Use of insulin or sulfonylureas as second-line treatment in adults with type 2 diabetes is associated with increased cardiovascular risk, whereas use of newer classes of glucose-lowering drugs is not, new real-world research from the United States indicates.
The findings, from a retrospective analysis of national administrative claims data, were published online published online in JAMA Network Open by Matthew J. O'Brien, MD, of the Division of General Internal Medicine and Geriatrics, Department of Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, and colleagues.
Among more than 130,000 insured adults with type 2 diabetes who required a second glucose-lowering agent after metformin, use of insulin or sulfonylureas was associated with consistent cardiovascular harm compared with dipeptidyl peptidase 4 (DPP-4) inhibitors, which have been shown to have a neutral cardiovascular effect.
On the other hand, glucagon-like peptide-1 (GLP-1) receptor agonists, sodium-glucose cotransporter 2 (SGLT2) inhibitors, and thiazolidinediones (TZDs) were not associated with cardiovascular harm compared with DPP-4 inhibitors, but they also didn't produce the significant cardiovascular benefit that has been demonstrated in randomized clinical outcome trials of these agents in patients with type 2 diabetes and established cardiovascular disease.
Such high-risk populations have typically been necessary for statistical power in US Food and Drug Administration-mandated cardiovascular outcomes trials(CVOTs), but aren't representative of the adult type 2 diabetes population as a whole, of whom just 18% have established cardiovascular disease. Moreover, CVOTs are conducted on just one drug compared with placebo.
New Study Targets Area of Significant Clinical Uncertainty
"To date, no studies have directly compared the cardiovascular effects of all contemporary [glucose-lowering drug] options among patients starting second-line therapy. By examining cardiovascular outcomes among patients initiating second-line [glucose-lowering drugs] in the real world, this study aimed to complement findings from individual drug trials and further inform [glucose-lowering drug] choices for the broad population of patients currently receiving these medications," the investigators say.
In an accompanying editorial, Alison Callahan, PhD, and Nigam H. Shah, MBBS, PhD, both of the Center for Biomedical Informatics Research, Stanford University School of Medicine, California, praise the study, noting that it "targets an area of significant clinical uncertainty with the potential to inform the treatment of millions of individuals with type 2 diabetes," and in doing so "makes an important contribution to this area."
Callahan and Shah note that the findings agree with their recent study examining the real-world impact of various classes of second-line glucose-lowering agents on glycemic control and complication rates, including myocardial infarction. This new study makes "a valuable contribution" by adding GLP-1 receptor agonists.
Both studies, they note, "leverage observational data that capture details of healthcare processes and patient outcomes for millions of lives, with significant longitudinal coverage."
Basal Insulin, Sulfonylureas Robustly Associated With CV Harm
The current study included 132,737 adults with type 2 diabetes enrolled in commercial or Medicare Advantage health insurance plans during 2011-2015. All had initiated a second-line glucose-lowering drug, mostly along with metformin. The data were analyzed from January 2017 to October 2018.
Overall, 5.5% had a history of cardiovascular events before starting treatment with the index second-line agent.
Of the prescription fills for those agents, 47.6% were sulfonylureas, 21.8% DPP-4 inhibitors, 12.2% basal insulin, 8.6% GLP-1 agonists, 5.6% TZDs, and 4.3% SGLT2 inhibitors.
The investigators established the DPP-4 inhibitor users as the comparison group because data have shown that class to have a neutral effect on cardiovascular outcomes.
The primary outcome was time to first cardiovascular event after starting the second-line agent, with events defined as hospitalization for congestive heart failure, stroke, ischemic heart disease, or peripheral artery disease. There were 3480 such events during 169,384 person-years of follow-up.
Relative to starting treatment with a DPP-4 inhibitor, and following adjustment for patient, prescriber, and health plan characteristics, the risk for composite cardiovascular events was 36% higher in the sulfonylurea group (HR, 1.36) and more than double with basal insulin (HR, 2.03).
This corresponds to numbers needed to harm during 2 years of treatment with sulfonylureas and basal insulin of 103 and 37, respectively.
Increased relative cardiovascular risk associated with use of sulfonylureas or basal insulin was observed across all individual cardiovascular outcomes and remained "robust" in sensitivity analyses, O'Brien and colleagues report.
