Abstract
Background
Inflammation is a central pathway leading to frailty but whether commonly used nonaspirin nonsteroidal anti-inflammatory drugs (NSAIDs) can prevent frailty is unknown.
Methods
Prospective cohort study of male physicians ≥60 who participated in the Physicians’ Health Study. Annual questionnaires collected data on NSAID use, lifestyle, and morbidity. Average annual NSAID use was categorized as 0 days/year, 1–12 days/year, 13–60 days/year, and >60 days/year. Frailty was assessed using a validated 33-item frailty index. Propensity score inverse probability of treatment weighting was used to address confounding by indication and logistic regression models estimated odds ratios (ORs) of prevalent frailty according to nonaspirin NSAID use.
Results
A total of 12 101 male physicians were included (mean age 70 ± 7 years, mean follow-up 11 years). Reported NSAID use was 0 days/year for 2 234, 1–12 days/year for 5 812, 13–60 days/year for 2 833, and >60 days/year for 1 222 participants. A total of 2 413 participants (20%) were frail. Higher self-reported NSAID use was associated with greater alcohol use, smoking, arthritis, hypertension, and heart disease, while less NSAID use was associated with coumadin use and prior bleeding. After propensity score adjustment, all characteristics were balanced. ORs (95% confidence intervals) of prevalent frailty were 0.90 (0.80–1.02), 1.02 (0.89–1.17), and 1.26 (1.07–1.49) for average NSAID use of 1–12 days/year, 13–60 days/year, and >60 days/year, compared to 0 days/year (p-trend < .001).
Conclusions
Long-term use of NSAIDs at high frequency is associated with increased risk of frailty among older men. Additional study is needed to understand the role of anti-inflammatory medication in older adults and its implication for overall health.
Keywords: Epidemiology, Inflammation, Pharmacoepidemiology
With the rapid aging of the population, multidimensional approaches to prevent frailty are urgently needed. Frailty is a state of depleted physiologic reserve or an accumulation of health-related deficits leading to impaired response to minor external stressors resulting in poor outcomes, including falls, institutionalization, and death (1–3). Proposed biologic mechanisms of frailty include chronic inflammation and oxidative stress (3–5). “Inflammaging” has been coined to explain the associations between chronically elevated inflammatory biomarkers that are associated with biologic aging and frailty (4,6–9). While there is some evidence that anti-inflammatory medications may lower levels of inflammatory biomarkers (10,11), there is limited data on the association between anti-inflammatory medications and frailty.
Frailty and cardiovascular disease (CVD) share a bidirectional relationship, which may be mediated by inflammation (12–15). Common nonsteroidal anti-inflammatory drugs (NSAIDs) such as aspirin have been proven to reduce the risk of CVD, in part by reducing inflammation and levels of C-reactive protein (16). This has led to the hypothesis that anti-inflammatory medications may lower the risk of frailty and related geriatric conditions. The Baltimore Longitudinal Study on Aging demonstrated in a cohort of 2 300 participants that long-term nonaspirin and aspirin NSAID use was associated with a lower risk of developing incident dementia (17). This raises the question of whether commonly used anti-inflammatory medications such as aspirin and other NSAIDs may also prevent frailty. In prior work in the Physicians’ Health Study (PHS), long-term aspirin use was associated with a 15% lower risk of frailty (18). Therefore, we sought to extend this hypothesis and examine the relation between long-term nonaspirin NSAID use and the risk of prevalent frailty using the well-defined PHS cohort.
Method
Study Population
The PHS I is a completed double-blinded, placebo-controlled randomized trial that evaluated the effects of aspirin and beta-carotene on CVD and cancer among 22 701 male physicians. A detailed description and rationale of PHS I has been previously published (19,20). Briefly, invitations were sent out to U.S. male physicians between the ages of 40 and 84 in 1982 (average age 60). Aspirin randomization was ended in 1988 and the randomized beta-carotene was continued until 1995. The physicians in PHS I were followed from 1982 through 2012. Information on functional status, health status, and mood were collected at the 16th annual follow-up questionnaire (from the start of the main trial in 1982) and were used for computing the frailty score. For the current study, 17 607 participants had responded to the questionnaire in 1999. We excluded 3 172 who were younger than 60 years old, 958 missing functional status, 183 missing mood information, 1 054 missing a frailty index (FI), and 139 missing data for the propensity score analysis described below. Thus, 12 101 participants met the inclusion criteria for this study.
Exposure: NSAID Assessment
Information about nonsteroidal anti-inflammatory agent use was self-reported on each annual follow-up questionnaire by the following question: “Over the past 12 months, on how many days have you taken platelet active or nonsteroidal anti-inflammatory agents?” Reponses included 0 days, 1–12 days, 13–30 days, 31–60 days, 61–90 days, 91–120 days, 121–180 days, and 180+ days of nonsteroidal anti-inflammatory agent (NSAID) use during the year. The average frequency of NSAID use was computed by summing the frequency through each questionnaire and dividing by the number of annual questionnaires from the end of the aspirin trial until frailty assessment (below). Average NSAID use was categorized into “0 days/year,” “1–12 days/year,” “13–60 days/year,” and “>60 days per year,” as was done previously (21). NSAID use was distinct from aspirin use which was asked in a separate question annually.
