Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: Short-acting insulin analogues reduce glycosylated hemoglobin more than regular human insulin but only in adults with type 1 diabetes


ACP J Club. 2005 May-Jun;142:63. doi:10.7326/ACPJC-2005-142-3-063

Related Content in this Issue
• Companion Abstract and Commentary: Review: Insulin monotherapy and insulin combined with oral hypoglycemic agents provide similar glycemic control

Clinical Impact Ratings

GIM/FP/GP: 6 stars

Endocrinology: 5 stars

Source Citation

Siebenhofer A, Plank J, Berghold A, et al. Short acting insulin analogues versus regular human insulin in patients with diabetes mellitus. Cochrane Database Syst Rev. 2004;(4):CD003287. [PubMed ID: 15495047]



In patients with diabetes, is a short-acting insulin analogue more effective than regular human insulin?


Data sources: 3 databases, references of studies and reviews, abstracts of major diabetology meetings, hand-searches of diabetes journals, the International Register of Clinical Trials Registers, the register of Current Science, 3 main pharmaceutical companies producing short-acting insulin analogues (Aventis, Eli Lilly, and Novo Nordisk), experts and approval agencies, and bibliographies of standard textbooks.

Study selection and assessment: Randomized controlled trials (RCTs) ≥ 4 weeks in duration that compared short-acting insulin analogues with regular human insulin in patients with type 1, type 2, or gestational diabetes. Methodological quality was assessed using the criteria in the Cochrane Handbook and the criteria of Schulz and Jadad.

Outcomes: Glycemic control, glycemic episodes, and quality of life (QOL).

Main results

42 RCTs (n = 7933) met the inclusion criteria. 25 RCTs were in patients with type 1 diabetes, 5 in patients with type 2 diabetes, 5 in a combined diabetes population, 3 in children, 1 in adolescents, 1 in pregnant women with type 1 diabetes, and 2 in women with gestational diabetes. 14 RCTs had parallel-group designs, and 28 had cross-over designs. 6 studies were of moderate methodological quality, and 36 were of poor quality. Glycated hemoglobin (HbA1c) was reduced more with short-acting insulin analogues than with regular human insulin in patients with type 1 diabetes; no difference was seen in patients with type 2 diabetes (Table). Hypoglycemic episodes did not differ between treatment groups in type 1 or type 2 diabetic patients (Table). In the RCT of pregnant women with type 1 diabetes, short-acting insulin analogues reduced HbA1c. Hypoglycemic episodes were reduced with short-acting insulin analogues in the RCT of adolescents (P = 0.02) and increased in the RCT of pregnant women with type 1 diabetes (P < 0.05). For the RCTs that assessed QOL with the Diabetes Treatment Satisfaction Questionnaire (the most-used instrument), 4 RCTs showed improvement with short-acting insulin analogues, and 3 RCTs showed no difference between groups.


In adults with type 1 diabetes, short-acting insulin analogue reduces glycosylated hemoglobin more than regular human insulin. No difference is seen for hypoglycemic episodes. In adults with type 2 diabetes, children, and women with gestational diabetes, insulin analogues and regular human insulin do not differ in effect.

Source of funding: Not stated.

For correspondence: Dr. A. Siebenhofer, Leopold Auenbrugger Medical University of Graz, Graz, Austria. E-mail andrea.siebenhofer@medunigraz-at.

Table. Short-acting insulin analogue vs regular human insulin*

Outcomes Diabetes patient group Number of comparisons Weighted mean difference (95% CI)
HbA1c change from baseline Type 1 21 −0.1% (−0.2 to −0.1)†
Type 2 4 −0.02% (−0.1 to 0.1)
Hypoglycemic episodes (mean episodes per patient-mo) Type 1 11 −0.2 (−1.2 to 0.9)
Type 2 5 −0.2 (−0.5 to 0.1)

*HbA1c = glycated hemoglobin. CI defined in Glossary. Mean follow-up was 3.6 months. A random-effects model was used.
†Difference favors short-acting insulin analogues.