Of concern, they point out, "Despite the observed cardiovascular harms associated with initiating sulfonylureas and basal insulin, prescriptions for these two [drug] classes were filled by 60% of patients in our nationwide analysis."
Newer Agents Don't Show Harm or Benefit in Real-World T2D Population
Among the newer agents, use of a GLP-1 agonist was associated with a significantly lower adjusted risk of composite cardiovascular events compared with DPP-4 inhibitor use (hazard ratio, 0.78; 95% CI, 0.63 - 0.96). However, that benefit lost significance in several sensitivity analyses.
The CV event rates after starting treatment with either SGLT2 inhibitors or TZDs weren't significantly different from those of DPP-4 inhibitors (HR, 0.81 and HR, 0.92, respectively).
O'Brien and colleagues say their work will be complemented by the ongoing randomized Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness (GRADE) study, which is comparing long-term glycemic efficacy of a sulfonylurea (glimepiride), a DPP-4 inhibitor (sitagliptin), a GLP-1 agonist (liraglutide), and basal insulin (glargine) added to metformin. Unfortunately, GRADE doesn't include any SGLT2 inhibitors.
In conclusion, the researchers say their new findings "raise concerns about the cardiovascular safety of sulfonylureas and basal insulin," compared with newer glucose-lowering drugs and suggest that short-term cardiovascular outcomes of newer glucose-lowering drug classes may be similar among patients starting second-line treatment.
In Summary; Therefore, clinicians may consider prescribing GLP-1 agonists, DPP-4 inhibitors, or SGLT2 inhibitors more routinely after metformin rather than sulfonylureas or basal insulin.
JAMA Network Open. Published online December 21, 2018.
Abstract and full text of the article free
December 21, 2018
Association of Second-line Antidiabetic Medications With Cardiovascular Events Among Insured Adults With Type 2 Diabetes
Matthew J. O’Brien, MD, MSc1,2; Susan L. Karam, MD3; Amisha Wallia, MD, MS1,3; et alRaymond H. Kang, MA1; Andrew J. Cooper, MSc2; Nicola Lancki, MPH2; Margaret R. Moran, MPH1,2,4; David T. Liss, PhD1,2; Theodore A. Prospect, FSA, MAAA5; Ronald T. Ackermann, MD, MPH1,2,3
JAMA Netw Open. 2018;1(8):e186125. doi:10.1001/jamanetworkopen.2018.6125
Question Are second-line antidiabetic medications (ADMs) associated with cardiovascular events among insured adult patients with type 2 diabetes who are initiating second-line therapy?
This cohort study of 132 737 adults with type 2 diabetes found that glucagon-like peptide 1 receptor agonists, dipeptidyl peptidase 4 inhibitors, and sodium-glucose cotransporter 2 inhibitors were associated with similar cardiovascular outcomes. Sulfonylureas and basal insulin were associated with comparatively higher cardiovascular risk.
Clinicians may consider prescribing newer ADM classes more routinely after metformin rather than sulfonylureas or basal insulin.
Understanding cardiovascular outcomes of initiating second-line antidiabetic medications (ADMs) may help inform treatment decisions after metformin alone is not sufficient or not tolerated. To date, no studies have compared the cardiovascular effects of all major second-line ADMs during this early decision point in the pharmacologic management of type 2 diabetes.
To examine the association of second-line ADM classes with major adverse cardiovascular events.
Design, Setting, and Participants Retrospective cohort study among 132 737 insured adults with type 2 diabetes who started therapy with a second-line ADM after taking either metformin alone or no prior ADM. This study used 2011-2015 US nationwide administrative claims data. Data analysis was performed from January 2017 to October 2018.
Dipeptidyl peptidase 4 (DPP-4) inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, sodium-glucose cotransporter 2 (SGLT-2) inhibitors, thiazolidinediones (TZDs), basal insulin, and sulfonylureas or meglitinides (both referred to as sulfonylureas hereafter). The DPP-4 inhibitors served as the comparison group in all analyses.
Main Outcomes and Measures
The primary outcome was time to first cardiovascular event after starting the second-line ADM. This composite outcome was based on hospitalization for the following cardiovascular conditions: congestive heart failure, stroke, ischemic heart disease, or peripheral artery disease.
