies described above or a basal or premixed insulin (Appendix Table 2). We excluded studies that did not specify adjunctive medications. We excluded acarbose because of its infrequent use and the absence of new key studies that would substantially change the conclusions from our original report (5, 8). Appendix Table 2. Priority Medication Comparisons Included for Each Key Question We included randomized, controlled trials (RCTs) that evaluated all-cause mortality, macrovascular outcomes, microvascular outcomes, intermediate outcomes, or safety (7). We also included observational studies that adequately accounted for confounding, although not for the intermediate outcomes. Data Extraction, Quality, and Applicability Assessment Using standardized forms, reviewers extracted information on the general study and participant characteristics, interventions, comparisons, and the outcome results. A second reviewer confirmed the abstracted data. Two independent reviewers assessed risk of bias in individual RCTs by using the criteria of Jadad and colleagues (9), as in our prior review (6). We used the Downs and Black tool for assessment of the risk of bias for the nonrandomized trials and observational studies (10). To assess study applicability, we evaluated whether the study population, interventions, outcomes, and settings were similar to usual care for people with type 2 diabetes in the United States. Data Synthesis and Analysis We created a set of detailed evidence tables. We conducted meta-analyses when data were sufficient (from at least 3 trials) and studies were sufficiently homogenous with respect to key variables (population characteristics, study duration, and medication dosing). When a trial had more than 1 study group, we included in the quantitative pooling the study group with drug doses and study durations most similar to the other studies for that comparison and outcome. We pooled the mean difference between groups for continuous outcomes and calculated pooled odds ratios for the dichotomous outcomes using the intention-to-treat denominator. We evaluated the heterogeneity among the trials by using the I 2 statistic (11). We generated summary treatment effects with the random-effects model estimated by using the DerSimonian and Laird method in settings of low heterogeneity (I 2< 50%) (12) and the profile likelihood estimate in settings of high heterogeneity (I 2 50%) (13). Grading of the Evidence Adapting an evidence grading scheme recommended in the AHRQ guide for conducting comparative effectiveness reviews (14), 2 reviewers sequentially graded the studies' limitations, consistency, directness, precision, and potential reporting bias for the evidence on each outcome and comparison. We graded the evidence separately for RCTs and observational studies. The final evidence grade and conclusion were based on the RCTs and could be strengthened by evidence from observational studies with few study limitations. High strength of evidence indicates that the evidence probably reflects the true effect; moderate strength indicates that further research may change the result; and low strength indicates low confidence that the evidence reflects the true effect, and further research is very likely to change the result. Insufficient evidence indicates that evidence is unavailable or the body of evidence has unacceptable deficiencies, precluding a conclusion. Role of the Funding Source The AHRQ reviewed the protocol and report but did not participate in the literature search, determination of study eligibility, analysis, interpretation of findings, or preparation, review, or approval of the manuscript for publication. Results Study and Quality Characteristics We included 204 studies, 116 of which are newly identified, in this updated review (Appendix Figure). Eighty-one percent were RCTs. Appendix Table 3 shows the number and design of studies, by outcome. Fifty studies were multicontinental; the others were conducted in Europe (55 studies), Asia (39 studies), and the United States (34 studies). Study durations ranged from 3 months to 8 years, but only 22 studies (7 RCTs) lasted longer than 2 years. Only 1 RCT specified a cardiovascular outcome as a primary outcome (15). Appendix Table 3. Number of Studies for Each Type of Outcome* Appendix Figure. Summary of evidence search and selection. FDA = U.S. Food and Drug Administration; RCT = randomized, controlled trial. * Total may exceed the number in the corresponding box because articles could be excluded for >1 reason at this level. Comorbid condition restrictions were end-stage renal disease, end-stage liver disease, cancer, new-onset diabetes after transplant, or a cardiovascular event within 3 mo (such as acute coronary syndrome, acute myocardial infarction, postcoronary artery bypass graft surgery, or drug-eluting stents). Study participants were overweight or obese men and women with baseline hemoglobin A1c levels between 7% and 9%. About 45% of the RCTs did not report race/ethnicity. When reported, only 10% to 30% of the enrolled population was of nonwhite race. Most studies excluded older persons and those with clinically significant comorbid conditions. Of the randomized trials, approximately one half reported on their randomization scheme. Sixty-five percent reported double-blinding, but most did not report steps taken to ensure double-blinding. Of the newly included RCTs, losses to follow-up exceeded 20% in the majority (>70%) of trials lasting 1 year or more and in 24% of those lasting less than 1 year. Most studies used the last-observation-carried-forward approach for analysis of intermediate outcomes. Of the newly included trials, use of rescue therapy was reported in 35%, but 42% did not report on this. All included observational studies were at low risk of bias. Sixty-seven percent of studies reported receiving funding from pharmaceutical companies. We did not identify substantive reporting bias that would have affected our findings. All-Cause Mortality and Macrovascular and Microvascular Outcomes Although we included 65 new studies (52 RCTs and 13 observational studies) for these outcomes in this review, the trials were largely 1 year or less in duration, with few or no events; the evidence was insufficient or of low strength for almost all comparisons for these outcomes. However, we found moderate strength of evidence that metformin monoth