Optimal glycemic control in neurocritical care patients: a systematic review and meta-analysis

Table 3 Subgroup analysis and meta-regression of studies assessing the efficacy of intensive glycemic control in neurocritical care patients

Comparison	Mortality			Poor neurological outcome
	Studies, number	Risk ratio (95% confidence intervals)	P -value for comparison	Studies, number	Risk ratio (95% confidence intervals)	P -value for comparison
Control group
Very loose	10	0.98 (0.80-1.20)		8	0.88 (0.79-0.98)
Moderate	6	1.00 (0.72-1.39)	0.89	5	0.99 (0.85-1.14)	0.04
Hypoglycemia†
Uncommon	7	1.00 (0.86-1.24)		5	0.94 (0.84-1.06)
Common	6	1.17 (0.78-1.76)	0.72	6	0.94 (0.81-1.08)	0.94
Duration of tight control
> 72 hours	12	0.99 (0.83-1.18)		11	0.92 (0.84-1.01)
< 72 hours	4	0.97 (0.56-1.67)	0.37	2	0.81 (0.57-1.15)	0.04
Risk of bias
Higher	11	1.03 (0.86-1.24)		10	0.94 (0.85-1.04)
Lower	4	1.00 (0.52-1.91)	0.76	3	0.94 (0.76-1.17)	0.17
Timing of nutrition
As soon as possible	6	1.07 (0.73-1.57)		5	0.93 (0.80-1.08)
Deferred > 24 hours	5	1.01 (0.81-1.26)	0.82	4	0.90 (0.79-1.03)	0.79
Definition
Hypoglycemia	8	0.94 (0.67-1.31)		8	0.88 (0.77-1.00)
56-80 mg/dl 55 mg/dl or below	6	1.01 (0.82-1.23)	0.72	4	0.91 (0.79-1.03)	0.72

^† As per a priori plan, patients were dichotomized based on the median prevalence of hypoglycemia across studies; common, hypoglycemia occurred in 33-100% of patients; uncommon, hypoglycemia occurred in 3 to 18% of patients.

ISSN: 1364-8535