Among 132 737 insured adult patients with type 2 diabetes (men, 55%; aged 45-64 years, 58%; white, 63%), there were 3480 incident cardiovascular events during 169 384 person-years of follow-up. Patients were censored after the first cardiovascular event, discontinuation of insurance coverage, transition from International Classification of Diseases, Ninth Revision (ICD-9) to end of ICD-9 coding, or 2 years of follow-up.
After adjusting for patient, prescriber, and health plan characteristics, the risk of composite cardiovascular events after starting GLP-1 receptor agonists was lower than DPP-4 inhibitors (hazard ratio [HR], 0.78; 95% CI, 0.63-0.96), but this finding was not significant in all sensitivity analyses.
Cardiovascular event rates after starting treatment with SGLT-2 inhibitors (HR, 0.81; 95% CI, 0.57-1.53) and TZDs (HR, 0.92; 95% CI, 0.76-1.11) were not statistically different from DPP-4 inhibitors.
The comparative risk of cardiovascular events was higher after starting treatment with sulfonylureas (HR, 1.36; 95% CI, 1.23-1.49) or basal insulin (HR, 2.03; 95% CI, 1.81-2.27) than DPP-4 inhibitors.
Conclusions and Relevance
Among insured adult patients with type 2 diabetes initiating second-line ADM therapy, the short-term cardiovascular outcomes of GLP-1 receptor agonists, SGLT-2 inhibitors, and DPP-4 inhibitors were similar.
Higher cardiovascular risk was associated with use of sulfonylureas or basal insulin compared with newer ADM classes. Clinicians may consider prescribing GLP-1 receptor agonists, SGLT-2 inhibitors, or DPP-4 inhibitors more routinely after metformin rather than sulfonylureas or basal insulin.
Editorial free 2 pages
Diabetes and Endocrinology
December 21, 2018
A Second Opinion From Observational Data on Second-line Diabetes Drugs
Alison Callahan, MISt, PhD1; Nigam H. Shah, MBBS, PhD1
Author Affiliations Article Information
Recommendations vary for second-line treatment of type 2 diabetes, with little consensus even among clinical guidelines1,2 in which the primary end point is maintenance of a blood hemoglobin A1c level less than 7%. Multiple prior studies also have contradictory results in terms of risk of adverse outcomes, including cardiovascular events and kidney disease.3,4
The study by O’Brien et al5 makes an important contribution to this area, by assessing the effectiveness of second-line treatment options for type 2 diabetes in reducing the risk of cardiovascular events (stroke, congestive heart failure, ischemic heart disease, and peripheral artery disease). Analyzing a large commercial claims data set, they found that risk of cardiovascular events did not differ between either sodium-glucose cotransporter 2 inhibitors or thiazolidinediones and dipeptidyl peptidase 4 inhibitors, and that risk was increased for sulfonylureas or meglitinides and basal insulin in comparison with dipeptidyl peptidase 4 inhibitors.
Their findings agree with our recent study6 examining the association between classes of second-line type 2 diabetes therapies and blood glucose control as well as risk of adverse events (myocardial infarction, kidney disorders, and eye disorders). O’Brien et al5 make a valuable contribution by examining an additional drug class, glucagon-like peptide 1 receptor agonists, and finding that, in some analyses, risk of cardiovascular events was lower for glucagon-like peptide 1 receptor agonists in comparison with dipeptidyl peptidase 4 inhibitors.
Both the study by O’Brien et al5 and our previous study leverage observational data that capture details of health care processes and patient outcomes for millions of lives, with significant longitudinal coverage. Such data are increasingly accessible via commercial claims databases available for purchase and data provider networks such as the Observational Health Data Sciences and Informatics initiative,7 which also provides a common data model and an associated suite of open source analytic tools. Observational studies carried out at the scale enabled by such resources make it possible to investigate the efficacy of an ever-growing pool of treatments and their associated risks of adverse outcomes. The increasing ease of conducting such studies with these large data sources makes it essential both to conduct rigorous analyses and to make the details of these analyses transparent.