Outcome
Frailty was defined primarily according to the cumulative deficit theory, first put forth by Rockwood et al. (2,22,23), and validated in PHS (24). Briefly, the PHS FI consists of 33 health deficits associated with aging, including comorbidities, cognition, function, and mental health variables (2). As an example, an individual with 4 deficits has an FI score of 4/33 = 0.12. Any score ≥0.21 was considered frail (24–26). In subgroup analyses, an FI score <0.1 was considered nonfrail, 0.1–0.2 prefrail, and ≥0.21 frail. All variables were taken from the questionnaire at the time of frailty assessment (1999).
In sensitivity analyses, we defined frailty according to an alternate theory developed by Fried and colleagues, based on the physical phenotype frailty model and further refined in the Study of Osteoporotic Fractures (SOF) cohort (1,27). Using available data in PHS, we used a validated modified SOF (mSOF) definition that includes 3 variables: weight loss >5% in the past year, inability to rise from a chair without assistance, and self-report of low energy (24,28). An individual with a score ≥2 was considered frail (24). In subgroup analyses, an mSOF score of 0/3 was considered nonfrail, 1/3 prefrail, and ≥2/3 frail. In total, 12 043 participants had data to calculate an mSOF score.
Other Clinical Covariates
Self-reported information on demographics and functional status was collected at the time of frailty index assessment. Comorbidities collected at the time frailty assessment included coronary heart disease, stroke, peripheral artery disease, bleeding history (gastrointestinal bleeding, peptic ulcer disease, gastritis, easy bruising/bleeding, epistaxis, hematuria, or hematemesis), arthritis, migraine or headache, atrial fibrillation, pulmonary embolism, deep vein thrombosis, hypertension (defined as hypertension medication use, systolic blood pressure > 140, or diabolic blood pressure > 90), diabetes mellitus, peripheral artery surgery, cancer, liver disease, and renal disease. In addition, information on anticoagulation use was collected from all prior annual questionnaires. Cumulative aspirin use was computed as average days of aspirin use per year.
Statistical Analysis
We classified participants into 1 of the following 4 categories of average NSAID use in days/year: 0, 1–12, 13–60, and >60 according to prior work (21). Propensity score inverse probability of treatment weighting was used to reduce confounding and balance covariate influences due to treatment effect. The propensity score estimated the probability of NSAID use given observed characteristics, including comorbidities related to NSAID use, smoking status, and alcohol consumption. We used a kernel density plot to assess the feasibility of performing propensity score analysis, which demonstrated good overlap between the 4 exposure groups (SupplementaryFigure 1). Baseline characteristics in Table 1 were weighted after adjusting by propensity score. We used logistic regression models to estimate the odds of prevalent frailty status by cumulative deficit definition (frailty vs nonfrail) according to average NSAID use over 16 years. Cubic splines were used to assess the linearity of average NSAID use and prevalent frailty status. Interaction terms for age, arthritis, heart disease, gastrointestinal bleed, and chronic kidney disease were each examined. We also conducted sensitivity analyses stratifying at age <75 or ≥75 years, <90 or ≥90 days of average aspirin use per year, relevant comorbidities, and body mass index. We repeated the main analysis using classic multivariate regression analysis. Additionally, we conducted a sensitivity analysis to exclude new users of NSAIDs (in the 3 years prior to frailty assessment) to ensure that long-term NSAID was assessed. Finally, analyses were repeated using the mSOF definition of frailty. A 2-tailed p value < .05 was considered statistically significant. All analyses were conducted in SAS version 9.4 (SAS Institute, Cary, NC).
Figure 1.
Unadjusted spline demonstrating the relationship between nonsteroidal anti-inflammatory drug (NSAID) use and odds of frailty. Frailty defined according to the frailty index.
Table 1.
