P0284renal Function After Discontinuation of Chronic Lithium Treatment: Findings From the Lisie Retrospective Cohort Study

Abstract

Background and Aims Current guidelines consider lithium a first-line treatment for bipolar disorder to prevent relapse and suicide. Lithium, however, has renal adverse effects that can accumulate over decades of treatment. Studies have found a positive association between duration of lithium use and incidence of chronic kidney disease and prevalence of end stage renal disease. When the kidney function declines the treating clinician is presented with a choice of either continuing lithium, despite the risk of renal failure, or discontinuing lithium, despite an increased risk of relapse of the underlying affective disorder. One recent study showed that concerns about reduced renal function may account for 9% of reasons for lithium discontinuation. Yet it remains unknown, whether after lithium discontinuation, decline of renal function will continue unchanged, become less steep, or whether renal function can even improve. The aim of our study was to investigate the effect of lithium discontinuation on renal function. Method A retrospective mirror-image study, examining the decline of renal function five years before and after lithium discontinuation. Our study was part of LiSIE, a retrospective cohort study of patients with bipolar disorder, schizoaffective disorder and depression in the Swedish county of Norrbotten 1997 - 2017. We included all patients who had discontinued lithium between 1997 and 2013. We excluded patients who (a) had renal diseases possibly affecting renal function acutely (e.g. glomerulonephritis) or (b) used lithium less than 90% of the 5 years of the pre-mirror period. The patients provided informed verbal consent. In accordance with the ethical approval, deceased patients were also included. Patients were followed for 5 years after lithium discontinuation, until they initiated renal replacement therapy, restarted lithium treatment or died. To evaluate renal function, we calculated eGFR by the CKD-EPI -formula from all creatinine measurements taken during out- or in-patient care in the county of Norrbotten. We censored values for episodes of acute kidney injury or values not reflective of the underlying kidney function, such as in terminal cancer and cachexia. All data was manually validated in the medical records. A mixed effects model was fitted including eGFR as dependent variable. Age, hypertension and diabetes mellitus were included as fixed effects. Results Of 1340 patients with lithium treatment 579 patients had discontinued lithium. 172 patients (55.2% women, 44.8% men) with 5335 creatinine samples were eligible for inclusion. Mean age at discontinuation was 61.5 (min. 24.5 max. 91.3). Mean duration of lithium treatment was 15.8 (min. 4.5, max. 42.5) years. 19.2% of patients had diabetes and 48.9% had hypertension diagnosed before or during the study period. There was a statistically significant change in annual decline of renal function (p<0.001), from -1.96 ml/min/year before (95% CI -2.22 to -1.70) and -0.22 ml/min/year (95% CI -0.50 to 0.07) after discontinuation. Diagnosed hypertension added -1.27 ml/min/year in decline of eGFR (p<0.001). Diabetes mellitus did not affect the decline. Conclusion Mean annual decline of renal function is significantly steeper during lithium treatment than after discontinuation. However, decline still occurs. In patients with hypertension, the decline of renal function is accelerated irrespective of lithium discontinuation. The implications for renal function in individual patients considering lithium discontinuation remain unclear. The decision of whether to discontinue lithium depends on a careful trade-off of potential psychiatric risks against potential renal benefits.