Association of Aspirin Use With Mortality Risk Among Older Adult Participants in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial

This cohort study uses data from the Prostate, Lung, Colorectal, and Ovarian Cancer Screen Trial to examine the associations among regular aspirin use, obesity, and all-cause and cancer mortality among older adults.


Introduction
Aspirin use of 10 years or more is estimated to reduce cancer incidence and mortality. [1][2][3][4] Across epidemiological studies, the most significant reductions in risk have been noted in gastrointestinal (GI) cancers. 2,4 Notably, results from the NIH-AARP study, 5 the Nurses' Health Study, 6 and Health Professionals Follow-up Study 7,8 demonstrated that long-term regular aspirin use was associated with reduced risk of obesity-and inflammation-associated cancers, particularly colorectal cancer (CRC). This result has been additionally illustrated in a 2010 meta-analysis of randomized clinical trials. 4 The association of aspirin with reduced cancer incidence and mortality has been consistently robust among GI cancers; however, there is limited evidence suggesting that aspirin use is associated with reduced mortality risk among other cancer types. 3,[9][10][11] The US Preventive Services Task Force (USPSTF) recommends low-dose aspirin use for the prevention of cardiovascular disease and CRC among average-risk individuals aged 50 to 59 years with a 10% or greater 10-year risk of cardiovascular disease; however, individuals aged 60 to 69 years should have an individualized approach, while evidence for use among individuals 70 years and older remains insufficient. 12 It is important to note that a 2015 systematic review of the literature by the USPSTF 11 found that individuals with daily aspirin dosing up to 1200 mg for a duration of 4 or more years exhibited a significant reduction in all-cause and cancer mortality, demonstrating a similar benefit of low-dose and standard aspirin.
Recently, the results of the Aspirin in Reducing Events in the Elderly (ASPREE) study 13 drew further attention of the potential cancer preventive effect of aspirin. The ASPREE study was a randomized clinical trial conducted in the United States and Australia investigating the efficacy of 100 mg of aspirin daily vs placebo in extending disability-free survival among elderly individuals (recruitment age 65 years or 70 years) during 4.7 years of follow-up. 13 Unexpectedly, an increased risk of all-cause and cancer-related death was observed in the group randomized to aspirin. 14 Thus, based on the results of the ASPREE randomized clinical trial that conflict with the preponderance of data showing a benefit of aspirin use, we examined the association of aspirin use with mortality risk among older adults enrolled in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. In addition, given data suggesting that aspirin may be less effective among individuals who are overweight or obese, we conducted exploratory analyses stratified by body mass index (BMI), calculated as weight in kilograms divided by height in meters squared. 1 had a valid baseline questionnaire (BQ) after study enrollment, and had reported information on aspirin use at baseline. Completion of a follow-up supplemental questionnaire (SQ) by participants between 2006 and 2008 was not required for inclusion into the study cohort. All responses to the BQ and SQ were ascertained by self-report. The full content of both questionnaires has been published previously. 17,18 Briefly, the BQ asked, "During the last 12 months, have you regularly used aspirin or aspirin-containing products, such as Bayer, Bufferin, or Anacin (Please do not include aspirin-free products such as Tylenol and Panadol)." If the respondent answered yes, the frequency categories were given as less than 2 per month, 2 to 3 per month, 1 per week, 2 per week, 3 to 4 per week, 1 per day, or 2 or more per day. The SQ asked, "During the last 12 months, about how often did you usually take aspirin (examples of aspirin include Bayer, Bufferin, Anacin, and baby aspirin)?" The frequency categories were given as none or less than once per month, 1 to 3 times per month, 1 to 2 times per week, 3 to 6 times per week, and 7 or more times per week. To allow for consistent values between BQs and SQs, we collapsed the aspirin use frequency variables into no aspirin use or less than once per month, 1 to 3 times per month, 1 to 2 times per week, and 3 or more times per week.
Usual dose information was not available on the BQ; therefore, this analysis does not consider dose.
We categorized BMI as less than 20, 20 to 24.9, 25 to 29.9, and 30 or higher.
Initial analysis of the PLCO Cancer Screening Trial was completed after 13 years of follow-up or on December 31, 2009, whichever came first. 15 Deaths were ascertained via annual study update questionnaire, individual reports, or death certificates. Beginning in 2011, participants were reconsented for follow-up. This allowed for linkage to the National Death Index, which extended mortality follow-up for up to 20 years after randomization. 19 For those reconsenting, the end of mortality follow-up was December 31, 2015 or their date of death, whichever came first. For the approximately 15% of individuals who were alive in 2011 but who declined extended follow-up, the end of mortality follow-up was their declining date, which was generally in 2011.

