PMH15 - Risk of hyponatremia among the incident users of antidepressants

Abstract

Newer antidepressants selective serotonin reuptake inhibitors (SSRIs) and serotonin norepinephrine reuptake inhibitors (SNRIs) are the most commonly prescribed antidepressants. This is due mostly to their better side effect profile when compared to older drugs like tricyclic antidepressants (TCAs). However these classes are not completely bereft of side effects. Hyponatremia is a condition in which the serum sodium level falls below normal; it has been proposed this occurrence is associated with the binding affinity to the 5-HT receptor. The objective of our study is to compare the incidence of hyponatremia among the newer classes of antidepressants. We used a cohort study design using Lifelink claims data to compare the incidence of hyponatremia in TCAs, SSRIs, and SNRIs. Incidence was reported per 10,000 person-years. Cox model was used to assess the risk of adverse events while adjusting for potential confounders. A total of 314,796 patients with an incident prescription for a TCA, SSRI, or SNRIs were identified and met study inclusion criteria. The unadjusted hazard ratio for hyponatremia in patients on SSRIs was 0.789 (C.I=0.729-0.854) and on SNRIs was 1.112 (C.I=0.976-1.268) when compared to TCAs. However, after adjusting for covariates using the Cox model the hazard ratios were found to be 1.013 (C.I=0.934-1.099) for SSRIs and 1.069 (C.I=0.937-1.221) for SNRIs again compared to TCAs. After doing stepwise regression to remove covariates that did not have significant effect the results remained similar, 1.017(C.I=0.939-1.101) for SSRIs and 1.074 (C.I=0.941-1.225) for SNRIs. No significant increase in the risk of hyponatremia for SSRIs or SNRIs over TCAs was found. The covariates we identified show significant interaction with hyponatremia. However, the results must be interpreted with caution as the ICD-9 code that identifies hyponatremia has been shown to have a sensitivity below 30%, thus our analysis potentially excludes a large number of cases.