Test-retest reliability of reflex receptive field mappings for objective evaluation of central sensitization

M.B. JENSEN; J.A. BIURRUN MANRESA; O.K. ANDERSEN

Buenos Aires

Congreso; 15th World Congress on Pain; 2014

Aim of Investigation: Recent studies indicate that assessment of the human nociceptive withdrawal reflex (NWR) is a useful tool for assessing central sensitization. Enlarged reflex receptive fields (RRF) have been detected in groups of chronic pain patients compared to control groups, most likely reflecting expansion of spinal neuronal receptive fields. The RRF of muscles in the lower extremities can be quantified by randomised, distributed electrical stimulation, typically on the sole of the foot. The NWR constitutes an objective measure to non-invasively probe the excitability of central nociceptive pathways. However, existing methodologies for quantification of RRF [1, 2] involve an element of subjectivity in the identification of pain thresholds to electrical stimulation. An alternative methodology based on detection of reflex thresholds has been developed which does not contain any subjective involvement from the subject or the investigator. In this study, the test-retest reliability of the novel reflex threshold methodology is compared to the existing methodologies. Methods: Reflexes were elicited by 5 sweeps of noxious electrical stimulation of 10 sites on the sole of the foot of 13 healthy subjects while the reflex responses were measured using EMG from the tibialis anterior muscle. Reflex sensitivity [1] and reflex probability [2] mappings were produced by interpolation of the reflex size (average root mean square value of the EMG signal) and reflex probability (percentage of sweeps eliciting a reflex) from each of the 10 stimulation sites. The stimulation intensities for each of the 10 sites were set to 1.3 times the pain thresholds identified using a staircase method. Reflex thresholds (i.e. the smallest stimulation intensity eliciting a reflex) were identified for each of the 10 stimulation sites using a fully automated staircase method based on conduction velocity analysis of the EMG signal [3]. Reflex threshold mappings were produced by interpolation of the average value of the inverse of the reflex threshold (one divided by the specific threshold) from each of the 10 stimulation sites. The area of the RRF for sensitivity and threshold mappings was quantified as the fraction of the foot with values exceeding a threshold defined as the peak value minus two standard deviations [1]. The RRF area of probability mappings was quantified as the fraction of the foot where a reflex was elicited by at least 60% of the stimulations [2]. The three different mappings were produced two times within one session (one hour interval - no reassessment of pain thresholds) and repeated once in a second session (48 ± 1 hour interval - subjective pain thresholds reassessed) to allow evaluation of both within and between session reliability. Test-retest reliability was assessed using Bland-Altman agreement analysis. The upper and lower limits of agreement (LOA) are expressed as bias ± coefficient of repeatability. Paired t-tests were applied to determine if the calculated biases were significantly different from zero. Results: The LOA for the within session comparison were similar for the areas derived from the sensitivity (-0.0042 ± 0.2047), probability (-0.0471 ± 0.2302) and threshold mappings (-0.0405 ± 0.2379). This was not the case for the between sessions comparison where the LOA for the areas derived from both the sensitivity (0.0606 ± 0.4415) and probability mappings (-0.0269 ± 0.3857) were wider than those of the threshold mappings (0.0163 ± 0.2971). None of the calculated biases were significantly different from zero. Conclusions: These results suggests that RRF areas derived from threshold mappings encompass less variability and are more stable between sessions compared to the existing methodologies. The larger variability observed for the sensitivity and probability mappings in the between sessions comparison compared to the within session comparison may likely be partly explained by the reassessment of subjective pain thresholds.[1] Neziri et al. J Neurosci Meth 2009[2] Manresa et al. J Electromyogr Kines 2011[3] Jensen et al. BMC Neurosci 2013