Abstract
INTRODUCTION: Lipohypertrophy (LH) is a common complication occurring in diabetes individuals. The most common methods used include palpation, visual examination and/or ultrasound (US). To date, there is limited information on the detection sensitivity among the different techniques used to identify LH. This systematic review aimed to identify studies that examined insulin-related LH using US detection to identify the prevalence, characteristics and morphology of LH, and to compare US and clinical palpation methods for detecting LH. METHODS: Three electronic databases were systematically searched for studies detecting LH using US in insulin users. Articles were screened for eligibility and included studies were appraised using quality assessment tools. The quality of the evidence was evaluated using Grading of Recommendations Assessment, Development and Evaluation, and the extracted data was synthesised narratively. RESULTS: Sixteen articles were included in the review providing data on 1722 patients. The prevalence of LH prevalence varied from 14.5% to 88% (median 56.6%). Identified risk factors for the development of included insulin injection behaviour such as a lack of injection site rotation and social factors such as low education level. Four studies compared LH detection by US to palpation, providing inconsistent results. One study showed that palpation detected 64% more LH, whilst two studies demonstrated that US identified 50% more sites and extended areas of LH (additional ~ 5 cm(2)). Another study provided comparable estimates between palpation and US in clinicians trained to detect LH (97%). CONCLUSION: The evidence highlights a lack of congruence in results pertaining to the detection sensitivity of US and palpation for LH sites. More research with robust study design is needed to verify whether clinically palpation is sufficient to detect LH, or whether US would increase the precision of LH assessment to help address this common clinically significant problem.