Cardiovascular Safety of Linagliptin in Type 2 Diabetes
Cardiovascular Safety of Linagliptin in Type 2 Diabetes
This comprehensive pooled analysis evaluated patient-level data from 9459 subjects with T2DM who participated in 19 clinical trials. It represents a collective 4421.3 patient-years of exposure to linagliptin—approximately twice as many years of exposure as previous analyses. In line with the conclusion of the previous analysis, this new analysis also suggests that linagliptin does not increase CV risk or CV outcomes versus combined comparator therapies or versus placebo, in patients with T2DM. Furthermore, the findings of this analysis also indicate that the CV risk profile was not influenced by a number of important factors associated with CV complications (age, gender, race, use of rescue medication, occurrence of hypoglycemia, Framingham CHD risk, renal disease, and microalbuminuria). The findings of the present analysis add to the existing evidence base for DPP-4 inhibitors, which shows the class to be generally well tolerated, but with less definitive evidence regarding CV safety.
To enable a relevant assessment of CV risk, studies evaluating the CV safety of glucose-lowering therapies should include patients at increased risk of CV events, such as those with relatively advanced disease, elderly patients, and those with some degree of renal impairment. In the present analysis, baseline data on the Framingham risk status of participants indicated that about 30% of patients had a 10-year risk score of >15%, that approximately half of all patients had been diagnosed with T2DM for >5 years, and that around 13% of patients were older than 70 years of age. In addition, some degree of renal impairment (eGFR based on MDRD staging of <90 mL/min) was present in just over half of patients analyzed (56.3% in the linagliptin and 57.9% in total comparator groups, respectively), with nearly a quarter (24.3% and 22.7%, respectively) having microalbuminuria. Thus, a sizeable proportion of the analyzed population may be deemed to exhibit moderately increased CV risk. Indeed, the observed incidence of rates of 4P-MACE (13.4 events and 18.9 events per 1000 patient-years for linagliptin- and active comparator-treated groups, respectively) is slightly higher than would be predicted from the Framingham risk scores. This observation, combined with the large number of patients exposed to linagliptin and the higher number of CV events accrued compared with previous analyses, therefore, supports the validity of the findings.
As noted above, the incidence of CV events reported in this analysis, per 1000 patient-years of exposure, was 13.4 for linagliptin and 18.9 for total comparators. Other, similar, analyses of phase 3 studies have reported incidence rates for custom MACE ranging from 5.8 to 14.6 for sitagliptin, saxagliptin, or vildagliptin, and 9.0 to 14.1 for pooled comparators. The most recent assessment of the CV safety of sitagliptin, based on pooled data from 25 double-blind studies, showed an incidence rate ratio, per 100 patient-years, of 0.83 (95% CI, 0.53–1.30) for the comparison of sitagliptin versus pooled comparators, and 1.01 (95% CI, 0.55–1.86) for sitagliptin versus placebo. These findings are in line with the results of the present analysis. Although the previous analyses of the CV safety of DPP-4 inhibitors differ in their methods, the findings nonetheless support the hypothesis that, in general, DPP-4 inhibitor therapy is not associated with increased CV risk. This was also supported by the neutral outcomes on 3P-MACE of the prospective studies SAVOR-TIMI 53 and EXAMINE, conducted in patients with T2DM at high CV risk.
EXAMINE was designed to investigate the CV risk of alogliptin compared with placebo in patients with T2DM and recent acute coronary syndrome. Similar rates of major CV AEs in addition to CV or all-cause mortality were reported for alogliptin- and placebo-treated patients. The SAVOR TIMI 53 trial evaluated the effects of saxagliptin on CV outcomes in patients with T2DM at high CV risk. No difference in the primary composite end point of CV death, MI, or ischemic stroke was found in saxagliptin- or placebo-treated patients. It should be noted that both of these trials were relatively short in duration (median follow-up, respectively, 2.2 and 1.5 years) and included patients predominantly, or exclusively, with manifest CV complications, 2 important considerations when assessing the potential CV risk modulation of any compound.
