Predisposing conditions include major surgery, circulatory failure, septic shock, severe coagulopathy, and HIV infections. Reports of incidence in the ICU range from 0 – 30% [J Intensive Care Med 4: 35, 1988; Crit Care Clin 7: 23, 1991]. In those in septic shock, rates are probably 25 – 40%, and mortality in these patients is twice as high as in those with normal adrenals. [Lancet 337: 582, 1991; Intensive Care Med 20: 489, 1994; Am J Med 98: 266, 1995]
Signs of adrenal insufficiency are extremely hard to tease out in critically-ill patients – hypoglycemia and eosinophilia are rare, possible signs of adrenal insufficiency and should be followed up. Hemodynamic instability in spite of volume resuscitation or vasopressors is another important clue. The remainder (electrolyte abnormalities, weakness, hyperpigmentation) are too common or too rare in the ICU to be of any use.
If random cortisol > 34 ug /dL, insufficiency is unlikely, but if random cortisol < 15 ug /dL during severe illness, it is likely. If 15 < x < 34, give high dose cosyntropin test – a 9 ug /dL increase in cortisol is cutoff for insufficiency [NEJM 348: 727, 2003]. Note that this is the “Rapid ACTH stimulation test” referred to by Marino – measure a plasma cortisol, give 250 ug synthetic ACTH, and remeasure in 30 – 60 minutes.
Plasma Cortisol (ug/dL) Probability of insufficiency Baseline Response@ < 15 n/a Very high 15-34 < 9 High 15-34 > 9 Low > 34 n/a Very low
@after receiving 250 ug synthetic ACTH [NEJM 348: 727, 2003]
Patients with documented SIRS, infection, and shock (SBP < 90, UOP < 0.5 mL/kg for 1 hour, PaO2/FiO2 < 280, and on the vent, MI, PE, advanced cancer excluded) who responded to 250 g tetracosactrin by < 9 µg /dL change in cortisol showed a 22% increase in survival at 28 days when given 50 mg IV hydrocortisone q6h and 50 µg fludocortisone qday for 7 days. Survival in non-responders was statistically insignificant, as was survival in both groups at 1 year. [JAMA 288: 862, 2002]
Note that the results of Annane et al’s study, which showed a significant reduction in the risk of death in septic patients deemed to be hypoadrenal (hazard ratio 0.67, 95% CI 0.47-0.95, p =0.02) [Annane D et al. JAMA 288: 862, 2002] were not replicated by Sprung et al. in the CORTICUS trial [Sprung CL et al. NEJM 358: 111, 2008]. Note that the CORTICUS trial used hydrocortisone only (Annane et al. used hydrocortisone and fludrocortisone) and had less restrictive entry criteria (the patients were not as sick), but did show an increase in the incidence of resistant infections in the steroid group
In patients in whom adrenal insufficiency is suspected, steroid therapy should begin immediately. Dexamethasone can be given and will not affect the cortisol assay or ACTH stimulation test (reference needed – controversial) – start with a 2 mg IV bolus. Methylprednisolone will not interfere with the cortisol assay provided that a radioimmunoassay is used [Passmore JM Clinics in Critical Care Med Vol. 5: Churchill Livingstone: 97, 1985] – start with 60 mg IV bolus. After the ACTH test is completed, empiric therapy can begin with hydrocortisone (Solu-Cortef) 50 mg IV q6h until test results are back. If the ACTH test is normal, the hydrocortisone can be abruptly stopped, otherwise continue dosing until the patient is unstressed, at which point 20 mg qday should suffice.
Glucocorticoid / Mineralocorticoid Potency
- Hydrocortisone: 1:1 (8h duration) – note that this is the synthetic version of cortisol
- Prednisone: 4:0.8 (16-36h duration) – less mineralocorticoid activity than hydrocortisone
- Methylprednisolone: 6:0.5 (18-40h duration) – MUCH less mineralocorticoid activity than hydrocortisone
- Dexamethasone: 60:0 (36-54h duration) – NO mineralocorticoid activity
- Fludrocortisone: 15:200 (24h duration) – predominantly mineralocorticoid activity
Note that 1) cytokines released during inflammation inhibit end-organ responsiveness to cortisol [NEJM 348: 727, 2003] 2) 250 µg is 100x a physiologic dose, thus a 1 µg stimulation may be more appropriate and 3) in cases of vasopressor-dependent septic shock, ACTH stimulation tests are not predictive of which patients will benefit from supplementation [Intensive Care Med 32: 1184, 2006]. For these reasons, Marino recommends not relying on the Rapid ACTH Stimulation Test and giving 50q6 hydrocortisone based on clinical suspicion alone.