Table 2 Trials’ calendar

2001

The first Leuven RCT (1,548 patients) reported a 34% relative risk reduction in hospital mortality with maintenance of BGL of between 80 and 110 mg/dL [4].

2003

Krinsley [49], in an observational study (1,826 patients), confirmed the survival benefit associated with protocolized IIT targeting a BGL of less than 140 mg/dL.

2006

The second Leuven RCT (1,200 patients) confirmed a 10% absolute reduction in hospital mortality for long-stay medical ICU patients with maintenance of BGL of between 80 and 110 mg/dL [11].

2008

De la Rosa et al. [9] RCT (504 patients) failed to show any survival benefit in mixed ICU patients targeting BGL between 80 and 110 mg/dL but with less effective control.

The VISEP study, with the same glycemic goals (537 patients with septic shock), was terminated prematurely because of an unacceptably high incidence of hypoglycemia (17.0% versus 4.1%; P < 0.001) and no evidence for survival benefit at 90 days (39.7% versus 35.4%; P = 0.31) [50].

Arabi et al. [8] RCT (523 patients) also failed to show survival benefit (adjusted hazard ratio 1.09, 95% confidence interval 0.70 to 1.72) and showed increased hypoglycemic rates (28.6% versus 3.1% of patients; P < 0.0001).

SPRINT (BGL goal of 72 to 110 mg/dL) is an observational study with historic control. Nutritional and insulin protocols provided less variable and tighter glucose control (standard deviation of blood glucose was 38% lower compared with the retrospective control) with subsequent improvement in organ failures and outcome for long-stay ICU patients: failure-free days were different (SPRINT = 41.6%; Pre-SPRINT = 36.5%; P < 0.0001) [51].

2009

The Glucontrol (1,101 patients) was stopped prematurely for unintended protocol violations. The IIT was associated with increased hypoglycemia (8.7% versus 2.7%; P = 0.0001) and a non-significant trend to higher mortality (15.3% versus 17.2%) while BGL was not optimally controlled [10].

2009

The NICE-SUGAR trial (6,104 mixed ICU patients) compared a strategy of BGL control of between 81 and 108 mg/dL versus a more liberal strategy (<180 mg/dL). This RCT found an increase in mortality with IIT (27.5 versus 24.9; P = 0.02) and increased incidence of hypoglycemia (6.8% versus 0.5%; P < 0.001) [7].

2010

COITTSS (509 patients with septic shock) compared a strategy of BGL control of between 80 and 110 mg/dL versus maintenance of BGL of less than 150 mg/dL. This trial did not find any difference in in-hospital mortality between the two strategies (45.9% versus 42.9%; P = 0.05) [8].