Low Back Pain in Pregnancy: Investigations, Management, and Role of Neuraxial Analgesia and Anaesthesia: A Systematic Review

Low back pain (LBP) is commonly experienced during pregnancy, affecting approximately 50% of pregnant women [1]. It is associated with significant morbidity in pregnant and post-partum patients [2, 3] with persistent pain lasting 2 years in a tenth of the patients [3, 4, 5, 6, 7]. Despite its prevalence, LBP during pregnancy is often considered a normal phenomenon. This benign neglect can contribute to physical inactivity during pregnancy, resulting in a higher incidence of obstetric complications [8] and caesarean sections [9, 10]. The epidemiology and the pathophysiology of LBP in pregnancy have been reviewed extensively [11, 12]. However, there is still much ambiguity with regard to the definition of LBP in pregnancy, identifying appropriate investigations, considering safe treatment options in pregnancy [13, 14, 15], and decision-making regarding the use of neuraxial analgesia and anaesthesia during labour and delivery [16, 17]. These considerations tend to differ from those for LBP in non-pregnant individuals. Choice of imaging modalities for investigating LBP during pregnancy is affected by the desire to avoid ionizing radiation exposure to the fetus. Concerns regarding maternal and fetal well-being also entail the use of conservative treatment regimens [18].

The primary objective of this review was to determine clinically useful and safe diagnostic investigations and management strategies for LBP in pregnancy. The secondary objective was to evaluate evidence and make recommendations for neuraxial analgesia and anaesthesia during labour and delivery in pregnant women with LBP.

We conducted a literature search using the OvidSP search platform in 3 bibliographic databases - Medline, Medline In-Process and Embase, using indexing terms related to pregnancy, LBP, and neuraxial techniques (online suppl. E-Table 1, see www.karger.com/doi/10.1159/000471764) to include articles indexed as of November 13, 2015. The search was limited to human data in pregnant and post-partum patients and restricted to the English language.

Studies

We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses recommendations during the preparation of this review [19]. We considered randomized controlled trials (RCTs), retrospective or prospective observational studies, case series or reports, conference proceedings and narrative reviews on investigative modalities, management strategies and the use of neuraxial (subarachnoid or epidural or combined spinal epidural analgesia/anaesthesia) techniques for pregnant women with LBP. We pre-specified eligibility criteria, which are as follows.

Participants

Pregnant subjects with new onset or worsening of pre-existing LBP of moderate-to-severe intensity (numerical rating score for pain of 4 or more on a 0-10 scale) were included. An analysis from pain diagrams drawn by pregnant patients reveals the predominance of pain in either the lower lumbar area or around the sacroiliac joints (SIJ) [2]. While the former has been addressed as “pregnancy-related LBP” (PLBP), the latter has been termed as “pregnancy-related pelvic girdle pain” (PGP) [3]. While PLBP represents low lumbar back pain between the 12th rib and the gluteal fold, PGP represents pain experienced between the posterior iliac crest and the gluteal fold [20]. Considering that there is considerable overlap between the 2 and that many women present with both, for purposes of this review, we included both PLBP and PGP under the term LBP.

Investigations and Interventions

(1) Diagnostic investigations for LBP in pregnancy included fluoroscopy (X-ray), CT, MRI, ultrasonography and electrodiagnostic studies such as electromyography (EMG) and nerve conduction studies (NCS).

(2) Therapies for treatment of LBP in pregnancy included physical therapy, patient education and psychological therapies, complimentary treatments (mechanical supports, osteopathic manipulative treatment [OMT], water gymnastics, yoga, acupuncture and transcutaneous electrical nerve stimulation [TENS]) and interventional techniques including injections (injection of steroids with or without local anaesthetics into the epidural space, sacro-iliac joint and symphysis pubis) and miscellaneous interventions (kyphoplasty, vertebroplasty, and laminectomy).

(3) Neuraxial analgesia and anaesthesia techniques for facilitating labour and delivery in parturients with LBP included spinal, epidural and combined spinal-epidural techniques. In this section, we also included patients with previous spine surgery or spinal implants (such as spinal cord stimulators [SCS]), because LBP is common in this cohort.

Comparator

Data on comparator group(s), when available, was collected and reported.

Outcomes

For diagnostic modalities, we looked at whether the modality was able to determine an underlying cause for LBP. For interventions, we looked at the efficacy in terms of impact on patient-reported pain scores or reduction in analgesic requirements and impact on functional ability.

Two authors (H.S. and R.D.) independently assessed the quality of included RCTs using the Cochrane Collaboration “Risk of Bias” tool for randomized trials [21]. Any disagreement was resolved through discussion or, if necessary, arbitration by a third reviewer (A.B.). The Cochrane tool assesses bias in the following domains: generation of the allocation sequence, allocation concealment, blinding of investigators and participants, blinding of outcome assessors, incomplete outcome data, selective outcome reporting and any other sources of bias. Each item is classified as low, unclear or high risk of bias. A decision to classify “overall bias” as low, unclear or high was made by the reviewers using the following method:

• High: any trial with a high risk of bias listed on 3 or more domains or significant methodological concerns that may have affected results.

