Paraspinal muscle oxygenation and mechanical efficiency are reduced in individuals with chronic low back pain

This study aimed to compare the systemic and local metabolic responses during a 5-min trunk extension exercise in individuals with chronic low back pain (CLBP) and in healthy individuals. Thirteen active participants with CLBP paired with 13 healthy participants performed a standardised 5-min trunk extension exercise on an isokinetic dynamometer set in continuous passive motion mode. During exercise, we used near-infrared spectroscopy to measure tissue oxygenation (TOI) and total haemoglobin-myoglobin (THb). We used a gas exchange analyser to measure breath-by-breath oxygen consumption (V̇O2) and carbon dioxide produced (V̇CO2). We also calculated mechanical efficiency. We assessed the intensity of low back pain sensation before and after exercise by using a visual analogue scale. In participants with CLBP, low back pain increased following exercise (+ 1.5 units; p < 0.001) and THb decreased during exercise (− 4.0 units; p = 0.043). Paraspinal muscle oxygenation (65.0 and 71.0%, respectively; p = 0.009) and mechanical efficiency (4.7 and 5.3%, respectively; p = 0.034) were both lower in participants with CLBP compared with healthy participants. The increase in pain sensation was related to the decrease in tissue oxygenation (R2 = − 0.420; p = 0.036). Decreases in total haemoglobin-myoglobin and mechanical efficiency could involve fatigability in exercise-soliciting paraspinal muscles and, therefore, exacerbate inabilities in daily life. Given the positive correlation between tissue oxygenation and exercise-induced pain exacerbation, muscle oxygenation may be related to persisting and crippling low back pain.


Population
Individuals with nonspecific CLBP were included in this study.They were invited by a physician to voluntarily participate in this study.To be included, they had to have had low back pain for at least 3 months.People with a body mass index over 25 kg m −2 or under 18.5 kg m −2 were not included in the study.Participants had to be physically active, in accordance with the World Health Organization 27 (i.e. at least 150 min of moderate-intensity aerobic physical activity throughout the week or at least 75 min of vigorous-intensity aerobic physical activity throughout the week or an equivalent combination of moderate-and vigorous-intensity activity).The level of physical activity was assessed in an interview, and then quantified using the Baecke questionnaire 28 .It assesses physical activity during work/occupational activities, during leisure time, during active commuting, and during sports activities.Each participant with CLBP was matched with a healthy volunteer in terms of age, weight, height, physical activity level and smoking status (based on the number of cigarettes consumed per).Any individual with a history of cardiovascular, metabolic, respiratory or neurological disease was excluded from the study.
The sample size was estimated with a paired t-test in the Sigmastat 3.5 software.We considered previously published results evaluating muscle oxygenation in individuals with CLBP and a control group.The values were 12.3 ± 3.0 and 9.3 ± 2.9, respectively 29 .The alpha level used was 0.05 and the fixed power level was 90%.Thus, the sample size for each group was estimated to be 12 participants.As the study protocol was spread over two visits, we anticipated that 10% of the participants would leave the study before the end of the protocol or withdraw their consent, so we aimed to include 14 participants in each group.

