2023 Volume 46 Issue 9 Pages 1217-1222
Heart failure is a prevalent comorbidity in patients with diabetes mellitus (DM). However, it is unclear whether the risk factors for heart failure in DM patients treated with dipeptidyl peptidase-4 (DPP-4) inhibitors are the same as those for the general population. In this study, we evaluated the factors of new-onset heart failure in working-age patients with diabetes who started DPP-4 inhibitor therapy. This study included 7938 working-age patients. The primary endpoint of the study was the proportion of patients developing heart failure within 12 months of starting DPP-4 inhibitor therapy, which was found to be 1.89% (n = 150). In these patients, risk factors of new-onset heart failure were aging, history of atrial fibrillation, and hypertension but not sex, smoking, high body mass index, weight gain of over 10 kg from 20 years of age, levels of low-density lipoprotein or glycated hemoglobin A1c (HbA1c), history of angina pectoris, myocardial infarction, and chronic kidney disease. We confirmed that cardiovascular comorbidities are risk factors for new-onset heart failure in patients with DM, while general risk factors are not. In conclusion, physicians and pharmacists need to carefully monitor working-age patients with cardiovascular history who start DPP-4 inhibitor therapy even if they do not exhibit general risk factors for heart failure.
Currently, there are 537 million patients with diabetes mellitus (DM) worldwide.1) One of the primary concerns in DM is controlling comorbidities, such as retinopathy, nephropathy, neuropathy, and heart diseases. The adequate management of comorbidities contributed to the prolongation of life span from 63.1 to 71.4 years for males and from 64.9 to 75.1 years for females between 1971–1980 and 2001–2010.2) In addition, DM patients with heart failure experience significantly higher mortality rates than DM patients without heart failure comorbidities (32.7 vs. 3.7 per 100 person-year).3)
Dipeptidyl peptidase-4 (DPP-4) inhibitors are widely used with a low risk of hypoglycemia regardless of the patient’s background. In 2022, the American Diabetes Association released a consensus report4) suggesting that DPP-4 inhibitors exerted positive and negative effects in patients with existing heart failure and recommended against the use of DPP-4 inhibitors in patients with any symptom of heart failure.5–8) Through a meta-analysis evaluating cardiovascular outcomes in 52520 DM patients treated with DPP-4 inhibitor or placebo until October 2020, Patoulias et al. revealed that DPP-4 inhibitors do not seem to be of benefit in cases with established cardiovascular disease or multiple risk factors.9) However, using rat and human heart failure tissues, Shigeta et al. reported that DPP-4 plays an important role in left ventricular diastolic dysfunction, suggesting the use of DPP-4 activity as a diagnostic surrogate marker or therapeutic target for chronic heart failure.10)
Recently, practical evidence from China revealed that compared with DPP-4 inhibitors, sodium-dependent glucose transporter 2 (SGLT2) inhibitors can reduce the proportion of new-onset heart failure cases.11) However, it is unclear whether the risk factors for heart failure in DM patients treated with DPP-4 inhibitors are the same as those for the general population. Therefore, we studied new-onset heart failure and its risk factors in working-age patients with DM who started DPP-4 inhibitor therapy.
The Japan Medical Data Center (JMDC) database contains anonymized data of patients. The cumulative dataset contains approximately 13 million subjects (inpatients, outpatients, and pharmacy claims) from approximately 90% of all hospitals in Japan. All patients in the JMDC database have “social insurance” that covers the working members and their families. As of September 2021, the database represented approximately 10.0% of the Japanese population.
Patients’ Identification ProcessUsing the JMDC database, we identified 57856 patients out of 3064012 participants, including family members, having social insurance and who were receiving DPP-4 inhibitor therapy from January 2005 to June 2017. Data collection for time 0 refers to the index month when patients were prescribed DPP-4 inhibitors for the first time. We considered it a washout window for DPP-4 inhibitors and “ −6 to −1 month from the index month” as a screening period. Furthermore, we considered the observation period until +11 months for a total of 12 months from the index month (Supplementary material). From the total population, we excluded patients 1) who could not be assessed during the 6 months of screening and 12 months of observation (n = 27611), 2) who had a history of heart failure (International Classification of Diseases-10 codes: ICD 10 code for I50) during the 6-month screening period and index month (n = 3029), 3) who were registered as a family member at the JMDC dataset cohort entry (n = 6064), and 4) who were without annual health-checkup data within the 6-month screening period and index month (n = 13214). In our dataset, laboratory data were limited in patients from annual health-checkup data because our data set from JMDC Inc. was based on claims data. Finally, we analyzed 7938 working-age patients in this study (Fig. 1).
