Trials in adult critical care that show increased mortality of the new intervention: Inevitable or preventable mishaps?

Annals of Intensive Care

Table 3 Frequency and statistical analyses of interim analyses, early stopping and prior Phase II RCTs in adult critical care RCTs showing increased mortality in the intervention groups

References	Interim analyses (frequency)	Interim statistical test	RCT D/C’d early	Prior Phase II RCT (n)	DSMB
Hayes [10]	Yes (every 50 patients)	Chi-square	Yes	Yes^e	No
Fisher [15]	No	NA	No	No	Not stated
Sloan [14]	Yes (after 10, 25, 50 and 75 % enrollment)	Possibly Log rank 28 day^a	Yes	Yes^f	Yes
Takala [13]^g	Yes (group sequential trials; 150 then every 40 patients)	Chi-square	Yes	Yes^c	Yes
Edwards^d [17]	Annually	Chi-square	Yes	Yes$	Yes
Perner [12]	Yes (after 400 patients)	Chi-square	No	Yes	Yes
Ferguson [16]	Yes (pilot phase after 94, 300, 500 and 700 for safety, 800 for efficacy)	Mantel–Haentszel Chi-square	Yes	Yes (N = 25)	Yes
Finfer [8]^b	Yes (after 1500 and 4000 patients)	Chi-square	No	Yes^g	Yes
Lopez [11]	Yes	Triangular test with Christmas tree correction at stopping boundaries	Yes	Yes (n = 312)	Yes
Gao Smith [9]	Yes (every 12 months)	95 % CI	Yes	Yes (n = 40)	Yes

Two parallel RCTs in Finland and in Europe (UK, the Netherlands, Belgium and Sweden) are reported together [13]. Because of slow recruitment due to the unexpectedly high incidence of exclusion criteria, the design was changed before the first interim analysis. The revised design was a fixed-sample analysis of 170 and 190 patients in the Finnish and multinational studies
D/C’d discontinued, NA not available in primary publication
^aThe interim analyses methods were not stated. The primary analysis was log rank to 28 days
^bPrior to the NICE SUGAR RCT [8], van den Berghe et al. [19] had published a similar RCT of intensive versus conventional insulin treatment in critically ill patients
^cPrior small RCTs in burns, postoperative surgical patients, trauma, sepsis and critically ill non-septic patients
^dMany prior small RCTs in head injury
^eSeveral prior RCTs in septic shock, high-risk surgical, trauma and critically ill patients
^fSeveral prior RCTs in trauma and critically ill patients
^gSeveral prior trials of intensive insulin