Procedures
The protocol used for the experiment consisted of two visits.During the first visit, the inclusion and exclusion criteria were checked by a physician and anthropometric measurements collected.Each participant performed the Sorensen test 30 to assess paraspinal muscle endurance.In this test, the participant is positioned prone on a table.The lower part of the body (below the iliac crest) is immobilised.The upper part of the body is out of the table.The test consists of keeping the upper body horizontal, aligned with the lower body.The test was stopped when the participant reached 150 s (corresponding to a higher value than the normative data) 31 .
The second visit was in the same week, at least 24 h after the first visit.During this visit, the participants performed trunk flexion and extension exercises on an isokinetic dynamometer (Con-trex TP-1000, CMV AG, Suisse).Each participant was placed on the dynamometer as described in the manufacturer's instructions: briefly, the participant was upright, with the knees slightly flexed and a popliteal pad directly behind the patella.The body was attached via a thigh pad, a tibial pad, a scapular pad and a pelvic belt.Once attached and upright, the trunk was tilted slightly forward or backward to define the anatomical zero position.This was defined as the position in which the participant felt neutral.The participant stood in this position during rest periods.
During the exercises, the axis of rotation was placed 3.5 cm below the top of the iliac crest.The range of motion was set at 70°, from − 5° extension to + 65° flexion.During each exercise, the participant was asked to perform trunk extensions, while flexions were passive (using the continuous passive motion mode).The velocity was set at 30° s −1 for passive flexion and 60° s −1 for trunk extension.After performing three maximal trunk extensions to assess peak torque (a relevant indicator of muscle strength) 32 , the participants were asked to perform a 5-min submaximal exercise session.The intensity was set at 80 Nm.Submaximal intensity dosage was via real-time visual feedback, which continuously showed the development of torque on a screen.The exercise duration and intensity were set at an absolute intensity achievable by all individuals in order to fully engage aerobic metabolism, reflected by a steady-state VȮ 2 during exercise.The absolute intensity was chosen to replicate an everyday lifting task.
During this exercise, the total work developed during trunk extension was calculated by the dynamometer.Before and after the exercise, a visual analogue pain scale was used to evaluate the intensity of pain sensation in the lower back (rated from 0 corresponding to 'No pain' to 10 corresponding to 'Worst pain imaginable').In addition, muscle oxygenation and the cardiorespiratory response were recorded and analysed using NIRS and a gas exchange analyser, respectively.

NIRS measurements
NIRS was used to assess paraspinal muscle oxygenation continuously as described previously 33 .This technique is based on the light absorption properties of oxy-haemoglobin/myoglobin (HbO 2 ), which absorbs light at 850 nm, and of deoxy-haemoglobin/myoglobin (HHb), which absorbs light at 760 nm.The NIRS device (Portamon Artinis, Zetten, the Netherlands) covered with plastic film to avoid sweat accumulation was used during exercise.The device was positioned vertically, parallel to the spine.The middle of the device was positioned at the level of the third lumbar vertebra.The device was positioned so that the light emitters and receiver were 3 cm from the spinous process of the vertebra.The distance between the light source and the receiver used was 40 mm.In accordance with the manufacturer's instructions and previous work 16,34 , the differential pathlength factor applied was 4. This factor takes into account the scattering of light within the tissues.
The values of HbO 2 and HHb were recorded for 2 min at rest, and then during the 5-min exercise.After recording, the HbO 2 and HHb values were used to calculate the total haemoglobin/myoglobin (THb = HbO 2 + HHb), which is a good reflection of the microvascular volume change.The values of HbO 2 , HHb and THb were kept for analysis.The values collected at rest were averaged to obtain HbO 2(0) , HHb (0) and THb (0) .All the data obtained during the exercise were normalised by the resting values to obtain ΔHbO 2 , ΔHHb and ΔTHb.
Normalisation of these data is necessary because they are influenced by individual factors (i.e.adipose tissue, skin perfusion, melanin contribution and the heterogeneity of blood flow in muscle) 35 .The tissue oxygenation index (TOI) was also analysed.The NIRS device calculates this index by utilising the spatially resolved spectroscopy technique.The TOI does not require normalisation with the rest value because it is expressed as a percentage, making it a good tool for inter-subject comparison 36,37 .ΔTOI (ΔTOI = TOI 5-min − TOI (0) ) was also calculated to determine a potential relationship between the change in the TOI and the change in the pain intensity (ΔPain = Pain post-ex − Pain pre-ex ) following exercise.

Pulmonary gas exchange measurements
A gas exchange analyser (Cortex Metamax 3B, Leipzig, Germany) coupled with a heart rate (HR) monitor (Polar Electro T31, Finland) were used to measure VȮ 2 , VĊO 2 , respiratory frequency (Fr), tidal volume (Vt) and HR.Before each measurement, the device was calibrated following the manufacturer's guidelines.After acquisition of the measurements, the minute ventilation (VĖ = Fr × Vt) and respiratory exchange ratio (RER = VĊO 2 /VȮ 2 ) were calculated, as well as the energy expenditure, according to the formula of Brouwer 38 : The mechanical efficiency was calculated as described by Moseley and Jeukendrup 38 : The V̇O2 values collected breath by breath were interpolated to produce one value per second.The onset kinetics was then calculated using a mono-exponential model, VȮ 2(t) = VȮ 2(0) + A(1 − e −(t−TD)/τ ), in which VȮ 2(t) represents the oxygen consumption at any time during exercise, VȮ 2(0) represents the oxygen consumption at rest, A represents the amplitude between VȮ 2(0) and VȮ 2 at steady state, t represent the time of exercise, TD represents the time delay, and τ represents the time constant (i.e. the time required to reach 63% of the steady state).The mean response time (corresponding to the sum of the delay and the time constant) and amplitude were collected.

