Discussion

We report the changes in PD and BPI outcomes of n=258 LBP patients who were positive to the BUAS test and hence were treated with gabapentin for their potential lumbar-RP condition. After the treatment, all prespecified endpoints of the study were reached. We argue that these results may be considered a piece of ex-adiuvantibus evidence for the ability of the BUAS test to detect lumbar-RP in LBP patients.
In a medical context, the term ex adiuvantibus (from Latin, meaning ”from that which helps”) refers to the method of making an inference about the presence/cause of a pathological conditionfrom the observed responses of the condition to a giventreatment . In our case, the potential pathological condition was lumbar-RP, as the BUAS test potentially diagnosed it.
We have recently proposed the BUAS-test as an adjunctive clinical sign for lumbar-RP 10. We have argued that positive BUAS test implies potential lumbar-RP, and vice-versa. Lumbar-RP, a known NeP condition, can be indirectly quantified by measures like the PD questionnaire and the BPI, which we here used to quantify the observed responses to the given treatment. We hypothesized that a congruent treatment for potential lumbar-RP (in our case, gabapentin) might lead to its clinical improvement. Such improvement is measurable with the tools mentioned above. Thus, subsequent significant improvement in PD and BPI outcomes may imply, ex adiuvantibus , that the underlying pathological condition was indeed lumbar-RP as the BUAS test initially implied it. Accordingly, we argued that should a significant improvement in the lumbar-RP condition occur, it would be reasonable to infer that the BUAS-test did detect the lumbar-RP. To further support this inference, we assumed that, after treatment, the observed responses might be considered ’significant improvement’ only should these responses fulfill restricted prespecified statistical and clinical endpoints, following the recommendations for achieving MCIC in pain research 14,15. These endpoints were: over 50% of the sample should show negative PD-V2 outcome, significant V2-V1 differences of the BPI-items’ mean-score, and that these differences should be of at least 2 points for each BPI item.
Lumbar-RP is considered the result of abnormal discharges arising from afferents of an injured dorsal root, ganglion, or inflamed nerve3. These abnormal discharges are the expression of nerve roots’ sensitization to mechanical stimulation after sustained and protracted compression (i.e., Mass Effect), or their response to the perineural-spread of inflammogenic material arising from the nucleus polposus (i.e., Chemical Radiculitis) 16. These mechanisms induce maladaptive neuronal responses that thus have the potential to generate abnormal responses from the affected nerve17. The differential diagnosis of LBP is complicated by the variability and possible coexistence of multiple pathophysiological and presentation patterns. A thorough history and clinical examination, also in primary care settings, are essential in establishing diagnosis and treatment in LBP patients. In particular, symptoms, an accurate neurological examination, imaging, and specific signs (e.g., SLRT and recently, the BUAS test) for lumbar-RP may guide differential diagnosis 1.
The rationale for the application of the BUAS test is explained elsewhere 10. Briefly, animal and human research assert that applying a strain on an injured nerve, sensitized or inflamed, may elicit heterospecific discharges, and hence subjective responses in multiple somatosensory afferents. Such elicited responses to the applied strain can confirm the presence of RP in various clinical conditions, including injured/inflamed lumbar dorsal roots and ganglion2,3,5,6,18–25.
With the BUAS test, we proposed to apply the strain on the sciatic nerve along its course in the buttock. Several reasons support the feasibility and efficacy of the BUAS test. First, the sciatic nerve, which contains axons from the L4-S3 levels, often involved in lumbar RP, is easily identifiable in the buttock. Second, applying the strain at this proximal sciatic-nerve course, the abnormal responses can be directly associated with proximal nerve injuries. Third, in the buttock, the sciatic nerve lies above stable anatomical structures allowing the strain to be effectively applied. Finally, the BUAS test showed, among LBP patients, satisfactory sensitivity (92%), specificity (100%), prior probability, and interrater reliability (Cohen’s Kappa, 0.911)10. It showed no associations with predictors like gender and age groups while it showed significant associations with predictors like the pain site (lumbar pain and sciatica), and the PD questionnaire. These findings imply that the BUAS test is not compromised by patients’ demographic features and supports its content validity. By contrast, the SLRT, widely used for the diagnosis of lumbar-RP, should be considered with caution as it shows crucial limitations 26. These limitations include the lack of standard procedure for carrying out and interpreting the SLRT, unclear underlying mode-of-action, compromised diagnostic reliability related to specific predictors (the patient’s age, gender, psychosocial factors, and diurnal variability), and lack of direct correlation with MRI findings, in case of a herniated disc. Finally, “the diagnostic value of the SLRT in detecting the presence of lumbar disc herniation may lie primarily in ruling out its presence because the test sensitivity (0.8) is far higher than its specificity (0.4)” 26 .
Treating lumbar-RP implies a step-wise approach from diagnosis to treatment (conservative, invasive, or potential surgery) to ameliorate or resolve the lumbar-RP condition and to permit patients to achieve a better functional status and quality of life 27. Based on moderate or high quality of evidence and efficacy, antidepressants and anticonvulsants (e.g., pregabalin and gabapentin) have strong GRADE recommendations for use in all NeP conditions and are thus proposed as first-line treatment for NeP 7,28,29. Several controlled studies have found gabapentinoids to be effective for NeP and lumbar-RP 10,30–34. We chose gababapentin both because of our positive experience with this drug, its known pharmacodynamic properties in lumbar-RP, and local favorable economic characteristics 9. Gabapentin is prescribed with a starting dose of 300 mg and may be titrated up to 1800-3600 mg/day with a “low and slow” incremental fashion 8,9,35. For optimal clinical efficacy gabapentin should be prescribed tidbecause of its short elimination half-life, and saturable absorption system.
