Adrenal insufficiency occurring during septic shock: incidence, outcome, and relationship to peripheral cytokine levels.

PURPOSE: In patients with septic shock, to (1) determine the incidence of adrenal insufficiency (AI), (2) observe the effects of glucocorticoid therapy on outcome in those with impaired adrenal function, and (3) investigate a possible correlation between adrenal function and peripheral cytokine levels. PATIENTS AND METHODS: Twenty-one patients admitted to the medical and surgical intensive care unit with septic shock and 11 healthy volunteers were studied. Cortisol, tumor necrosis factor-alpha (TNF-alpha), and interleukin-6 (IL-6) levels were measured before and after infusion of low (1 microgram) and standard doses (250 micrograms) of adrenocorticotropic hormone (ACTH) within 24 hours of the diagnosis of septic shock. Patients with subnormal adrenal responses to ACTH were treated with stress doses of steroids. Hormone, cytokine, and survival data in patients with normal response were compared to those with subnormal adrenal function. RESULTS: Five patients (23.8%) exhibited AI by ACTH stimulation testing. Three of them received steroid supplementation with rapid improvement in hemodynamic parameters. Autopsies of 2 patients with AI revealed intact adrenal cortices. Sixteen patients had adequate adrenal responses (AAR) to the standard-dose ACTH infusion. TNF-alpha levels were inversely correlated with mean arterial pressure (MAP) (r = -.52, P = 0.038) in AAR but not AI. There was no difference in mean peripheral TNF-alpha levels between AAR and AI. There was no correlation between TNF-alpha levels and mortality or adrenal function in those with septic shock. A trend toward lower IL-6 levels in AI suggests a link between reduced IL-6 levels and understimulation of the pituitary-adrenal axis in this group. Mortality in patients with AI was 80% at 4 weeks as compared with 43.8% in the group with normal adrenal response. CONCLUSIONS: Adrenal hyporesponsiveness is a feature of septic shock in some patients. Its etiology is probably complex. Steroid supplementation appeared to improve short-term survival when AI occurred, although these patients' overall mortality was worse than that of patients with septic shock and AAR. The standard-dose (250 micrograms) rapid ACTH infusion test was adequate for detecting AI. Adrenal insufficiency should be suspected in patients with septic shock who do not respond to conventional treatment. Performing the ACTH infusion test and initiating a trial of stress doses of glucocorticoids pending the results is a reasonable strategy in this situation.

Circadian distribution of in-hospital cardiopulmonary arrests on the general medical ward.

BACKGROUND: A circadian pattern has been convincingly demonstrated for the onset of many outpatient ischemic vascular events. A morning peak exists for the onset of acute myocardial infarction, sudden cardiac death, reversible myocardial ischemia, and ischemic stroke. Data regarding circadian patterns of disease in hospitalized patients, however, are lacking. We examined in-hospital cardiopulmonary arrest (CPA) occurring on the general medical ward to determine if a circadian distribution existed in time of onset. METHODS: All CPAs that occurred during a 9-month period and met entry criteria were included for study. The day was divided into 4-hour intervals and analyses were performed for evidence of periodicity in time of onset. The CPAs were then divided into those that were "expected" and those that were "unexpected," and further analyses of periodicity were performed. RESULTS: For the total study population (137 patients), a primary peak frequency of CPA occurred during the interval from 4 to 7:59 AM, and a secondary peak frequency occurred during the 8 to 11:59 PM interval. A minimum frequency occurred during the midnight to 3:59 AM interval. The onset of unexpected CPA peaked during the 4 to 7:59 AM interval, and expected CPA followed no circadian pattern. CONCLUSIONS: Our analysis of CPA occurring in patients hospitalized on the general medical ward demonstrated a circadian pattern of onset that favored the early-morning hours. This pattern is predominantly due to unexpected CPA. If further study confirms our observations, changes in the prophylaxis of in-hospital CPA and adjustments in staff responses to its occurrence may be indicated.