Statistical analysis
The Sigmastat 3.5 software was used for statistical analysis.The data are expressed as the mean (standard deviation [SD]), except for ΔHbO 2 , ΔHHb, ΔTHb and TOI, which are expressed as the mean (standard error [SE]) because of the large SDs.A p value < 0.05 was considered to indicate a statistically significant difference.
The holding time during the Sorensen test, the peak torque, the mechanical efficiency and the VȮ 2 onset kinetics (i.e.amplitude and mean response time) were compared between groups using a paired t-test or a Wilcoxon signed-rank test (if the data did not follow a normal distribution).When the data had a normal distribution, the 95% confidence interval (CI) for the difference in means was calculated.The group effect sizes for the t-tests are described with Cohen's d, calculated as the difference in the means divided by the pooled SD.The effect size is considered small when d = 0.2, medium when d = 0.5 and large when d = 0.8.
Changes in pain sensation intensity, ΔHbO 2 , ΔHHb, ΔTHb, TOI and cardiorespiratory measures (i.e.VȮ 2 , VĊO 2 , HR, VĖ and RER) during exercise were compared using a two-way analysis of variance (group effect × time effect).When there were differences, a post hoc Bonferroni test was applied.Group effect sizes for analyses of variance are described using the eta squared (η 2 ) 39 , calculated as the sum of squares of an effect divided by the total sum of squares.The effect size is small when η 2 = 0.01, medium when η 2 = 0.06 and large when η 2 = 0.14.
The relationship between ΔPain and ΔTOI was analysed with Pearson correlation coefficients.

Participants
Fourteen individuals with CLBP were included and paired with healthy individuals, but one of them did not participate in all the testing sessions, resulting in the exclusion of both the individuals with CLBP and the paired healthy individual from the study.All participants were physically active in their leisure time (walking, weightlifting, swimming, cycling, motorcycling, fitness) or at work (forklift driver, roofer).Two Baecke questionnaires were not completed correctly; therefore, the questionnaire data reported in Table 1 includes 11 participants with CLBP and 11 healthy participants.One participant with CLBP smoked; therefore, he was paired with a healthy participant who smoked the same number of cigarettes per day (12 cigarettes per day).The anthropometric data are reported in Table 1.
For TOI, there were significant time (p < 0.001; η 2 = 0.12) and group (p = 0.009; η 2 = 0.20) effects, but the group × time interaction was not significant (p = 0.764; η 2 = 0.002).The pairwise comparisons showed that in both groups, the TOI was lower from the first minute until the last minute of exercise compared with resting values (p < 0.001).The TOI was lower in participants with CLBP compared with the healthy participants at rest (p = 0.042) and during exercise (first minute: p = 0.007; second minute: p = 0.008; third minute: p = 0.023; fourth minute: p = 0.012; fifth minute: p = 0.09) (Fig. 3).www.nature.com/scientificreports/For all the cardiorespiratory variables (VȮ 2 , VĊO 2 , HR, VĖ and RER), there was a time effect (p < 0.001) but not a group effect.The mechanical efficiency was lower in the participants with CLBP (p = 0.034; d = 0.61).There was no difference between the groups for VȮ 2 onset kinetics variables (Table 2).

