When added to metformin, dipeptidyl peptidase 4 (DPP-4) inhibitors and sulfonylureas reduce hemoglobin A1c to a similar degree, but other differences may favor the former, new research suggests.
The findings, derived from a novel approach that analyzed real-world data from more than 246 million patients, were published online August 24 in JAMA Open, by Rohit Vashisht, PhD, from the Center for Biomedical Informatics Research, Stanford University School of Medicine, California, and colleagues with the Observational Health Data Sciences and Informatics in New York City.
None of the drugs raised the risk for kidney disorders, according to the analysis, which examined the effects of sulfonylureas, DPP-4 inhibitors, and thiazolidinediones added to metformin.
However, sulfonylureas were associated with a small increased risk for myocardial infarction and eye disorders compared with DPP-4 inhibitors.
"Large-scale characterization of the effectiveness of type 2 diabetes therapy via an open collaborative research network suggests DPP-4 inhibitors over sulfonylureas in patients with diabetes for whom metformin was the first-line treatment," principal investigator Nigam H. Shah, MBBS, PhD, also from the Center for Biomedical Informatics Research, Stanford University School of Medicine, told Medscape Medical News.
Asked to comment, M. Sue Kirkman, MD, professor of medicine and medical director of the Diabetes Care Center Clinical Trials Unit at the University of North Carolina School of Medicine in Chapel Hill, noted, "For clinicians, this study supports prior thinking that most oral agents lower HbA1c about the same amount on average. The concerns about sulfonylureas being associated with cardiovascular disease are again raised."
Kirkman added, "For now, clinicians need to continue to individualize therapy beyond metformin, taking into account outcomes of importance to patients, such as cost, side effects, hypoglycemia, and weight gain, and incorporating what we know from cardiovascular outcome trials for patients with known cardiovascular disease."
Novel Approach Allows for Examination of Heterogeneous Data
For the study, the authors used patient data from eight healthcare systems in three countries. To allow incorporation into a single dataset and subsequent analysis, they standardized the data with regard to terminology and structure using the Observational Medical Outcomes Partnership Common Data Model (OMOP CDM).
Shah said that the platform allows for large datasets to inform clinical decision-making. "This study is an example of a large multinational open collaborative research network, which can produce evidence at scale and is made feasible via the adoption of a common data model and open-source analytical tools."
The platform is being used in many areas of medicine, including assessment of treatments for hypertension, fracture prevention, and thyroid conditions.
Kirkman commented, "Since we cannot do randomized controlled trials to answer every question regarding medical therapy of type 2 diabetes, this type of observational big data analysis is very important. However, concerns about unmeasured confounders are always present."
For example, she said, "In the United States, sulfonylureas are often prescribed for patients with limited financial means, so I am always concerned that the sulfonylurea findings in nonrandomized trials may be confounded by socioeconomic status."
She also notes that although the data sources included one from France and another from South Korea, the majority of patients were American.
The sources used were Truven MarketScan Commercial Claims and Encounters; Columbia University Medical Center; IQVIA Disease Analyzer (France); Truven MarketScan Medicare; Mount Sinai Icahn School of Medicine; Optum Clinformatics Data Mart; Ajou University School of Medicine (South Korea); and Stanford University.
Specific combinations of drugs, diagnostic codes, and laboratory test results from those records were used to identify patients with type 2 diabetes who had received a second-line treatment in addition to metformin.
Only sulfonylureas, DPP-4 inhibitors, and thiazolidinediones were examined because not enough patients had been prescribed glucagon-like peptide-1 (GLP-1) receptor agonists at some of the sites. In addition, the authors excluded sodium-glucose cotransporter type 2 (SGLT2) inhibitors because "some of the collaborators on this open science project were not allowed to work on those drugs given their employment regulations," Shah said.
Kirkman called those omissions "unfortunate...since both classes have been shown to provide cardiovascular benefit in high-risk individuals."
From the Optum dataset, one-to-one propensity matching based on pretreatment drug prescriptions, disease diagnosis, procedure, and demographics yielded 24,777 participants in each drug treatment group. Although the uncalibrated value suggested that HbA1c reduction was greater with sulfonylureas, after calibration of the P value using negative controls, the difference was nonsignificant (P = .81).
In a meta-analysis across all datasets comprising 246,558,805 patients, there was no significant difference between sulfonylureas versus DPP-4 inhibitors in HbA1c reduction to 7% or less (consensus hazard ratio [HR], 0.99).
There was a small increased risk for myocardial infarction and eye disorders for sulfonylureas compared with DPP-4 inhibitors (HR, 1.12 and 1.15, respectively) in the meta-analysis, although the calibrated P values showed that the associations were not significant at individual sites. No differences were seen regarding kidney disorders (consensus HR, 1.09).
Comparisons of sulfonylureas and DPP-4 inhibitors versus thiazolidinediones showed no differences in achievement of HbA1c levels of 7% or less, myocardial infarction, kidney disorders, or eye disorders, either after P value recalibration or in the meta-analysis.
Kirkman said the ongoing National Institutes of Health-funded Glycemia Reduction Approaches in Diabetes (GRADE) study is designed to address a similar question. It is a large randomized trial comparing the addition one of four drug classes — sulfonylureas, DPP-4 inhibitors, GLP-1 receptor agonists, or basal insulin — to metformin, in patients with a broad age range and significant racial diversity. Results, expected in 2021, should help answer the second-line drug question in terms of HbA1c reduction, she said, although the study isn't powered to examine differences in long-term complications.
The study was supported by grants from the National Library of Medicine and National Institute of General Medical Sciences. Additional support is listed in the article. Co-authors reported receiving consulting fees or honoraria, serving as an advisor, or holding equity in Janssen, GlaxoSmithKline, AstraZeneca, Hoffman-La Roche, LAM Therapeutics, NuMedii, Ayasdi, and Ontomics, Hebta, Melax Technologies, and More Health. Two co-authors are employees of Janssen Research and Development. Kirkman's institution has received research support from Novo Nordisk and Theracos for studies of type 2 diabetes medications.
JAMA Network Open. 2018;1:e181755. Full text
Medscape Medical News © 2018 WebMD, LLC
Send comments and news tips to email@example.com.
Cite this: Big Data Confirm Type 2 Diabetes Treatment Approach - Medscape - Aug 31, 2018.