Aim  Sensory diabetic neuropathy, determined by nerve conduction studies, is common in children with Type 1 diabetes. Diabetic neuropathy diagnoses are rarely made in paediatric daily care because they are asymptomatic, vibration detection is mostly normal and nerve-conduction testing is impractical. The present study aims to: (1) describe somatosensory dysfunction in children with diabetes, (2) test whether diabetes duration and HbA1c are related to somatosensory dysfunction and (3) identify the best screening test for large-fibre dysfunction, as indicated by nerve conduction studies.

Methods  Forty-five children (age 13.2 ± 2.5 years) with Type 1 diabetes for 6.7 ± 2.5 years and matched control subjects were assessed by neurological examinations, nerve conduction tests and quantitative sensory testing on the feet using the protocol of the German Research Network on Neuropathic Pain. Abnormal nerve conduction was used as gold standard to define neuropathies.

Results  We found a high prevalence of mechanical (38%) and thermal (24%) hypoesthesia often associated with hyperalgesia (47%). Tactile hypoesthesia (33%) was more frequent than pallhypaesthesia (11%). Only cold detection and mechanical pain thresholds were related to HbA1c. Tactile hypoesthesia had the highest sensitivity (75%), specificity (89%) and positive (75%) and negative (89%) predictive values for neuropathies defined by nerve conduction tests (31% abnormal).

Conclusions  Almost half of the children with diabetes have subclinical large- and small-fibre neuropathies. Tactile detection was better than vibration for neuropathy assessment. Quantitative sensory testing is a valuable tool for assessment of neuropathy as well as a target of interventional studies in children with diabetes.