Patterns of comorbidity and their effects on mortality in rheumatoid arthritis: a study of 5658 UK Biobank participants

Problem

Rheumatoid arthritis (RA) is a debilitating, chronic autoimmune disease characterised by inflammation of the synovial joints. To date, few studies have quantified the predominance or type of comorbid long-term conditions (LTCs) in rheumatoid arthritis, nor examined their effect on mortality. This work aims to address these gaps using a large prospective cohort study.

Approach

Using data from UK Biobank (N=502,543), we sampled participants who self-reported RA. Comorbidity was measured in this group by examining 42 non-RA LTCs self-reported by participants. We examined lifestyle factors including smoking status, frequency of alcohol intake, body mass index (BMI) and physical activity level alongside demographic factors such as sex, age and socioeconomic status. Kaplan-Meier plots were used to examine differences in proportion of deaths between patients with RA and those without. Cox proportional hazards models were used to examine all-cause mortality outcomes in patients with RA. This model controlled for number of LTCs in addition to lifestyle and demographic factors listed above.

Findings

5658 of the 502,543 of participants in Biobank self-reported RA (1.1%). 74.7% of participants reported at least one LTC besides RA (30.1% one LTC, 22.1% two LTCs, 12.1% three LTCs, 6.2% four LTCs, 4.2% five or more LTCs). In comparison, only 63.8% of participants without RA reported one or more LTC. The most predominant physical health-based LTC in patients with RA was hypertension (35.6%; 26.4% in population without RA), whilst the most predominant mental health based LTC was depression (7.0%; 5.7% in population without RA). 311 participants (5.4%) self-reporting RA died in the follow up period (median=7 years) compared with 2.8% who reported no RA. Participants with RA experienced a 39% increase in all-cause mortality compared with patients without (hazard ratio (HR) 1.39, 95% confidence intervals (CI) 1.24-1.57) when using a Cox’s proportional hazards model controlling for sex, age, socioeconomic status (Townsend score), smoking status, alcohol intake, BMI, physical activity level and number of LTCs.

Consequences

Participants with RA had an increased proportion of long-term conditions compared to those without RA. Further, participants with RA had an increased rate of all-cause mortality after controlling for demographic factors, lifestyle factors and numbers of long-term conditions. Further work will focus on identifying other factors that influence this increased mortality.