The transverse abdominal muscle is excessively active during active straight leg raising in pregnancy-related posterior pelvic girdle pain: an observational study

Low back pain (LBP) and pelvic girdle pain (PGP) or a combination of both, lumbopelvic pain (LPP) are common during pregnancy. Although a large proportion of women recover within one month after delivery, a substantial percentage (5–8.5%) has persisting complaints even up to 2 years after delivery [1]. Many studies suggest that impairment of motor control is the mechanical component of the pathogenesis of painful disorders in the lumbo-sacral region; [2] however, that theory is still unproven and the recommendations for intervention remain questionable.

The theory of compromised motor control to explain the development of LPP proposes that many activities create (small) movements in the lumbopelvic region, which subsequently induce strain on the ligaments and pain. For PGP this comprises the sacroiliac joints and the pubic symphysis, and the ligaments in the pelvic ring. Biomechanical and anatomical studies have shown that transversely-oriented muscles of the abdominal wall, especially the transverse abdominal muscle (TrA), in co-contraction with the pelvic floor, are the most suitable muscles to achieve compression of both innominate bones against the sacrum (‘force closure’ of the sacroiliac joints) and could, theoretically, reduce movement in the sacroiliac joints and the strain on the engaged ligaments [3‐6]. The theory implies that the more efficiently individuals contract those muscles, the better they protect themselves against the strain on the ligaments in the pelvic ring, resulting in a lower risk to develop PGP and a greater chance to recover from it.

The theory would gain strength if the contraction of TrA during various tasks was shown to be smaller in patients with painful conditions than in healthy controls. However, the results of case-control studies in LPP are not consistent. A review from 2009 included three case-control studies that support the theory [7]. However, some studies published after that review (e.g. Beazell et al., 2011; Himes et al., 2012; Pinto et al., 2011) found no significant difference in TrA contraction between patients with LPP and controls [8‐11]. Some limitations of previous studies include: heterogeneity of the study group with respect to gender, severity, duration and cause of LPP. Moreover, bias can be introduced if participants in the control group are not well defined (e.g. ‘convenient samples’, ‘colleagues’, etc.) due to differences in level of fitness (e.g. sports, work, absence from work, unemployment, fatigue, etc.). In addition, colleagues might have detailed knowledge of the objective of the study and this knowledge could influence ‘spontaneous, automatic recruitment’. Furthermore, in most studies, maximal voluntary muscle contraction was investigated instead of spontaneous, automatic recruitment of muscles during a well-defined task.

As a consequence, the recommendations for intervention are not unequivocal. For example, Teyhen et al. showed that contraction of the TrA during active straight leg raising (ASLR) is reduced in a well-defined subgroup of patients with LPP [12]. Therefore, they suggested that ‘patients with unilateral lumbopelvic pain who have a positive ASLR test may benefit from motor control exercises that specifically target activation of the deep abdominal musculature’ [12]. O’Sullivan and Beales suggested that motor control impairments in long-lasting PGP show a large variation: ‘non-specific’ PGP disorders are represented by a number of sub-groups with different underlying pain mechanisms rather than a single entity.’ [13, 14] These authors recommend adapting the therapeutic intervention based on this sub-classification. Nevertheless, in physiotherapy for PGP, contraction of the TrA is emphasised, implying that the role of the muscle to compress both innominate bones against the sacrum is diminished. In contrast, our clinical experience is that contraction of those muscles appears to be excessive in PGP.

The contraction pattern of many muscles has been investigated in various subgroups of patients with LPP. In view of these earlier findings, it is important to establish the spontaneous activity of the muscles that provide a compressive force during a load transfer task in patients with long-lasting PGP. More specifically, during ASLR, the present study examines TrA thickness increase in a) women with long-lasting pregnancy-related posterior PGP compared with that in b) women from the same age group who had delivered at least once and have no current PGP (control group).

A cross-sectional observational study was performed to address the research question. During ASLR, measurement of TrA thickness was performed by means of rehabilitative ultrasound imaging (RUSI). Increase of TrA thickness was regarded as contraction of the muscles, and decrease of thickness as a diminished contraction or relative relaxation. Thickness increase of the TrA during ASLR was computed with respect to rest and expressed in percentages of thickness at rest. Results of participants with PGP were compared with those of parous women without PGP. All participants were selected in a private orthopaedic clinic specialised in long-lasting pregnancy-related PGP.

After applying the first three inclusion criteria, 77 women with pregnancy-related posterior pelvic pain of more than 6 months duration were selected. Of these, 31 could not be included due to a negative ASLR test (12 times) and/or a negative P4 test (18 times) and/or the absence of a positive effect of a pelvic belt on ASLR (15 times). In addition, two women were excluded because of ‘post-surgery bony abnormalities’ and one woman because her last delivery was less than 6 months ago. Thus, 43 participants were available for analysis. The characteristics of the 43 patients and 39 controls are presented in Table 1. On average, patients with PGP were about 4 years younger than controls, and duration since postpartum was also shorter by about 4 years. Of the 43 women with PGP, the pain started during pregnancy in 36 (84%) and within 3 weeks after delivery in 7 (16%). Posterior pelvic pain was felt at both sides in 31 patients (72%) and was unilateral in the remaining 28%.

Intra-rater reproducibility of TrA thickness measurements expressed as ICC_2,1 for the three conditions ranged from 0.80–0.88 in patients with PGP and from 0.79–0.83 in controls; all these values could be labelled as ‘excellent’.

