#3118 ERECTILE DYSFUNCTION IS A HARBINGER OF CHRONIC KIDNEY DISEASE

Background and Aims Chronic kidney disease (CKD) shares similar cardiovascular risk factors with erectile dysfunction (ED). ED is an early warning sign for future cardiovascular events, however it is not yet known whether it also behaves as a harbinger of future CKD. Method We identified cases of ED and CKD using diagnostic codes from primary care records, hospital episode statistics, Death Register records, and self-reported diagnoses in a prospective cohort of 94,955 male participants of UK Biobank who had linkage to primary care records, from 01/01/2000 until 01/09/2017. We excluded participants with CKD diagnosed prior to this period as well as before the age of 35. We compared risk of CKD using cause-specific Cox proportional hazards models with competing risk of death, using age as time scale and adjusted for time-fixed covariates (birth cohort, ethnicity, body mass index [BMI], Townsend deprivation index [TDI], and smoking status) and time-dependent covariates (hypertension, diabetes mellitus, ischaemic heart disease, and ischaemic stroke/transient ischaemic attack). We imputed missing values for BMI, TDI, and smoking status (percentage missing ranging between 0.16-0.71%) using multiple imputation with predictive mean matching. As a sensitivity analysis, we repeated the analyses in a propensity score-matched cohort. Results Among 15,848 men diagnosed with ED, 830 developed CKD (71.9/10,000 person-years) median 5.9 years later (IQR 2.8-10.4), compared with 3,452 among 79,107 men without ED (39.6/10,000 person-years). Men with a diagnosis of ED had higher BMIs than men without (28.7 vs 27.7 kg/m2) and were more likely to have other diagnoses such as hypertension (38% at time of diagnosis of ED/56% at any time vs 40% at any time) and diabetes mellitus (21% at time of diagnosis of ED/30% at any time vs 12% at any time). A diagnosis of ED (without hypertension or diabetes mellitus) carried an increased risk of subsequent CKD (Figure 1), with an adjusted cause-specific hazard ratio of 1.53 (95% CI 1.33-1.77). There was significant interaction with hypertension (p<0.001), i.e. the increase in hazard of CKD by ED was smaller (adjusted hazard ratio 1.10, 95% CI 1.00-1.20), whereas there was no significant interaction with diabetes mellitus; i.e. men with a diagnosis of ED had a similar hazard ratio for CKD regardless of presence of diabetes mellitus (p = 0.33 for interaction). These results were numerically similar in a propensity score matched analysis. Conclusion Men who are diagnosed with ED are at a higher risk of developing CKD, independent of comorbidities such as diabetes mellitus. A diagnosis of ED should therefore prompt screening for CKD and its risk factors.