Reflections on the diagnostic accuracy of the Upper Limb Neurodynamic Test 1

We congratulate Bueno-Gracia and colleagues for their study on the diagnostic accuracy of the Upper Limb Neurodynamic Test 1 (ULNT1), which stressed the target of identifying clear and shared criteria for positive and/or negative results (Vanti et al., 2012). This study reinforces that ULNT1 “must be interpreted in combination with other clinical findings and diagnostic studies” (Vanti et al., 2011). Nevertheless, some methodological aspects deserve comment, as they are able to affect estimates of sensibility and specificity of a diagnostic procedure. The usefulness of an index test is determined by comparison to a reference standard test that has previously been shown through experience to establish a definitive diagnosis of the disease under study. Scientific evidence states that electrodiagnostic studies should be used to confirm a clinical diagnosis of CTS. Bueno-Gracia and colleagues recruited patients in the presence of “hand, wrist or forearm symptoms referred to “X” for median nerve NCS”; this criterion seems too comprehensive to be considered a clinical diagnosis of CTS. In fact, stricter consensus criteria have been proposed in literature (Rempel et al., 1998). Furthermore, criteria for abnormal nerve conduction findings are usually defined in statistically computed terms (e.g., range and mean ± 2 SD) from a reference population formed of individuals believed to be free of the disease and comparable to subjects on whom the protocol is being applied (Jablecki et al., 2002). These criteria should be clearly described, to allow the replication of the research protocol. Particular attention should also be paid when defining cut-offs for specific conditions that require the use of normative data collected from a subset of patients with specific medical conditions (e.g. diabetes, which accounts for more than 17% in the study group) (Werner, 2013). The definition of positive and negative results for the index test is also detrimental to the investigation of a diagnostic test. An evocation of symptoms during ULNT1 in areas different from the typical complaints of those suffering from CTS (e.g. the neck), is likely unrelated to a compression at the carpal tunnel. Otherwise, an induction of symptoms typical of median nerve compression at the wrist, but dissimilar to the patients' own symptoms, could be interpreted as a simple response to the nerve traction, having poor diagnostic usefulness. According to their criterion A, Bueno-Gracia and colleagues considered ULNT1 positive when patients' clinical symptoms were reproduced and changed during structural differentiation. Which symptoms had to be reproduced? Each kind of symptom on hand, wrist or forearm? In our opinion, this could be too approximate to diagnose CTS and risk yielding false positive results. The adherence to rigorous and unambiguous criteria, both for the reference and the index test, may help researchers in the correct evaluation of diagnostic procedure performance. We thank the Editor for giving the Authors the possibility to comment on this study, and hope to have provided productive feedback.