Discussion
The aim of this study was to investigate aerobic metabolic adaptations in individuals with CLBP during a trunk extension exercise by measuring systemic and paraspinal muscle responses.For the first time, we have shown reduced mechanical efficiency, in addition to reduced muscle endurance and strength in individuals with CLBP compared with matched healthy individuals.In addition, the TOI was lower in participants with CLBP at rest and during exercise.Moreover, microvascular local blood volume (i.e.THb) decreased during the 5-min trunk extension exercise only in participants with CLBP.These results were independent of physical inactivity or sedentary lifestyle, as the participants were matched on the basis of their physical activity level.Some outcomes were not different between participants with CLBP and healthy participants, including cardiorespiratory parameters, corroborating the preliminary results discussed in the study by Vrana and colleagues 40 .They suggested that variability in paraspinal muscle metabolic responses to exercise may exist without variability in systemic responses.There was also no difference in VȮ 2 kinetics.Because VȮ 2 onset kinetics is mainly dependent on the oxidative capacity of the muscle 41 , our results suggest that it is not altered in the paraspinal muscles of individuals with CLBP.This is reinforced by the lack of difference in ΔHHb during exercise between groups, as it is an indicator of O 2 extraction 37 .

Alteration in mechanical efficiency and tissue oxygenation
As previously demonstrated, people with CLBP have reduced holding times during the Sorensen test 14 and reduced peak torques during trunk extension 42 , demonstrating poor low back muscle endurance and strength.In addition, we have shown for the first time that paraspinal muscle oxygenation is reduced in individuals with CLBP, even at rest.The lower TOI in the standing position may be associated with higher muscle O 2 consumption.This alteration could be related to functional limitations in usual activities, even in very-low intensity tasks.Table 2. Cardiorespiratory variables and muscle efficiency at the end of exercise and VȮ 2 onset kinetics.The data are presented as the mean ± standard deviation.The time effect is only reported for variables that were analysed minute by minute.The 95% confidence interval (CI) is only reported for variables that followed a normal distribution.www.nature.com/scientificreports/During exercise, TOI was even lower, and we found reduced mechanical efficiency.This shows an increase in oxygen cost to perform tasks.An increased metabolic cost may be associated with an impaired physical ability to perform tasks in individuals with CLBP.The combination of these outcomes may explain, at least in part, the reduced functional capacity to perform daily tasks.Moreover, the reduction in the TOI even at rest may be related to functional limitations in usual activities, even during very-low intensity tasks.
Several hypotheses may explain the decrease in mechanical efficiency and the TOI.First, these may be affected by muscle activation.Individuals with CLBP are known to have particular motor patterns, such as altered flexionrelaxation responses to trunk flexion and/or increased activation of the low back muscles while standing 43,44 .
In addition, the level of activation of co-agonist and antagonist muscles may differ 17 .These phenomena may be secondary to pain or kinesiophobia, and may alter muscle solicitations, and thus metabolic needs.The increased contribution of the low back muscles in the standing position could explain both the decrease in oxygenation at rest-denoted by the lower TOI-and the decrease in mechanical efficiency.
Second, mechanical efficiency may be affected by haemodynamic responses to exercise.On the one hand, previous work has shown that blood flow restriction can increase energy expenditure 45 , thereby reducing mechanical efficiency.On the other hand, previous studies have shown that contraction of low back muscles increases lower back intramuscular pressure, which can restrict blood flow due to blood vessel crushing 12,46 .In our study, microvascular blood volume restriction in participants with CLBP is supported by a reduction in ΔTHb during exercise.Impaired mechanical efficiency in participants with CLBP could be secondary to increased lower back intramuscular pressure.This view is supported by previous studies reporting lumbar compartment syndrome in individuals with CLBP during exercise 13 .This would contribute to reduced perfusion in the lower back and induce altered muscle oxygenation, as evidenced by the reduction in the TOI in the participants with CLBP, which is influenced by local blood flow 47 .Although exercise-induced intramuscular pressure may explain exerciseinduced low back pain, such a chronic phenomenon that persists even at rest has rarely been described, and it appears to be a very rare factor associated with persistent low back pain 48 .

Relationship between paraspinal pain and oxygenation during exercise
Our results suggest that impaired muscle oxygenation may be associated with low back pain.In response to exercise, ΔTOI may increase due to increased O 2 consumption, but it may also increase secondarily to blood flow restriction 47 , as suggested by the decrease in ΔTHb in this study.Previous work has indicated the association between blood flow restriction and substance P accumulation in an animal model 46 , and hypoperfusion and hypoxia have already been related to pain sensation in humans 49,50 .The potential relationship between pain and altered haemodynamics and/or oxygenation is supported by the negative correlation between ΔTOI and ΔPain, and it deserves further investigation.

