Prevalence of Hyponatremia in Elderly Patients

To the Editor: We read with interest the high prevalence of hyponatremia in older people quoted by Miller in his recent review.1 Most of the studies giving figures for prevalence of hyponatremia in an elderly population, and frequently quoted in reviews of hyponatremia, were conducted in the 1970s and 1980s or earlier. These may be inaccurate, particularly given the sharp increase in use of medications well known to cause hyponatremia in older people. These include thiazide diuretics for hypertension2 or secondary prevention of stroke and selective serotonin reuptake inhibitors for depression.3,4 Recent work performed in our own department, and presented at the spring meeting of the British Geriatrics Society, Gateshead 2006, supports the theory that there is a higher prevalence in elderly hospitalized patients than the 5% previously reported.5 We conducted a retrospective study of unselected, consecutive admissions to our geriatric assessment ward over a 3-month period from January to March 2005 (N=103, 59% female, mean age 82). We found that 18% of patients were hyponatremic (serum sodium concentration (Na+) <135 mmol/L) on admission. Furthermore, an additional 24% became hyponatremic at some point during their admission. Lower admission Na+ (P=.01) and larger drop in Na+ during admission (P<.001) were independently associated with longer length of stay after correcting for potential confounders such as age. We wonder if the higher prevalence in our patient cohort reflects a more widespread increase in hyponatremia in the elderly population, perhaps caused by different prescribing patterns than 20 years ago, as described above. An alternative explanation is that we studied an unusually frail cohort, although our cohort was unselected and consisted mostly of referrals from the local accident and emergency department or directly from the community by a general practitioner. Furthermore, 78% of these patients were admitted from their own home, with the remainder coming from institutional care facilities. Miller rightly points out that hyponatremia is associated with poorer prognosis, although it is unclear whether this is a causal relationship or hyponatremia is a surrogate marker of disease severity. Nevertheless, the high prevalence of hyponatremia should be of concern to clinicians looking after older people. Financial Disclosure: None. Author Contributions: All authors equally contributed to this letter. Sponsor's Role: None.