A randomized, controlled trial comparing insulin lispro with human soluble insulin in patients with Type 1 diabetes on intensified insulin therapy


Professor Edwin Gale, Diabetes and Metabolism, Medical School Unit, Southmead Hospital, Bristol BS10 5NB, UK. E-mail: Edwin.Gale@bristol.ac.uk
Members of the UK Trial Group: H. Alban-Davies (West Hill Hospital, Dartford), R.W. Bilous (Diabetes Care Centre, Middlesborough General Hospital, Middlesborough), C. Bradley (Diabetes Research Group, Royal Holloway College, University of London), D. Cunnah (Broomfield Hospital, Chelmsford), P.J. Grant (General Infirmary, Leeds), C.S. McIntosh (Queen Mary's Hospital, Roehampton, London), R. Rowe (St Bartholomew's Hospital, London), F.P. Vince (Coventry and Warwickshire Hospital, Coventry), S. Walford (New Cross Hospital, Wolverhampton), P.J. Watkins (Diabetes Department, King's College Hospital, London), D.J. Wayne (James Paget Hospital, Great Yarmouth)


Aims Despite considerable experience with insulin lispro, few blinded comparisons with soluble insulin are available. This study compared insulin lispro with human soluble insulin in patients with Type 1 diabetes mellitus on multiple injection therapy who inject shortly before meals.

Methods Glucose control, frequency of hypoglycaemia and patient preference were examined in the course of a prospective, randomized, double-blind, crossover comparison, with a 6-week run-in period and 12 weeks on each therapy. Ninety-three patients took part, all on multiple daily doses of insulin, with soluble insulin before meals and NPH (isophane) insulin at night. The main outcome measures were self-monitored blood glucose profiles, glycated haemoglobin, frequency of hypoglycaemic episodes, patient satisfaction and well-being and patient preference.

Results Blood glucose levels were significantly lower after breakfast and lunch, but higher before breakfast, lunch and supper, in patients taking insulin lispro. Levels of HbA1c were 7.4 ± 1.1% on Humulin S and 7.5 ± 1.1% on insulin lispro (P = 0.807). The overall frequency of symptomatic hypoglycaemia did not differ, but patients on insulin lispro were less likely to experience hypoglycaemia between midnight and 6 a.m., and more likely to experience episodes from 6 a.m. to midday. Questionnaires completed by 84/87 patients at the end of the study showed that 43 (51%) were able to identify each insulin correctly, nine (11%) were incorrect, and 32 (38%) were unable to tell the insulins apart. No significant preference emerged: 35 (42%) opted for insulin lispro, 24 (29%) opted for Humulin S, while the remainder had no clear preference.

Conclusions Substitution of insulin lispro for soluble insulin in a multiple injection regimen improved post-prandial glucose control at the expense of an increase in fasting and pre-prandial glucose levels. Patients who already injected shortly before meals expressed no clear preference for the fast-acting analogue, and did not improve their overall control as a result of using it. Nocturnal hypoglycaemia was however, less frequent on insulin lispro, and may emerge as a robust indication for its use.