Our inhouse protocol for first-line treatment of potential lumbar-RP (i.e., patients positive to the BUAS test and/or the SLRT), is a 7 days titration to 900 mg of gabapentin daily (300 mg tablets tid). Therapy outcomes and side effects are evaluated in the follow-up visits where doses may be tapered or increased according to the patients’ clinical responses. In our sample no patient stopped the therapy because of side effects even though 1.2% patients reported initial dizziness and somnolence during the gabapentin titration period, which diminished over time.
The observed responses to the gabapentin treatment in this paper were retrieved from the PD and BPI questionnaires. The PD explores the likelihood of the NeP component in LBP patients. The final score of the questionnaire yields the three PD outcomes: negative, uncertain, and positive 11–13. In our sample, PD-V1 outcomes were positive (45%) or uncertain (55%), while PD-V2 outcomes were mostly negative (77%). These results imply that over 50% of the sample had a negative PD-V2 outcome and hence fulfilling the first endpoint of the study. Interestingly, while the BUAS+/SLART- subset showed only negative PD-V2 outcome, in the BUAS+/SLART+ subset, several PD-V2 outcomes were positive (20%) or uncertain (37%). A possible explanation to these results may be that in the BUAS+/SLART+ subset, the underlying lumbar-RP condition was somehow severer. Indeed, unlike the BUAS+/SLART- subset, patients in the BUAS+/SLART+ subset were positive to both SLRT and BUAS tests. Further, it is possible that in these patients, the gabapentin dose (300 mg tid ) was not sufficient to manage the severer lumbar-RP condition. Indeed, recommended daily dose of gabapentin is higher (1800-3600 mg daily) 8,9,35. Interestingly, we found that in the BUAS+/SLART+ subset, and compared to PD-V1, the mean of the positive PD-V2 final-scores decreased from 23.8 to 19.0 (the lower cutoff of the negative PD outcome) in all patients, and from 18.0 to 15.9 for the uncertain outcome. These results indirectly show that the treatment did affect the severer lumbar-RP condition but yet, it was not clinically sufficient. Congruently, it is reasonable to speculate that the BUAS test may diagnose the presence of potential lumbar-RP also in the absence of positive SLRT. The presence of both signs may imply a severer lumbar-RP condition and possibly the necessity of higher doses of gabapentin or more invasive approach for its treatment. Indeed, in most patients who did not achieve a negative PD-V2 outcome, gabapentin dose was increased, or epidural steroid injections were programmed.
We used the BPI to evaluate changes in patients’ pain intensity and its interference with quality of life, given the clinical importance of pain as a primary patient-reported outcome in pain research14. This tool permits to follow the key recommendations for reporting pain research outcomes given its established content validity, reliability, and the ability to detect MCIC. In this study, we retrieved information about MCIC from the comparison between V2 and V1 of the BPI items’ mean score and their absolute difference. Following literature recommendations for achieving a prespecified MCIC in pain research 14,15, we have established that for this study the MCIC should include statistically significant differences of BPI mean-scores between V2 and V1 and that such differences will be ≥2 points for each BPI item.
BPI outcomes showed, both in the sample and in the study’s subsets, significant V2-V1 differences of the BPI-items’ mean-score, and these differences were of ≥2 points for each of the BPI item. These results thus fulfill the second and the third endpoints of the study.
Interestingly, although not significant, V2-V1 absolute differences of the BPI-items’ mean-score in the BUAS+/SLART+ subset were slightly higher than in the BUAS+/SLART- subset. In particular, in the BUAS+/SLART+, the BPI pain items (item 1 to 5) showed absolute differences of roughly 3 points or higher; for the QoL items the trend of higher absolute difference was less pronounce but still present. These results indirectly show that the treatment did affect the pain outcomes of the severer lumbar-RP condition, and thus confirming its presence as yielded by the BUAS test.
Finally, no significant associations were found between the PD-V2 outcomes and the independent predictors gender, age groups, and pain sites. These findings imply that the BUAS test was not compromised by patients’ demographic or clinical features and thus supports its reliability and content validity. Significant associations were found between the PD-V2 outcomes and the predictors SLRT-V1 and PD-V1, respectively. In particular, the presence of positive SLRT at V1 was associated with the uncertain PD-V2 outcome while its absence with the negative PD-V2 outcome. These findings imply that, as mentioned above, the co-presence of both signs may denote a severer lumbar-RP condition. In such a condition, the treatment was able to shift part of the positive PD-V1 outcome cases to the uncertain PD-V2 outcome, but not to a negative one like it could in the absence of positive SLRT. Similarly, positive and uncertain PD-V1 outcomes were associated, respectively, with uncertain and negative PD-V2 outcomes. Also, the MLR analysis confirms that for ‘negative’ SLRT-V1 cases, the risks for ‘positive’ or ‘uncertain’ PD-V2 outcomes to occur decrease versus the Reference outcome class PD-V2 ‘negative’. Albeit, the results mentioned above imply that in the presence of positive BUAS test, with or without positive SLRT, the given treatment (i.e., gabapentin) did affect the lumbar-RP condition.
Study limitations. The main limitation of the study is its observational and retrospective nature. The study’s ecological and external validity may be questioned as to the study design of the indirect assessment of the BUAS-test diagnostic ability. Indeed, using the effect of gabapentin on PD and BPI outcomes modifies the direct evaluation method. A direct evaluation would have implied, for example, a comparison of the BUAS test outcomes with invasive neurological tests or imaging studies whose outcomes are not always reliable and conclusive for lumbar-RP diagnosis and often yield ex-adiuvantibus  treatments. In this study, data were collected within a setting of routine work. The external validity of such observational study comes from its strong relevance to practice and the ability to highlight meaningful pain assessment solutions.