A meta-analysis and meta-regression on the prevalence of lipohypertrophy in diabetic patients on insulin therapy

Abstract

This study aimed to assess the prevalence of lipohypertrophy (LH) in diabetes mellitus (DM) patients on insulin therapy and identify factors influencing the prevalence rates through a meta-analysis and meta-regression. Databases of PubMed, Embase, Scopus, and Google Scholar were screened from inception to 20th Aug 2020. Data on prevalence was transformed using the logit transformation for pooling the proportions using the DerSimonian-Laird meta-analysis model. Random-effects meta-regression analysis was performed to assess the influence of the following moderators on the pooled prevalence: male gender, mean age, type of DM, DM duration, duration of insulin therapy, needle re-users, rotation of injection site, HbA1c, insulin dose, ≥8mm needle users and mean number of injections/day. Forty-five studies were included. LH was diagnosed by observation and palpation by all studies. On meta-analysis of data from 26,865 participants, the overall prevalence of LH was found to be 41.8% (95% CI: 35.9% to 47.6%). On region-wise analysis, pooled prevalence of LH in Europe was 44.6% (95% CI: 37.5% to 51.8%) in Africa was 34.8% (95% CI: 16.9% to 52.8%) and in Asia was 41.3% (95% CI: 27.2% to 55.3%). The pooled prevalence of LH in studies only on type 1 and type 2 DM patients was 39.9% (95% CI: 28.3% to 51.6%) and 45.9% (95% CI: 29.5% to 62.4%) respectively. Only insulin duration was found to significantly influence the prevalence of LH on meta-regression. Our study indicates that the prevalence of LH in DM patients on insulin therapy is estimated to be 41.8%. The longer the duration of insulin therapy, the higher is the prevalence of LH.