Introduction

Low back pain (LBP) is the condition causing most disability globally [1]. Only about 10% of all LBP problems are represented by disc herniation or spinal stenosis but they represent the most common conditions in spinal surgery [2, 3]. In these groups, the outcome is questioned since 20–35% is doubtful or dissatisfied with the results at one-year follow-up [3]. Challenges are to develop screening identifying those with increased risk of a poor prognosis and to select interventions based on prognostic factors for improvement of outcomes [2, 4]. Socio-demographic, clinical, work-related and psychological risk factors may partly explain poor outcomes after spinal surgery [5]. Another reported risk factor is insufficient pain treatment pre- and post-surgery [6]. It has been suggested that screening of somatosensory function may contribute to the understanding of pain mechanisms involved [7]. Somatosensory profiles can be assessed with quantitative sensory testing (QST), by measuring amongst others pressure-, cold- and heat pain thresholds (PPT, CPT and HPT). Lowered pain thresholds in QST measurement is described as an indication for central hyperexcitability [8]. This alteration in sensory thresholds can either be localized or widespread including a non-affected body regions [9].

Thermal thresholds in an affected body region were normalized in patients with spinal degenerative disease, who had recovered 6 months post-surgery, while they were not normalized in patients with remaining pain post-surgery [10]. Lowered PPT in non-affected body regions was associated with higher pre-surgery and 12 months post-surgery pain severity, but not with change in pain from pre-surgery to follow-up after hip- or knee replacement [11]. A previous analysis of the present cohort of patients with degenerative lumbar spine disorders selected for surgery showed that an altered sensory profile in both affected and non-affected body regions pre-surgery were associated with higher back- and leg pain intensity and lower mental health pre-surgery (Lindbäck Y et al. submitted manuscript 2016). However, prospective studies are needed to further examine whether these dimensions can be useful as a screening tool for outcome post-surgery. The purpose was to investigate the association between pain sensitivity in the hand pre-surgery, and patient-reported outcomes (PROs) in function, pain, and health pre- and post-surgery in patients with disc herniation or spinal stenosis. Our hypothesis was that patients more sensitive to pain in the hand, a sign of widespread altered pain response, will have higher pain, lower function, and health pre- and post-surgery.

Method

Study design

This is a prospective cohort study. The study conforms to the STROBE statement checklist. The study was approved by the Regional Ethics committee (Dnr 2013/410-31). The patients received oral and written information about the study and they all signed an informed consent before the measurement. Forty-seven patients (57%) were even included in Prepare, a study investigating the effect of pre-surgery physiotherapy (Clintrials.gov: NCT02454400).

Setting

The patients were consecutively recruited at the University Spine Clinic, Linkoping, Sweden, between September 2013 and December 2014. The somatosensory function was investigated with QST 1–2 weeks pre-surgery by one single investigator at the Spine Clinic. The patients filled out PROs pre- and 3 months post-surgery.

Participants

The inclusion criteria were: age 25–80 years; the presence of leg pain with or without back pain due to lumbar disc herniation or lumbar spinal stenosis, diagnosis confirmed by magnetic resonance imaging; failure of conservative intervention (without further specification) and pain level high enough to indicate surgical intervention (discectomy or decompression surgery with or without fusion).

Variables

A standardized QST protocol was used [12]. During QST, the patients were sitting comfortably in a quiet room with an air temperature of 22 °C. Cold and heat pain threshold (CPT and HPT) were measured by a thermic stimulator (Somedic, Hörby, Sverige). A thermode containing a peltier element was used with a stimulating area of 25 × 50 mm. The baseline temperature was + 32 °C and for the thermal measures the temperature decreased or increased with 1 °C/s within a range of +10 and + 50 °C. The thermode was held on the dominant hand on the thenar eminence muscle. The instruction was to push the stop button when the cold/heat sensation was perceived as painful.

PPT was measured with a handheld electrical pressure algometer (Somedic, Hörby, Sweden). The pressure was applied at a rate of 30 kPa/s, with a 1 cm diameter probe. The maximal pressure was 700 kPa and at that point the applied pressure was released. The patient was instructed to tell when the pressure on the hand started to become painful and at that point the applied pressure would release.

The dependent variables were the patient-reported measures pre- and post-surgery: Oswestry disability index (ODI) [13], back and leg pain intensity (VAS) [14], Health-related quality of life (EQ-5D) [15], Hospital Anxiety and Depression Scale (HADS) [16], and Self-Efficacy Scale (SES) [17]. The independent variables were PPT, CPT and HPT pre-surgery.