The unexpected finding of an increased risk of CHF in these trials merits further careful evaluation. A recent study of echocardiograms from 254 patients with T2DM and existing CHF (New York Heart Association [NYHA] class I to III), who received vildagliptin or placebo for 52 weeks, found that vildagliptin therapy did not change the cardiac ejection fraction (the primary end point of the study), and was not associated with worsening of CHF (confirmed by a blinded adjudication committee); worsening CHF occurred in 22 patients in the placebo group compared with 23 in the vildagliptin group. However, patients taking vildagliptin, compared with those taking placebo, showed unexpected increases in left ventricular end-diastolic volume (P = .007), end-systolic volume (P = .06), and stroke volume (P = .002). The possible mechanisms underlying these observations are not fully understood. Two recent meta-analyses of available data from randomized clinical trials of DPP-4 inhibitors have indicated that these drugs could be associated with an increased risk of acute heart failure. However, both meta-analyses included data from the SAVOR-TIMI 53 trial, the results of which strongly influence the overall findings. Furthermore, a new analysis of data from EXAMINE, presented at the Scientific Sessions of the American College of Cardiology in March 2014, indicated that alogliptin did not increase the risk of new-onset heart failure or the risk of readmission in patients with T2DM and a history of heart failure. In the linagliptin clinical trial program, adjudication of hospitalization for CHF was implemented during phase 3b. Given the small number of reported cases of CHF (for the overall comparator analysis, n = 12 and n = 9; and for the comparison with placebo only, n = 9 and n = 5), the reported HR of 1.04 (95% CI, 0.43–2.47) in the overall comparator analysis, or 1.29 (95% CI, 0.43–3.87) in the placebo-only analysis must be interpreted with some caution. Similarly, in the additional analysis of investigator-reported AEs in 24 placebo-controlled studies, the occurrence of events suggestive of heart failure was low (0.5% [n = 26 of 5488] and 0.2% [n = 8 of 3290] for linagliptin- and placebo-treated patients, respectively), and was within the expected background incidence for this population. Therefore, this database is currently too small to allow firm conclusions to be made regarding the impact of linagliptin on the risk of heart failure. So far, there has been no explanation for the increase in risk of CHF observed in the trials mentioned above, and data are currently limited. However, it has been noted that trials with reported CV outcomes tend to include patients who are older, have a longer duration of T2DM, more CV risk factors, lower renal function, more comorbidities, and often treated with a greater number of antidiabetic drugs, including insulin, when compared with other studies of patients with T2DM. These factors might be relevant in the identification of subpopulations who could be at increased risk of CHF or other CV outcomes with DPP-4 inhibitor therapy. Reassuringly, a review of safety data for DPP-4 inhibitors has demonstrated the safety and tolerability of these agents in fragile populations such as elderly patients and individuals with renal impairment. Similarly, a recent pooled analysis of 6 clinical trials demonstrated the safety and tolerability of linagliptin in a vulnerable subpopulation of patients at high risk of renal or CV disease. A recent systematic review and network meta-analysis of 10 clinical trials of DPP-4 inhibitors in patients requiring third-line therapy for T2DM showed no difference between these agents and placebo in the incidence of adverse events, including CVD. This analysis thus provides further evidence of the safety and tolerability of DPP-4 inhibitors in patients with T2DM that is difficult to control.
The ongoing Trial Evaluating Cardiovascular Outcomes with Sitagliptin (TECOS), designed to test the hypothesis that sitagliptin added to usual diabetes care does not increase CV risk in patients with existing CVD (estimated mean trial duration, approximately 4 years), will also provide an evaluation of DPP-4 inhibitor therapy over a longer period than SAVOR-TIMI 53 or EXAMINE, and should provide additional information on the CHF issue.
An interesting finding of the present pooled analysis is the significant reduction in cerebrovascular end points observed in linagliptin-treated patients, compared with the other groups, albeit limited by small number of observations. This finding would need to be further evaluated, in particular because similar results were not observed in the recent SAVOR-TIMI 53 and EXAMINE trials. Whether linagliptin could offer benefits following stroke is another possible area for further research in light of preclinical findings that have demonstrated a glucose-independent neuroprotective effect of linagliptin in the diabetic mouse brain model, possibly as a result of neural stem cell proliferation. Potential improvements in endothelial function with DPP-4 inhibition, as indicated in some studies, could have implications for cerebrovascular outcomes. The impact of linagliptin on cerebrovascular events and post-stroke function is currently under investigation in 2 outcome trials, as discussed below.
