Dipeptidyl Peptidase-4 Inhibitors and Joint Pain: A Retrospective Cohort Study of Older Veterans with Type 2 Diabetes Mellitus

Glycemic control continues to be the hallmark of diabetes management. Among adults with type 2 diabetes mellitus who do not achieve their glycemic goals with metformin alone, other classes of oral antidiabetes drugs are consequently added.¹ Many organizations now recommend dipeptidyl peptidase (DPP)-4 inhibitors as a second-line treatment for the management of type 2 diabetes.^2-4 In a systematic review and meta-analysis of 19 randomized controlled trials (RCTs) that compared DPP-4 inhibitor monotherapy, metformin monotherapy, and combination therapy, DPP-4 inhibitor treatment resulted in improved glycemic control, without weight gain.⁵

A recent meta-analysis of 16 studies (9 RCTs and 7 observational studies) that examined the safety and effectiveness of DPP-4 inhibitors among older adults (aged ≥65 years) concluded that the overall health benefits of DPP-4 inhibitors were unclear and inconsistent, with some studies reporting increased risks for sepsis, hospitalizations for heart failure, and heart failure outcomes.⁶

Emerging evidence from case reports also suggests a possible association of DPP-4 inhibitors with joint pain and rheumatoid arthritis.^7-10 Based on 3 case reports, Crickx and colleagues reported that the use of DPP-4 inhibitors was associated with inflammation of the joints.⁹ A case-control study also showed an association between DPP-4 inhibitors and joint inflammation.¹¹ Consequently, the US Food and Drug Administration issued a warning that the use of DPP-4 inhibitors for the treatment of type 2 diabetes may cause severe joint pain.¹²

A systematic review and meta-analysis of 67 RCTs (including a total of 79,110 adults with diabetes) suggested an increased risk (relative risk ratio, 1.13; 95% confidence interval [CI], 1.04-1.22; P = .003) for any joint pain with the use of DPP-4 inhibitors.¹³ However, a population-based cohort study of 52,933 Taiwanese patients with type 2 diabetes did not show an association between DPP-4 inhibitors and severe joint pain.¹⁴ That study only included adults aged ≥18 years.¹⁴ Despite the increased prevalence of diabetes and joint pain and the associated functional limitations among older adults,¹⁵ there is no robust evidence regarding the association of joint pain with the use of DPP-4 inhibitors in this age-group.

For this article, we selected older veterans with diabetes, because 1 of 5 veterans has diagnosed diabetes.¹⁶ Among veterans with uncomplicated diabetes who are receiving care from the Veterans Health Administration (VHA), an oral antidiabetes drug is often added to metformin treatment within the first year of diagnosis.¹⁷ In addition, some subgroups of veterans continue to have poor glycemic control, that is, hemoglobin (Hb)A1c ≥9%,¹⁸ which suggests that they may benefit from the addition of DPP-4 inhibitors to their treatment regimen.

Analyzing the relationship between DPP-4 inhibitors and joint pain is important for weighing the harms and benefits of adding a DPP-4 inhibitor as an oral antidiabetes drug for veterans. Therefore, the primary objective of this study was to examine the association between newly prescribed DPP-4 inhibitors and joint pain within 1 year using a retrospective cohort of joint pain–free VHA clinic users with type 2 diabetes.

We adapted an integrated conceptual framework derived from the Health in All Policies approach and the Andersen and Newman behavioral model to guide the selection of variables for our study.^19,20 Under this integrated framework, health may be influenced by social, physical, and economic environments, which are collectively referred to as the “determinants of health.”¹⁹

Based on our adapted model, health (in this case, joint pain) can be affected by biological (sex, age, race, ethnicity), sociocultural (marital status), socioeconomic, behavioral (tobacco use, body mass index [BMI]), and diabetes disease– and diabetes treatment–related factors (HbA_1c levels, Diabetes Complications Severity Index [DCSI], low-density lipoprotein [LDL] cholesterol levels, insulin use, oral antidiabetes agents); access to healthcare services (Veterans Affairs [VA] priority status, Medicaid indicator); external environment (urban or rural residence, census region); and coexisting medical and mental health conditions (eg, arthritis, asthma, dementia).

Methods

We used a retrospective cohort design with baseline and follow-up periods. The cohort comprised older (aged ≥66 years) veterans with type 2 diabetes who were enrolled in the VHA. Calendar years 2008 and 2009 served as the baseline period, and 2010 served as the follow-up period. Because of a time lag in the release of updated data from the VA, no later data were available for this study. To ensure that the patients had joint pain after they had a prescription for a DPP-4 inhibitor, we identified a cohort of people with joint pain–free diabetes and no DPP-4 inhibitor use during baseline.

