UC Irvine

Background

Hyponatremia is the most common electrolyte abnormality in clinical practice, yet little is known about its frequency in patients with cancer or its impact on their clinical outcomes.

Study design

Retrospective analysis of prospectively collected data.

Setting & participants

Patients with cancer admitted to the University of Texas M.D. Anderson Cancer Center in 2006 for 3 months.

Predictor

Serum sodium levels categorized as eunatremia (serum sodium, 135-147 mEq/L) and mild (134-130 mEq/L), moderate (129-120 mEq/L), and severe (<120 mEq/L) hyponatremia.

Outcomes

(1) Length of hospital stay and (2) 90-day mortality.

Results

In 4,702 admissions in 3,357 patients with cancer, hyponatremia (serum sodium <135 mEq/L) was noted in 47% of admissions. It was mild in 36%, moderate in 10%, and severe in 1%. Hyponatremia was acquired during the hospital stay in 24%. Using the first admission data, mean length of stay was 5.6 ± 5.0 days for patients with eunatremia and 9.9 ± 9.2, 13.0 ± 14.1, and 11.5 ± 12.6 days for those with mild, moderate, and severe hyponatremia, respectively. The respective HRs in the multivariate Cox model for longer hospital stay, using patients with eunatremia as reference, were 1.92 (95% CI, 1.75-2.13; P < 0.01), 2.94 (95% CI, 2.56-3.45; P < 0.01), and 2.32 (95% CI, 1.32-4.00; P = 0.01). 283 (8.4%) deaths occurred during 90 days, and in the multivariate model, the respective HRs for 90-day mortality for mild, moderate, and severe hyponatremia were 2.04 (95% CI, 1.42-2.91; P < 0.01); 4.74 (95% CI, 3.21-7.01; P < 0.01), and 3.46 (95% CI, 1.05-11.44; P = 0.04). These findings were consistent when analyses were repeated with sodium levels in tertiles.

Limitations

Observational study, retrospective, inability to adjust for all comorbid conditions.

Conclusion

Hyponatremia in patients with cancer is associated with longer hospital stay and higher mortality. Whether long-term correction of hyponatremia would improve these outcomes remains to be determined.