REVIEW OF: Bosi A et al, JAMA Netw Open 2023;6(7):e2322056

STUDY TYPE: Retrospective cohort study

Studies have had mixed results regarding the relationship between lithium and kidney injury. A 2012 meta-analysis of lithium and kidney injury was inconclusive, and studies since then have suffered from limitations and conflicting results. In an effort to clarify this issue, researchers once again looked at the risk of kidney injury in patients taking lithium—this time comparing it to those taking valproate.

This retrospective cohort study analyzed data from about 11,000 Stockholm residents who began lithium or valproate between 2007 and 2018 and had no prior history of kidney transplant or maintenance dialysis. Using data from Stockholm’s Creatinine Measurements database, the study tracked these patients for up to 10 years, covering roughly 5,300 individuals per medication group, with a median follow-up duration of 4.5 years. The primary outcome was progression of chronic kidney disease (CKD), incidence of acute kidney injury (AKI), changes in estimated glomerular filtration rate (eGFR), and the onset of new albuminuria. Researchers measured both lithium and valproate levels, specifically looking at outcomes at lithium levels above and below 1.0 mmol/L.

The study found no significant difference in CKD progression between patients starting on lithium or valproate, with about 3.5% of individuals in each group developing CKD. There were also no significant differences in non-CKD eGFR reduction, AKI risk, or albuminuria between the two groups. Surprisingly, the overall risk of AKI over the 10-year study period was actually 3.2% lower for lithium than valproate (95% confidence interval [CI] -5.6 to -1.1). 

However, there was some concerning news about lithium. For the subgroup of lithium-treated patients with blood levels higher than 1.0 mmol/L, the risk of CKD was almost triple that of patients with levels below 1.0. Even a level of more than 0.8 mmol/L significantly increased the risk of AKI (hazard ratio 2.56, 95% CI 1.67–3.92).

CARLAT TAKE 

Despite its reputation, lithium does not increase the risk of acute or chronic kidney injury any more than valproate. The overall risk of kidney injury in both groups was low, and keeping lithium below 0.8 mmol/L (when clinically feasible) is safest renally. In this study, valproate was more likely to cause AKI than lithium, and there are case reports of kidney injury in patients starting valproate (Anguissola G et al, Pediatr Nephrol 2023;38(6):1725–1731). No guidelines call for checking kidney function any time a patient starts valproate—but we should be aware of the rare possibility of AKI.