TEE screening allowed identification of ACP in one-fourth of patients submitted to protective ventilation for early moderate to severe ARDS. PFO shunting was less frequent and never large or extensive. ACP and PFO were not related to outcome.

Our data suggest that in a selected population of patients with severe septic shock single plasma cortisol determination has no predictive value. The short ACTH stimulation test performed within the first 24 h of onset shock can neither predict outcome nor estimate impairment in adrenocortical function in patients with high basal cortisol level. Adrenal insufficiency is rare in septic shock and should be suspected when cortisol level is below 15 micrograms/dl and then confirmed by a peak cortisol level lower than 18 micrograms/dl during the short ACTH stimulation test.

Delayed ICU admission due to a full unit at first referral is associated with increased mortality.

Diagnosis of acute adrenocortical insufficiency must be considered if clinical manifestations are present suggesting septic shock without any obvious infectious cause in patients having undergone considerable intravenous fluid therapy as an initial course of treatment.

BackgroundThe potential influence of bed availability on triage to intensive care unit (ICU) admission is among the factors that may influence the ideal ratio of ICU beds to population: thus, high bed availability (HBA) may result in the admission of patients too well or too sick to benefit, whereas bed scarcity may result in refusal of patients likely to benefit from ICU admission.MethodsCharacteristics and outcomes of patient admitted in four ICUs with usual HBA, defined by admission refusal rate less than 11 % because of bed unavailability, were compared to patients admitted in six ICUs with usual low bed availability (LBA), i.e., an admission refusal rate higher than 10 % during a 90-day period.ResultsOver the 90 days, the mean number of days with no bed available was 30 ± 16 in HBA units versus 48 ± 21 in LBA units (p < 0.01). The proportion of admitted patients was significantly higher in the HBA (80.1 %; n = 659/823) than in the LBA units [61.6 %: n = 480/779; (p < 0.0001)]. The proportion of patients deemed too sick to benefit from admission was higher in LBA (9.0 %; n = 70) than in the HBA (6.3 %; n = 52) units (p < 0.05). The HBA group had a significantly greater proportion of patients younger than 40 years of age (22.5 %; n = 148 versus 14 %; n = 67 in LBA group; p < 0.001) and higher proportions of patients with either high or low simplified acute physiologic score II values.ConclusionsBed availability affected triage decisions. Units with HBA trend to admit patients too sick or too well to benefit.