Data were extracted from linked VHA electronic health records (ie, inpatient, outpatient, laboratory, extended care, and prescription drugs data) and Medicare claims (ie, inpatient, outpatient, skilled-nursing facility, home healthcare, and prescription drugs data) of VHA and Medicare dually enrolled users with type 2 diabetes and county-level Area Health Resource File. Rural residence was determined by linking the veterans’ ZIP codes with Rural-Urban Commuting Area Codes.²¹

The study cohort was constructed by identifying all veterans aged ≥66 years with type 2 diabetes over a 2-year period (2008-2009). We identified veterans with type 2 diabetes using a validated algorithm of 1 inpatient or 2 face-to-face outpatient healthcare encounters in the VHA records or in Medicare claims, using International Classification of Diseases, Ninth Revision (ICD-9) code 250.xx.22 We selected community-dwelling (ie, no hospital stays of >180 days, no skilled-nursing facility stays, and no hospice admissions) veterans who were dually enrolled in the VHA system and in Medicare Parts A and B throughout the study period and were active healthcare users (with >1 face-to-face outpatient visits and a pharmacy record) during the baseline period.

We excluded older veterans who did not have a prescription filled for any oral antidiabetes drug during the baseline period (2008-2009), had a prescription filled for a DPP-4 inhibitor in 2008, reported joint pain (ICD-9 codes 710-719) during the baseline period, or died during the study period (2008-2010). The final study population size was 134,488 veterans with diabetes who satisfied all the study inclusion and exclusion criteria.

Study Measures

We used ICD-9 codes 710 to 719 to identify joint pain during the follow-up period (ie, calendar year 2010). Older veterans with at least 1 healthcare encounter record (in the inpatient or outpatient setting) with a joint pain–related ICD-9 code were considered to have diagnosed joint pain.

DPP-4 inhibitor use was measured during calendar year 2009. Drugs in the DPP-4 inhibitor class were identified using RxNorm.²³ RxNorm, which is provided by the US National Library of Medicine, includes the names of drugs from various software, including the National Drug File from the VHA, and provides the standardized names of clinical drugs for research.²³

With the exception of DPP-4 inhibitor use, all other independent variables were measured during the 2 calendar years baseline period (2008-2009). The biological factors consisted of sex, age in years (66-74, 75-79, or ≥80), and race or ethnicity (white, black, Hispanic, or other). The environmental factors consisted of residence rurality (urban, rural, highly rural, or unknown/foreign) and census region (Northeast, Midwest, South, West, or territories). Marital status was defined as a social factor. Tobacco use, alcohol and drug use, and BMI (underweight/normal, overweight, obese, morbidly obese) defined the behavioral factors.

Access to healthcare services was measured by VA priority status (low-income, severely disabled, moderately disabled, or copays required) and Medicaid indicator. The disease-related factors included glycemic control, DCSI, LDL cholesterol levels, insulin use, and the number of oral antidiabetes drug prescriptions.

DCSI was calculated based on a validated algorithm24 and did not include laboratory values. The coexisting conditions included the medical and mental health conditions of asthma, cancer, chronic obstructive pulmonary disease, dementia, infectious disease, anxiety (including posttraumatic stress disorder), and serious mental illness.

Statistical Analysis

Subgroup differences in newly prescribed DPP-4 inhibitors were examined with chi-square tests. The unadjusted association between DPP-4 inhibitors and newly documented joint pain was tested with chi-square and logistic regressions. We conducted several models to analyze the adjusted association between DPP-4 inhibitor use and joint pain.

In model 1, we adjusted only for DPP-4 inhibitors. In model 2, we adjusted for biological characteristics (sex, age, and race/ethnicity). In model 3, we adjusted for biological characteristics, DCSI, marital status, and VHA priority status. In model 4, we adjusted for biological characteristics, DCSI, marital status, VHA priority status, rural residence, and census region. The fully adjusted model (model 5) controlled for the biological, socioeconomic, sociocultural, coexisting conditions, behavioral factors, access to healthcare, and external environment characteristics as described above.

