Reply to comment on: Hyponatremia-associated hospital visits are not reduced by early electrolyte testing in older adults starting antidepressants

Abstract

We are grateful for Dr. Akkar and colleagues' interest in and succinct commentary on our study examining the association between early electrolyte monitoring and hospitalizations with hyponatremia among older adults starting antidepressants. Our sample's median serum sodium concentration was 132 mmol/L at the time of hospital admission, in keeping with relatively mild hyponatremia. We disagree that information regarding symptoms would have clarified the role of hyponatremia in necessitating hospitalization. The symptoms of hyponatremia are extremely nonspecific in all but the most severe cases, and our datasets do not include information regarding symptoms in any event. In contrast, a laboratory finding of hyponatremia is unambiguous.

We further disagree that censoring 3945 people—a mere 0.9% of our sample—during the 60-day follow-up period is a source of selection bias in our study. We note that rather than being excluded from our cohort, these individuals were censored at the time of death because there was no information on hospitalization with hyponatremia before their death.¹

We share Akkar and colleagues' interest in the approximately 25% of patients in our study whose hospitalization with hyponatremia lasted more than 12 days. In Table 2, we showed that the medians and interquartile ranges for length of stay were comparable among those who did and did not get early electrolyte testing.²

Given that most cases of hyponatremia can be effectively treated within a week of admission,³ we posit that the hospitalizations lasting greater than 12 days in our cohort reflected chronic conditions associated with hyponatremia, such as heart failure and cirrhosis. Our clinical experience leads us to suspect that some older, frailer patients in our cohort experienced deconditioning in hospital and were unable to return promptly to their preadmission settings, resulting in longer hospital stays while awaiting rehabilitation or alternate living arrangements.⁴

NEL wrote the reply, and all coauthors have contributed to and approved the final version of the reply.

None of the authors have any financial or personal conflicts or potential conflicts of interest.

Study sponsors had no role in preparing this reply.