Methods

Settings and patients

This retrospective and observational study was held at the acute and CP center of Bologna’s Teaching Hospital, Italy. This is an anesthesiology-based pain program that provides outpatient consultation to primary care physicians and specialty services for inpatients. The sample for the present study, according to the inclusion/exclusion criteria (see below), includes n=258 consecutive non-cancer LBP outpatients.

Proceedings and instruments

Routinely, upon first visit (V1) and before the clinical examination, patients complete three questionnaires: Short Portable Mental Status Questionnaire (SPMSQ), the Brief Pain Inventory (BPI), and the painDETECT (PD) questionnaire; BPI and PD are administered to patients also upon follow-ups. Focused clinical history and physical examination for LBP patients include Straight Leg Raising Test (SLRT) and the BUAS test. The rationale, description, and reliability of the BUAS test for detecting lumbar-RP in LBP patients were earlier reported10. Briefly, “it consists of identifying, with the patient in the prone position, the Posterior Superior Iliac Spine (PSIS) with the examiner’s index or middle finger. Hence, with a full extended hand, the examiner’s thumb reaches the greater trochanter; keeping the index or middle finger on the PSIS, the thumb is moved towards halfway between the greater trochanter and the sacrococcygeal symphysis. The thumb thus exerts pressure (ie, the strain) against the underlying bony structures hitting the sciatic nerve. The test is considered positive if the patient reports pain exacerbation different from that of the mere pressure, or reproduction of his/her pain” 10.
In our practice, all patients, positive to the BUAS test or the SLRT, are prescribed (if no counterindications exist) with gabapentin 300 mg tablets. Our inhouse gabapentin-prescription protocol foresees one tablet at bedtime for three days, hence, one tablet bid (in the morning and at bedtime) for three days, and afterwards one tablet tid. Therapy outcomes are evaluated in the follow-up visits where gabapentin dose may be tapered or increased according to the patients’ clinical responses. All retrieved clinical and therapy information is stored in the patients’ chart and at the clinic’s database.
Inclusion criteria for this study were: outpatients with at least one follow-up visit after V1, ≥18 years of age, with chronic LBP (≥3 months), positive, at V1, for the BUAS test with or without positive SLRT, having had, at the end of V1, gabapentin-protocol prescription, and no pain procedures before the first follow-up visit (V2), and who signed informed consent. Exclusion criteria were: SPMSQ score <8, history or diagnosis of cancer or diabetes upon V1 or later, positive clinical signs for the piriformis syndrome, ongoing therapy, at V1, with anticonvulsants (gabapentin or pregabalin) or antidepressants, and finally, withdrawal from the gabapentin protocol, before V2, for severe side effects.

Demographic and clinical predictors

Demographic predictors are: (I) gender: male/female; (II) age groups [in order to avoid unbalanced over representation in wider age interval groups, patients over 35 years of age are divided into 15-year interval subsets: 36–50; 51–65; 66–80 and ≥81 years (classes B- E, respectively); the only subset having a 17-year interval is that of young adults ie 18–35 years of age (class A)].
Pain-related predictors were retrieved from the BPI. In the latter, a human body image allows the topographical location of the pain site. Pain site categories were: lumbar spine, lumbar spine and sciatica, lumbar spine associated with diffused pain, and lumbar spine associated with other specific pain sites (cervix, groin, dorsal spine, shoulder, and headache). The following five items assess, using a 0–10 numerical rating scale (NRS), the highest, lowest, and mean pain intensity in the past 24 hours, and the actual pain both under static and dynamic conditions. Item-6, estimates the percentage of pain relief with the ongoing pain therapy (0%, no relief – 100%, optimal relief). Finally, items 7-13 assess how much, in the past 24 hours, pain interferes with the patient’s QoL (activity in general, mood, ability to walk, work activity, social interaction, sleep, and the enjoyment of life; estimation is made using a 0–10 NRS (0 = no interference to 10 = maximum interference).
The PD is a clinician-administered and patient-reported screening questionnaire to reveal the likelihood of a neuropathic pain component in LBP patients; it consists of seven items that address neuropathic-pain symptoms’ quality with a final score between –1 and 38 (a score of ≤12 implies no neuropathic component, a score of ≥19 implies the presence of neuropathic component, while a score of 13-18 implies that the result is uncertain 11–13.
Finally, based on the results of the SLRT and BUAS test, the sample includes two subsets of patients: those who, at V1, were positive for the BUAS test and negative for the SLRT (BUAS+/SLART-), and those who were positive to both tests (BUAS+/SLART+).

Ethics

This study was approved and authorized by the Hospital Ethics Committee (235/2013/O/Oss). The study was conducted according to the Helsinki Declaration and the International Association for the study of Pain (IASP)’s guidelines for pain research in animals and humans. All participants were personally and thoroughly informed by the investigators on the aims of the study. Patients were informed that participation was voluntary, anonymous and would not affect their care; hence, an informed consent was obtained.

Study questions, data presentation and statistical analysis

Fig. 1 describes the flowchart of the study. We sought to quantify, at V2, the changes in PD outcomes and BPI scores, in patients who were, at V1, positive for the BUAS test, with or without being positive for the SLRT, and thus were prescribed with the gabapentin protocol to treat their potential lumbar-RP condition. We have hypothesized that, at V2, positive BUAS-test patients will report significant improvement in PD outcomes and BPI scores. We have established that significant improvement should include the following elements: over 50% of the sample will present, at V2, negative PD outcome, statistically significant differences of BPI mean-scores between V2 and V1, and that such differences will be ≥2 points for each BPI item. We have established these elements following literature recommendations for achieving a prespecified Minimal Clinically Important Change (MCIC) in pain research 14,15. We assumed that in case of affirmative results, these might be considered a piece ofex-adiuvantibus evidence for the capacity of the BUAS test to detect lumbar-RP in LBP patients. Finally, we planned to look for associations between patients’ independent variables and PD outcomes at V2 (PD-V2) in order to uncover variables that may undermine the capacity of the BUAS test to detect lumbar-RP.
Continuous data are reported as the mean (± SD, standard deviation); category data are expressed as absolute numbers and percentages. The dependence of PD-V2 outcomes upon independent variable categories was determined using χ2 analysis. Independent variables were gender, age group, pain localization, SLRT-V1, and BUAS-test-V1; when significant, a post-hoc cell contribution analysis was performed, and major contributions for the association were reported. Multinomial Logistic Regression (MLR) was used to classify subjects based on a set of predictor variables. Dependent variables for MLR were the PD-V2 outcome classes and the ‘negative’ class (the most numerous) was the reference outcome class. Independent variables were those used for the above mentioned χ2 analysis. Differences of BPI-items’ mean scores between V2 and V1 are reported as absolute difference (Δ) and 95% upper and lower confidential intervals (CI); t -test was used to compare the BPI-items’ scores between V1 and V2. Statistical significance was defined as p<0.05. When appropriate, p values were rounded to three decimals.