JMDC, Japan Medical Data Center; DPP-4, dipeptidyl peptidase-4; n, number.
The primary endpoint of this study was evaluating the proportion of heart failure (I50) within the 12-month observation period without the index month to avoid the overestimation of new-onset heart failure after starting DPP-4 inhibitor therapy. Heart failure (I50) is equal to ICD 9 code for 428.xx, which was validated in a previous report.12,13) Risk factors were identified to assess the factors for short-term risks of new-onset heart failure in patients with DM who started DPP-4 inhibitor therapy.
DPP-4 inhibitors were defined by combining the name of each drug with the respective WHO-Anatomical Therapeutic Chemical code as follows: Sitagliptin (A10BH01), Vildagliptin (A10BH02), Saxagliptin (A10BH03), Alogliptin (A10BH04), Linagliptin (A10BH05), Teneligliptin (A10BH), Anagliptin (A10BH), Trelagliptin (A10BH), and Omarigliptin (A10BH). We did not include combination medicines with DPP-4 inhibitors. We collected the following data: 1) annual health-checkup data for the screening and index months (−6 to 0 months), 2) comorbidities in the screening month (−6 to −1 month), and 3) patient demographics such as sex, age, and co-administered DM drugs with DPP-4 inhibitors in the index month (0 month) as baseline data (Supplementary material). Baseline comorbidities were defined as follows: angina pectoris (I20), myocardial infarction (I21–22), atrial fibrillation (I48), chronic kidney disease (N18), and hypertension (I10–15).
Statistical AnalysisUnivariate analyses (Student’s t-, Mann–Whitney U-, chi-square, and Fisher’s exact tests) of the groups designated as “with heart failure” and “without heart failure” during the 12-month observation period without the index month were performed using 7938 patients’ baseline demographics.
The risk factors for heart failure were assessed using the baseline age (older population are at higher risk),14,15) sex (males are at higher risk), weight gain, body mass index (BMI) (higher BMI indicates higher risk),16) low-density lipoprotein (LDL) cholesterol (higher LDL cholesterol is a known cardiac risk factor),17,18) smoking status,19) glycated hemoglobin A1c (HbA1c) levels (higher blood sugar is a known cardiac risk factor),20) and baseline comorbidities clinically related to heart failure, including angina pectoris, myocardial infarction, atrial fibrillation, hypertension, and chronic kidney disease. We followed standard methods to estimate the sample sizes for multiple logistic regression. A minimum of 10 outcomes was required for each independent variable to avoid overfitting.
Data are presented as medians with ranges or means ± standard deviations. Data analysis was performed using JMP® 16 (SAS Institute Inc., Cary, NC, U.S.A.).
Ethics ApprovalThe commercially available JMDC database used in this study contains anonymized information processed based on Japan’s Personal Information Protection Law, and individual informed consent is not required for the provision and use of this information. In addition, according to the ethical guidelines for clinical research in Japan, reviewal by an ethics committee is not required for studies using anonymized information. Therefore, informed consent was not obtained for this study since patient data were anonymized before accessing.
In our study, a total of 7938 (with/without new-onset heart failure during the observation period without the index month: 150/7788) patients were analyzed (Table 1). The average age of the “with heart failure” group was higher than that of the “without heart failure” group (53.8 ± 7.6 vs. 51.9 ± 7.8 years, p = 0.0009). Significant differences were not observed in LDL levels (129.7 ± 39.8 vs. 129.9 ± 33.7 mg/dL, p = 0.7733), number of patients with smoking behavior (59/150 vs. 3058/7788, p = 0.9866), BMI (26.7 ± 5.1 vs. 26.7 ± 4.6 kg/m2, p = 0.7722), and HbA1c levels (7.8 ± 1.5 vs. 8.0 ± 1.7%, p = 0.2720) between the “with” and “without heart failure” groups. Weight gain of over 10 kg from 20 years of age was lower in the “with heart failure” group than in the “without heart failure” group (89/150 vs. 5002/7788, p = 0.2158). Comorbidities, including angina pectoris (16/150 vs. 351/7788, p = 0.0004), atrial fibrillation (6/150 vs. 45/7788, p < 0.0001), myocardial infarction (3/150 vs. 26/7788, p = 0.0008), chronic kidney disease (3/150 vs. 39/7788, p = 0.0122), and hypertension (84/150 vs. 3115/7788, p < 0.0001), were more common in the “with heart failure” group than in the “without heart failure” group.