The study by O’Brien et al5 and the experience in the Observational Health Data Sciences and Informatics network has surfaced the following essential properties of rigorous and meaningful observational studies: (1) clearly stating the type of question and study; (2) exploratory analyses of the data source to ensure that the data are appropriate to answer the question of interest in terms of existence of necessary variables (eg, hemoglobin A1cblood test results) and years covered, especially when studying drugs that enter the market in specific years; (3) reporting precise and reproducible cohort definitions; (4) conducting sensitivity analyses and reporting their results in the main conclusions; (5) empirically calibrating significance thresholds8 and confidence intervals9; and (6) emphasizing study replication across sites and data sources.10
Study replication is especially important because if multiple researchers ask the same clinical question over time, across data sources and settings, and with differing study designs, the conclusions drawn from them are collectively stronger. The study by O’Brien et al5 targets an area of significant clinical uncertainty with the potential to inform the treatment of millions of individuals with type 2 diabetes. The study states the question type, clearly reports cohort definitions, and conducts sensitivity analyses (eg, note that after performing sensitivity analyses, they found that the seemingly lower risk of composite cardiovascular events with glucagon-like peptide 1 receptor agonists was no longer significant).
To advance a learning health system as a community, clinical researchers need to take advantage of observational data sources. To generate findings that can be trusted, clinical researchers need to adopt good practices in the analyses of observational data. JAMA Network Open, with its emphasis on ensuring rigorous, transparent, and reproducible studies, is leading the way in making the design and conduct of such research a norm.
Marathe PH, Gao HX, Close KL. American Diabetes Association standards of medical care in diabetes 2017. J Diabetes. 2017;9(4):320-324. doi:10.1111/1753-0407.12524PubMedGoogle ScholarCrossref
Garber AJ, Abrahamson MJ, Barzilay JI, et al. Consensus statement by the American Association of Clinical Endocrinologists and American College of Endocrinology on the comprehensive type 2 diabetes management algorithm—2017 executive summary. Endocr Pract. 2017;23(2):207-238. doi:10.4158/EP161682.CSPubMedGoogle ScholarCrossref
Bennett WL, Maruthur NM, Singh S, et al. Comparative effectiveness and safety of medications for type 2 diabetes: an update including new drugs and 2-drug combinations. Ann Intern Med. 2011;154(9):602-613. doi:10.7326/0003-4819-154-9-201105030-00336PubMedGoogle ScholarCrossref
Wilkinson SV, Tomlinson LA, Iwagami M, Stirnadel-Farrant HA, Smeeth L, Douglas I. A systematic review comparing the evidence for kidney function outcomes between oral antidiabetic drugs for type 2 diabetes. Wellcome Open Res. 2018;3:74. doi:10.12688/wellcomeopenres.14660.1PubMedGoogle ScholarCrossref
O’Brien MJ, Karam SL, Wallia A, et al. Association of second-line antidiabetic medications with cardiovascular events among insured adults with type 2 diabetes. JAMA Netw Open. 2018;1(8):e186125. doi:10.1001/jamanetworkopen.2018.6125ArticleGoogle Scholar
Vashisht R, Jung K, Schuler A, et al. Association of hemoglobin A1c levels with use of sulfonylureas, dipeptidyl peptidase 4 inhibitors, and thiazolidinediones in patients with type 2 diabetes treated with metformin: analysis from the Observational Health Data Sciences and Informatics initiative. JAMA Netw Open. 2018;1(4):e181755. doi:10.1001/jamanetworkopen.2018.1755ArticleGoogle ScholarCrossref
Hripcsak G, Duke JD, Shah NH, et al. Observational Health Data Sciences and Informatics (OHDSI): opportunities for observational researchers. Stud Health Technol Inform. 2015;216:574-578.PubMedGoogle Scholar
Schuemie MJ, Hripcsak G, Ryan PB, Madigan D, Suchard MA. Robust empirical calibration of P-values using observational data. Stat Med. 2016;35(22):3883-3888. doi:10.1002/sim.6977PubMedGoogle ScholarCrossref
Schuemie MJ, Hripcsak G, Ryan PB, Madigan D, Suchard MA. Empirical confidence interval calibration for population-level effect estimation studies in observational healthcare data. Proc Natl Acad Sci U S A. 2018;115(11):2571-2577. doi:10.1073/pnas.1708282114PubMedGoogle ScholarCrossref
Hripcsak G, Ryan PB, Duke JD, et al. Characterizing treatment pathways at scale using the OHDSI network. Proc Natl Acad Sci U S A. 2016;113(27):7329-7336. doi:10.1073/pnas.1510502113PubMedGoogle ScholarCrossref
The original article in summary
Cardiovascular disease is the leading cause of morbidity and mortality in type 2 diabetes, and reducing its burden is an important goal of antidiabetic medications (ADMs).1,2 Metformin, which may have cardiovascular benefits, is widely recommended as first-line therapy.2,3However, there is a lack of consensus about choosing subsequent ADMs among patients who do not achieve adequate glycemic control with metformin or do not tolerate it.2,4 Comparing cardiovascular outcomes of second-line ADMs during this early transition in diabetes pharmacotherapy may help improve treatment decisions after metformin or in place of it.