Baseline Characteristics of 12 101 PHS Participants According to Average Annual NSAID Use After Propensity Score Weighting
| Average Annual NSAID Use Categories | ||||
|---|---|---|---|---|
| 0 Day n = 2 234 |
1–12 Days n = 5 812 |
13–60 Days n = 2 833 |
≥60 Days n = 1 222 |
|
| Age (mean, SD) | 72.3 ± 7.5 | 69.9 ± 6.9 | 69.9 ± 7.1 | 71.0 ± 7.1 |
| Randomized to aspirin % | 50.0 | 50.6 | 50.7 | 51.0 |
| Smoking history | ||||
| Never | 49.2 | 48.2 | 47.6 | 45.8 |
| Past | 45.7 | 47.9 | 49.5 | 51.2 |
| Current | 5.1 | 3.9 | 2.8 | 3.0 |
| Alcohol consumption | ||||
| Daily | 41.7 | 41.7 | 42.0 | 43.4 |
| Weekly | 37.7 | 37.1 | 36.7 | 35.3 |
| Monthly | 3.7 | 4.3 | 3.9 | 3.8 |
| Rarely | 16.9 | 17.0 | 17.4 | 17.6 |
| Arthritis | 30.6 | 29.5 | 29.8 | 30.2 |
| Coronary heart disease | 18.2 | 16.4 | 16.5 | 17.9 |
| Stroke | 3.1 | 3.0 | 3.0 | 3.0 |
| Atrial fibrillation | 11.9 | 10.9 | 11.1 | 10.5 |
| Hypertension | 56.4 | 55.5 | 55.8 | 54.5 |
| Diabetes | 8.8 | 8.5 | 8.5 | 8.5 |
| Peripheral artery disease | 1.0 | 0.9 | 0.9 | 1.0 |
| Cancer | 13.3 | 13.3 | 13.3 | 14.1 |
| Liver disease | 13.4 | 12.8 | 13.0 | 13.3 |
| Renal disease | 7.5 | 6.9 | 6.9 | 7.9 |
| Migraine | 14.7 | 12.6 | 12.9 | 12.7 |
| Headache | 53.5 | 51.6 | 51.6 | 55.2 |
| Pulmonary embolus | 0.9 | 1.1 | 1.1 | 1.3 |
| Deep vein thrombosis | 1.8 | 2.0 | 2.0 | 1.7 |
| Coumadin use | 6.7 | 6.5 | 6.5 | 6.1 |
| Gastritis | 47.3 | 46.5 | 46.2 | 46.7 |
| Melena | 3.4 | 3.8 | 3.7 | 3.6 |
| Bleeding/easy bruising | 40.4 | 39.3 | 39.4 | 39.1 |
| Peptic ulcer disease | 5.4 | 5.5 | 5.7 | 5.1 |
| Epistaxis | 20.3 | 19.3 | 19.7 | 20.7 |
| Hematuria | 14.8 | 14.7 | 14.6 | 14.8 |
| Hematemesis | 0.5 | 0.4 | 0.4 | 0.5 |
| GI bleeding | 11.5 | 12.4 | 12.3 | 11.7 |
| Cumulative days aspirin use/year (days/year) | 136 ± 69 | 142 ± 66 | 148 ± 67 | 153 ± 76 |
| FI score (mean, SD) | 0.1 ± 0.1 | 0.1 ± 0.1 | 0.2 ± 0.1 | 0.2 ± 0.1 |
| % frail by FI score | 20.5 | 19.0 | 20.9 | 24.5 |
| % frail by mSOF score | 8.0 | 9.1 | 11.8 | 14.5 |
Notes: FI = frailty index; GI = gastrointestinal; mSOF score = modified Study of Osteoporotic Fracture score; NSAID = nonsteroidal anti-inflammatory drug; PHS = Physicians’ Health Study; SD = standard deviation.
Ethics
Each participant gave informed consent and the Institutional Review Board at Brigham and Women’s Hospital approved the study protocol.
Results
There were 12 101 participants available for inclusion in this study using the FI definition of frailty. At the time of frailty assessment, mean age was 70.5 ± 7.1 years (range 60–101). Over a mean 11 ± 0.6 years of follow-up, reported NSAID use was 0 days/year for 2 234, 1–12 days/year for 5 812, 13–60 days/year for 2 833, and >60 days/year for 1 222 participants. Twenty percent of participants (n = 2 413) were frail. Before the propensity score was applied, those with greater NSAID use were more likely to report drinking alcohol daily, having smoked previously, and having arthritis, hypertension, and heart disease. On the other hand, less NSAID use was associated with greater coumadin use and prior bleeding (Supplementary Table 1). Characteristics of 1 986 participants ≥60 years excluded due to insufficient data to calculate frailty are shown in Supplementary Table 2.
Unadjusted cubic splines shown in Figure 1 demonstrate an increased risk of frailty according to average NSAID use, largely driven by those taking NSAIDs for >60 days per year. The same pattern of increased risk of frailty with NSAID use was seen using the mSOF definition of frailty (Supplementary Figure 2).
After propensity score adjustment, all covariates were balanced (Table 1; Supplementary Tables 2 and 3). The odds ratios (ORs; 95% confidence intervals [CIs]) of prevalent frailty were 0.91 (0.80–1.03), 1.02 (0.89–1.17), and 1.26 (1.07–1.49) for average NSAID use of 1–12 days/year, 13–60 days/year, and >60 days/year, compared to 0 days/year (p-trend < .001; Table 2). When results were analyzed using the mSOF alternative definition of frailty, 10% were identified as frail, and results showed an even stronger association between regular long-term NSAID use and increased odds of frailty.
Table 2.
Odds Ratios (ORs) of Prevalent Frailty by Average Annual NSAID Use Among 12 101 Physicians’ Health Study Participants
| Model | OR of Frail vs Nonfrail* | 95% CI | p-Trend |
|---|---|---|---|
| Frailty determined using frailty index score ≥0.21 | |||
| 0 days/year | Ref | — | <.001 |
| ≤12 days/year | 0.91 | 0.80–1.03 | |
| 13–60 days/year | 1.02 | 0.89–1.17 | |
| >60 days/year | 1.26 | 1.07–1.49 | |
| Frailty determined using mSOF score ≥2 (n = 12 028) | |||
| 0 days/year | Ref | — | <.001 |
| ≤12 days/year | 1.15 | 0.96–1.37 | |
| 13–60 days/year | 1.53 | 1.27–1.85 | |
| >60 days/year | 1.95 | 1.56–2.43 |
Notes: CI = confidence interval; mSOF score = modified Study of Osteoporotic Fracture score; NSAID = nonsteroidal anti-inflammatory drug.