As with all pooled analyses, the present analysis has several limitations; in particular, the relatively short and different durations of the included studies limit the extent of interpretations that can be made. Furthermore, despite a large cumulative patient exposure to linagliptin, individual patient exposure was for a maximum of 2.2 years, so the time available for the development of CV events, or modulation of CV risk, was limited. As might be expected, only a relatively small proportion of patients received triple therapy or insulin therapy at baseline, suggesting that a relatively limited number of patients in the study population had advanced T2DM. However, the robustness of the findings of the present study is supported by its pre-specified design, which incorporated independent prospective, blinded adjudication of CV events, and by the consistency of the results, both across the individual trials and across the different pools of results. However, none of the individual studies included in our analysis was powered or designed to assess CV risk or events.
Two ongoing CV outcome trials of linagliptin will provide a more definitive answer on the CV safety profile of linagliptin. One of these trials, the CARdiovascular Outcome Study of LINAgliptin versus Glimepiride in Patients with Type 2 Diabetes (CAROLINA®) (NCT01243424) started in 2010 and has randomized 6041 patients with early T2DM and predominantly medium CV risk, to treatment with either linagliptin or glimepiride. CAROLINA® is the first head-to-head CV outcome trial of a DPP-4 inhibitor versus active comparator that is sufficiently powered to demonstrate potential differences in CV events between treatment groups. CAROLINA® will allow an assessment of the impact of long-term linagliptin therapy in a population at lower overall CV risk, with the possibility of demonstrating a CV benefit compared with the active comparator glimepiride.
The second CV outcome study, CARdiovascular Safety & Clinical outcoME with LINAgliptin (CARMELINA®), will compare the CV and renal safety of linagliptin versus placebo, when added to standard care in approximately 8300 patients with T2DM at high CV and renal compromise, and is the only large outcome study dedicated to the evaluation of tangible renal outcomes with a DPP-4 inhibitor in comparison with placebo. The CARMELINA® trial was initiated in 2013 and results are expected between 2017 and 2018.
Discussion
This comprehensive pooled analysis evaluated patient-level data from 9459 subjects with T2DM who participated in 19 clinical trials. It represents a collective 4421.3 patient-years of exposure to linagliptin—approximately twice as many years of exposure as previous analyses. In line with the conclusion of the previous analysis, this new analysis also suggests that linagliptin does not increase CV risk or CV outcomes versus combined comparator therapies or versus placebo, in patients with T2DM. Furthermore, the findings of this analysis also indicate that the CV risk profile was not influenced by a number of important factors associated with CV complications (age, gender, race, use of rescue medication, occurrence of hypoglycemia, Framingham CHD risk, renal disease, and microalbuminuria). The findings of the present analysis add to the existing evidence base for DPP-4 inhibitors, which shows the class to be generally well tolerated, but with less definitive evidence regarding CV safety.
To enable a relevant assessment of CV risk, studies evaluating the CV safety of glucose-lowering therapies should include patients at increased risk of CV events, such as those with relatively advanced disease, elderly patients, and those with some degree of renal impairment. In the present analysis, baseline data on the Framingham risk status of participants indicated that about 30% of patients had a 10-year risk score of >15%, that approximately half of all patients had been diagnosed with T2DM for >5 years, and that around 13% of patients were older than 70 years of age. In addition, some degree of renal impairment (eGFR based on MDRD staging of <90 mL/min) was present in just over half of patients analyzed (56.3% in the linagliptin and 57.9% in total comparator groups, respectively), with nearly a quarter (24.3% and 22.7%, respectively) having microalbuminuria. Thus, a sizeable proportion of the analyzed population may be deemed to exhibit moderately increased CV risk. Indeed, the observed incidence of rates of 4P-MACE (13.4 events and 18.9 events per 1000 patient-years for linagliptin- and active comparator-treated groups, respectively) is slightly higher than would be predicted from the Framingham risk scores. This observation, combined with the large number of patients exposed to linagliptin and the higher number of CV events accrued compared with previous analyses, therefore, supports the validity of the findings.