Use of ADMs has increased owing to the rising prevalence of diabetes and a proliferation of novel therapeutic classes. Recent placebo-controlled trials of dipeptidyl peptidase 4 (DPP-4) inhibitors generally found no cardiovascular benefits or harms.5-7 However, some trials of glucagon-like peptide 1 (GLP-1) receptor agonists and sodium-glucose cotransporter 2 (SGLT-2) inhibitors reported reductions in composite cardiovascular outcomes and some individual cardiovascular events.8-14 These trials specifically recruited participants with a high burden of cardiovascular disease who were taking multiple ADMs. Therefore, it is not known whether their findings apply to the broader population with diabetes. Older research on second-line ADMs suggested cardiovascular harms associated with sulfonylureas, thiazolidinediones (TZDs), and insulin.15 To date, no studies have directly compared the cardiovascular effects of all contemporary ADM options among patients starting second-line therapy.
Our study investigated the comparative effectiveness of all major second-line ADM classes on major adverse cardiovascular events among insured adult patients with type 2 diabetes. By examining cardiovascular outcomes among patients initiating second-line ADMs in the real world, this study aimed to complement findings from individual drug trials and further inform ADM choices for the broad population of patients currently receiving these medications.16,17
In this large observational analysis of insured patients with type 2 diabetes who initiated second-line therapy, GLP-1 receptor agonist use was associated with significant reductions in the composite primary outcome compared with DPP-4 inhibitor use. However, this finding was not significant in some sensitivity analyses.
There was a direction toward cardiovascular benefit among patients starting treatment with SGLT-2 inhibitors compared with DPP-4 inhibitors that did not achieve statistical significance.
This study found consistent cardiovascular harms associated with use of basal insulin or sulfonylureas compared with DPP-4 inhibitors. Collectively, these findings raise concerns about the cardiovascular safety of sulfonylureas and basal insulin compared with newer ADMs and suggest that short-term cardiovascular outcomes of newer ADM classes may be similar among patients starting second-line treatment.
After metformin, current guidelines recommend selecting ADMs based on expected glycemic improvements, potential risks, and other factors, such as their effect on body weight.2,4Some experts suggest that clinicians also consider cardiovascular benefits and harms when prescribing second-line ADM therapy.15 However, limited cardiovascular data are currently available for the large population of patients starting second-line ADMs after metformin alone is not sufficient or not tolerated. Our analysis provides preliminary evidence needed by patients, clinicians, insurance plans, and pharmacy benefit managers to weigh the comparative cardiovascular harms and benefits of each second-line ADM class in this understudied population.
Recent randomized clinical trial results (eTable 8 in the Supplement) provide some context for interpreting our findings. Although the trials used the highest level of methodologic rigor, they have limitations. First, the major cardiovascular outcome trials compared a single medication to placebo, hindering comparisons across ADM classes. Second, most participants had known cardiovascular disease, which affects only 18% of all US adults with diabetes.30Therefore, trial findings may not apply to patients with a lower short-term risk of cardiovascular events.31 Future cardiovascular outcome trials in this broader population are unlikely because of the resources required, limited interest among funders, and the lack of a regulatory requirement to conduct such trials for older ADMs.32 Randomized clinical trials also may overestimate medication effects observed in clinical practice by selecting highly adherent participants and other study procedures that are not replicable during routine care.33
This context highlights an important role for observational research comparing the cardiovascular effectiveness of ADMs in more typical patients receiving medication in real-world conditions.34,35
Our findings will be complemented by an ongoing randomized clinical trial comparing the glucose-lowering effects of all ADMs studied here, except TZDs and SGLT-2 inhibitors, among participants who initiate second-line treatment after metformin monotherapy.36
When available, results from that ongoing trial will provide evidence about the glycemic efficacy of these drugs among adherent research volunteers and under ideal conditions.