Statistical methods

Statistical analysis was performed using IBM SPSS statistics version 23. The level of significance was set to 0.05. Patient demographics were presented as mean and standard deviation (SD) or numbers and frequencies. For between-group comparisons the unpaired Student’s t test, Mann–Whitney U test or the Chi-square test were used. To assess normal distribution and outliers Kolmogorov–Smirnov, Shapiro–Wilk, skewness, kurtosis, and plots for observed value and expected or deviation from normal, respectively, was used for each variable. Some variables were deviant, but judged as sufficient in visual assessment. Linear regression was used to analyze the associations between pain thresholds (independent variables) and patient-reported measures (dependent variables). All models were adjusted for age and gender. To assure adequate sample size, each analysis had a cases-to-independent variables ratio of about ten cases for every independent variable in the model. Missing data were few, and imputation was not used for missing data.

Results

Descriptive data

Patients with spinal stenosis (n = 53) were older and had more frequently a pain duration of more than 1 year than patients with disc herniation (n = 29) (Table 1). There were no significant differences between the patient groups in gender distribution or in the pre- and post-surgery PROs (Table 1). On group level, there were no signs of anxiety or depression; on an individual level, there were eight patients with signs of anxiety and four patients with signs of depression according to HADS.

Table 1 Demographic characteristics and self-reported function, pain and health of patients with disc herniation (n = 29) and spinal stenosis (n = 53), respectively at pre- and post-surgery
Full size table

Lumbar disc herniation

PPT pre-surgery versus patient-reported measures pre- and post-surgery

In patients with disc herniation, a negative association existed between PPT and ODI pre-surgery (P = 0.006) (Table 2). Such negative association also existed between PPT and the two subscales of HADS pre-surgery (anxiety and depression) (P = 0.043 and P = < 0.001, respectively) and positively with SES pre-surgery (P = 0.017).

Table 2 Linear regression between pain thresholds (PPT, CPT and HPT) in the hand pre-surgery (independent variables) and self-reported function, pain and health pre- and post-surgery (dependent variables) in patients with disc herniation (n = 29)
Full size table

A negative association also existed between PPT pre-surgery and ODI post-surgery (P = 0.020) (Table 2). Furthermore, PPT pre-surgery was negatively associated with pain intensity in the leg post-surgery (P = 0.036) and positively with self-efficacy (SES) post-surgery (P ≤ 0.001).

CPT pre-surgery versus patient-reported measures pre- and post-surgery

In patients with disc herniation there was a positive association between CPT pre-surgery and ODI pre-surgery (P = 0.017) (Table 2). Furthermore, there were also positive correlations between CPT pre-surgery and ODI post-surgery (P = 0.002) and higher back pain and leg pain in VAS post-surgery (P = 0.001 and P = 0.009, respectively). Furthermore, a negative association between CPT pre-surgery and self-efficacy (SES) post-surgery (P < 0.001) also existed.

HPT pre-surgery versus patient-reported measures pre- and post-surgery

There were no associations between HPT pre-surgery and patient-reported measures pre- and post-surgery (Table 2).

Lumbar spinal stenosis

PPT pre-surgery versus patient-reported measures pre- and post-surgery

In patients with spinal stenosis, there were no significant associations between PPT and the pre-surgery patient-reported measures (Table 3). A negative association existed between PPT pre-surgery and HADS (anxiety) post-surgery (P = 0.001).

Table 3 Linear regression between pain thresholds (PPT, CPT and HPT) in the hand pre-surgery (independent variables) and self-reported function, pain and health pre- and post-surgery (dependent variables) in patients with spinal stenosis (n = 53)
Full size table

CPT pre-surgery versus patient-reported measures pre- and post-surgery

There were positive correlations between CPT pre-surgery, and HADS (anxiety) (P = 0.014) pre- and also post-surgery (P = 0.025) (Table 3).

HPT pre-surgery versus patient-reported measure pre- and post-surgery

There were no associations between HPT pre-surgery and patient-reported measures pre- nor post-surgery (Table 3).

Discussion

The results showed that patients with disc herniation in lumbar spine who were more sensitive to pressure pain in the hand pre-surgery had lower function, more signs of anxiety and depression, and lower self-efficacy pre-surgery. Prospectively, more sensitivity to pressure pain pre-surgery was associated with lower function, higher pain intensity and lower self-efficacy post-surgery. The pattern of associations was very similar for cold pain in this group of patients. Hence, our hypothesis that patients with sign of widespread altered pain response would have worse PROs pre- and post-surgery was mainly confirmed in patients with disc herniation. However, it was not confirmed in patients with spinal stenosis, since the only associations found were that patients who were more sensitive to cold pain had more signs of anxiety pre- and post-surgery, and those more sensitive to pressure pain had more signs of anxiety post-surgery. Another interesting result was that there were no associations between HPT and patient-reported measures in either of the diagnostic groups.