Practical implications
People with CLBP showed reduced endurance-time, mechanical efficiency and muscle oxygenation, without cardiorespiratory alteration to exercise.The muscle oxygenation index was negatively correlated with the pain sensation intensity in the low back muscles.All of these adaptations suggest that physical disability could be associated with altered muscle metabolic adaptations to exercise.It may be necessary to stimulate muscle aerobic metabolism to reduce muscle fatigue and muscle pain.High-volume resistance training could be relevant to improve local aerobic adaptations.This approach could be more effective than traditional resistance exercise, traditional aerobic exercise or even a combination of the two 51 .To our knowledge, the effects of high-volume resistance training on muscle oxygenation have never been described.And, the benefits for chronic low back pain have never been investigated.

Study limitations
Our study has several strengths: we used multiple specific techniques simultaneously to assess the responses to exercise, we employed a standardised exercise and we compared the population with CLBP with a well-matched control group, notably concerning the physical activity level.However, some limitations must be underlined.First, because of the variability of the participants with CLBP-due to the multifactorial causes of the pathology-our results should be considered with caution.We did not consider psychosocial and/or behavioural factors in this study.Moreover, we did not assess central sensitisation, which is commonly associated with CLBP 52 .The skin fold thickness could not be indicated in this study because it was not correctly reported during the experiments.However, it is a factor to be carefully considered when performing NIRS measurements 37 .Furthermore, we did not assess muscle activity during the protocol, whereas specific motor patterns have often been associated with CLBP 8,17,44 .Motor patterns may influence paraspinal muscle involvement to exercise and then influence the oxygenation needs in paraspinal muscle to exercise.Finally, we did not perform muscle imaging; the composition of the low back muscles may differ between the groups and partly explain the differences revealed in this study.Further investigation is needed to conclude the causes involving the metabolic differences revealed in our results.Although we found significant differences with medium and large effect sizes, the lack of information about the minimum clinically important differences in mechanical efficiency and tissue oxygenation should be considered and could be the subject of future work.

Conclusions
We evaluated aerobic metabolic responses in the paraspinal muscles of people with CLBP.The results showed lower back weakness and reduced mechanical efficiency in participants with CLBP, even though they were physically active.We also found altered muscle oxygenation and reduced microvascular blood volume to exercise.These changes could be related to pain, and improving the aerobic metabolic response of the paraspinal muscles www.nature.com/scientificreports/ to exercise could be a way to reduce the muscle pain and fatigue that characterise individuals with CLBP.Further investigation is needed to explore the potential causal relationship between muscle metabolic responses and symptoms of CLBP.

Figure 1 .
Figure 1.Violin plots showing the holding time during the Sorensen test and peak torque.Bold lines represent the mean of each group; thin lines represent individual values.Significant group effect: * p < 0.05 ** p < 0.001.

Figure 2 .
Figure 2. Violin plots showing the intensity of pain sensation in the lower back before and after exercise.Bold lines represent the mean of each group; thin lines represent individual values.Significant group effect: * p < 0.05 ** p < 0.001.Significant time effect: † † p < 0.001.

Figure 3 .
Figure 3. Changes in oxy-haemoglobin, deoxy-haemoglobin, total haemoglobin and the tissue oxygenation index in low back muscles during exercise.The values are presented as the mean ± standard error.Significant group effect: * p < 0.05 ** p < 0.001.Different from rest: † p < 0.05 † † p < 0.001.

Table 1 .
holding time during the Sorensen test and peak torque developed during the maximal extension exercise were lower in participants with CLBP compared with healthy participants (Sorensen test: 87.23 ± 40.96 and Energy expenditure (J) = [(3.869×VO 2 ) + (1.195 × VCO 2 )] × (4.186/60) × 1000Mechanical efficiency (%) = Work rate J s −1 /Energy expended J s −1 × 100 Anthropometric characteristics.The data are presented as the mean ± standard deviation.There were no significant differences between the groups.