Statistical Analysis
The purpose of this analysis was to assess the associations among aspirin use, BMI, and mortality of all causes, any cancer, gastrointestinal (GI) cancer, and CRC among individuals in the PLCO Cancer Screening Trial who survived to age 65 years and older (Figure 1). Participant follow-up time was measured from the time of cohort entry, which was either the date of PLCO Cancer Screening Trial  The current study cohort included participants who provided the baseline questionnaire with completed aspirin frequency and survived until age 65 years. Static variables included in the models were sex (male or female), race/ethnicity (white non-Hispanic, black non-Hispanic, or other), and randomization arm (intervention or control). We additionally stratified by BMI to evaluate the association of aspirin use and BMI with mortality.

Causes of death were determined by International Classification of Diseases, Ninth Revision
(ICD-9) codes. 20 All-cause mortality was defined as a reported death of any cause. Any cancer mortality was defined as a reported cause of death by cancer based on the official PLCO Cancer Screening Trial definitions, derived from standard ICD-9 groupings. Cause of death by GI cancer was determined by ICD-9 codes for esophagus (150), gastric (151), and colorectal (153 and 154) cancer, whereas CRC was determined by ICD-9 codes for CRC only (153 and 154). All statistical analyses were performed in SAS statistical software version 9.4 (SAS Institute). P values were 2-tailed, and statistical significance was set at less than .05.

Results
The eligibility criteria for the study analysis are outlined in Figure 1. , and 814 individuals (6.6%) who died of CRC. Demographic characteristics of the cohort are summarized in Table 1.
The results of the Cox proportional hazards regression model are presented in Table 2. In our model, any aspirin use was associated with reduced all-cause and cancer-specific mortality. Reduced risk of all-cause mortality was associated with using aspirin 1 to 3 times per month (HR, 0.84; 95%   To explore the potential association of aspirin use and BMI with mortality risk, we stratified the cohort according to BMI. The results are shown in Figure 2. Compared with no aspirin use and among individuals with a BMI of 20 or higher, aspirin use was associated with reduced risk of all-cause mortality when used 1 to 3 times per month (BMI 20-24.9: HR, 0.83; 95% CI, 0.76-0.89; P < .001;  No aspirin use was used as the reference for all comparisons. The model was adjusted for sex, race/ethnicity, randomization group, smoking status, ibuprofen use 3 times or more per week, and history of myocardial infarction, stroke, hypertension, and diabetes. Body mass index was calculated as weight in kilograms divided by height in meters squared.  Figure 2B). No statistically significant risk reduction was observed for cancer mortality among participants who used aspirin 1 to 2 times per week or 1 to 3 times per month, except for aspirin use 1 to 3 times per month among individuals with BMI 20 to 24.9 (HR, 0.83; 95% CI, 0.72-0.95; P = .007) ( Figure 2B).
We additionally examined the association between reported aspirin use at baseline and at the time of the SQ with risk of mortality. Participants who had reported aspirin use at follow-up only or consistent users were associated with a reduced risk of mortality ( Table 3)

