Acetaminophen Safety in the Elderly and Under-recognized Risks

Many clinical guidelines for osteoarthritis (OA) recommend acetaminophen as the first line oral pharmacologic, because of perceived safety over other other analgesics, such as non-steroidal anti-inflammatory drugs (NSAIDs).  However, acetaminophen can cause COX-dependent side effects similar to NSAIDs and have unclear efficacy in OA.  Given previous studies of acetaminophen safety are post-marketing, observational studies, a channelling bias may be observed, where individuals at higher risk of gastrointestinal and cardiovascular adverse events are less likely to be given NSAIDs, and more likely to receive acetaminophen.  To better address the gap in knowledge on acetaminophen safety and mitigate such channeling bias, Kaur et al performed a cohort study to compare acetaminophen exposure versus non-exposure to any analgesics for major adverse events in the general population and a subgroup with OA (1).

The objectives of the study were to (1) examine the incidence of major adverse events, including gastrointestinal (GI), cardiovascular (CV) and renal, in those prescribed acetaminophen compared to those not on analgesics, and (2) examine the dose-response relationship between acetaminophen and specific adverse events.  This was a population cohort study with data obtained from a large United Kingdom database (Clinical Practice Research Datalink, CPRD).  The CPRD has anonymized patient data from over 700 general practices, covering 17 million UK residents.  The study included older adults aged 65 years or older at the index date, the date of the first acetaminophen prescription between 1998 and 2018, who had been registered for at least 12 months with a general practice.  They excluded participants with a diagnosis of the outcome of interest before the follow-up period.

The authors defined exposure to acetaminophen as participants issued at least two acetaminophen prescriptions within 6 months, not in combination with other analgesics such as codeine, excluding occasional use for acute conditions (such as headache or influenza).  The control group (unexposed) included participants aged 65 years or older with less than two acetaminophen prescriptions within six months during the study period.  The controls were matched to acetaminophen with propensity scoring.  The outcomes of interest were incident diagnoses of GI conditions (peptic ulcer bleed, uncomplicated ulcers, lower GI bleeding), CV (hypertension, myocardial infarction, heart failure), and renal (chronic renal failure).  The researchers calculated the propensity score for acetaminophen prescription, and used inverse probability treatment weighted (IPTW) propensity score and matched analyses to account for confounding.  Potential covariates used to calculate the propensity score included age, sex, Charlson comorbidity index, opioids, NSAIDs, aspirin, H2-receptor blockers, proton pump inhibitors (PPI), body mass index.   Cox proportional hazards regression modeling was used.

From an initial group of nearly 2.7 million, the study included approximately 180,000 acetaminophen users and 400,000 non-users in this UK population of adults 65 years and older.  Acetaminophen use was associated with an increased risk of peptic ulcer bleeding (aHR 1.24), uncomplicated peptic ulcers (aHR 1.20), lower gastrointestinal bleeding (aHR 1.36), heart failure (aHR 1.09), hypertension (aHR 1.07), and chronic kidney disease (aHR 1.19), compared to the unexposed participants not on acetaminophen.  In the dose-response analysis, association between developing peptic ulcer bleed, uncomplicated ulcers, and chronic renal failure increased with number of acetaminophen prescriptions.  In the OA subgroup, acetaminophen exposure was also associated with increased incidence of lower GI bleeding (aHR 1.20), hypertension (aHR 1.06), and chronic renal failure (aHR 1.15), compared to those with OA not on acetaminophen.

This novel study demonstrated acetaminophen use was associated with increased incidence of gastrointestinal, cardiovascular and renal adverse side effects.  The authors argue the association is strong with robust evidence, given similar risk was observed across different models and subgroup analyses.  They note most randomized controlled trials of acetaminophen may not have found these adverse effects due to trials designed for efficacy (not adverse events), solely report short-term effects, and had healthier and younger participants.  The precise mechanism by how acetaminophen may result in these adverse events is not entirely clear.  Finally, there are some limitations to the study, including (1) potential unknown confounders, (2) the selection of an older age group at higher risk of GI, CV and renal adverse events at baseline, and (3) the fact acetaminophen is often used episodically for varied indications, making “users” difficult to define.

Ultimately, the authors conclude this population-based cohort study provides robust evidence of potential safety concerns of acetaminophen use in older adults.  While the incidence of side effects with acetaminophen may be lower than that of NSAIDs, the authors suggest that we should still challenge whether acetaminophen is retained as first-line oral analgesic for conditions like osteoarthritis, given the potential adverse effect profile and unclear efficacy.

References:

[1] Kaur J, Nakafero G, Abhishek A et al. Incidence of side effects associated with acetaminophen in people aged 65 years or more: a prospective cohort study using data from the Clinical Practice Research Datalink. AC&R.  Novemeber 2024; doi: 10.1002/acr.25471.