Effect of Aspirin on CKD Progression in Older Adults: Secondary Analysis From the ASPREE Randomized Clinical Trial

To the Editor:

Aspirin is a commonly prescribed and ‘over-the-counter’ therapy in older persons. While its use in the secondary prevention of cardiovascular disease (CVD) events is well established¹, aspirin is not recommended for primary prevention of CVD in adults aged 60 years or older.² Low dose aspirin increases the risk of bleeding in older persons³, but whether it has any effect on kidney function is not clear^{4, 5}.

We sought to investigate the effect of low dose aspirin on kidney function in healthy older persons enrolled in the ASPirin in Reducing Events in the Elderly (ASPREE) trial (Clinicaltrials.gov NCT01038583).⁶ ASPREE was a large double-blind, randomized, placebo-controlled trial designed to assess whether daily treatment with 100 mg of enteric-coated aspirin could extend the duration of life free of dementia and persistent physical disability.

The aims of the present study were to compare the trajectory of kidney function, as measured by estimated glomerular filtration rate (eGFR) or urinary albumin-creatinine ratio (UACR), in participants randomized to aspirin treatment or placebo from the trial’s commencement until its cessation.

In brief, 19,114 healthy community-dwelling individuals aged ≥70 years (aged ≥65 years for African-American and Hispanic participants in the US) were recruited in Australia and in the US. Recruitment took place between March 2010 and December 2014, with annual assessments conducted from randomization until the intervention period ended in June 2017 (median follow-up 4.7 years). Participants were randomly assigned to receive a 100 mg tablet of enteric-coated aspirin or matching placebo daily in double-blind fashion. For this analysis, 7 participants with stage G5 CKD ⁷ were omitted, as were 1349 participants missing baseline kidney measures. Full details, including the ASPREE trial protocol and main results are reported in detail elsewhere and in the Supplementary Methods (Item S1).^{6, 8, 9}.

Exposure was randomization to aspirin or to placebo. Outcome measures were change in kidney function, assessed as annual decline in eGFR and, separately, annual increase in UACR. No participant was documented as commencing dialysis or receiving a kidney transplant during the intervention period, and participants reaching CKD 5 (eGFR<15) during the trial period were not removed from the analysis.

Linear mixed models were used, which included the group (randomised aspirin versus placebo, i.e. intention to treat), annual visit number (0 [baseline], 1, 2, 3, 4, 5, and 6 years; referred to as “time”), a participant-specific intercept (baseline eGFR or UACR), and a participant-specific slope describing change in eGFR or UACR over time (per annual visit). A treatment-by-time interaction was included to examine whether the trajectory of eGFR or UACR for an average participant differed between the treatment groups. As the distribution of UACR was skewed, it was log (base 2) transformed in all models. To account for dropout due to death, a sensitivity analysis was performed using a shared random effect joint model for longitudinal eGFR or UACR and the overall survival outcome ¹⁰. The survival component was modelled using a Cox proportional hazards model, adjusted for baseline age, sex, aspirin, diabetes and time-dependent value of eGFR or UACR.

The primary analysis consisted of 17,758 participants (Figure S1). Baseline characteristics of the participants were well matched across treatment arms (Table 1). The mean age of the cohort was 75.1 years (SD 4.5) and 56.4% were female. The median number of eGFR and UACR measures per patient were 5 (range, 1 to 7) and 4 (range, 1 to 7), respectively. In total 983 deaths occurred among participants included in the primary analysis (523 in the aspirin group, 460 in the placebo group).

Summary measures over time for eGFR and UACR, by treatment assignment, are presented in Figure 1. Results of the mixed models are shown in Table S1 . Mean annual decline in eGFR was not different in participants randomized to aspirin (−0.97 mL/min/1.73 m², 95% confidence intervals [CI] −1.02, −0.92) compared with those randomized to placebo (−0.99 mL/min/1.73 m2, 95% CI −1.04, −0.94, interaction p-value 0.60). Likewise, annual increase in UACR was similar in participants randomized to aspirin (mean log₂ UACR 0.055, 95% CI 0.050, 0.059) compared with placebo (0.051, 95% CI 0.046, 0.056, interaction p-value 0.30). Results of the joint longitudinal and survival models for both outcomes are presented in Table S2. Results were all consistent with the results of the main analysis models, with no evidence of an effect of aspirin treatment on either eGFR decline or UACR increase over time, allowing for loss to follow-up due to mortality.

In summary, we found no evidence of an effect of aspirin on the trajectories of kidney function, as assessed separately by eGFR and UACR, in healthy community-dwelling older persons, over an average of nearly 5 years of follow-up. The results of our study, the largest available trial of older individuals receiving aspirin compared with placebo, suggests that fears over decline in kidney function associated with low-dose aspirin among older individuals may not be justified.

Supplementary Material