*All models were run after the propensity score weighting was applied.
Table 3 shows stratified results. There was no significant interaction for age, arthritis, gastrointestinal bleed, heart disease, chronic kidney disease, regular use of aspirin or a composite pain variable (headache, migraine, or arthritis). Models were also stratified by body mass index category and results did not change (not shown).
Table 3.
Odds Ratios (ORs) of Prevalent Frailty by NSAID Category, Stratified by Age, Major Comorbidities, and Aspirin Use
| Variable | Model* | OR of Frailty vs Nonfrail | 95% CI | p-Interaction |
|---|---|---|---|---|
| Age | ||||
| <75 N = 8 942 |
≤12 days/year | 1.18 | 0.98–1.43 | .51 |
| 13–60 days/year | 1.65 | 1.35–2.02 | ||
| >60 days/year | 3.12 | 2.49–3.91 | ||
| ≥75 N = 3 159 |
≤12 days/year | 1.21 | 1.00–1.47 | |
| 13–60 days/year | 1.69 | 1.36–2.10 | ||
| >60 days/year | 2.57 | 1.99–3.33 | ||
| Coronary heart disease | ||||
| No N = 10 112 |
≤12 days/year | 1.00 | 0.86–1.17 | .08 |
| 13–60 days/year | 1.42 | 1.21–1.68 | ||
| >60 days/year | 2.66 | 2.20–3.21 | ||
| Yes N = 1 989 |
≤12 days/year | 1.02 | 0.78–1.34 | |
| 13–60 days/year | 1.24 | 0.93–1.67 | ||
| >60 days/year | 1.68 | 1.22–2.32 | ||
| Arthritis | ||||
| No N = 8 529 |
≤12 days/year | 0.97 | 0.83–1.14 | .20 |
| 13–60 days/year | 1.12 | 0.92–1.36 | ||
| >60 days/year | 1.77 | 1.36–2.30 | ||
| Yes N = 3 572 |
≤12 days/year | 0.68 | 0.53–0.87 | |
| 13–60 days/year | 0.78 | 0.60–0.995 | ||
| >60 days/year | 1.08 | 0.83–1.40 | ||
| Gastrointestinal bleeding | ||||
| No N = 10 602 |
≤12 days/year | 1.01 | 0.88–1.16 | .80 |
| 13–60 days/year | 1.38 | 1.19–1.62 | ||
| >60 days/year | 2.49 | 2.09–2.98 | ||
| Yes N = 1 499 |
≤12 days/year | 0.93 | 0.67–1.30 | |
| 13–60 days/year | 1.48 | 1.03–2.12 | ||
| >60 days/year | 2.53 | 1.68–3.82 | ||
| Renal disease | ||||
| No N = 11 260 |
≤12 days/year | 1.00 | 0.87–1.14 | .08 |
| 13–60 days/year | 1.40 | 1.20–1.62 | ||
| >60 days/year | 2.61 | 2.20–3.10 | ||
| Yes N = 841 |
≤12 days/year | 1.09 | 0.73–1.64 | |
| 13–60 days/year | 1.41 | 0.90–2.19 | ||
| >60 days/year | 1.69 | 1.01–2.82 | ||
| Average aspirin use | ||||
| <90 days/year N = 2358 |
≤12 days/year | 0.84 | 0.65–1.10 | .24 |
| 13–60 days/year | 1.20 | 0.89–1.62 | ||
| >60 days/year | 1.94 | 1.38–2.73 | ||
| ≥90 days/year N = 9743 |
≤12 days/year | 1.07 | 0.92–1.24 | |
| 13–60 days/year | 1.49 | 1.27–1.75 | ||
| >60 days/year | 2.75 | 2.29–3.31 | ||
| Pain** | ||||
| No N = 4136 |
≤12 days/year | 0.97 | 0.76–1.17 | .58 |
| 13–60 days/year | 1.05 | 0.80–1.38 | ||
| >60 days/year | 1.85 | 1.27–2.70 | ||
| Yes N = 7965 |
≤12 days/year | 0.91 | 0.77–1.07 | |
| 13–60 days/year | 1.26 | 1.06–1.50 | ||
| >60 days/year | 2.15 | 1.77–2.61 |
Notes: CI = confidence interval; NSAID = nonsteroidal anti-inflammatory drug.
*Reference = no NSAID use.
**Pain: migraine, headache, or arthritis.
Results were similar using the mSOF definition (Supplementary Table 4). Repeating the main analysis using classic multivariate regression and using the propensity score as a covariate in the model or as a doubly robust covariate did not change the results (Supplementary Tables 5–7). Additionally, excluding individuals who reported new NSAID use in the 3 years prior to frailty assessment did not change the results (not shown).
