The literature search retrieved 26 studies, in which insulin glargine was compared with NPH insulin (13 studies), ultralente (one study), or CSII (five studies) in adults (). Of these, four studies versus NPH and two studies versus CSII in adults were shorter than 12 weeks and were subsequently excluded. Six studies compared insulin glargine with NPH insulin/lente insulin in children and adolescents (); one of these studies was excluded because the duration of treatment was <12 weeks.
Randomized Controlled Trials Comparing Insulin Glargine with NPH Insulin, Ultralente and Continuous Subcutaneous Insulin Infusion
Randomized Controlled Trials in Children/Adolescents Comparing Insulin Glargine with NPH Insulin, Ultralente and Continuous Subcutaneous Insulin Infusion
Clinical efficacy and safety with insulin glargine
Insulin glargine versus NPH insulin
The search identified nine randomized controlled trials that compared insulin glargine with NPH insulin as part of an MDI regimen, ranging in duration from 12 weeks10
to 1 year11
(). As summarized in this table, insulin glargine consistently provided significantly10–13
greater baseline-to-endpoint improvements in FBG than NPH insulin, and this was evident in studies ranging in duration from 12 weeks to 1 year. This difference was apparent both in studies where NPH was administered once daily and in studies where NPH was administered twice daily. The improvement in FBG with insulin glargine over NPH insulin was approximately 1–2
mmol/L. However, two studies reported improvements in FBG >3
mmol/L with insulin glargine.15,16
The improvements in FBG were associated with small improvements in A1C with both insulins, although A1C increased slightly in one study.14
In terms of the magnitude of A1C improvement, four studies showed no difference between the two insulins,12,13,17,18
whereas four studies showed significantly greater improvements with insulin glargine compared with NPH insulin,10,11,13,18
consistent with improvements in FBG. However, three of these studies used NPH insulin once daily,10,11,16
and greater improvements in A1C may have been possible if NPH insulin was used twice daily or if there was an option to add a second daily dose of NPH insulin in these studies. For example, in the 30-week study by Fulcher et al.,16
insulin glargine achieved greater improvements in A1C and FBG than once-daily NPH insulin. In contrast, in the study by Bolli et al.,17
insulin glargine elicited significantly greater improvements in FBG than twice-daily NPH insulin, but the magnitude of A1C improvement was identical in both groups. Thus, it is possible that, although twice-daily NPH insulin does not seem to affect FBG, it may affect other parameters of glucose control, such as postprandial BG, to improve A1C.
Twice-daily insulin glargine has been tested in a crossover study in which 20 patients with type 1 diabetes were given once-daily insulin glargine at dinnertime or twice-daily insulin glargine (half doses at breakfast and dinner).19
Over the 4-week treatment period, the twice-daily regimen resulted in lower BG levels after breakfast, lunch, and before dinner and was also associated with lower mean 24-h BG levels (7.1 vs. 8.8
0.031) and less intraday variability in BG levels (P
0.044). The authors concluded that for patients who experience late afternoon increases in BG levels, twice-daily insulin glargine may be a suitable alternative that does not require an increase in insulin dosage, although it is important to note that administering insulin glargine more than once daily is not currently approved by regulators and represents off-label use.
Another outcome that should be considered in the treatment of type 1 diabetes is the prevalence of hypoglycemia. Hypoglycemia is often debilitating and may lead to adverse outcomes. Therefore, in an effort to avoid hypoglycemia, patients may inadequately titrate their insulin, which may ultimately result in suboptimal glycemic control. Reviewing these nine studies, the authors found that insulin glargine was associated with a significantly reduced risk of symptomatic hypoglycemia10,11,13,17,18
or no difference in risk12,14–17
compared with NPH insulin. These differences were most marked in the studies that administered NPH insulin once daily. This difference in hypoglycemia between once-daily NPH and insulin glargine is reflected in the recommendations for a dose reduction of 20–30% when transitioning from once-daily NPH insulin to insulin glargine, while affording the same or better overall efficacy at lower doses.