The effective use of insulin in clinical practice remains an important clinical challenge for many practitioners. Indeed, insulin treatment is often delayed or withheld in patients with type 2 diabetes, and most practitioners appreciate that insulin-treated patients often have poorer glycemic control than those receiving oral agents. Two issues key to insulin use are examined in the reviews of Goudswaard and Siebenhofer and their colleagues: What is the potential benefit of combining OHAs with insulin in type 2 diabetes, and what is the benefit of such rapid-acting insulin analogues as lispro, aspart, and glulisine?

The review by Goudswaard and colleagues concludes that no significant clinical advantage exists with the use of combination OHA-insulin therapy compared with insulin monotherapy. However, this result is not entirely unexpected because many of the trials used combination sulfonylurea–insulin therapy. These 2 insulin-providing treatments might not be expected to have substantial advantages over insulin injection alone. Most studies of OHA combination suggest that the use of both insulin-providing and insulin-sensitizing therapy (e.g., adding metformin and/or thiazolidinedione [TZD] to sulfonylurea) specifically targets the metabolic defects that give rise to hyperglycemia and may in fact achieve superior control. Given this, an assessment of combined insulin–metformin or insulin–TZD therapy might reveal distinctly different results. Further, many of the included trials used only once- or twice-daily insulin injections, whereas optimal insulin therapy may require ≥ 3 injections in many patients.

The review by Goudswaard and colleagues simply confirms that use of either sulfonylurea or metformin with low-dose once- or twice-daily insulin is of limited benefit. The broad conclusion that insulin monotherapy and insulin combined with OHAs provide similar glycemic control should be viewed with caution. Further assessment of oral agents added to optimized multidose insulin and specific analysis of combination insulin-sensitizer therapy may show distinctly different results and could lead to other advantages, such as lower rates of hypoglycemia, a need for fewer injections, or the use of a lower insulin dose. Further, metformin may limit the weight gain often seen with insulin use. Independent of the results reported, the primary goal for patients and practitioners is to achieve the best and safest control possible. The review certainly does not exclude the possibility of a benefit from combinations of OHAs and insulin, and further assessment of such an approach must be done.

Rapid-acting insulin analogues are widely used in both type 1 and type 2 diabetes. However, even after nearly a decade of experience, optimal use of such insulins is not well understood. Are they most useful in multidose regimens? Are they equally effective for type 1 and type 2 diabetes? Do they offer a safety advantage? Do they improve either glycemic control or QOL?

Clinical experience and several smaller studies suggest that mealtime administration of rapid-acting insulin may improve postmeal blood glucose control and is associated with a lower risk for hypoglycemia, particularly at night. Further, meal-time dosing of rapid-acting insulin is convenient for many patients. The review by Siebenhofer and colleagues suggests that only modest improvements in glycemic control are achieved with rapid-acting insulin. However, even these small benefits may underestimate the potential clinical advantage of such insulin. Rapid-acting insulin offers greater convenience, can allow patients to alter meal plans more readily, and may be associated with less weight gain. However, these benefits would not be expected to improve glucose control. For many, hypoglycemia risk limits the ability to intensify insulin. Is it possible that use of insulin analogues in a real-life setting would result in greater improvements in HbA1c levels? This unanswered question underscores the need for well-designed clinical trials that include careful individualization of treatment regimens and carefully measure outcomes, including safety, satisfaction, flexibility, and convenience. In lieu of such trials, these reviews provide useful guidance for clinical decisions on insulin use, as always to be tempered by individual patient needs and wishes.

David M. Kendall, MD
International Diabetes Center
Minneapolis, Minnesota, USA


1. Chase HP, Lockspeiser T, Peery B, et al.the impact of the diabetes control and complications trialand humalog insulin on glycohemoglobin levels and severe hypoglycemia in type 1 diabetes. Diabetes Care. 2001;24:430-4. [PubMed ID: 11289463]