. 2010 Nov;12(11):835–846. doi: 10.1089/dia.2010.0135

Table 1.

Randomized Controlled Trials Comparing Insulin Glargine with NPH Insulin, Ultralente and Continuous Subcutaneous Insulin Infusion

					Change from baseline		Frequency of hypoglycemia (episodes/patient–month, % patients)
Reference	Design	No. patients	Trial Duration	Treatments	A1c (%)	FBG (mmol/L)	Symptomatic	Nocturnal	Comment
vs NPH insulin
Ratner 2000¹³ (Study 3004)	Open-label, multicenter, parallel-group study	534	28 weeks	Bedtime glargine NPH (od or bd)	−0.16% (baseline 7.7%) −0.21% (baseline 7.7%) P = 0.44	−1.67 (FPG) −0.33 (FPG) P = 0.0145	39.9% 49.2% P = 0.0219	18.2% 27.1% P = 0.016	Lower FPG levels with fewer episodes of hypoglycemia with glargine compared with od- or bd NPH as part of a basal–bolus regimen
Raskin 2000¹²	Open-label, multicenter, parallel-group study	619	16 weeks	Glargine + lispro (od) NPH (od or bd) +lispro	−0.1% (baseline 7.6%) −0.1% (baseline 7.7%) P = NS	−30.6 −10.8 P = 0.0001	90.6% 90.6% P = 0.84	69.0% 63.1% P = 0.06	Glargine appears to be as safe and at least as effective as using NPH as basal–bolus treatment with lispro
Rossetti 2003¹⁰	Open-label, parallel-group study	51	12 weeks	Dinnertime glargine Bedtime glargine NPH (od) + lispro	−0.4% (baseline 7.0%) −0.4% (baseline 6.8%) + 0.1% (baseline 6.9%) P < 0.04	7.6 ± 0.1^a 7.6 ± 0.2^a 8.1 ± 0.2^a P < 0.03	8.1 ± 0.8^b 7.7 ± 0.9^b 12.2 ± 1.3^b P < 0.04 vs NPH	1.7 ± 0.2 2.0 ± 0.19 3.6 ± 0.4 P < 0.05	Decreased A1c and hypoglycemia with glargine. In contrast to NPH, which should be given at bedtime, glargine can be administered at dinner time without deteriorating BG control
Porcellati 2004¹¹	Open-label, parallel-group study	121	52 weeks	Dinnertime glargine + lispro NPH (od) + lispro	−0.4% (baseline 7.0%) 0.0% (baseline 7.0%) P < 0.05	7.6 ± 0.11^a 8.1 ± 0.22^a P < 0.05^a	7.2 ± 0.5^b 13.2 ± 0.6^b P < 0.05^b	0.3^c 2.1c P < 0.05^c	Glargine decreased A1c and reduced the frequency of hypoglycemia versus NPH
Hershon 2004¹⁸ (Study 3004)	Subgroup analysis of Ratner study (2000)⁶ of patients treated with NPH bd	394	28 weeks	Glargine (od) NPH (bd)	−0.09% −0.19% P = NS	−1.17 −0.56 P = 0.015	73.3%^d 81.7%^d P = 0.02	Severe 36.6%^e Severe 46.2%^e P = 0.003	Glargine was at least as effective as NPH in improving FBG control with fewer symptomatic hypoglycemic events
Home 2005¹⁴ (Study 3001)	Open-label, multicenter, parallel study	585	28 weeks	Glargine + lispro NPH (od or bd) +RHI	+ 0.21 ± 0.05% (baseline 7.9%) + 0.10 ± 0.05% (baseline 8.0%) P = NS	−1.17 −0.89 P = 0.07	89.0%^g 84.6%^g P = NS	61.0%^g 61.1%^g P = NS	Bedtime, glargine was as least as effective as NPH, without an increased risk of hypoglycemia
Fulcher 2005¹⁶ (Study 4010)	Single-blinded, multicenter, parallel-group study	125	30 weeks	Glargine (od) +lispro NPH (od) + lispro	−1.04% (baseline 9.2%) −0.5% (baseline 9.7%) P < 0.01	−3.46^h −2.34^h P = 0.032	Severe 0.87ⁱ Severe 0.99ⁱ P = NS	Severe: 0.22ⁱ Severe: 0.37ⁱ P = 0.02	Significantly lower A1c and FBG levels and less severe nocturnal hypoglycemia with glargine
Chatterjee 2007¹⁵ (Study 6004)	Open-label, single-center, two-period crossover study using a BBT regimen	53	36 weeks	Glargine + aspart NPH (bd) + aspart	−0.46% (baseline 8.53%) −0.26% (baseline 8.53%) P = 0.04	−3.08 −0.08 P < 0.01	80.7%^j 77.2%^j P = 0.63	NR NR	Lower A1c and mean FPG (–54 mg/dL, P = 0.002), and greater satisfaction with glargine compared with NPH (DTSQ, P = 0.001)
Bolli 2009¹⁷ (Study 4019)	Parallel, open-label, multicenter study of patients switched from NPH	175	4-week run-in, 24 weeks treatment	Dinnertime glargine + lispro NPH (bd) + lispro	−0.56% (baseline 7.82%) −0.56% (baseline 7.82%) P = NS	−28.0 mg/dL −9.8 mg/dL P = 0.0064	+ 0.26^k + 0.21^k P = 0.953	Serious (42 mg/dL) –0.19^k −0.10 (<42 mg/dL)^k P = 0.383	Lower FBG, lower BG variability and reduced nocturnal hypoglycemia with glargine. Similar changes in A1c, FBG, PPBG and incidence of hypoglycemia. Greater satisfaction and lower cost with glargine