Insulin glargine versus ultralente
The authors found only one study that compared insulin glargine with ultralente insulin. In this study, once-daily insulin glargine was shown to be associated with greater improvements in A1C and FBG and fewer episodes of hypoglycemia compared with once-daily ultralente.20
However, this was a relatively small study, with only 22 patients, and should be interpreted with caution.
The “dawn phenomenon” and insulin glargine: implications on dose timing
The term “dawn phenomenon” describes hyperglycemia that occurs in the early morning.21
Its causes are still not fully understood, but it seems to be related to a combination of the waning of the effects of intermediate-acting insulins, such as NPH insulin, before the next dose and to surges in circulating levels of other hormones, including growth hormone and cortisol, and a reduction in free insulin-like growth factor-1 (IGF-1) levels resulting in greater insulin resistance.22–24
It has been suggested that as many as 54% of patients with type 1 diabetes experience the dawn phenomenon.21
Patients with marked hyperglycemia are generally managed with CSII programmed to deliver an increased basal rate at an appropriate time during the night,23
although there is some evidence to suggest that long-acting insulin analogs could also provide better control of early morning BG compared with NPH insulin. A nonrandomized study of 48 Japanese patients with type 1 diabetes treated with insulin glargine or NPH insulin as part of an MDI regimen, or CSII at a constant rate, evaluated nocturnal and early morning BG and free IGF-1 levels.25
A total of 60% of patients receiving NPH insulin experienced the dawn phenomenon, whereas BG levels in the glargine and CSII groups were more stable, and few patients experienced the dawn phenomenon.25
These differences corresponded with a marked reduction in free IGF-1 with NPH insulin, whereas IGF-1 levels remained relatively stable with CSII and insulin glargine. The authors concluded that the more constant insulin bioavailability with CSII or insulin glargine was effective in managing early-morning BG increases.
Pharmacokinetic and pharmacodynamic studies have revealed that insulin glargine has a longer duration of action than NPH insulin. However, the duration of action of insulin glargine may not reach 24
h in some people,26,27
which may be reflected by hyperglycemia shortly before the next administration. In one clinical study that compared the effects of timing of insulin glargine administration (lunchtime, dinnertime, and bedtime) (), plasma insulin levels tended to wane shortly before the injection, corresponding to a small increase in plasma glucose levels.28
This effect was most notable for the dinnertime injection. In a similar study, the FBG levels did not change during the 24-week treatment period when insulin glargine was injected at breakfast, similar to the dawn phenomenon.29
The results of these studies suggest that changing the time of insulin glargine injection to lunchtime or bedtime should avoid hyperglycemia before the next insulin glargine injection. This may be explained by the shorter intervals between breakfast and lunch and between dinner and bedtime, compared with that between lunch and dinner, thus providing sufficient insulin cover from the prandial insulin to overcome any waning of insulin glargine.28
Studies Evaluating Timing of Insulin Glargine Administration
Combining insulin glargine with a rapid-acting insulin
As outlined in the previous section, the use of insulin glargine versus other long- and intermediate-acting insulins as part of an MDI regimen in type 1 diabetes has been extensively studied. Almost all of the studies of insulin glargine to date either used insulin glargine in both groups and compared the efficacy of short-acting insulins at mealtimes or used the same short-acting insulin in both groups to compare the efficacy of basal insulin (as described above). The results of studies that compared short-acting insulins alone are beyond the scope of the present review. To the authors' knowledge, only one study in type 1 diabetes has compared insulin glargine plus a rapid-acting analog (insulin lispro) versus NPH insulin in combination with RHI.30
In that 32-week, two-way crossover study (16 weeks per treatment period), insulin glargine plus insulin lispro achieved a significantly lower A1C compared with NPH insulin plus RHI (7.5% vs. 8.0%; P
0.001). This was associated with significantly lower 24-h glucose area under the curve (AUC) (187 vs. 203
0.037), plasma glucose AUC >7
mmol/L (47 vs. 62
0.017), and postprandial plasma glucose AUC (75 vs. 88
mmol/L/h), although not nighttime plasma glucose AUC or plasma glucose AUC <3.5
mmol/L. The total rate of symptomatic hypoglycemia was comparable (1,277 vs. 1,327 episodes), but the rate of symptomatic nocturnal hypoglycemia was significantly lower with insulin glargine plus insulin lispro (0.66
0.02 vs. 1.18
0.02 episodes/month; P
0.001). Most episodes of nocturnal hypoglycemia occurred at 06:00–0:700
h with insulin glargine plus insulin lispro versus 00:00–04:00
h with NPH insulin plus RHI.
