The short corticotropin stimulation test is widely used to assess adrenocortical function in critically ill patients [1-4]. Previous studies reported that an impaired response to corticotropin was associated with poor outcome and justified hydrocortisone therapy to reduce mortality in patients with severe sepsis [1, 2]. Because patients with septic shock have widely varying cortisol levels and disruption of pulsatile cortisol secretion [5], the interpretation of the result of a short corticotropin stimulation test can be questionable.

We prospectively did two consecutive short corticotropin stimulation tests in 22 patients with septic shock. Patients met the following criteria: positive blood cultures or an identified site of infection, inotropic or vasopressor treatments to maintain systolic blood pressure after adequate fluid resuscitation, and one or more manifestations of inadequate organ perfusion (such as confusion, oliguria, and acidosis). The patients had no hormonal disorders and were not receiving corticosteroids. A short corticotropin stimulation test was done within 24 hours of the onset of shock (day 1) and 48 hours later (day 2). Blood samples were taken just before and 60 minutes after 0.25 mg of tetracosactrin was given intravenously.

All patients had basal cortisol levels greater than 275 nmol/L; no significant difference was seen between survivors and nonsurvivors on days 1 and 2. The mean cortisol responses (difference between the 60-minute cortisol level and the baseline value) did not significantly differ on days 1 and 2 (218 nmol/L [95% CI, 125 nmol/L to 312 nmol/L] compared with 234 nmol/L [CI, 163 nmol/L to 304 nmol/L]; Wilcoxon test for paired samples) and between survivors and nonsurvivors. The cortisol response on day 1 (Figure 1) was not correlated with the cortisol response on day 2 (correlation, 0.03; P = 0.9; Spearman rank correlation test).

View larger version (22K): [in this window] [in a new window]   Figure 1. Cortisol response to corticotropin in patients with septic shock on days 1 and 2 after shock.

If a cortisol level increase of more than 200 nmol/L above the baseline is considered an adequate response to the short corticotropin stimulation test [1, 2, 4], 10 patients (45%) had widely varying responses on day 1 and day 2. Six of these patients died. Six (50%) of the 12 patients with a response of less than 200 nmol/L on day 1 had an adequate response on day 2 (Figure 1). However, when a cortisol level greater than 500 nmol/L achieved 60 minutes after corticotropin injection is considered a normal response [3], the responses of 3 patients (13.6%) differed between days 1 and 2. Two of these patients died. The peak cortisol level on day 1 was positively correlated with the peak cortisol level on day 2 in survivors (correlation, 0.9; P = 0.008) and nonsurvivors (correlation, 0.7; P = 0.01).

These data suggest that the cortisol response to the short corticotropin stimulation test is not reproducible in patients with septic shock. The use of the peak cortisol level seems to be more reliable for detecting absolute adrenocortical insufficiency [3]. Nevertheless, the absence of any significant difference between the results of the short corticotropin stimulation test in survivors and nonsurvivors, whatever the criteria used, suggests that adrenocortical impairment is not a major problem in patients with septic shock.