Reference	Design	No. pts	Trial duration	Treatments	A1c change from baseline (%)	24-h PG AUC (mmol/L/h)	PG AUC > 7.0 mmol/L (mmol/L/h)	PPPG AUC	Frequency of nocturnal hypoglycemia (episodes/patient-month)	Comments
Ashwell 2006³⁰ (Study 4006)	Open-label, multicenter, two-way crossover study	56	32 weeks (two 16-week periods)	Glargine +lispro NPH (od or bid) + RHI	−0.5% (baseline 8.0%) 0.0% (baseline 8.0%) P < 0.001	187 203 P = 0.037	47 62 P = 0.017	75 88 P = 0.002	0.66 episodes 1.18 episodes P < 0.001	Glargine + lispro improved A1c and 24-h PG monitoring compared with NPH + RHI, with a reduction in nocturnal hypoglycemia

					Change from baseline		Frequency of hypoglycemia (episodes/patient–month, % patients)
Reference	Design	No. patients	Trial Duration	Treatments	A1c (%)	FBG (mmol/L)	Symptomatic	Nocturnal	Comment
vs ultralente
Kudva 2005²⁰	Partially-blinded, multicenter, crossover study	22	16 weeks of treatment	Glargine (bedtime) + aspart Ultralente (bedtime) + aspart	−0.12% (baseline 6.94%) + 0.08% P = 0.03	155 (at endpoint) 191 (at endpoint) P = 0.047	27.5 events 31.5 events P = 0.0455	3 2 P = NS	Lower A1c, nocturnal glycemic variability and hypoglycemia with glargine
vs CSII
Lepore 2003³²	Open, parallel-group study in patients previously on NPH	32	52 weeks	CSII with lispro + boluses at meals (n = 16) Lispro + dinner or bedtime glargine (n = 16)	−1.0% (baseline 9.2%), P < 0.001 −0.6% (baseline 8.5%), P < 0.001 P = NS	−3.8% (baseline 11.9), P < 0.001 −1.8 (baseline 12.3), P < 0.001 P = NS	Severe: 0.12 vs 0.37 on NPH, P < 0.05 Severe: 0.18 vs 0.43 episodes per patient/year, P < 0.05 P = NS	NR NR	Both CSII and MDI with lispro + glargine improved metabolic control and reduced severe hypoglycemia in patients not controlled on NPH as baslin insulin
Bruttomesso 2008³⁶	Multicenter, open-label, randomized. crossover study	39	6 months	CSII with lispro (n = 24) MDI with glargine + lispro (n = 15)	−0.2% (baseline 7.3%) −0.17% (baseline 7.3%) P = NS	BG variability 5–12% vs MDI^L NR P < 0.05	Moderate: 1.1 ± 1.7 episodes per patient 1.3 episodes per patient P = 0.327	Severe: 0.1 ± 0.3 0.1 ± 0.4 P = 0.710	Glucose variability was lower with CSII, with better glycemic control and higher treatment satisfaction than with glargine
Bolli 2009³⁷ (Study 4036)	Randomized, open, multicenter study of people previously on NPH	50	24 weeks	CSII (lispro) Glargine + lispro	−0.7 ± 0.7% (baseline 7.7%) −0.6 ± 0.8% (baseline 7.8%) P = NS	−3.3 (baseline 10.1) −2.7 (baseline 10.4) P = NS	1152 events by 23 of 28 patients (82%) 1022 events by 27 of 29 patients (93%) P = NS	3 events per patient 5 events per patient P = 0.34	In people naïve to CSII or glargine, glycemic control is similar with glargine vs the more expensive CSII therapy

^aMean daily blood glucose; ^bMild hypoglycemia (self-assisted episodes, blood glucose ≤ 4.0 mmol/L) episodes/patient-month; ^cEpisodes occurred at an earlier time at night (01:00–04:59); ^dPercentage of patients with at least one symptomatic event confirmed by BG of < 2.8 mmol/L; ^ePercentage of patients with at least one symptomatic event confirmed by BG of < 2.0 mmol/L; ^fMean total number of episodes per patient; ^gPercentage of patients with at least one episode; ^hAdjusted least-squares mean change from baseline. Unadjusted A1c change was −0.89% in glargine group and −0.67% in NPH group (P < 0.05); ⁱEvents per 100 patient-days; ^jMean incidence of both severe and non-severe hypoglycemia; ^kMean change in overall events (diurnal and nocturnal) episodes/patient/month; FBG = fasting blood glucose; NPH = neutral protamine Hagedorn insulin; od = once daily; bd = twice daily; NS = not significant; FPG = fasting blood glucose; RHI = regular human insulin; NR = not reported; BBT = basal–bolus therapy; DTSQ-diabetes treatment satisfaction questionnaire; BG = blood glucose; PPBG = postprandial blood glucose; CSII = continuous subcutaneous insulin infusion; MDI = multiple daily injections; PG = plasma glucose; AUC = area under curve; CGMS = continuous glucose monitoring system; glargine = insulin glargine.