Results

The study population included 134,488 older veterans with diabetes who were dually enrolled in Medicare and in the VHA system. The study population characteristics are presented in Table 1. A majority of the older veterans with diabetes were men (98.8%), white (86.9%), aged ≥75 years (50.3%), and lived in urban areas (58.3%). In all, 26.8% of the patients reported having 3 or more diabetes complications based on DCSI scores. Of the total patients, 41.7% were prescribed 3 or more oral antidiabetes drugs other than a DPP-4 inhibitor and 18.7% were prescribed insulin; among the 100,255 patients with HbA_1c results available, only 4.9% had an HbA_1c level of >9% and 8.3% had an HbA_1c level between 8.1% and 9%.

We found significant subgroup differences in DPP-4 inhibitor use (P <.05; see Appendix). The percentage of DPP-4 inhibitor use was significantly higher among African-Americans than among whites (10.2% vs 8.2%, respectively), in those living in urban areas than in their rural counterparts (9.0% vs 6.4%, respectively), and in those with DCSI scores of 3 or higher than in patients with a score of 0 (9.3% vs 7.2%, respectively). Among the patients with available HbA_1c values, a higher percentage of older veterans with an HbA_1c level of >9% were receiving a DPP-4 inhibitor than those with an HbA_1c level of <7% (12.1% vs 9.3%, respectively).

In the study cohort, 11.7% of the 134,488 patients had newly documented joint pain during the follow-up period (ie, 2010; Table 2). In unadjusted analyses, we found significant differences in joint pain between DPP-4 inhibitor users and nonusers (12.9% vs 11.6%, respectively; P <.0001). We also observed significant differences in joint pain by almost all covariates, except for alcohol or drug use (P = .1030), LDL level (P = .1349), insulin use (P = .8205), and oral antidiabetes drug use (P = .4690).

Table 3 presents the unadjusted odds ratios (OR), adjusted OR, and the 95% CIs from logistic regressions for the association between DPP-4 inhibitor use and joint pain.

In model 1, without any adjustments, veterans with at least 1 filled prescription for a DPP-4 inhibitor in 2009 had significantly higher odds of having joint pain during the follow-up period than veterans without any DPP-4 inhibitor prescription (unadjusted OR, 1.13; 95% CI, 1.07-1.20). In model 2, when we adjusted for biological factors (ie, age, sex, and race/ethnicity), the association between DPP-4 inhibitors and joint pain was still significant (adjusted OR, 1.13; 95% CI, 1.07-1.20). When we added marital status, DCSI, VHA priority status, Medicaid census region, and geographic residence, the association between DPP-4 inhibitors and joint pain did not change (adjusted OR, 1.14; 95% CI, 1.07-1.20).

When we added smoking, drug or alcohol use, and BMI, the association between DPP-4 inhibitors and joint pain remained significant (adjusted OR, 1.13; 95% CI, 1.07-1.20). In the fully adjusted model, after controlling for biological, environmental, social, and behavioral factors; access to healthcare services; disease-related factors; and coexisting chronic conditions, patients with a filled prescription for a DPP-4 inhibitor had higher odds of joint pain than those without a prescription for a DPP-4 inhibitor (adjusted OR, 1.17; 95% CI, 1.10-1.24).

Discussion

To our knowledge, this is the first study to examine the association between DPP-4 inhibitor use and joint pain among older adults with diabetes. In our sample of older veterans, who were predominantly white men, we found a small but significant association between DPP-4 inhibitor use and subsequent joint pain. Although many mechanisms have been proposed for the role of DPP-4 inhibitors in inducing joint pain,¹⁰ the exact mechanism remains unknown. One possible mechanism invokes an increase in certain cytokine levels by DPP-4 inhibitors.^10,25 Levels of these inflammatory markers are known to increase with aging.^26,27 Therefore, the addition of DPP-4 inhibitors in the treatment regimen for type 2 diabetes in elderly patients may adversely alter cytokine levels and, more notably in this age-group, contribute to joint pain.

Our study’s findings are not consistent with an observational study of Taiwanese adults with diabetes.¹⁴ The differences in the findings between our study and that study can be attributed to differences in the study cohorts, the definition of outcome, and statistical adjustments. For example, the Taiwanese study included only adults, focused only on severe joint pain, did not control for glycemic control or BMI, and used a propensity score–matched sample.

Our data preclude insights into the actual clinical decision to prescribe a DPP-4 inhibitor, but we observed that a greater percentage of DPP-4 inhibitor users in our VA population had severe complications of type 2 diabetes (ie, higher DCSI scores with 3 or more complications), and almost 50% of DPP-4 inhibitor users had an HbA_1c level of >8% (among those with available data on HbA_1c).

Clearly, physicians and patients need to weigh the need to control type 2 diabetes complications through better glycemic control relative to the possible elevated risk for joint pain among older adults from the addition of a DPP-4 inhibitor to treatment.