| Total | With heart failure | Without heart failure | p-Value | |
|---|---|---|---|---|
| Number (male/female) | 7938 (7347/591) | 150 (141/9) | 7788 (7206/582) | 0.4960 |
| Age, years [S.D.] | 51.9 [7.8] | 53.8 [7.6] | 51.9 [7.8] | 0.0009 |
| Diabetes drug category | ||||
| Biguanide, number (%) | 2034 | 40 (26.7) | 1994 (25.6) | 0.7677 |
| Sulfonylurea, number (%) | 1505 | 26 (17.3) | 1479 (19.0) | 0.6080 |
| Alpha-glucosidase inhibitor, number (%) | 945 | 27 (18.0) | 918 (11.8) | 0.0200 |
| Thiazolidinedione, number (%) | 722 | 16 (10.7) | 706 (9.1) | 0.4993 |
| Insulin, number (%) | 543 | 19 (12.7) | 524 (6.7) | 0.0043 |
| Glinide, number (%) | 229 | 2 (1.3) | 227 (2.9) | 0.3296 |
| Sodium-glucose cotransporter 2 inhibitor, number (%) | 171 | 0 (0) | 171 (2.2) | 0.0800 |
| Weight gain of over 10 kg from 20 years of age, number (%) | 5091 | 89 (59.3) | 5002 (64.2) | 0.2158 |
| LDL, mg/dL [S.D.] | 129.9 [33.8] | 129.7 [39.8] | 129.9 [33.7] | 0.7733 |
| Smoking behavior, number (%) | 3117 | 59 (39.3) | 3058 (39.3) | 0.9866 |
| BMI, kg/m2 [S.D.] | 26.7 [4.6] | 26.7 [5.1] | 26.7 [4.6] | 0.7722 |
| HbA1c, % [S.D.] | 8.0 [1.7] | 7.8 [1.5] | 8.0 [1.7] | 0.2720 |
| Comorbidity, number (%) | ||||
| Angina pectoris [I20] | 367 | 16 (10.7) | 351 (4.5) | 0.0004 |
| Atrial fibrillation [I48] | 51 | 6 (4.0) | 45 (0.6) | <0.0001 |
| Myocardial infarction [I21, 22] | 29 | 3 (2.0) | 26 (0.3) | 0.0008 |
| Chronic kidney disease [N18] | 42 | 3 (2.0) | 39 (0.5) | 0.0122 |
| Hypertension [I10–15] | 3199 | 84 (56.0) | 3115 (40.0) | <0.0001 |
S.D., standard deviation; LDL, low-density lipoprotein cholesterol; BMI, body mass index; HbA1c, glycated hemoglobin A1c.
The incidence of heart failure within 12 months after starting DPP-4 inhibitor therapy, which was the primary endpoint, was 1.89% (n = 150).
Risk factors were used to determine factors for the onset of heart failure within 12 months in patients with aging (odds ratio for +1 year and 95% confidence interval (CI), 1.03 [1.00–1.05], p = 0.0022), atrial fibrillation (odds ratio and 95% CI, 4.78 [1.94–11.79], p = 0.0007), and hypertension (odds ratio and 95% CI, 1.66 [1.16–2.37], p = 0.0053).
Previously reported risk factors for new-onset heart failure, such as male sex (p = 0.4922), higher BMI (p = 0.4343), weight gain of over 10 kg from 20 years of age (p = 0.0800), higher LDL levels (p = 0.1983), smoking behavior (p = 0.5787), higher HbA1c levels (p = 0.6493), angina pectoris (p = 0.1009), myocardial infarction (p = 0.0510), and chronic kidney disease (p = 0.1448), did not increase the risk of new-onset heart failure in working-age patients with DM who started DPP-4 inhibitor therapy in this study.
