Table 3—

Outcome measures (pairwise comparisons) using insulin detemir or NPH insulin as basal insulin therapy in type 1 diabetes

Measure	Group 1	Group 2	Mean difference (95% CI)	P
HbA_1c (%)	IDet_combined	NPH	−0.2 (−0.34 to −0.02)	0.027
Clinic FPG (mmol/l)	IDet_12h	NPH	−1.5 (−2.51 to −0.48)	0.004
	IDet_morn+bed	NPH	−2.3 (−3.32 to −1.29)	<0.001
	IDet_12h	IDet_morn+bed	0.8 (−0.18 to 1.80)	NS
Self-monitored fasting blood glucose
Mean	IDet_12h	NPH	−0.8 (−1.32 to −0.22)	0.006
	IDet_morn+bed	NPH	−0.8 (−1.34 to −0.25)	0.004
	IDet_12h	IDet_morn+bed	0.0 (−0.51 to 0.56)	NS
Within-patient SD*	IDet_12h	NPH		<0.001
	IDet_morn+bed	NPH		<0.001
	IDet_12h	IDet_morn+bed		NS
Body weight change (kg)	IDet_12h	NPH	−0.8 (−1.44 to −0.24)	0.006
	IDet_morn+bed	NPH	−0.6 (−1.23 to −0.03)	0.040
	IDet_12h	IDet_morn+bed	−0.2 (−0.80 to 0.37)	NS
CGMS glucose profiles deviation from average (mmol/l · h)
>24 h	IDet_combined	NPH	−6.45 (−13.3 to 0.40)	0.065
Overnight	IDet_combined	NPH	−1.72 (−4.05 to −0.62)	NS
Hypoglycemia in final 12 weeks [RR (95% CI)]†
Minor-any time	IDet_12h	NPH	0.75 (0.56–1.00)	0.046
	IDet_morn+bed	NPH	0.68 (0.56–0.84)	0.002
Minor-nocturnal	IDet_12h	NPH	0.74 (0.50–1.08)	NS
	IDet_morn+bed	NPH	0.47 (0.36–0.62)	<0.001

Pairwise testing was only performed where allowed by the protocol (see research design and methods); accordingly only limited comparisons are given.
*
↵* Data distribution did not allow calculation of mean difference and CIs for within-patient variability;
†
↵† estimates of the relative risk of hypoglycemic episodes during the maintenance period calculated using a recurrent event γ-frailty model. IDet_combined, insulin detemir groups combined.