Hofste et al (2021)

Characteristics of Lumbar multifidus muscle in subgroups with low back pain

Lower range of motion and muscle thickness is observed in people with low back pain when they are compared to healthy controls.

The greatest difference in muscle thickness was seen between the healthy controls and the subacute low back pain group, suggesting atrophy develops in the first 12 weeks of a back pain episode.

Chronic low back pain patients had higher disability compared to subacute low back pain patients.


The lumbar multifidus muscle is frequently studied in relation to low back pain and reductions in cross-sectional area have been found in patients with chronic low back pain. Yet, these observations took place in laboratory settings in small, homogeneous populations, which limits the generalizability to other patients with low back pain. Therefore, this study aimed to examine the functional and morphological features of the lumbar multifidus muscles in primary care and to compare these muscle characteristics in low back pain of different durations and in healthy controls.



This multicentric cross-sectional study used a case-control design. Patients were recruited from a Dutch “Spine Network” encompassing more than 100 physiotherapists. Eligible patients were between 18 and 65 years and had non-specific low back pain. Serious pathologies, radicular syndrome, previous back surgery, pregnancy, psychiatric disorders, and body mass index above 30 were exclusion criteria applied. Healthy controls (without low back pain in the previous 6 months) were recruited through social networks.

Eligible patients from which informed consent was obtained were referred to one of four physiotherapy practices where protocol-trained physiotherapists performed the measurements. Muscle characteristics in low back pain that were studied include participants’ muscle function (assessed with surface EMG), morphology of the lumbar multifidus (evaluated through ultrasound), and the function of their low back (measured with a 3D kinematic device).

Surface EMG of the lumbar multifidus was obtained when the participant performed the Biering Sorensen test to assess the isometric muscle endurance. For this test, the participant is lying prone on the examination table with only the lower extremity strapped on the bench. The test is performed with the participant brought back in a horizontal position without upper extremity support. This position had to be held for 60 seconds.

With the ultrasound measurement, the left and right lumbar multifidus muscles were assessed at rest and with submaximal contraction. A pillow to reduce lumbar lordosis was placed under the participants’ belly and submaximal contraction was achieved with the left and right contralateral arm lift test for 15 seconds.

Three dimensional kinematics were examined using an inertial measurement tool placed at the thoracolumbar junction. Participants were asked to maximally flex and extend the spine without bending the knees and without moving the hips respectively. Maximal lateral flexion to both sides was performed as well.

Secondary outcomes were personal characteristics, body mass index, pain intensity (numeric rating scale), and disability (Oswestry Disability Index). Disability was rated from 0-100 where 0-20 refers to minimal limitations, 21-40 to moderate limitations, 41-60 to obvious limitations, 61-80 to most limitations and 81-100 to bedridden patients.



A total of 161 participants were included. The healthy control group included 50 participants. 52 participants had subacute low back pain and 59 had chronic low back pain. On average, the participants in the control group had significantly lower body weight compared to both low back pain groups. The low back pain groups had similar levels of pain, but the chronic low back pain group had higher disability scores than the subacute group.

The range of motion of the trunk was larger in all directions in the healthy controls compared to all participants with low back pain, except for right lateral flexion. Healthy controls had thicker lumbar multifidus muscles except in the relaxed condition on the right side. The lumbar multifidus muscles were thickest in the healthy controls and lowest in the chronic low back pain group. Regarding the endurance of the lumbar multifidus muscles, surface EMG revealed no differences between the groups. Yet the data revealed significant heterogeneity, so caution is required.


Schermafbeelding 2021 09 28 om 22.14.45
From: Hofste et al (2021)


Questions and thoughts

Groups differed significantly on all participant characteristic outcomes, except for body height. This means that these groups were not comparable at baseline. However, analyses of the relation of gender, age, and weight as confounders showed only little impact on the primary outcomes.

No group differences were found regarding muscle endurance, measured by surface EMG. However, only 130 participants were included in the analysis since 21 failed to reach the 60-second hold in the Biering Sorensen test. Interestingly, only 1 out of 50 participants in the healthy controls failed, compared to 13 out of 52 in the subacute and 17 out of 59 in the chronic low back pain group. This seems to indicate that there may have been found a difference in muscle endurance between groups, if the endurance test would have been of shorter duration, enabling all participants to complete the test. However, as it is an endurance test, a sufficient amount of time should be examined and the authors based the 60-second hold on a paper that showed that patients at high risk of complaints have an endurance of fewer than 58 seconds on average.>

The largest decrease in muscle thickness was found in the first 12 weeks of a low back pain episode (difference between healthy controls and subacute low back pain group). The authors think that this can be explained by disuse of the lumbar multifidus with an episode of low back pain. This atrophy in the first episode of low back pain has been found in other studies as well. We think it may be interesting to target these muscles in the prevention of transition to chronicity, yet no proof for this thought is available.


Talk nerdy to me

Good aspects of this study include the use of validated devices (Gyko 3D kinematics) and procedures (surface EMG) in standardized conditions. Only 5 data points were missing and these were imputed using the Monte Carlo Markov chain method. The relationship between gender, weight, and age with the primary outcomes was assessed as potential confounders. However, we think that there may have been more possible confounding variables. Think of leisure or occupational physical activity levels, for example.

Some aspects endanger the conclusions. The authors state that a sample size calculation was not possible. Instead, they used a “generic calculation” and included 50 participants per group. They refer to an article to support their procedure; however, upon review, it appears that this article examines sample size calculation in a completely different domain (pediatric neuropsychology). As a result, the assumption of 50 subjects per group does not appear to be supported by any scientific source within the domain under investigation and therefore this is a potential limitation.


Take home messages

This study examining muscle characteristics in low back pain found patients with low back pain had less range of motion and lower thickness of the lumbar multifidus muscles compared to healthy controls. Flexion range of motion was reduced by 15° and extension and lateral flexion range of motion was reduced by 5°. The difference in muscle thickness was approximately 1cm, which is nearly one-third of the thickness of the muscle in healthy controls. Patients with chronic low back pain had more disability compared to patients with subacute low back pain.



Hofste, A., Soer, R., Groen, G. J., van der Palen, J., Geerdink, F. J., Oosterveld, F. G., … & Hermens, H. (2021). Functional and morphological lumbar multifidus characteristics in subgroups with low back pain in primary care. Musculoskeletal Science and Practice55, 102429.


5 absolutely crucial lessons you won’t learn at university that will improve your care for patients with low back pain immediately without paying a single cent


back pain course
Download our FREE app