In a review of cross-sectional studies of patients with heterogeneous diagnoses of spinal pain, it was concluded that pain threshold had little or no correlation with disability and pain intensity [18]. In comparison, our study included homogenous groups of patients with specific diagnosis who also went through a decision making process for surgery. Our study is the first to report prospective results showing associations between high sensitivity to pressure and cold in a non-affected body region pre-surgery and lower function, higher pain intensity and lower self-efficacy post-surgery in patients with LBP. However, the results were only found in the disc herniation group. A possible reason for the different results for patients with disc herniation may be the presence of inflammation, which often causes more on-going nociceptive stimuli and constant pain experience that could cause more risk of plasticity changes in the nervous system [6]. In disc herniation, the sciatic pain is described to have both a neuropathic component with the mechanical-stimuli causing the nerve root compression and also an inflammatory component [19].

In the patients with disc herniation, PPT on the hand, as a screening tool for signs of widespread altered pain response in pain threshold, had associations with the PROs. PPT is supposed to measure deep pain and muscle sensitivity [20] and pressure pain hyperalgesia is sign of peripheral sensitization of C-fibers in tissue injury [20]. It has not yet been fully demonstrated that PPT captures central pain processing. Although in single studies hyperalgesia to pressure pain in a non-affected body region has been reported in non-specific LBP [21] and after hip- and knee replacement concluding that it might indicate central pain processing [11, 21].

PPT and CPT both measure A-delta and C-afferents, but with mechanical, respectively, thermal stimuli [7]. In this study, CPT had no associations with psychological factors pre-surgery, while lowered PPT was associated with more anxiety- and depression signs and lower self-efficacy. Further, more sensitivity to cold pain had an association with higher back pain intensity post-surgery; where PPT also had association to pain intensity but to leg pain. Despite these smaller differences in PPT and CPT associations pre-surgery, this study showed that PPT and CPT mainly had similar associations with the post-surgery outcomes. PPT is the QST measure described as most sensitive for pain excitability in patients with osteoarthritis [22] as well as in chronic LBP [23]. While CPT in patients with whiplash-associated disorders (WAD) had moderate evidence to be a predictor for outcomes in pain and disability [24]. Our result support to further analyze PPT and CPT also in degenerative lumbar spine disorders scheduled for surgery, when more sensitivity for pressure- and cold pain in a non-affected body region pre-surgery were associated with lower function, higher pain intensity and lower self-efficacy post-surgery.

The current study has its strengths and weaknesses. QST results can be influenced by psychological factors [12]. Our material was too small to adjust the regressions for possible confounding factors, e.g., psychological factors. Further, temporal summation or other dynamic QST could have been of interest, since these have been recommended to assess pain experience [7]. The difference in age between patients with disc herniation and spinal stenosis does probably not explain the differences in results since a strength of our study is that regression model was adjusted for age. Moreover, pain thresholds do not change with age [25]. Another strength of our study is that the regressions were adjusted for gender to handle the well-known sex differences in pain thresholds [20].

The results from this study should be interpreted with some caution since the study samples were relatively small. Despite this, we have compared the profile on patient-reported characteristics for the groups with the Swedish national spine register [3] and they have similar characteristics suggesting that results can be generalized to the larger Swedish population and potentially international populations with a similar health system context. Further larger prospective studies are needed to confirm our findings and be able to perform multivariate analyses. Prospective studies with validated outcomes were recently requested in a QST review concerning LBP [8]. With the associations found in this study, measuring PPT hand, a non-affected body region pre-surgery, might be a screening tool potentially capable of identifying patients with suboptimal functional outcome, e.g., ODI.

Barriers to use QST in clinical practice are that the measurements take long time and advanced equipment is required. Development of less time consuming QST protocols are needed [7]. The handheld electrical pressure algometer used for PPT measurement is more or less standard equipment and more feasible in Spine Clinics. To investigate only a non-affected body region for screening of widespread altered pain response can help to overcome barriers for clinical use and collection of larger patients’ groups is needed to further analyze the value of PPT.

Conclusion

Measuring pressure pain threshold in a non-affected body region, as a sign for widespread altered pain response, showed that patients with lumbar disc herniation who were more sensitive to pressure pain in the hand pre-surgery had lower function, more signs of anxiety and depression and lower self-efficacy pre-surgery. Prospectively, more sensitivity to pressure pain pre-surgery was associated with lower function, higher pain intensity and lower self-efficacy post-surgery. The results for cold pain were similar except there were no associations with psychological variables pre-surgery. In patients with spinal stenosis, there was lack of associations. Based on the associations post-surgery, further analyses of PPT in the hand as a sign of widespread altered pain response pre-surgery in patients with disc herniation are suggested.