Our primary model showed significantly lower risk of composite cardiovascular events associated with GLP-1 receptor agonist use compared with DPP-4 inhibitor use. Improved cardiovascular outcomes among those who initiated treatment with GLP-1 receptor agonists who did not take metformin previously, relative to those who did, may reflect time-lag bias with a longer duration of diabetes among metformin users. Recent placebo-controlled trials of semaglutide and liraglutide therapy also showed reductions in composite fatal and nonfatal cardiovascular events.12,13 Like the semaglutide trial,13 this study found a significant decrease in the incidence of stroke among patients starting treatment with GLP-1 receptor agonists.
Reductions in stroke risk approached statistical significance in other GLP-1 receptor agonist trials despite not being designed to detect this outcome.10-12,37 Our analysis likely had greater statistical power because there was follow-up of many more patients receiving treatment than past cardiovascular outcome trials. Interestingly, none of the patients in our GLP-1 receptor agonist group filled a prescription of semaglutide (data not shown). Although preliminary, these findings suggest a potential class effect of GLP-1 in reducing stroke risk that requires confirmation.
Compared with DPP-4 inhibitors, we found no statistically significant cardiovascular benefit from SGLT-2 inhibitor therapy. Our analysis included relatively few data on SGLT-2 inhibitors because of their introduction in 2013. Therefore, including longer follow-up in this group may have increased statistical power to detect significant cardiovascular benefits from these medications. Treatment with SGLT-2 inhibitors and GLP-1 receptor agonists has shown improvements in mortality and some individual cardiovascular events among trial participants with a high burden of cardiovascular disease.14 Prior observational analyses that included patients taking sulfonylureas or insulin in their comparison groups have generally reached similar findings.38-42 If sulfonylureas and insulin are more harmful than other ADMs, this finding may have accentuated estimates of comparative cardiovascular benefit associated with GLP-1 receptor agonists and SGLT-2 inhibitors. Furthermore, observational analyses examining the subgroup of patients without prior cardiovascular disease reported smaller or no benefits on composite cardiovascular outcomes when initiating treatment with these 2 drug classes, which is more similar to our findings.38,39 Finally, most previous research included Europeans, whose cardiovascular outcomes may differ from this US cohort.
Sulfonylureas constituted almost half of the second-line ADM prescriptions in our cohort, which is comparable with prior reports.43 We found increased cardiovascular risk with sulfonylureas compared with DPP-4 inhibitors, which was robust across all analyses. This risk is consistent with a large number observational studies describing cardiovascular harms of sulfonylureas, many of which were conducted before statins were widely used in diabetes care.44 Our comparison of sulfonylureas with all newer ADMs in the poststatin era shows continued widespread use despite comparatively greater cardiovascular harms.
We also found that basal insulin treatment was associated with higher cardiovascular risk than DPP-4 inhibitor treatment. Recent trials studying insulin as part of intensive diabetes treatment have reported conflicting findings on cardiovascular outcomes.45 However, a large body of research reveals common mechanisms linking basal insulin and sulfonylureas with increased cardiovascular risk, namely, hyperinsulinemia, weight gain, and hypoglycemia.45-47Hypoglycemia may be most important for short-term cardiovascular outcomes.48 Despite the observed cardiovascular harms associated with initiating sulfonylureas and basal insulin, prescriptions for these 2 ADM classes were filled by 60% of patients in our nationwide analysis. The cardiovascular effects of TZDs have been debated for many years,49-53 and we found no significant associations in this group.
