The Syndrome of Inappropriate Antidiuresis

Hyponatremia is the most common electrolyte abnormality encountered in clinical practice. The reported frequency of the disorder is determined by a number of factors, including the definition of hyponatremia, the frequency of testing, the healthcare setting, and the patient population. This review focuses on the incidence and prevalence of hyponatremia. In acute hospital care, particular attention is given to admission versus hospital-acquired hyponatremia. Although less well studied, the epidemiology of hyponatremia in the ambulatory-based setting and the geriatric/nursing home population is also summarized. Finally, the frequency of hyponatremia occurring in special clinical conditions--including congestive heart failure, cirrhosis, pneumonia, and acquired immunodeficiency syndrome--as well as in marathon runners will be reviewed. Substantial additional work is still required to determine the true occurrence of hyponatremia in the various clinical settings. Beyond the phenomenologic value, advances in the epidemiology of hyponatremia should also provide insights in the prognostic implications as well as the preventive and management strategies of the disorder in various clinical settings.

The study objective was to determine the eventual consequences (falls, unsteadiness, and cognitive impairment) of mild chronic hyponatremia, which is generally considered as asymptomatic. In a case-control study, we focused on the incidence of falls among 122 patients (mean age 72+/-13 years) with asymptomatic chronic hyponatremia (mean serum sodium concentration [SNa] 126+/-5 mEq/L), who were admitted to the medical emergency department, compared with 244 matched controls. To explore the mechanisms of the excess of falls, we prospectively asked 16 comparable patients (mean age 63+/-15 years; SNa+/-2 mEq/L) to perform 8 attention tests and a gait test consisting of 3 steps "in tandem," in which we measured the "total traveled way" by the center of pressure or total traveled way. Thereafter, the patients were treated and tested again (50% of the patients were tested first with normal SNa to avoid learning biases). Epidemiology of falls: Twenty-six patients (21.3%) of 122 were admitted for falls, compared with only 5.3% of the control patients (adjusted odds ratio: 67; 95% confidence: 7.5-607; P <.001). The frequency of falls was the same regardless of the level of hyponatremia. Gait: The total traveled way by the center of pressure significantly increased in hyponatremia (1336+/-320 mm vs 1047+/-172 mm with normal SNa; P=.003). Attention tests: The mean response time was 673+/-182 milliseconds in hyponatremia and 615+/-184 milliseconds in patients with normal SNa (difference: 58 milliseconds, P <.001). The total error number in hyponatremia increased 1.2-fold (P=.001). These modifications were comparable to those observed after alcohol intake in 10 volunteers. Mild chronic hyponatremia induces a high incidence of falls possibly as the result of marked gait and attention impairments. Treating these patients might prevent a considerable number of hospitalizations.

There are no controlled experimental data that assess the accuracy of the commonly used correction factor of a 1.6 meq/L decrease in serum sodium concentration for every 100 mg/dL increase in plasma glucose concentration. The purpose of this study was to evaluate experimentally the hyponatremic response to acute hyperglycemia. Somatostatin was infused to block endogenous insulin secretion in 6 healthy subjects. Plasma glucose concentrations were increased to >600 mg/dL within 1 hour by infusing 20% dextrose. The glucose infusion was then stopped and insulin given until the plasma glucose concentration decreased to 140 mg/dL. Plasma glucose and serum sodium concentrations were measured every 10 minutes. Overall, the mean decrease in serum sodium concentration averaged 2.4 meq/L for every 100 mg/dL increase in glucose concentration. This value is significantly greater than the commonly used correction factor of 1.6 (P = 0.02). Moreover, the association between sodium and glucose concentrations was nonlinear. This was most apparent for glucose concentrations >400 mg/dL. Up to 400 mg/dL, the standard correction of 1.6 worked well, but if the glucose concentration was >400 mg/dL, a correction factor of 4.0 was better. These data indicate that the physiologic decrease in sodium concentration is considerably greater than the standard correction factor of 1.6 (meq/L Na per 100 mg/dL glucose), especially when the glucose concentration is >400 mg/dL. Additionally, a correction factor of a 2.4 meq/L decrease in sodium concentration per 100 mg/dL increase in glucose concentration is a better overall estimate of this association than the usual correction factor of 1.6.