Discussion
This study aimed to understand the association of long-term nonaspirin NSAIDs and prevalent frailty. In a cohort of 12 101 male physicians with 16 years of information related to NSAID use, we found that long-term use of NSAIDs was associated with an increased odds of frailty. Similar results were observed when using an alternative definition of frailty and remained consistent when stratifying for age and other key comorbidities. If inflammation underlies the frailty syndrome, these findings are intriguing when considering prior findings from the same cohort of an association between long-term aspirin use and lower prevalence of frailty after 11 years of follow-up.
As pain management has moved away from the use of opiates, NSAIDs are increasingly first-line pharmacologic therapy and have an important role in care plans (29). However, there are known risks associated with NSAIDs, including bleeding, renal impairment, and increased CVD risk, all of which increase with aging and frailty (30). As the population rapidly ages, accurately determining the long-term impact of NSAIDs is of critical importance. NSAIDs are primarily taken for inflammatory conditions such as arthritis and pain syndromes which are themselves associated with an increased risk of frailty through shared inflammatory pathophysiology (31). Moreover, the adverse effects of NSAIDs may lead to an increased risk of frailty. Most studies have focused on the relation between NSAID use and risk of a single condition such as peptic ulcer disease, hypertension, heart failure, CVD, and acute renal failure (32,33). These chronic conditions are both risk factors for frailty and likely fall on the mechanistic pathways through which NSAIDs increase the risk of frailty. Finally, frailty itself is associated with greater analgesic use (34).
Much of the existing literature on NSAIDs and frailty has focused on short-term adverse reactions and drug–drug interactions (35). Studies specifically examining the association between NSAID use and prevalent or incident frailty are sparse. In 14 208 participants in the Health and Retirement Study, a nationally representative study of older Americans, prescription pain medication use was significantly associated with an increased risk of frailty over an average of 3 years, with a hazard ratio 1.58 (95% CI 1.4–1.8); however, type of pain medication was not specified (36). A cohort study of 2 238 individuals in Taiwan and another study of 605 community-dwelling older adults in Finland both reported a significant association between analgesic use and increased risk of prevalent frailty; however, analyses were done in aggregate and did not separate NSAID use (34,37).
While, theoretically, NSAIDs should lower the inflammatory milieu that leads to frailty, there are several reasons that may explain the increased risk associated with frailty found in this study, particularly in the context of prior work which found a protective association between long-term aspirin use and frailty. Unlike NSAIDs, aspirin is typically prescribed for prevention of CVD. Aspirin exerts its effects via an irreversible inhibition of the cyclooxygenase (COX) enzymes with a higher affinity for COX-1 over COX-2 (38). In platelets, there is a permanent inhibition of the COX-1-mediated thromboxane A2 synthesis throughout the entire dosing interval (39), leading to the antithrombotic effects of aspirin, with the greatest anti-inflammatory effects seen at higher doses. NSAIDs, however, reversibly inhibit COX enzymes only during their dosage interval, with only transient inhibition of platelet aggregation (40). The increased CVD risk associated with NSAIDs is in part due to inhibition of COX-2-mediated PGI-2 (prostaglandin-2), which has a cardioprotective effect through vasodilation and platelet inhibition (39). Increase in blood pressure and the blunting of antihypertensive drugs is another well-known adverse effect of NSAIDs. Given the bidirectional relationship between frailty and CVD (5), it is conceivable that the increased risk of frailty with NSAIDs may be due in part to the increased cardiovascular risk.
Increasing attention has focused on discovery of specialized proresolving lipid mediators such as lipoxins, resolvins, and protectins in the resolution of inflammation. These substances are normally synthesized by the COX-2-mediated arachidonic acid (AA) metabolism (41). Use of aspirin, which preferentially inhibits COX-1, leads to COX-2-mediated synthesis of 15(R)-hydroxyeicosatetraenoic acid from AA, which is further metabolized into aspirin-triggered lipoxins (ATLs). ATLs exert numerous anti-inflammatory effects such as inhibition of neutrophil recruitment and the inhibition of synthesis of reactive oxygen species and pro-inflammatory cytokines (41). Aspirin, even at low doses, has been shown to increase the production of ATLs, suggesting that it might contribute to its cardiovascular benefit (41). However, this effect is not shared by other NSAIDs which may additionally explain the reverse association seen in the current study.
The many side effects of NSAIDS, including bleeding, renal dysfunction, and cardiovascular risk, make this a class of medications to potentially avoid in older adults (42). The increased prevalence of frailty found in our study has implications for the shared decision-making conversations clinicians and patients have regarding choice and duration of analgesics. When necessary, alternatives to opiates are always preferred; however, this work adds yet an additional reason to explore alternative therapies and use NSAIDs for the shortest time possible.
Our study has several strengths. The PHS is a well-characterized cohort with well-defined covariates and details about duration of NSAID use. We were able to define frailty according to 2 leading theories with similar results. We used propensity score methods to account for confounding by indication.
