ABSTRACT
AIMS: A systematic review of the literature was carried out to examine whether published evidence suggests a difference in the frequency and awareness of hypoglycaemia induced by 'human' and animal insulin. METHODS: The review identified randomized controlled trials and studies of other designs including observational comparisons, case series and case reports in which the use of 'human' insulin was compared to animal insulin in people with diabetes. These were identified from bibliographic databases and hand-searches of key journals. The main outcome measures were frequency, severity, awareness and symptoms of insulin induced hypoglycaemia. RESULTS: Fifty-two randomized controlled trials, 37 of double-blind design, were identified which included one or more of the relevant outcome measures. Of these, 21 specifically investigated hypoglycaemic frequency and awareness as primary outcomes (six in people with previously reported reduced hypoglycaemic awareness). The remainder of the identified trials reported hypoglycaemic outcomes as a secondary or incidental outcome during comparative investigations of efficacy or immunogenicity. Seven of the double-blind studies reported differences in frequency of hypoglycaemia or awareness of symptoms, although none of the studies which selected subjects on the basis of previously reported impaired awareness demonstrated significant differences between insulin species. Four of the unblinded trials reported differences in hypoglycaemia. This reached statistical significance in two of the studies. A further 56 studies of other designs and case reports were considered. In addition to the 10 case reports describing individuals with impaired hypoglycaemic awareness, nine studies reported differences in the incidence and manifestation of hypoglycaemia during 'human' insulin treatment. Notably, none of the four population time trend studies found any relationship between the increasing use of 'human' insulin and hospital admission for hypoglycaemia or unexplained death among those with diabetes. The largest case series could find no support for the hypothesis that an influence of treatment with 'human' insulin on hypoglycaemia had contributed to any of the 50 deaths investigated. When all types of studies considered are ranked in order of rigour (according to the accepted 'hierarchy of evidence'), it is the least rigorous which lend most support to the notion that treatment with 'human' insulin has an effect on the frequency, severity or symptoms of hypoglycaemia. CONCLUSIONS: Evidence does not support the contention that treatment with 'human' insulin per se affects the frequency, severity or symptoms of hypoglycaemia. However, a number of studies, mainly those of less rigorous design, describe an effect when people are transferred from animal insulin to 'human' insulin. It is not possible to state how common this is or whether the phenomenon is specific to 'human' insulin or an effect resulting from stricter glycaemic control (perhaps compounded, in some cases, by neurological complications in long-standing diabetes). This remaining uncertainty makes it essential that insulin from animal sources continues to be available so that clinicians and patients may retain this choice of treatment.
