The main findings in the present study highlights that all patients could perform the experimental protocol both pre- and postoperatively. Furthermore, both EMG and NIRS methods were shown to be able to measure the paraspinal muscle activity and changes of oxygenation in response to isometric trunk extensions in patients pre- and postoperatively.
The dynamic trunk flexion-extension test was considered mainly as a warm up for the test subject. The isometric Ito test applied in our experimental protocol was easy to perform, safe and comfortable for both pre- and postoperative patients. Loading time was set to 60 seconds which is well within the endurance time values that Ito et al. reported for female (mean 70.1 seconds) and male (mean 85.1 seconds) with chronic low back pain (CLBP) (13). The majority of the patients were thus believed to be able to perform the test without reaching their endurance limit. However, this finding is not in line with the findings of Demoulin et al. (2008), who showed that the Ito test was less comfortable and more difficult to standardize. The Biering-Sorenser test is often suggested for measuring trunk extensor muscles, particularly in endurance tests when fatigue of the muscles is investigated. The Ito test share the same principals, but use greater trunk support and in doing so, become more user friendly in a cohort of LBP patients that might exhibit fear of movement.
Muscle activity as recorded by EMG and presented as RMS in all 4 measuring sites displayed an activity level within the same range as previous studies, using similar loading mode and protocols (15, 16). In general, a strong valid EMG recording was possible from all sensors, but highly dependent on a good reference position (C7) as well as its conductivity. The RMS values recorded varied, dependent on location of sensor, and independent of pre- or postoperative measuring session or Ito test. EMG sensor 1 and 2 were placed more lateral and at a lower spinal level than sensor 3 and 4, thereby recording different anatomical structures. The level of muscle activity is also specific to the loading mode as reported by Tucker K et al (17). The Ito test used in this study is viewed mainly as an upper body extension movement without any intentional lateral bending or rotation.
Muscle oxygenation responses in the paraspinal muscles showed a rapid decline during the Ito test and an increase toward the baseline value during recovery from the Ito test (Figure 8). These results indicate an acute imbalance between oxygen supply and oxygen demand in the working muscles during the Ito tests (18). Furthermore, increased intramuscular pressure during muscle contractions reduces muscle blood supply and oxygen delivery to the active muscles (19). The rapid increase in muscle oxygenation during the immediate post-exercise period counteracts the reduced oxygen supply due to muscle contractions. The muscle oxygenation trends observed in this study were similar for the pre- and postoperative patients across both right and left sides and our results are consistent with previous studies (9, 20). There were large variations shown for muscle oxygenation levels among patients because the level of change may reflect a variation in work intensity and may also be influenced by differences in the location of their pain.
Since the primary objective of the present study was to propose and validate a protocol suitable for evaluation of lumbar muscle functionality during isometric loading, the cohort was limited in terms of numbers of patients. This has an adverse effect on the possibility to draw any statistical significant conclusions from the material. However, trends or associations could be identified and the cohort can provide information for a future power analysis.
Left – right symmetry
Muscle activity as recorded by EMG and oxygenation are plausible properties that would be able to detect differences in left – right symmetry. No significant EMG difference was shown to occur in observed patients in neither the pre- nor postoperative measurements between left and right side of the lumbar spine in any load block. However, a weak trend was noted in the postoperative group at time event t=30, 40, 50 and 60 in test load block 2 and 3.
Pre – postoperative differences
In the present study, the time dependent behavior of the muscle activity as recorded by EMG exhibited pre- and postoperative differences. The postoperative EMG data displayed a more uniform expression since in all Ito tests the values decreased as a function of time. Preoperatively a more inconsistent behaviour was present, in test 1 the values decreased, but in test 2 and 3 the values increased as a function of time. This could speculatively be an expression of fatigue, namely that the RMS may increase over time. A permanent change in spinal muscular morphometry and substance composition has been observed in patients postoperatively. (21-24). MRI and other image generating tools seems to be the dominating means to evaluate such changes, but EMG offers the opportunity to evaluate the functionality of the muscles and particularly time and load dependent properties such as fatigue.
The moderate isometric load level reached in a Ito test may not be sufficient to initiate a marked fatigue development within 60 seconds and fatigue is therefore not as well detectable as it would with a load level in the vicinity of maximal voluntary contraction (MVC.) Changes in MF and RMS values of the EMG signal as a function of time during isometric loading has been previously suggested as being a reflection of muscle fatigue (4, 7). However, the observed phenomenon that MF decrease and RMS increase as a function of time appears not to be consistent in the literature. Plausible causes for this could be low load levels, short loading time events or that the investigated muscles exhibit an unknown behaviour, such as pato-physiological behaviour should not be ruled out.
In the present study the cohort where spinal patients diagnosed with spinal stenosis. They are subjected to a fairly low load compared to the MVC, but are still within the ranges that an individual may be subjected to during daily life, i.e 20-40 % of MVC (25). A notable difference between pre- and postoperative MF and RMS inter-relationship was present.
Normalization of data
In order to evaluate individual properties and compare between individuals a normalizing of data is proposed. The value at 5 seconds is used as a normalizing reference value for all parameters in order to determine a relative expression in percentage as a function of time. The main focus is thereby shifted to the time dependent properties of the parameters.
Sensors EMG 1 and EMG 2 were placed lateral to the NIRS sensors, this means that the muscle activity recorded was from a slightly different muscle bulk than the EMG 3 and EMG 4 sensors. A consistently lower RMS value was recorded for EMG 1 and EMG 2 as compared to EMG 3 and EMG 4 throughout the study, indicating a reduction in lateral muscle activity compared with the more centralized positioning of the sensors. This finding has previously been shown by others and is attributed to the anatomy of the erector spinae and multifidus muscle (26, 27). The EMG 1 and EMG 2 sensors appeared to be more consistent in picking up electrical environment noise than EMG 3 and EMG 4. Consequently, more of their recordings were rejected in the analysis. Based on these two major findings, it was elected to remove these from the final analysis and propose only use of EMG 3 and EMG 4 sensors in the future studies.
Thickness of the skin and subcutaneous tissue
The thickness of the skin and subcutaneous tissue ranged from 8.6 to 20 mm in the present study. We found no correlations between the thickness of the skin and subcutaneous tissue and MrSO2 values.
The location of EMG sensors were defined and dependent on where the NIRS sensors were placed with reference to the dorsal midline of the subject. Previous test-retest studies of EMG sensors placement yield a signal variation expressed as ICC on the order of 0.52 – 0.91 which is to be considered when comparisons between different test occasions are made (5, 28).
Electrical environmental noise was present mainly on the EMG 1 and EMG 2 sensors which were located lateral to the NIRS sensors and often occurred on slightly high BMI subjects where speculatively the muscle fat ratio is low and more susceptible to electrical interference, therefore creating a poor signal to noise ratio. Quality control of EMG were performed on data derived from the initial resting phase to assure a minimum of electrical noise superimposed on the true muscle signal. Data containing 50 Hz noise were omitted from the analysis as noted.