There was no significant difference in mean TrA thickness at rest between women with and without PGP (Table 2). In PGP, the median increase of the right TrA during ipsilateral ASLR and contralateral ASLR was 31% for both sides; this increase was significantly more than the 11% and 13%, respectively, of the controls (p < 0.001 and p = 0.005, respectively).

The relation between TrA thickness increase during ASLR and possible confounders was checked per group per side (ipsilateral/contralateral) for all the variables listed in Table 1. This revealed no significant correlations (analyses are shown in the Additional file 1: Table S1).

The main finding of the present study is that an excessive contraction of the TrA is present during ASLR in patients with long-lasting pregnancy-related posterior PGP. These results were independent of the duration or severity of their complaints, or their score on ASLR.

Strengths of the present study are: i) the homogeneity of the study group for gender and the cause of pain, ii) the size of the study group and the method used to study the spontaneous automatic muscle recruitment in response to a load transfer task, and iii) the use of a well-defined control group.

Although two earlier studies on motor control in pregnancy-related PGP also showed enlarged muscle activity during ASLR, TrA activity was not measured. One of these studies, investigating pregnant women with PGP, showed increased activity of the psoas major, external oblique (EO) and rectus abdominis during ASLR compared with controls [25]. In another study investigating long-lasting pregnancy-related PGP, indications were found for increased muscle tone of the pelvic floor, both at rest and during ASLR [26]. However, in two other studies a reduced muscle activity during ASLR was found. In a case-control study, Teyhen et al. showed that TrA and internal oblique (IO) were less active in patients with unilateral low back pain and a positive sacroiliac test [12]; however, in that study, the patients’ pain did not start during pregnancy or delivery (personal communication with Teyhen). Shadmehr et al. found a smaller recruitment of various muscles (EO, biceps femoris, gluteus maximus and erector spinae) during ASLR in women with sacroiliac pain compared with controls [27]; the participants of that study were not pregnant and at least 6 months postpartum. No information was available as to whether or not the pain started in relation to pregnancy (personal communication with Shadmehr).

It is a challenge to explain the difference between the excessive TrA contraction in the patients of the present study, and the reduced TrA contraction of the patients in the study of Teyhen et al. [11]. Nevertheless, one explanation could be the difference between the two study groups. In contrast with the present study, in the study of Teyhen et al. the pain of the participants did not start during pregnancy or delivery. In pregnancy-related PGP, enlarged mobility of the joints in the pelvic ring is well documented and this could be the key to the explanation (see below) [28].

It is also a challenge to explain why, in the present study, the enhanced contraction of the TrA in patients with PGP does not seem to help them to compress the pelvis sufficiently to perform the ASLR without difficulty. After all, in patients selected for this study the pelvic belt did help them to raise their leg.

We present two hypotheses. The first is related to a possible delay of TrA contraction during ASLR. A delay of TrA contraction is well-documented in electromyographic studies of patients with various painful disorders in the lumbopelvic region and in experimental pain [29‐31]. It is feasible that, in case of enlarged mobility (and not in case of normal mobility), a small movement in the sacroiliac joints takes place during ASLR before contraction of the TrA. A radiographic study demonstrated the forward rotation of the innominate at the ipsilateral side during ASLR in pregnancy-related PGP [32]. It is possible that muscular bracing of the sacroiliac joints does not need a large TrA contraction, but rather a timely contraction (or a pelvic belt fastened before the ASLR) to prevent painful sacroiliac movements. It is also possible that pain caused by the small displacement and/or fear for more pain, induce not only a strong (but late) TrA contraction but also contraction of other muscles. A study in healthy volunteers demonstrated that excessive activity of IO, EO, rectus abdominis, biceps femoris and latissimus dorsi during ASLR could be induced by experimental pelvic pain [33]; unfortunately TrA activity was not measured in that study. Richardson et al. demonstrated how a large TrA contraction can be inefficient in stabilizing the pelvic ring [34]. They showed that the effect of TrA contraction on force closure is smaller when TrA contracts together with the IO and EO, than in selective TrA contraction; the authors introduced the term ‘bracing’ to describe the large, but inefficient, contraction of many trunk and hip muscles.

A second hypothesis is that TrA activity with RUSI is assessed halfway the iliac crest and ribcage (thus in the middle region of the TrA) and is not representative for TrA contraction of the pelvic part of the TrA. In healthy volunteers it has been shown that TrA activity can vary per region [35, 36]. Theoretically, excessive activity of the middle region of the TrA may exist in combination with insufficient activity of the pelvic part of the TrA. Contraction of the middle part of the TrA might support the lumbar spine and the lower part of the pelvic ring. Moreover, excessive contraction of the TrA in the middle region might have an adverse effect on the pelvic girdle due to increased intra-abdominal pressure (IAP). IAP increase during ASLR in patients with pelvic girdle pain is well documented [37]. The possible adverse effect of increased IAP on the pelvic ring has been demonstrated in a biomechanical study [38].

Significant excessive contraction of the TrA is present during ASLR in patients with long-lasting pregnancy-related posterior PGP. The results of this study do not support the idea that contraction of the TrA is decreased in long-lasting pregnancy-related PGP. This implies that there is no rationale for the prescription of exercises to enhance contraction of the TrA in patients with long-lasting pregnancy-related PGP.