Insulin glargine-based MDI therapy versus CSII
The authors found three clinical studies that compared insulin glargine-based MDI with CSII, all of which used insulin lispro in the CSII group.31–37
As summarized in , these studies show comparable or marginally greater improvements in glycemic control and lower rates of hypoglycemia with CSII versus MDI therapy with insulin glargine, indicating that CSII may be more effective than MDI therapy with insulin glargine. However, most of these studies were relatively small in size, limiting the ability to detect differences in either regimen. In addition, CSII allows the delivery of multiple basal rates, which might help mitigate the dawn phenomenon. Larger, better-designed studies with appropriate powering may help to better understand the relative impact of CSII and insulin glargine-based MDI on glycemic control and hypoglycemia in patients with type 1 diabetes.
Implications of insulin analogs for the treatment of type 1 diabetes in adults
As described above, insulin analogs provide improved pharmacokinetic and pharmacodynamic characteristics relative to the respective human insulin. Since their introduction, it has been proposed that these properties confer advantages for the treatment of diabetes, particularly in terms of reduced risk of hypoglycemia, which has been demonstrated in meta regression analyses for insulin glargine6
and insulin detemir.9
The reduced risk of hypoglycemia with insulin glargine relative to NPH insulin may enable more patients to achieve treatment targets, through more aggressive titration of the insulin dose. Studies show that basal insulin analogs provide potentially important improvements in glycemic control that should reduce the risk of diabetes-related complications with long-term intensive therapy. The flexibility of insulin glargine dosing in relation to timing of administration has also been demonstrated (and approved by the Food and Drug Administration)—in particular, the opportunity for administration at breakfast, dinnertime, or bedtime, providing the timing of daily injection is constant.28,29,38
As described previously in this review, insulin analogs provide clinically important improvements in glycemic control within a tightly controlled clinical trial setting; but how is this evidence reflected in everyday clinical practice? Observational studies are generally regarded as the best approach to assess the actual health outcomes of patients in routine care.39,40
This is because the level of care patients receive in clinical trials is often of a different standard and not representative of that seen in daily clinical practice, particularly with respect to patient populations and medication adherence.41,42
In addition, clinical trials may include a limited scope of titration with other glucose-lowering drugs (or other concomitant treatments), a relatively short observational period, and potential for population bias, which may prevent extrapolation of findings to everyday practice.43,44
Therefore, what is the evidence in less rigorously controlled settings, where patients may receive less support from their clinician? There is ample evidence from everyday clinical practice to demonstrate the efficacy of insulin analogs.45–61
Switching to insulin analog-based MDI regimens was associated with marked improvements in glycemic control. For example, in an observational study of 1,942 patients who were switched from NPH insulin to insulin glargine, mean A1C declined by 0.8% over 6 weeks of treatment, from 8.0% at baseline.56
In a second study of longer duration, 1,447 patients who were switched from various insulin regimens to basal–bolus therapy with insulin glargine and insulin glulisine experienced a significant mean reduction in A1C of 1% over 6 months from a baseline of 8.0%.54
These findings support the evidence gained in randomized controlled trials. Nevertheless, one must interpret such studies with care, because of the lack of a comparator group, and the potential for bias through patient selection, and limited data collection—hypoglycemia, for example, is often under-reported in such studies.