Further prospective studies are needed to explore the causal relationship between DPP-4 inhibitor use and joint pain. Furthermore, because we restricted the study population to older (aged ≥65 years) patients, most of the veterans included in this analysis were from the Vietnam and Korean Wars, although that information was not included in the database.

The strength of our findings lies in this analysis being a large population-based study with data from real-world clinical practices. In our study, we used a joint pain–free diabetes cohort without DPP-4 inhibitor prescriptions during the baseline period, and thus we were able to ensure the proper temporal relationship between DPP-4 inhibitor use and subsequent joint pain that is potentially attributable to this medication. Furthermore, because the data were derived from multiple sources (ie, Medicare, the VHA, and the Area Health Resource File), we were able to control for a comprehensive list of variables that can affect the association between DPP-4 inhibitor use and joint pain among older veterans with diabetes.

Limitations

Secondary data, as used in this study, have inherent limitations that should be considered when interpreting our study’s results. First, we restricted our population to veterans who are enrolled in the VHA system and receive care from an integrated system; therefore, the findings may not be generalizable to patients who receive care from other health systems.

In addition, a filled prescription does not indicate that the medication has been taken by the patients. We might have overestimated the use of DPP-4 inhibitors (eg, not all filled prescriptions might have been used by the veterans). Given the differences between DPP-4 inhibitor users and nonusers, we cannot rule out the role of selection bias.

We captured only diagnosed joint pain as indicated by ICD-9 codes; any reports of joint pain in physicians’ notes or in other provider documentation were not captured. Any undiagnosed or undocumented joint pain was also not included in our study, which might have biased our finding of an association between DPP-4 inhibitor use and joint pain toward the null.

The data used in this study were old (before 2010), which might have led to an underestimation of the extent of DPP-4 inhibitor use today. New DPP-4 inhibitors have been added to the market since 2010, and their use is more common today than our data may reflect. However, this further suggests that our findings linking DPP-4 inhibitors and joint pain are valid and make an impor­tant contribution.

Conclusion

Our findings show that in a cohort of older veterans with diabetes and without a diagnosis of joint pain at baseline, the initiation of DPP-4 inhibitor therapy was significantly associated with newly diagnosed joint pain. This finding is important and relevant, despite the use of relatively old data, which might have underestimated the use of DPP-4 inhibitors and did not include all the newer DPP-4 inhibitors that are available today.

Future prospective cohort studies that include a more diverse population of older adults are needed to confirm this association. It may also be interesting to explore whether the joint pain reverses when treatment with a DPP-4 inhibitor is stopped.

Author Disclosure Statement

Ms Rai, Dr Dwibedi, Ms Rowneki, Dr Helmer, and Dr Sambamoorthi have no conflicts of interest to report.

Source of Funding

This work was supported by the Veterans Health Administration (VHA) Health Services Research & Development (grant number IIR 12-401). The content is solely the responsibility of the authors and does not necessarily represent the official views of the VHA, West Virginia University, or other affiliated organizations. Support for the Department of Veterans Affairs (VA)/Centers for Medicare & Medicaid Services data is provided by the VA, Veterans Health Administration, Office of Research and Development, Health Services Research and Development, and VA Information Resource Center (project numbers SDR 02-237 and 98-004).