Discussion
This cohort study found that aspirin use among individuals 65 years and older was associated with a lower risk of mortality. This observation was consistent across all causes of mortality (ie, all-cause, cancer, GI cancer, and CRC); however, the greatest reduction in risk was noted for CRC mortality among individuals who used aspirin 3 or more times per week. Additionally, our exploratory analyses investigating the potential associations among aspirin use, BMI, and mortality risk suggest that the efficacy of aspirin as a cancer preventive agent may be associated with BMI. Participants in the PLCO Cancer Screening Trial who were underweight (ie, BMI <20) had no observable benefit associated with aspirin use, while those with BMI 20 or higher were associated with reduced mortality risk, particularly with aspirin use 3 or more times per week. reduced risk of CRC mortality was only associated with individuals with BMI 20 to 29.9 who reported aspirin use 3 or more times per week.
The efficacy of prophylactic aspirin use for prevention of cancer incidence and mortality has been debated; however, the most evidence from prospective cohorts and secondary analyses from clinical trials indicates a protective association with aspirin use. A 2016 systematic analysis of primary and secondary cardiovascular prevention trials 21 found reduced CRC incidence 10 to 19 years after aspirin use initiation. This association persists among investigations of aspirin use and cancer mortality. A 2011 systematic analysis of 8 clinical trials 22 found that daily aspirin use was associated with reduced risk of death of several cancers, with increased benefit associated with long-term use. A similar association was observed in a 2018 cohort study 23 of veterans in which aspirin users were associated with reduced risk of CRC mortality compared with nonusers. This association was additionally examined in the Cancer Prevention Study-II Nutrition Cohort, 24 in which daily aspirin use was associated with reduced overall cancer mortality. These observations are in contrast with data from the ASPREE trial. 13 However, the interpretation of the ASPREE results is limited owing a lack of an association of aspirin with cancer and CRC incidence and the short duration of follow-up. 25 With additional follow-up, an association of aspirin with lower cancer incidence and death may have emerged. 26  The observation that BMI may be associated with efficacy of aspirin in individuals 65 years and older is notable; however, our findings require further confirmation. Increasing rates of overweight and obesity globally may substantially alter the population-based efficacy of cancer prevention prophylatics. 32 Studies have suggested that aspirin has reduced effectiveness as a primary prevention modality among individuals who are obese owing to decreased bioavailability and antithrombotic efficacy; however, this study did not find an association of overweight or obesity with decreased efficacy. 5,33-35 Therefore, although aspirin use is associated with benefit as a cancer preventive agent, the changing characteristics of the global population may alter its efficacy and must be considered along with age and risk of bleeding before recommending aspirin for cancer prevention.

Limitations and Strengths
Our study has several limitations. First, this is a secondary analysis of a randomized intervention (screening) trial with self-reported data on aspirin use. When collecting self-reported data, one must be aware of the limitations to these data, such as underreporting parameters of interest and the potential effect of this error in the analyses, such as biasing toward the null. For example, it is possible that aspirin use was underreported if the participant was unaware that a given agent, based upon its trade name, was classified as aspirin. Next, the information on aspirin use may be subject to participant interpretation and measurement error. Regular use of aspirin was not defined in the BQ, so participants could then have interpreted regular aspirin use differently, which may have affected their categorization (eg, none, monthly, weekly, daily). Furthermore, the information collected regarding aspirin dosage was limited. These issues may, in addition to confounding by unmeasured covariates, help explain the lack of a dose-response association for mortality risk in the observed data. Though there is substantial evidence documenting the association of frequent aspirin use with cancer prevention, the biological effect of low frequency use (eg, 1-3 times per month) is uncertain.
Therefore, the noted risk reduction associated with this group may be due to unmeasured confounders. Along similar lines, it remains possible that individuals who reported low or no use of aspirin were unable to tolerate aspirin use for an extended length of time, perhaps owing to risk of bleeding or gastrointestinal injury, and may be inherently different than those who reported aspirin use. 36 Additionally, our analysis was focused on individuals in the PLCO Cancer Screening Trial cohort who survived to and began follow-up at or after age 65 years. However, only a small percentage (5.5%) of the PLCO cohort was ineligible for our study.

JAMA Network Open | Public Health
Although this study had several limitations, the research questions investigated examining the associations among aspirin use, BMI, and mortality adds to our knowledge of modifiable factors associated with cancer prevention efficacy. We were able to explore this question in a large, prospective study with an extended follow-up period.

Conclusions
In this cohort study, we found a significant association of aspirin use with reduced all-cause, any cancer, GI cancer, and CRC mortality among individuals 65 years and older in the PLCO Cancer Screening Trial. Aspirin use was associated with reduced risk of all-cause and any cancer mortality, and aspirin use 3 or more times per week was associated with reduced risk of GI cancer mortality and CRC mortality when stratified by BMI. Future studies should further examine the association of BMI with the efficacy of aspirin as a cancer preventive agent to adapt to the changing global obesity trends.