However, there are important limitations to acknowledge as well. First, as only male participants were included in PHS, this study should be repeated in women. Second, drug use was from self-report and we did not have information on type of NSAID or dose, as it is possible that different NSAIDs may have differing associations with frailty. We did not have direct measures of frailty, although we did use 2 well-validated measures based on questionnaire data. There may still be residual confounding, although we used propensity score methods to account for confounding by indication. Because frailty was only assessed at 1 time, we could not investigate whether NSAID use increases the risk of incident frailty. Finally, even with robust propensity score methods reverse causation remains possible.
Conclusion
In this prospective, observational cohort we found that long-term NSAID use was associated with an increased prevalence of frailty in older men, even after consideration of multimorbidity and health behaviors. This work should be repeated in women. Additional study is needed to better understand the role of anti-inflammatory medication in older adults and its implication for overall health.
Supplementary Material
Acknowledgments
The authors gratefully acknowledge the patients and research staff who participated in the Physicians’ Health Study.
Contributor Information
Ariela R Orkaby, New England GRECC (Geriatric Research, Education, and Clinical Center), VA Boston Healthcare System, Boston, Massachusetts, USA; Division of Aging, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA.
Rachel Ward, New England GRECC (Geriatric Research, Education, and Clinical Center), VA Boston Healthcare System, Boston, Massachusetts, USA; Massachusetts Veterans Epidemiology and Research Information Center (MAVERIC), VA Boston Healthcare System, Boston, Massachusetts, USA.
Jiaying Chen, Division of Aging, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA.
Akshay Shanbhag, Department of Gerontology, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.
Howard D Sesso, Division of Aging, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA; Division of Preventive Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA.
J Michael Gaziano, Division of Aging, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA; Massachusetts Veterans Epidemiology and Research Information Center (MAVERIC), VA Boston Healthcare System, Boston, Massachusetts, USA.
Luc Djousse, Division of Aging, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA; Massachusetts Veterans Epidemiology and Research Information Center (MAVERIC), VA Boston Healthcare System, Boston, Massachusetts, USA.
Jane A Driver, New England GRECC (Geriatric Research, Education, and Clinical Center), VA Boston Healthcare System, Boston, Massachusetts, USA; Division of Aging, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA.
Funding
This work was supported by an Eleanor & Miles Shore/Brigham and Women’s Faculty Career Development award to A.R.O. A.R.O. is also supported by Veterans Affairs CSR&D CDA-2 IK2-CX001800 and R03-AG060169. The Physicians’ Health Study is funded by grants CA-34944, CA-40360, and CA-097193 from the National Cancer Institute and grants HL-26490 and HL-34595 from the National Heart, Lung, and Blood Institute, Bethesda, MD. The funders played no role in study design, collection, analysis, interpretation of data, writing of the report, or in the decision to submit the paper for publication.
Conflict of Interest
None declared.
Author Contributions
A.R.O. (lead investigator) conceived the study; obtained peer reviewed funding; developed the analysis plan; interpreted study results; and wrote the first draft of the manuscript. R.W. and J.C. developed and performed the statistical analysis; interpreted study results. A.S. interpreted study results. H.D.S., J.M.G., and L.D. obtained peer reviewed funding, provided study oversight; interpreted study results; J.A.D. conceived the study; developed the analysis plan; interpreted study results; provided study oversight; and wrote the first draft of the manuscript. All authors critically revised and approved the final manuscript. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. A.R.O. is the guarantor of the study and affirms that the manuscript is an honest, accurate, and transparent account of the study being reported.
References
- 1. Fried LP, Tangen CM, Walston J, et al. ; Cardiovascular Health Study Collaborative Research Group . Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146–M156. doi: 10.1093/gerona/56.3.m146 [DOI] [PubMed] [Google Scholar]