As noted above, the incidence of CV events reported in this analysis, per 1000 patient-years of exposure, was 13.4 for linagliptin and 18.9 for total comparators. Other, similar, analyses of phase 3 studies have reported incidence rates for custom MACE ranging from 5.8 to 14.6 for sitagliptin, saxagliptin, or vildagliptin, and 9.0 to 14.1 for pooled comparators. The most recent assessment of the CV safety of sitagliptin, based on pooled data from 25 double-blind studies, showed an incidence rate ratio, per 100 patient-years, of 0.83 (95% CI, 0.53–1.30) for the comparison of sitagliptin versus pooled comparators, and 1.01 (95% CI, 0.55–1.86) for sitagliptin versus placebo. These findings are in line with the results of the present analysis. Although the previous analyses of the CV safety of DPP-4 inhibitors differ in their methods, the findings nonetheless support the hypothesis that, in general, DPP-4 inhibitor therapy is not associated with increased CV risk. This was also supported by the neutral outcomes on 3P-MACE of the prospective studies SAVOR-TIMI 53 and EXAMINE, conducted in patients with T2DM at high CV risk.
EXAMINE was designed to investigate the CV risk of alogliptin compared with placebo in patients with T2DM and recent acute coronary syndrome. Similar rates of major CV AEs in addition to CV or all-cause mortality were reported for alogliptin- and placebo-treated patients. The SAVOR TIMI 53 trial evaluated the effects of saxagliptin on CV outcomes in patients with T2DM at high CV risk. No difference in the primary composite end point of CV death, MI, or ischemic stroke was found in saxagliptin- or placebo-treated patients. It should be noted that both of these trials were relatively short in duration (median follow-up, respectively, 2.2 and 1.5 years) and included patients predominantly, or exclusively, with manifest CV complications, 2 important considerations when assessing the potential CV risk modulation of any compound.
The unexpected finding of an increased risk of CHF in these trials merits further careful evaluation. A recent study of echocardiograms from 254 patients with T2DM and existing CHF (New York Heart Association [NYHA] class I to III), who received vildagliptin or placebo for 52 weeks, found that vildagliptin therapy did not change the cardiac ejection fraction (the primary end point of the study), and was not associated with worsening of CHF (confirmed by a blinded adjudication committee); worsening CHF occurred in 22 patients in the placebo group compared with 23 in the vildagliptin group. However, patients taking vildagliptin, compared with those taking placebo, showed unexpected increases in left ventricular end-diastolic volume (P = .007), end-systolic volume (P = .06), and stroke volume (P = .002). The possible mechanisms underlying these observations are not fully understood. Two recent meta-analyses of available data from randomized clinical trials of DPP-4 inhibitors have indicated that these drugs could be associated with an increased risk of acute heart failure. However, both meta-analyses included data from the SAVOR-TIMI 53 trial, the results of which strongly influence the overall findings. Furthermore, a new analysis of data from EXAMINE, presented at the Scientific Sessions of the American College of Cardiology in March 2014, indicated that alogliptin did not increase the risk of new-onset heart failure or the risk of readmission in patients with T2DM and a history of heart failure. In the linagliptin clinical trial program, adjudication of hospitalization for CHF was implemented during phase 3b. Given the small number of reported cases of CHF (for the overall comparator analysis, n = 12 and n = 9; and for the comparison with placebo only, n = 9 and n = 5), the reported HR of 1.04 (95% CI, 0.43–2.47) in the overall comparator analysis, or 1.29 (95% CI, 0.43–3.87) in the placebo-only analysis must be interpreted with some caution. Similarly, in the additional analysis of investigator-reported AEs in 24 placebo-controlled studies, the occurrence of events suggestive of heart failure was low (0.5% [n = 26 of 5488] and 0.2% [n = 8 of 3290] for linagliptin- and placebo-treated patients, respectively), and was within the expected background incidence for this population. Therefore, this database is currently too small to allow firm conclusions to be made regarding the impact of linagliptin on the risk of heart failure. So far, there has been no explanation for the increase in risk of CHF observed in the trials mentioned above, and data are currently limited. However, it has been noted that trials with reported CV outcomes tend to include patients who are older, have a longer duration of T2DM, more CV risk factors, lower renal function, more comorbidities, and often treated with a greater number of antidiabetic drugs, including insulin, when compared with other studies of patients with T2DM. These factors might be relevant in the identification of subpopulations who could be at increased risk of CHF or other CV outcomes with DPP-4 inhibitor therapy. Reassuringly, a review of safety data for DPP-4 inhibitors has demonstrated the safety and tolerability of these agents in fragile populations such as elderly patients and individuals with renal impairment. Similarly, a recent pooled analysis of 6 clinical trials demonstrated the safety and tolerability of linagliptin in a vulnerable subpopulation of patients at high risk of renal or CV disease. A recent systematic review and network meta-analysis of 10 clinical trials of DPP-4 inhibitors in patients requiring third-line therapy for T2DM showed no difference between these agents and placebo in the incidence of adverse events, including CVD. This analysis thus provides further evidence of the safety and tolerability of DPP-4 inhibitors in patients with T2DM that is difficult to control.
