Professor Vincent, in an eloquent commentary in Critical Care, calls for a further trial into supraphysiological cortico-steroid therapy in vasopressor-resistant shock 1. Together with editorials in several of the intensive care journals, he has pointed out many of the shortcomings in the Corticus trial 2. We would like to add to this chorus by posing a further question to the authors and putting forward some suggestions. Regrettably, we are prohibited from addressing these directly due to the Letters policy of the journal in which the original paper was published.

Data from numerous sources suggest that the earlier shock is reversed, the better the outcome – be it mortality, morbidity, length of stay or other surrogate endpoints. In the Corticus study, the median time to shock reversal was 2 to 3 days shorter in the hydrocortisone group (see Table 1). Despite this, no outcome improvement was demonstrated. No investigation, or explanation, of this apparent discrepancy has been forthcoming. Sequential Organ Failure Assessment scores were performed at the time of study enrolment, but no serial Sequential Organ Failure Assessment score data are presented. If available, these data would be intriguing.

Following the publication of the Corticus data, a consensus statement regarding the diagnosis and management of corticosteroid insufficiency in critically ill adult patients has been published 3. Together with a detailed review by Dickstein and Saiegh 4, this statement suggests a working diagnostic paradigm. However, we would like to suggest the following three pragmatic definitions of functional hypoadrenalism, which future trial designers might with wish to consider and which we currently employ.

First, patients with septic shock requiring high-dose vaso-pressors – defined as requiring ≥ 0.2 μg/kg/minute norepinephrine (or equivalent), who are not volume responsive (defined as a ≥ 10% increase in stroke volume following a 3 ml/kg fluid bolus administered in ≤ 5 min) and who are hyperdynamic (defined as a cardiac index ≥ 2.8 l/min/m²).

Patients with evidence of acute myocardial depression or chronic insufficiency should be considered separately.

Second, patients who, having been stable for ≥ 2 hours on a dose of vasopressor, develop increasing dose requirements (≥ 20% increase), are unresponsive to a volume bolus (as above) and are hyperdynamic (as above).

Third, patients whose dose of vasopressor cannot be weaned ≥ 24 hours following initiation of appropriate broad-spectrum antimicrobial therapy and/or effective source control.

Furthermore, this therapy should be withdrawn from patients who fail to demonstrate a ≥ 20% decrease in vasopressor requirement to maintain the same mean arterial pressure 60 minutes after the initial dose of hydrocortisone. Due to the pharmacokinetics of hydrocortisone, we favour a 100 mg intravenous bolus followed immediately by initiation of a 10 mg/hour intravenous infusion.

Finally, we would like to promote two recently published papers that offer useful insights into the pharmacodynamics of supraphysiological steroid therapy in vasopressor-resistant shock. Firstly, Druce and colleagues make a convincing argument that the principal effect of hydrocortisone is as a mineralocorticoid and not as an anti-inflammatory 5.

Secondly, Kaufman and colleagues 6 found that hydrocortisone administered as described above does have some arguably clinically valuable anti-inflammatory effects but, in addition, enhances neutrophil phagocytosis. This has led us to conclude that, in the absence of a systemic form of fludrocortisone and with the unreliable enteral absorption of drugs, systemic hydrocortisone monotherapy at optimal mineralocorticoid doses should be the therapy of choice.

Authors' response

Competing interests