In our study, the incidence of new-onset heart failure was 1.89% in working-age patients within 12 months of starting DPP-4 inhibitor therapy (Table 1). The risk factor was the presence of cardiovascular-related comorbidities; however, previously known risk factors found in the general populations, such as higher BMI, high LDL, HbA1c levels, and smoking behavior, were not risk factors for heart failure in working-age patients who started DPP-4 inhibitors in this study.
Japanese patients with heart failure accounted for approximately 5% of patients with heart failure globally,21) and the number of deaths caused by heart failure represented 6.1% of all deaths reported in Japan in 2020.22)
Notably, compared with patients without DM, the onset of heart failure is two and five times higher in males and females with DM, respectively.23) This indicates that compared with patients without DM, patients with DM have a higher risk of critical health problems, such as heart failure. Moreover, 15% of patients with type 2 DM have been hospitalized due to heart failure exacerbations within an average observation period of 5.2 years.24) To prevent the progression of heart failure in patients with DM, good glycemic control is necessary. Preventing the onset of heart failure in patients with DM is essential because 23 million patients develop heart failure every year worldwide, which costs 39.2 billion dollars.21) Therefore, we studied real-world data on heart failure onset in working-age DM patients, whose information on lethal health risks is lacking.
In our study, sex, smoking, weight gain, BMI, HbA1c, and LDL levels were not related to heart failure onset in working-age DM patients within 12 months of starting DPP-4 inhibitor therapy (Table 2). In our study, cardiovascular comorbidities in patients who started/selected DPP-4 inhibitor therapy were associated with heart failure onset during a 12-month observation period (Table 2). This suggests that our study results are applicable to patients who exhibit factors determined in this study as baseline characteristics, which can be closely monitored to identify the early-stage symptoms of heart failure onset.
| With heart failuren = 150 | Without heart failuren = 7788 | Adjusted odds ratio (95% CI) | p-Value | ||
|---|---|---|---|---|---|
| Age, year [S.D.] | 53.8 [7.6] | 51.9 [7.8] | 1.03 (1.00–1.05) | 0.0022 | |
| Sex, number | |||||
| Female | 9 | 582 | Reference | — | |
| Male | 141 | 7206 | 1.27 (0.64–2.54) | 0.4922 | |
| Weight gain of over 10 kg from 20 years of age, number (%) | 89 (59.3) | 5002 (64.2) | 0.79 (0.39–1.57) | 0.0800 | |
| LDL, mg/dL [S.D.] | 129.7 [39.8] | 129.9 [33.7] | 1.00 (1.00–1.01) | 0.1983 | |
| Smoking behavior, number (%) | 59 (39.3) | 3058 (39.3) | 1.10 (0.78–1.54) | 0.5787 | |
| BMI, kg/m2 [S.D.] | 26.7 [5.1] | 26.7 [4.6] | 1.02 (0.98–1.06) | 0.4343 | |
| HbA1c, % [S.D.] | 7.8 [1.5] | 8.0 [1.7] | 0.98 (0.88–1.09) | 0.6493 | |
| Comorbidity, number | |||||
| Angina pectoris [I20] | Without | 134 | 7437 | Reference | — |
| With | 16 | 351 | 1.62 (0.91–2.87) | 0.1009 | |
| Atrial fibrillation [I48] | Without | 144 | 7743 | Reference | — |
| With | 6 | 45 | 4.78 (1.94–11.79) | 0.0007 | |
| Myocardial infarction [I21, 22] | Without | 147 | 7762 | Reference | — |
| With | 3 | 26 | 3.52 (0.99–12.43) | 0.0510 | |
| Chronic kidney disease [N18] | Without | 147 | 7749 | Reference | — |
| With | 3 | 39 | 2.50 (0.73–8.56) | 0.1448 | |
| Hypertension [I10–15] | Without | 66 | 4673 | Reference | — |
| With | 84 | 3115 | 1.66 (1.16–2.37) | 0.0053 | |
DPP-4, dipeptidyl peptidase-4; CI, confidence interval; S.D., standard deviation; n, number; LDL, low-density lipoprotein cholesterol; BMI, body mass index; HbA1c, glycated hemoglobin A1c.