The ongoing Trial Evaluating Cardiovascular Outcomes with Sitagliptin (TECOS), designed to test the hypothesis that sitagliptin added to usual diabetes care does not increase CV risk in patients with existing CVD (estimated mean trial duration, approximately 4 years), will also provide an evaluation of DPP-4 inhibitor therapy over a longer period than SAVOR-TIMI 53 or EXAMINE, and should provide additional information on the CHF issue.
An interesting finding of the present pooled analysis is the significant reduction in cerebrovascular end points observed in linagliptin-treated patients, compared with the other groups, albeit limited by small number of observations. This finding would need to be further evaluated, in particular because similar results were not observed in the recent SAVOR-TIMI 53 and EXAMINE trials. Whether linagliptin could offer benefits following stroke is another possible area for further research in light of preclinical findings that have demonstrated a glucose-independent neuroprotective effect of linagliptin in the diabetic mouse brain model, possibly as a result of neural stem cell proliferation. Potential improvements in endothelial function with DPP-4 inhibition, as indicated in some studies, could have implications for cerebrovascular outcomes. The impact of linagliptin on cerebrovascular events and post-stroke function is currently under investigation in 2 outcome trials, as discussed below.
As with all pooled analyses, the present analysis has several limitations; in particular, the relatively short and different durations of the included studies limit the extent of interpretations that can be made. Furthermore, despite a large cumulative patient exposure to linagliptin, individual patient exposure was for a maximum of 2.2 years, so the time available for the development of CV events, or modulation of CV risk, was limited. As might be expected, only a relatively small proportion of patients received triple therapy or insulin therapy at baseline, suggesting that a relatively limited number of patients in the study population had advanced T2DM. However, the robustness of the findings of the present study is supported by its pre-specified design, which incorporated independent prospective, blinded adjudication of CV events, and by the consistency of the results, both across the individual trials and across the different pools of results. However, none of the individual studies included in our analysis was powered or designed to assess CV risk or events.
Two ongoing CV outcome trials of linagliptin will provide a more definitive answer on the CV safety profile of linagliptin. One of these trials, the CARdiovascular Outcome Study of LINAgliptin versus Glimepiride in Patients with Type 2 Diabetes (CAROLINA®) (NCT01243424) started in 2010 and has randomized 6041 patients with early T2DM and predominantly medium CV risk, to treatment with either linagliptin or glimepiride. CAROLINA® is the first head-to-head CV outcome trial of a DPP-4 inhibitor versus active comparator that is sufficiently powered to demonstrate potential differences in CV events between treatment groups. CAROLINA® will allow an assessment of the impact of long-term linagliptin therapy in a population at lower overall CV risk, with the possibility of demonstrating a CV benefit compared with the active comparator glimepiride.
The second CV outcome study, CARdiovascular Safety & Clinical outcoME with LINAgliptin (CARMELINA®), will compare the CV and renal safety of linagliptin versus placebo, when added to standard care in approximately 8300 patients with T2DM at high CV and renal compromise, and is the only large outcome study dedicated to the evaluation of tangible renal outcomes with a DPP-4 inhibitor in comparison with placebo. The CARMELINA® trial was initiated in 2013 and results are expected between 2017 and 2018.
Source...