References

1. American Diabetes Association. Standards of Medical Care in Diabetes—2018 abridged for primary care providers. Clin Diabetes. 2018;36:14-37.
2. Sinclair A, Morley JE, Rodriguez-Mañas L, et al. Diabetes mellitus in older people: position statement on behalf of the International Association of Gerontology and Geriatrics (IAGG), the European Diabetes Working Party for Older People (EDWPOP), and the International Task Force of Experts in Diabetes. J Am Med Dir Assoc. 2012;13:497-502.
3. American Diabetes Association. 9. Pharmacologic approaches to glycemic treatment: Standards of Medical Care in Diabetes—2019. Diabetes Care. 2019;42(suppl 1):S90-S102.
4. 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 – 2019 executive summary. Endocr Pract. 2019;25:69-100. Erratum in: Endocr Pract. 2019;25:204.
5. Karagiannis T, Paschos P, Paletas K, et al. Dipeptidyl peptidase-4 inhibitors for treatment of type 2 diabetes mellitus in the clinical setting: systematic review and meta-analysis. BMJ. 2012;344:e1369. Errata in: BMJ. 2012;345:e4959; BMJ. 2012;345:e5129.
6. Schott G, Martinez YV, Ediriweera de Silva RE, et al. Effectiveness and safety of dipeptidyl peptidase 4 inhibitors in the management of type 2 diabetes in older adults: a systematic review and development of recommendations to reduce inappropriate prescribing. BMC Geriatr. 2017;17(suppl 1):226.
7. Sasaki T, Hiki Y, Nagumo S, et al. Acute onset of rheumatoid arthritis associated with administration of a dipeptidyl peptidase-4 (DPP-4) inhibitor to patients with diabetes mellitus. Diabetol Int. 2010;1:90-92.
8. Yokota K, Igaki N. Sitagliptin (DPP-4 inhibitor)-induced rheumatoid arthritis in type 2 diabetes mellitus: a case report. Intern Med. 2012;51:2041-2044.
9. Crickx E, Marroun I, Veyrie C, et al. DPP4 inhibitor-induced polyarthritis: a report of three cases. Rheumatol Int. 2014;34:291-292.
10. Mascolo A, Rafaniello C, Sportiello L, et al. Dipeptidyl peptidase (DPP)-4 inhibitor-induced arthritis/arthralgia: a review of clinical cases. Drug Saf. 2016;39:401-407.
11. Saito T, Ohnuma K, Suzuki H, et al. Polyarthropathy in type 2 diabetes patients treated with DPP4 inhibitors. Diabetes Res Clin Pract. 2013;102:e8-e12.
12. US Food and Drug Administration. FDA Drug Safety Communication: FDA warns that DPP-4 inhibitors for type 2 diabetes may cause severe joint pain. August 28, 2015. www.fda.gov/Drugs/DrugSafety/ucm459579.htm. Accessed June 2, 2018.
13. Men P, He N, Song C, Zhai S. Dipeptidyl peptidase-4 inhibitors and risk of arthralgia: a systematic review and meta-analysis. Diabetes Metab. 2017;43:493-500.
14. Hou WH, Chang KC, Li CY, Ou HT. Dipeptidyl peptidase-4 inhibitor use is not associated with elevated risk of severe joint pain in patients with type 2 diabetes: a population-based cohort study. Pain. 2016;157:1954-1959.
15. Molton IR, Terrill AL. Overview of persistent pain in older adults. Am Psychol. 2014;69:197-207.
16. Liu Y, Sayam S, Shao X, et al. Prevalence of and trends in diabetes among veterans, United States, 2005-2014. Prev Chronic Dis. 2017;14:E135.
17. Gatwood JD, Chisholm-Burns M, Davis R, et al. Disparities in initial oral antidiabetic medication adherence among veterans with incident diabetes. J Manag Care Spec Pharm. 2018;24:379-389.
18. Thorpe CT, Gellad WF, Good CB, et al. Tight glycemic control and use of hypoglycemic medications in older veterans with type 2 diabetes and comorbid dementia. Diabetes Care. 2015;38:588-595.
19. Rudolph L, Caplan J, Ben-Moshe K, Dillon L. Health in All Policies: A Guide for State and Local Governments. Washington, DC and Oakland, CA: American Public Health Association and Public Health Institute: 2013.
20. Andersen R, Newman JF. Societal and individual determinants of medical care utilization in the United States. Milbank Mem Fund Q Health Soc. 1973;51:95-124.
21. United States Department of Agriculture Economic Research Service. Rural-urban commuting area codes. 2010. www.ers.usda.gov/webdocs/DataFiles/53241/ruca2010.xlsx?v=0. Accessed April 18, 2018.
22. Miller DR, Safford MM, Pogach LM. Who has diabetes? Best estimates of diabetes prevalence in the Department of Veterans Affairs based on computerized patient data. Diabetes Care. 2004;27(suppl 2):B10-B21.
23. US National Library of Medicine. Unified Medical Language System (UMLS): RxNorm. 2018. www.nlm.nih.gov/research/umls/rxnorm/overview.html. Accessed June 11, 2019.
24. Young BA, Lin E, Von Korff M, et al. Diabetes Complications Severity Index and risk of mortality, hospitalization, and healthcare utilization. Am J Manag Care. 2008;14:15-24.
25. Drucker DJ. Dipeptidyl peptidase-4 inhibition and the treatment of type 2 diabetes: preclinical biology and mechanisms of action. Diabetes Care. 2007;30:1335-1343.
26. Michaud M, Balardy L, Moulis G, et al. Proinflammatory cytokines, aging, and age-related diseases. J Am Med Dir Assoc. 2013;14:877-882.
27. Singh T, Newman AB. Inflammatory markers in population studies of aging. Ageing Res Rev. 2011;10:319-329.