Strengths and Limitations
To our knowledge, this is the first study comparing cardiovascular outcomes among initiators of all major second-line ADM. Examining a large, population-based, nationwide cohort enabled the ascertainment of sufficient cardiovascular events to perform this current analysis. Our study design provided adequate power to examine stroke as an individual outcome, which was not possible in clinical trials with smaller numbers of participants receiving treatment. Our rates of nonfatal cardiovascular events were comparable with those observed in the UK Prospective Diabetes Study and prior analyses of similar populations.54-56 The consistency of this study’s findings across multiple cardiovascular outcomes, modeling approaches, and subpopulations supports its internal validity. Our observational design also captured real-world conditions that randomized clinical trials do not, which promotes generalizability of the results in diverse US practice settings. This analysis is also timely, providing data on all major ADMs in a contemporary context where prescribers have little available cardiovascular evidence to inform initiation of second-line therapy.
Our study also has limitations. Patients were required to fill the index ADM at least twice, which was chosen to represent a level of exposure that has been associated with cardiovascular outcomes.8,9,12,13 This requirement may have introduced bias if reasons for discontinuation after the first fill were correlated with cardiovascular outcomes.57 Although this is likely not the case for most ADMs studied here,2 the systematic exclusion of some patients with early discontinuation of sulfonylureas and basal insulin owing to hypoglycemia may have underestimated the cardiovascular harms reported.58,59 We analyzed administrative claims data, which may be prone to misclassification and do not include granular information used by clinicians when prescribing second-line ADMs. Medical record data, which may contain some of these factors, have been used in similar studies but were not available in our data source.60 Nonrandomized studies like ours are also prone to selection bias. Multivariable adjustment for observed differences in cardiovascular risk among ADM groups, as well as falsification testing to address potential unobserved confounders, were approaches used to mitigate selection bias.61 Propensity score matching was not used because it would have limited statistical power for comparing 6 ADM groups,62 especially those with relatively smaller numbers of patients. Like similar observational studies, we had no body weight data. Although we used diagnosis codes to ascertain obesity identically for all ADM groups, this strategy underestimates its true prevalence and lacks detail about obesity severity.63
Our data sources included HbA1c values for one-third of the study population, which represents a proportion comparable to similar cohort studies using claims data.64 Compared with patients who had no HbA1c data, those with available values were younger with a higher burden of other cardiovascular risk factors (eTable 9 in the Supplement). However, the cardiovascular effectiveness of ADM classes was similar in these 2 subpopulations (Table 3). Furthermore, adjustment for HbA1c in the primary model did not change our findings, which supports a growing consensus that short-term cardiovascular events are not primarily mediated by glycemic control and its association with atherosclerosis progression.65 A large body of research has identified other mechanisms that may explain the short-term association of ADM with cardiovascular outcomes.48,50,66-70 Although the mean duration of follow-up in our was 1.3 years, this represents a comparable or longer time interval than other recent observational analyses of second-line ADMs.38,39,41,71
We also had no information on diabetes duration, which is an important determinant of cardiovascular risk.72 However, our selection of patients who were early in their medical treatment of diabetes was intended to compare those at a similar disease stage. In addition, all analyses adjusted for individual cardiovascular risk factors and evidence of existing microvascular and macrovascular complications that are a proxy for diabetes duration. Because we selected patients at a relatively early point in their pharmacologic management of diabetes, our findings may not apply to patients who have a longer duration of diabetes, are already taking multiple ADMs, or have a higher risk of cardiovascular events. Such populations have been studied elsewhere.5-14,38-42,71 Our data source did not allow us to examine mortality, hindering comparisons with studies that assessed fatal cardiovascular events. We did not analyze individual drugs within ADM classes or total costs of diabetes care. These are important directions for future research.
Among the large population of insured adult patients with type 2 diabetes who initiated second-line antidiabetic medication therapy, an increased risk of cardiovascular events was associated with starting sulfonylureas or basal insulin treatment compared with newer ADM alternatives.
Therefore, clinicians may consider prescribing GLP-1 receptor agonists, DPP-4 inhibitors, or SGLT-2 inhibitors more routinely after metformin rather than sulfonylureas or basal insulin. Furthermore, our findings may suggest a role for these newer ADMs in managing cardiovascular risk among patients with type 2 diabetes who either are taking metformin alone or have received no ADM previously. Although our findings should be interpreted with some caution due to the observational design of this study, they were robust to several rigorous sensitivity analyses and are supported by prior mechanistic and clinical evidence. Future research should compare ADM classes on glycemic effectiveness and additional metabolic end points.
www red DiabetologNytt