Insulin analogs in children and adolescents
Several studies have been performed to evaluate the efficacy of insulin glargine in children and adolescents. These studies are important because type 1 diabetes is usually diagnosed at a young age, and pediatric treatment represents an important aspect of managing the condition. The literature search identified six studies (seven publications)62–68
in children, ranging in duration from 9 weeks to 32 weeks (). However, the small sample size (<50 patients) in most of these studies limits their validity. The other two studies enrolled 349 and 175 patients and compared insulin glargine with either NPH insulin (once or twice daily)65,66
or NPH insulin/lente insulin.62
In the study by Schober et al.65,66
with children and adolescents 5–16 years of age, insulin glargine was associated with significantly greater improvements in FBG, although this did not translate into improvements in A1C. In the study by Chase et al.62
in adolescents and teenagers 9–17 years of age, there were no differences in the magnitude of improvement in A1C. However, after adjusting for baseline A1C, the change in A1C was significantly greater with insulin glargine than with NPH insulin/lente insulin. In terms of hypoglycemia, the study by Schober et al.65,66
revealed no difference in the rate of hypoglycemia, whereas the study by Chase et al.62
revealed higher rates of confirmed hypoglycemia with BG <70
mg/dL (116 vs. 94 events/patient-year, P
Treatment flexibility and treatment satisfaction
Insulin analogs provide several functional advantages that may increase treatment satisfaction for patients with diabetes. Clinical studies show that insulin glargine provides flexible basal insulin control with the option of once-daily administration at any time of day.28,29,38
Insulin glargine, compared with NPH insulin, is also associated with lower rates of hypoglycemia, a side effect that patients find particularly distressing. Rapid-acting insulin analogs provide flexibility, with the possibility of injecting immediately after a meal constituting an important advantage because of the potential to administer a dose of insulin appropriate for the meal size/content. It has been reported that insulin analogs are associated with greater treatment satisfaction relative to NPH insulin in patients with type 1 diabetes.69,70
Ashwell et al.69
compared quality of life (QoL) and treatment satisfaction with MDI regimens based on insulin glargine plus insulin lispro versus NPH insulin plus RHI. Over 32 weeks of treatment, insulin glargine plus insulin lispro significantly improved treatment satisfaction and patient QoL compared with the NPH plus RHI regimen.
However, data on the impact of insulin therapy on patient satisfaction and QoL remain limited in type 1 diabetes. These outcomes should be evaluated in more detail in future clinical trials because there is some evidence that such factors not only influence patients' perceptions of their condition, but also their adherence to treatment. Indeed, patients reporting poor QoL or poor treatment satisfaction may be less likely to adhere to their treatment.71–73
Furthermore, as reported by Samann et al.,74
flexible, intensive insulin therapy with dietary freedom can achieve significant improvements in glycemic control without increasing the risk of severe hypoglycemia. Such treatment flexibility with insulin analogs75
may be especially important to reduce the incidence of hypoglycemia in children and adolescents, particularly considering their unpredictable lifestyles.
The authors performed a comprehensive literature search to retrieve reports describing the use of insulin glargine versus NPH insulin for the treatment of type 1 diabetes. However, some limitations should be discussed. First, the search was limited to EMBASE, MEDLINE, and Web of Science. However, it is possible that some studies published in journals not indexed in any of these databases were missed, particularly non-English journal articles. Second, the present analysis was limited to randomized controlled trials; this is recommended to avoid possible bias associated with non-randomized cohort studies or retrospective reviews of medical databases. However, the authors identified a large number of such studies involving several thousand patients that may greatly influence the overall interpretation, as such studies may better reflect everyday clinical practice and patient expectations for treatment. Third, although QoL is an important factor in the holistic effects of insulin therapy, few studies to date have assessed QoL or reported data if the studies did assess QoL. Moreover, the reports that did include QoL did not routinely use the same questionnaire(s). Therefore, a meaningful assessment of QoL could not be provided by this review.