Non-invasive assessment of vibration perception and protective sensation in people with diabetes mellitus: inter- and intra-rater reliability.

Background: Testing of protective sensation and vibration perception are two of the most commonly used non-invasive methods of screening for diabetes-related peripheral neuropathy (DPN). However, there is limited research investigating the reliability of these tests in people with diabetes. The aim of this study was to determine the inter- and intra-rater reliability of methods used to test vibration perception and protective sensation in a community-based population of adults with type 2 diabetes.

Methods: Three podiatrists with varying clinical experience tested four- and 10-site, 10 g monofilament and vibration perception threshold (VPT). In a separate cohort, the reliability of a graduated tuning fork as well as two methods of conventional tuning fork (on/off method and dampening method) was undertaken by a new graduate podiatrist and podiatrist with one-year's clinical experience. The intra- (Cohen's К) and inter-rater (Cohen's or Fleiss' К) reliability of each test was determined.

Results: Fifty participants (66% male, 100% type 2, 32% with DPN) underwent monofilament and neurothesiometer testing with 44 returning for the retest. Twenty-four participants (63% male, 100% type 2, 4% with DPN) underwent tuning fork testing and returned for retest. All tests demonstrated acceptable inter-rater reliability ranging from moderate (10-site monofilament, К: 0.54, CI: 0.38-0.70,  = 0.02) to substantial (graduated tuning fork, К: 0.68, CI: 0.41-0.95,  < 0.01). The 10-site monofilament (К: 0.44-0.77) outperformed the 4-site test (К: 0.34-0.67) and the dampened tuning fork method (К: 0.41-0.49) showed lower intra-rater reliability compared to both conventional (К: 0.52-0.57) and graduated methods (К: 0.50-0.57).

Conclusion: We support the current recommendations of using more than one test to screen and monitor progression of DPN. Four- and 10-site 10 g monofilament testing have similarly acceptable levels of reliability and the neurothesiometer is the most reliable method of assessing vibration perception function. Use of a graduated tuning fork was slightly more reliable than other methods of tuning fork application however all had substantial reliability. Years of clinical experience only marginally affected test reliability overall and due to subjective nature of the tests we suggest that testing should be performed regularly and repetitively.