This study analyzed a sub-cohort of 4,110 community-dwelling adults aged 60 years or older drawn from the SABE Colombia survey, covering the Antioquia, Caldas, Risaralda, and Quindío regions. At baseline, polypharmacy (defined in this study as 5–9 concurrent medications) was present in 20.2% of participants, while excessive polypharmacy (10 or more concurrent medications) was present in 2.1%.
Vital status for cohort members was adjudicated using the National Health System Resources Administrator (ADRES) database. Mean follow-up time was 79 months, during which 1,092 deaths (26.6% of the sample) were recorded. The investigators report that the study received ethics approval from the Ethics Committee of the Universidad Pontificia Bolivariana and complied with national data-use guidelines. The manuscript is presented as a preprint and has not undergone peer review.
Medication burden categories used in the analysis were: no/few medications (reference), polypharmacy defined as 5–9 medications, and excessive polypharmacy defined as 10 or more medications. These thresholds were applied to baseline medication data from the survey to classify exposure.
The authors used extended Cox proportional hazards models to estimate associations between medication burden categories and all-cause mortality. Models adjusted for sociodemographic variables, measures of multimorbidity, and functional dependency. The approach accounts for time-to-event data and permitted estimation of hazard ratios (HRs) with corresponding 95% confidence intervals and p values.
Across a mean 79-month follow-up, 1,092 deaths (26.6%) occurred in the 4,110-person cohort. After multivariable adjustment, moderate polypharmacy (5–9 medications) was associated with a statistically significant increase in mortality risk (HR 1.17; 95% CI 1.02–1.31; p = 0.029). Excessive polypharmacy (≥10 medications) showed a larger association with mortality (HR 1.82; 95% CI 1.34–2.47; p < 0.001). The authors highlight that the markedly higher hazard at the ≥10 medication threshold suggests a non-linear relationship between medication count and survival.
The pattern of results indicates that increased medication burden correlates with worse survival independent of measured sociodemographic factors, multimorbidity, and functional status. While moderate polypharmacy carried a modest increased risk, the nearly doubled hazard associated with excessive polypharmacy implies that risk may accelerate beyond a certain medication count rather than increase linearly.
The authors recommend structured medication review and targeted deprescribing interventions to reduce high pharmacological accumulation among older adults, particularly in resource-constrained healthcare systems similar to the study setting. They position medication burden as a modifiable exposure that could potentially affect older adult survival if addressed through system-level and clinical interventions.
Strengths reported within the study include a population-based, longitudinal design with a relatively long mean follow-up (79 months) and linkage to a national registry for mortality ascertainment. Limitations inherent to the report include the observational design (which precludes causal inference) and any residual confounding despite multivariable adjustment; the source text did not provide further detail on measurement of medication exposure over time or potential changes in prescribing during follow-up. The data that underpin the analyses are available from the corresponding author on request, subject to privacy and ethical restrictions.
In this SABE Colombia sub-cohort, baseline polypharmacy and especially excessive polypharmacy were independent predictors of higher all-cause mortality over a mean follow-up of 79 months. The findings support prioritizing medication review and deprescribing strategies for older adults, and they call for further research—ideally peer-reviewed replication studies and interventional trials—to determine whether reducing medication burden improves survival or other patient-centered outcomes. As this report is a preprint, its findings should be interpreted cautiously pending peer review and additional confirmatory evidence.