Test-retest reliability of nerve and muscle morphometric characteristics utilizing ultrasound imaging in individuals with unilateral sciatica and controls

Fifteen patients, aged 30–50 years with the complaint of LBP with unilateral radiculopathy (sciatica) lasting for a minimum of three consecutive months participated voluntarily in this study (Table 1). Diagnosis was made by a neurosurgeon based on the criteria recommended by Nijs et al. (2015) (LBP with at least one symptom of pain, numbness, or tingling radiating down to the leg and / or foot) confirmed by associated disc bulging or herniation and nerve root compression between the vertebral level at L4-L5 and L5-S1 on MRI [1]. In addition, 15 asymptomatic individuals were recruited from the local population as the control group with no history of LBP during the preceding 6 months, or dysfunction in the low back, thoracic, pelvis, or lower extremities. The control group was matched with the patient group regarding age and gender. Patients completed the Persian version of the Oswestry Disability Index (ODI) [16], Tegner Activity Scale [17] and the Numerical Pain Rating Scale (NPRS) [18]. Tegner Activity Scale is a graduated list of activities of daily living, recreation, and competitive sports; and the score scales from 0 to 10 where 0 represents sick leave or disability pension, and 10 is participation in competitive sports [17]. Patients were excluded if they had tumor / malignancy or bony defects in the lumbar region on MRI, systemic myopathy/neuropathy, previous surgery in the region of assessment, or evidence of central sensitization in the mechanism of pain [1]. All participants signed a consent form and the study was approved by the Human Ethics Committee at Tehran University of Medical Sciences, Tehran, Iran. The data were collected from March to July 2017.

A diagnostic ultrasound imaging unit set in B-mode (Affinity 50 Philips-Netherland) with a linear-array probe with 7–12 MHz band frequency was used to record the images. The gain was set at 48% of the range, dynamic range was maintained at 93 dB, and time compensation was kept at the same (neutral) position for all images’ depths. The depth setting was adjusted for each muscle to visualize their superior and inferior margins. Images were recorded as JPEG files and stored on a computer for later processing. ImageJ software (Version 1.48v, National Institutes of Health, Bethesda, MD, USA) was utilized to calculate the muscle thickness and echo intensity as well as the sciatic nerve CSA and echo intensity.

Intrarater within-day reliability of ultrasound measurements of contraction ratio and echo intensity was assessed in all the participants. First, the examiner performed all the measurements, and then, repeated the measurements after 60 min in a random order (condition names picked from a bowl) with the same procedure. To assess the intrarater reliability, three ultrasound images were acquired in the transverse and longitudinal views of four muscles, i.e. multifidus, biceps femoris, medial gastrocnemius, soleus. Between each scan, the probe was moved away from the nerve and muscle and then placed back again over the same area of the nerve and muscle for the next scan. All images were recorded at the end of expiration and they were captured from two sides within one session in the transverse and longitudinal views. Muscle thickness was measured as the largest distance between the superficial and deep fasciae, identified by their hyperechoic appearance in the longitudinal scan (Fig. 1). Two different ROIs were selected from the rest position to measure the echo intensity. First, maximum ROI was drawn for each scan to include as much of the muscle as possible, avoiding bone and surrounding fasciae in two scan views (Figs. 2 and 3). Second, to calculate max rectangular ROI, a rectangular ROI (as large as possible) was positioned over the inner region of muscle image in two scan views (Figs. 4 and 5). Echo intensity was then defined as the mean level of gray within the ROI in 8-bit resolution images (gray levels from 0 to 255, where black = 0 and white = 255) [7].

Direct and indirect methods were used to measure the CSA of the sciatic nerve. In the direct method, the inner border of the perineal echogenic rim that surrounds the hypoechoic sciatic nerve was traced (Fig. 6). The indirect method employed the formula for calculating an ellipsoid area (major diameter × minor diameter × 3.14/4) (Fig. 7) [19]. The major diameter is defined as the longest line between two points of the nerve that passes through the center, whereas the minor diameter is the line through the center of the nerve perpendicular to the major diameter. All images were saved and exported for further analysis and the examiner was blind to the images’ group allocation during processing the images by ImageJ software.

Data were presented as means ± standard deviations (SDs). All statistical analyses were performed with SPSS statistical software version 24 (IBM Corporation, Chicago, IL, USA) software package. Intra-session reliability for the average of three measures in contraction ratio and echo intensity was assessed by the intraclass correlation coefficient (ICC 3, 1; method: alpha, two-way mixed, consistency). The ICCs are classified as follow: < 0.69, poor correlation; 0.70–0.79, fair correlation; 0.80–0.89 good correlation; 0.90–1.00 high correlation [23]. The standard error of measurement (SEM) and minimal detectable change (MDC) were also calculated to make a judgment about the degree that measurements vary for an individual. The SEM values indicate the precision of the measurement and were calculated based on the ICC and the SD of the mean of differences between the two measurements [SEM = SD√1 - ICC]. The MDC represents the smallest change in a score within an individual that can be considered as a real change above measurement error, which was determined using the formula: [MDC = 1.96 × √2 × SEM] [24].