This study has several limitations. First, we used annual health check-up data during the screening period and index month. Health check-up data have differences in the timing of measurement for each subject. Second, we set a screening period as a washout window for new users of DPP-4 inhibitors in this study. This is based on the definition of “DPP-4 inhibitor new user” in this study. However, there is a possibility that the patients were administered DPP-4 inhibitors before the screening period. Third, thiazolidinedione derivatives and metformin affect the incidence and survival of patients with heart failure, respectively.25,26) Herein, between the with and without heart failure groups, no significant difference was found in the proportion of the co-administration of DPP-4 inhibitors and thiazolidinedione (10.7 vs. 9.1%, p = 0.4993) or biguanide (26.7 vs. 25.6%, p = 0.7677) at the baseline. However, these co-administration drugs probably affected the new onset of heart failure in this study. In addition, SGLT2 inhibitors are indicated for heart failure per package inserts. We could not assess the effects of SGLT2 inhibitors on the new onset of heart failure in our study. However, all patients co-administered with SGLT2 inhibitors were observed in the without heart failure group, which needs further investigation. Fourth, our study patients are different from general DM patients. Our study recruited working-age DM patients who undergo annual health check-ups and have social insurance. Therefore, socio-economic differences may exist between our study patients and general DM patients. Fifth, given the limitations of our study design, we cannot directly compare the incidence of new-onset heart failure between patients initiated on DPP-4 inhibitors and those initiated on other anti-diabetic drugs. Additionally, we were unable to assess the extent to which DPP-4 inhibitors reduce the risk of new-onset heart failure because we did not set control patients treated without DPP-4 inhibitors.
We confirmed that 1.89% of working-age DM patients developed heart failure within 1 year of starting DPP-4 inhibitor therapy. We focused on the new onset of heart failure within 1 year after starting DPP-4 inhibitor therapy. Thus, the new onset of heart failure in patients using DPP-4 inhibitors for the long term should be studied in the future. Risk factors for heart failure in these patients are cardiovascular comorbidities but not general risk factors such as male sex, smoking, weight gain, higher BMI, or having high HbA1c or LDL levels. In conclusion, physicians and pharmacists must carefully monitor working-age patients with cardiovascular histories who started DPP-4 inhibitor therapy even if they do not show general risk factors for heart failure.
A part of this study was supported by 2022 Scholarship Fund for Young/Women Researchers, Promotion and Mutual Aid Corporation for Private Schools of Japan, and Grant in Showa University translational research.
All authors meet the ICMJE recommendations. AM and KM especially contributed to the study conception, drafted the manuscript, and collected the raw data; AW created and confirmed the definitions of the analysis; HK, YK, KT, and TS provided clinical interpretation; and TS provided the final approval. All authors participated in the discussions during the manuscript preparation. All authors have agreed to publish the final version of the manuscript.
KM, TS (Showa University), and JMDC Inc. collaborate on other project according to the collaborative research agreement; JMDC Inc. did not intervene in the implementation of the data analyzed in this study; KM received honorarium fees for presentations from JMDC Inc.; TS received an honorarium for presentation at Daiichi Sankyo Co., Nipro Corp., Nichi-Iko Pharmaceutical Co., Ltd., Pfizer Inc., Sandoz, Mylan, and Meiji Seika Pharma Co., Ltd.; TS is an advisor at Torii Pharmaceutical Co., Ltd.; KM and TS received travel reimbursement from Abbvie Inc. to attend their conference. Department of Hospital Pharmaceutics, School of Pharmacy, Showa University received funding from Ono Pharmaceutical Co., Ltd. with a contract research project according to the collaborative research agreement.
As a potential conflict of interest, AM has second degree relatives who are employed and own shares in Terumo Corp.; KM received an honorarium for Nippon-Kayaku, and Abbvie Inc.; YK received an honorarium for presentation at Daiichi Sankyo Co.; KT received an honorarium for presentation at AYUMI Pharmaceutical Corporation, Daiichi Sankyo Co., Terumo Corp., Chugai Pharmaceutical Co., Ltd., Towa Corp., and Taisho Pharmaceutical Holdings; HK owns shares of Meiji Holdings, Alfresa, Shionogi, Chugai, Eisai, Ono, Rohto, Towa, Daiichi Sankyo, Sawai, FUJIFILM, and Nipro.; Hospital Pharmaceutics received a research grant from Daiichi Sankyo Co., Mochida Pharmaceutical Co., Ltd., Shionogi Inc., Ono Pharmaceutical Co., Ltd., Taiho Pharmaceutical Co., Ltd., and Nippon-Kayaku; The other authors declare no conflict of interest.
This article contains supplementary materials.