- 2. Searle SD, Mitnitski A, Gahbauer EA, Gill TM, Rockwood K. A standard procedure for creating a frailty index. BMC Geriatr. 2008;8:24. doi: 10.1186/1471-2318-8-24 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Clegg A, Young J, Iliffe S, Rikkert MO, Rockwood K. Frailty in elderly people. Lancet. 2013;381(9868):752–762. doi: 10.1016/S0140-6736(12)62167-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Liu CK, Lyass A, Larson MG, et al. Biomarkers of oxidative stress are associated with frailty: the Framingham Offspring Study. Age (Dordr). 2016;38(1):1. doi: 10.1007/s11357-015-9864-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Afilalo J, Alexander KP, Mack MJ, et al. Frailty assessment in the cardiovascular care of older adults. J Am Coll Cardiol. 2014;63(8):747–762. doi: 10.1016/j.jacc.2013.09.070 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Hubbard RE, Woodhouse KW. Frailty, inflammation and the elderly. Biogerontology. 2010;11(5):635–641. doi: 10.1007/s10522-010-9292-5 [DOI] [PubMed] [Google Scholar]
- 7. Franceschi C, Campisi J. Chronic inflammation (inflammaging) and its potential contribution to age-associated diseases. J Gerontol A Biol Sci Med Sci. 2014;69(suppl 1):S4–S9. doi: 10.1093/gerona/glu057 [DOI] [PubMed] [Google Scholar]
- 8. López-Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G. The hallmarks of aging. Cell. 2013;153(6):1194–1217. doi: 10.1016/j.cell.2013.05.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Walston J, Hadley EC, Ferrucci L, et al. Research agenda for frailty in older adults: toward a better understanding of physiology and etiology: summary from the American Geriatrics Society/National Institute on Aging Research Conference on Frailty in Older Adults. J Am Geriatr Soc. 2006;54(6):991–1001. doi: 10.1111/j.1532-5415.2006.00745.x [DOI] [PubMed] [Google Scholar]
- 10. Rosenstein ED, Kunicka J, Kramer N, Goldstein G. Modification of cytokine production by piroxicam. J Rheumatol. 1994;21(5):901–904. [PubMed] [Google Scholar]
- 11. Beyer I, Njemini R, Bautmans I, Demanet C, Mets T. Immunomodulatory effect of NSAID in geriatric patients with acute infection: effects of piroxicam on chemokine/cytokine secretion patterns and levels of heat shock proteins. A double-blind randomized controlled trial (ISRCTN58517443). Cell Stress Chaperones. 2012;17(2):255–265. doi: 10.1007/s12192-011-0304-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Afilalo J, Karunananthan S, Eisenberg MJ, Alexander KP, Bergman H. Role of frailty in patients with cardiovascular disease. Am J Cardiol. 2009;103(11):1616–1621. doi: 10.1016/j.amjcard.2009.01.375 [DOI] [PubMed] [Google Scholar]
- 13. Klein BE, Klein R, Knudtson MD, Lee KE. Frailty, morbidity and survival. Arch Gerontol Geriatr. 2005;41(2):141–149. doi: 10.1016/j.archger.2005.01.002 [DOI] [PubMed] [Google Scholar]
- 14. Bandeen-Roche K, Xue QL, Ferrucci L, et al. Phenotype of frailty: characterization in the Women’s Health and Aging Studies. J Gerontol A Biol Sci Med Sci. 2006;61(3):262–266. doi: 10.1093/gerona/61.3.262 [DOI] [PubMed] [Google Scholar]
- 15. Woods NF, LaCroix AZ, Gray SL, et al. ; Women’s Health Initiative . Frailty: emergence and consequences in women aged 65 and older in the Women’s Health Initiative Observational Study. J Am Geriatr Soc. 2005;53(8):1321–1330. doi: 10.1111/j.1532-5415.2005.53405.x [DOI] [PubMed] [Google Scholar]
- 16. Ridker PM, Cushman M, Stampfer MJ, Tracy RP, Hennekens CH. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med. 1997;336(14):973–979. doi: 10.1056/NEJM199704033361401 [DOI] [PubMed] [Google Scholar]
- 17. Waldstein SR, Wendell CR, Seliger SL, Ferrucci L, Metter EJ, Zonderman AB. Nonsteroidal anti-inflammatory drugs, aspirin, and cognitive function in the Baltimore Longitudinal Study of Aging. J Am Geriatr Soc. 2010;58(1):38–43. doi: 10.1111/j.1532-5415.2009.02618.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Orkaby AR, Yang L, Dufour AB, et al. Association between long-term aspirin use and frailty in men: the Physicians’ Health Study. J Gerontol A Biol Sci Med Sci. 2021;76(6):1077–1083. doi: 10.1093/gerona/glaa233 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Steering Committee of the Physicians’ Health Study Research Group. Final report on the aspirin component of the ongoing Physicians’ Health Study. N Engl J Med. 1989;321:129–135. doi: 10.1056/NEJM198907203210301 [DOI] [PubMed] [Google Scholar]
- 20. Sesso HD, Christen WG, Bubes V, et al. Multivitamins in the prevention of cardiovascular disease in men: the Physicians’ Health Study II randomized controlled trial. JAMA. 2012;308(17):1751–1760. doi: 10.1001/jama.2012.14805 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Driver JA, Logroscino G, Lu L, Gaziano JM, Kurth T. Use of non-steroidal anti-inflammatory drugs and risk of Parkinson’s disease: nested case-control study. BMJ. 2011;342:d198. doi: 10.1136/bmj.d198 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Mitnitski AB, Mogilner AJ, Rockwood K. Accumulation of deficits as a proxy measure of aging. ScientificWorldJournal. 2001;1:323–336. doi: 10.1100/tsw.2001.58 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Rockwood K, Andrew M, Mitnitski A. A comparison of two approaches to measuring frailty in elderly people. J Gerontol A Biol Sci Med Sci. 2007;62(7):738–743. doi: 10.1093/gerona/62.7.738 [DOI] [PubMed] [Google Scholar]