To the best of our knowledge, this is the first study to evaluate test-retest reliability contraction ratio of the lower back and lower limb muscles affected by sciatica. Measuring muscle thickness using rehabilitative ultrasound imaging may be useful as an indicator of muscle function or neuromuscular motor control in the assessment of sciatica due to lumbar disc herniation. The contraction ratio had high reliability and may be used to assess the alterations of muscles morphology and function in patients with sciatica.

Multifidus activation by the abdominal hollowing and contralateral arm rising maneuver has been studied in individuals with and without LBP [29‐32]. However, these studies have not used the prone hip extension maneuver to assess multifidus activation. Thus, this method of activation can be employed to contract multifidus in clinical or research settings.

The contraction ratio of measured muscles ranged from 0.87 to 1.27 with large variability. It may be due to the variability of motor control or muscle recruitment patterns in the participants. For example, reduced multifidus muscle contraction ratio on the affected side may result from pain or reflex inhibition due to the nerve root compression. It is also plausible that reduced soleus contraction ratio was a result of individual differences in preferential motor activation and/or co-contraction muscle synergy or antagonist into the muscle chain [33]. Reliability of architectural properties of the medial gastro-soleus had been widely studied previously [34]; however, there is no report regarding the reliability of the muscle contraction ratio of these muscles.

Ruas et al. (2017) also reported utilizing ultrasound imaging as a highly reliable method for the measuring of hamstrings muscle thickness [26]. Individuals may develop a protective motor strategy and adaptation (e.g., co-contraction of agonist-antagonist muscles, or increased activity of synergistic muscles) to pain provocation movement [35]. Due to the high reliability of the muscle contraction ratio, it can be considered as a potential outcome measure when assessing neuromotor function in clinical practice and in research. All the SEMs and the corresponding MDCs were low; thus making the contraction ratio a promising measure for future studies.

The reliability aspect of measuring CSA and echo intensity of the sciatic nerve with ultrasound in patients with LBP with unilateral radiculopathy (sciatica) has received little attention in the literature. This study showed that the sciatic nerve CSA had good to high test-retest reliability, while echo intensity had moderate reliability. The sciatic nerve was not always ellipsoid, therefore the tracing method was a slightly more reliable method than the ellipse method, which is comparable to the previous finding on the measurement of tibial nerve CSA [25]. Therefore, we suggest the direct method to determine the CSA of the sciatic nerve in this population. To our knowledge, little is known about the magnitude of SEM or MDC of the sciatic nerve in these patients. The SEM and MDC of CSA by tracing method were lower than ellipsoid in the controls. A similar trend was observed in the patient group. In line with our findings, two previous studies had reported acceptable reliability when measuring the sciatic nerve CSA [14, 27]. We found moderate reliability for the nerve echo intensity. However, no study has investigated the reliability of the echo intensity of the sciatic nerve in this population. Variability in the MDC measurements of sciatic nerve echo intensity may be due to the echogenic properties of surrounding tissues [13].

Muscle echo intensity values might offer important insight into the muscle changes caused by disease / pathological disorders [36], especially because it is more objective and possibly more reliable than a simple visual assessment of ultrasound images [37]. We observed moderate reliability of echo intensity using max rectangular ROI in the transverse scan. This finding is in line with Varanoske et al. (2017), who stated echo intensity may be heterogeneous when examining a portion of individual muscles in the transverse plane [38].

One of our aims was to determine the effect of ROI size on the reliability of ultrasound muscle echo intensity measures. Caresio et al. (2015) reported ICCs of 0.54–0.86 for within-session tibialis anterior and gastrocnemius muscle echo intensity depending on ROI size, with larger ROIs being associated with higher reliability [7]. In our study, using max rectangular ROI displayed lower SEM values for the longitudinal scan but larger SEM values for the transverse scan. Our results also showed that echo intensity reliability was generally lower than the contraction ratio nearly in all the assessed muscles. Lower echo intensity reliability can be due to variability in the US probe placement and potential image-to-image differences in background brightness, which may in return affect the ultrasound absorption and reflection of echo signals [39].

One of the limitations of this study was calculating only intrarater within-day reliability. However, we decided to measure within-day reliability because changes in the hydration level, posture and muscle relaxation between the sessions may affect between-day reliability. Second, this study was performed in the middle-aged population with 1–3 activity level, which limits generalizability to other age groups with higher activity level (e.g., athletes). Another potential limitation was having only one rater, so consistency across multiple raters is unknown.