- 24. Orkaby AR, Hshieh TT, Gaziano JM, Djousse L, Driver JA. Comparison of two frailty indices in the Physicians’ Health Study. Arch Gerontol Geriatr. 2017;71:21–27. doi: 10.1016/j.archger.2017.02.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Hoover M, Rotermann M, Sanmartin C, Bernier J. Validation of an index to estimate the prevalence of frailty among community-dwelling seniors. Health Rep. 2013;24(9):10–17. [PubMed] [Google Scholar]
- 26. Pajewski NM, Williamson JD, Applegate WB, et al. ; SPRINT Study Research Group . Characterizing frailty status in the systolic blood pressure intervention trial. J Gerontol A Biol Sci Med Sci. 2016;71(5):649–655. doi: 10.1093/gerona/glv228 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ensrud KE, Ewing SK, Cawthon PM, et al. ; Osteoporotic Fractures in Men Research Group . A comparison of frailty indexes for the prediction of falls, disability, fractures, and mortality in older men. J Am Geriatr Soc. 2009;57(3):492–498. doi: 10.1111/j.1532-5415.2009.02137.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Ensrud KE, Ewing SK, Taylor BC, et al. Comparison of 2 frailty indexes for prediction of falls, disability, fractures, and death in older women. Arch Intern Med. 2008;168(4):382–389. doi: 10.1001/archinternmed.2007.113 [DOI] [PubMed] [Google Scholar]
- 29. Kolasinski SL, Neogi T, Hochberg MC, et al. 2019 American College of Rheumatology/Arthritis Foundation guideline for the management of osteoarthritis of the hand, hip, and knee. Arthritis Care Res. 2020;72:149–162. doi: 10.1002/acr.24131 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Wongrakpanich S, Wongrakpanich A, Melhado K, Rangaswami J. A comprehensive review of non-steroidal anti-inflammatory drug use in the elderly. Aging Dis. 2018;9(1):143–150. doi: 10.14336/AD.2017.0306 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Nessighaoui H, Lilamand M, Patel KV, et al. Frailty and pain: two related conditions. J Frailty Aging. 2015;4(3):144–148. doi: 10.14283/jfa.2015.53 [DOI] [PubMed] [Google Scholar]
- 32. Schjerning AM, McGettigan P, Gislason G. Cardiovascular effects and safety of (non-aspirin) NSAIDs. Nat Rev Cardiol. 2020;17(9):574–584. doi: 10.1038/s41569-020-0366-z [DOI] [PubMed] [Google Scholar]
- 33. Marcum ZA, Hanlon JT. Recognizing the risks of chronic nonsteroidal anti-inflammatory drug use in older adults. Ann Longterm Care. 2010;18(9):24–27. [PMC free article] [PubMed] [Google Scholar]
- 34. Koponen MP, Bell JS, Karttunen NM, Nykänen IA, Desplenter FA, Hartikainen SA. Analgesic use and frailty among community-dwelling older people: a population-based study. Drugs Aging. 2013;30(2):129–136. doi: 10.1007/s40266-012-0046-8 [DOI] [PubMed] [Google Scholar]
- 35. Zazzara MB, Palmer K, Vetrano DL, Carfì A, Graziano O. Adverse drug reactions in older adults: a narrative review of the literature. Eur Geriatr Med. 2021;12(3):463–473. doi: 10.1007/s41999-021-00481-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Cil G, Park J, Bergen AW. Self-reported prescription drug use for pain and for sleep and incident frailty. J Am Geriatr Soc. 2019;67(12):2474–2481. doi: 10.1111/jgs.16214 [DOI] [PubMed] [Google Scholar]
- 37. Chen CY, Wu SC, Chen LJ, Lue BH. The prevalence of subjective frailty and factors associated with frailty in Taiwan. Arch Gerontol Geriatr. 2010;50(suppl 1):S43–S47. doi: 10.1016/S0167-4943(10)70012-1 [DOI] [PubMed] [Google Scholar]
- 38. Ornelas A, Zacharias-Millward N, Menter DG, et al. Beyond COX-1: the effects of aspirin on platelet biology and potential mechanisms of chemoprevention. Cancer Metastasis Rev. 2017;36(2):289–303. doi: 10.1007/s10555-017-9675-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Brune K, Patrignani P. New insights into the use of currently available non-steroidal anti-inflammatory drugs. J Pain Res. 2015;8:105–118. doi: 10.2147/JPR.S75160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Koch M, Dezi A, Ferrario F, Capurso I. Prevention of nonsteroidal anti-inflammatory drug-induced gastrointestinal mucosal injury. A meta-analysis of randomized controlled clinical trials. Arch Intern Med. 1996;156(20):2321–2332. doi: 10.1001/archinte.1996.00440190069008 [DOI] [PubMed] [Google Scholar]
- 41. Costa AC, Reina-Couto M, Albino-Teixeira A, Sousa T. Aspirin and blood pressure: effects when used alone or in combination with antihypertensive drugs. Rev Port Cardiol. 2017;36(7–8):551–567. doi: 10.1016/j.repc.2017.05.008 [DOI] [PubMed] [Google Scholar]
- 42. By the 2019 American Geriatrics Society Beers Criteria® Update Expert Panel. American Geriatrics Society 2019 updated AGS Beers Criteria® for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2019;67(4):674–694. doi: 10.1111/jgs.15767 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.