• Unclear: any trial with a high risk of bias listed on 2 domains or moderate methodological concerns that may have affected study results.

• Low: any trial with a high risk of bias on none or 1 domain and with no significant methodological concerns that may have affected the study results.

Studies on pregnant and post-partum patients reporting on one or more of our outcomes of interest were included in this review. Reference data, populations, and outcomes were extracted from the articles into pre-specified tables using a standardized data extraction procedure by 2 of the authors (H.S. and A.B.). They extracted information on studies' general characteristics (including design, number of arms and primary outcomes), participants (characteristics of the populations, sample size), interventions (diagnostic modality or therapeutic technique or neuraxial techniques), comparator (if any), parameters used for assessing efficacy of the intervention and summary of main outcomes.

The search results and study selection flowchart is reported in Figure 1. A total of 6,803 records were identified through database searching and 61 through citation tracking. After excluding duplicates and screening titles and abstracts, 78 studies were found to be relevant to the objectives of this review and these were included for synthesis and discussion. This included 32 RCTs, 2 non-randomized CTs, 3 review of literatures, 3 cohort studies, 2 prospective case series, 5 retrospective case series, 29 case reports, 1 letter to the editor and 1 conference proceeding.

The reviewers' consensus assessment of the risk is detailed in Figure 2. The overall risk of bias across the studies was moderate. Most studies adequately stated the methods used for randomization and allocation concealment barring 4 RCTs [22, 23, 24, 25]. Blinding of participants and personnel was considered “unclear” in most studies barring “low risk” in only one study [26] and high risk in 3 other studies [27, 28, 29]. The blinding of outcome assessors was judged “unclear risk” in 17 studies and “low risk” in 12 studies [26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36]. All trials barring one study [37] were judged at “low” or “unclear” risk of attrition bias and selective reporting. Most studies were determined to be at “unclear risk” of other bias mainly due to lack of adequate details regarding blinding or outcome assessment. Finally, the overall bias was adjudged as “low” in 15 studies and as “unclear” in the remaining 14 studies.

Of the 32 studies concerning diagnostic investigations, there were 3 prospective and 2 retrospective observational studies, 26 case reports, and 1 review of literature (Table 1). Most patients were investigated when the intensity of pain was moderate or severe, or associated with neurological compromise.

MRI was most commonly used modality (17 publications) for diagnosing the cause of LBP, especially when LBP was associated with neurological symptoms in the lower limb such as radicular pain, paraplegia and/or cauda equina [18, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51]. The underlying pathologies associated with these presentations included herniation of intervertebral discs [18, 38, 39, 40, 41, 42, 43, 44, 45, 46], vertebral hemangiomas [51], spinal body tumour [44], and tuberculosis [44] (Table 1). MRI was also used to investigate LBP in pregnant women with co-existing hip pain, with 3 studies reporting inflammation of the SIJ as the underlying cause [52, 53, 54] and 10 studies describing sacral fractures [55, 56, 57, 58, 59, 60, 61, 62, 63, 64].

Six publications reported the use of CT scan for evaluating LBP during pregnancy or postpartum. The pathologies evaluated included vertebral collapse [47] and LBP associated with hip, pelvic or buttock pain [52, 54, 57, 58, 65]. While some of these scans were done during pregnancy [52, 65], most were done during the postpartum period including a therapeutic CT-guided hip joint aspiration in a postpartum woman with buttock pain [54].

We identified one study that used Doppler ultrasound to identify SIJ laxity [66] and another prospective study that utilized trans-abdominal pelvic floor ultrasound to identify pelvic floor muscle dysfunction in parturients with LBP [67].

One study reported the use of EMG in conjunction with MRI for evaluating4 pregnant patients with lumbar radiculopathy in the third trimester [46]. No publications reporting the use of NCS were identified.

There were 32 RCTs among the 57 publications outlining management strategies for LBP in pregnancy (Table 2). The following therapies were identified:

Six RCTs and one non-randomized trial studied multimodal management (education, exercise, and/or others) often as integrated antenatal classes for management of LBP in pregnancy (Table 2a). Bastiaenen et al. [68] studied the effectiveness of a tailor-made program modifying biopsychosocial factors (i.e., targeting patients' anxieties) in women with PLBP, and found it to be effective in improving the limitations in activities as compared to the control group (usual care). While 3 other studies reported better analgesic and functional outcomes when compared with the standard care for LBP [25, 69, 70], the rest reported no difference [24, 37, 71].

Conservative measures such as exercises were evaluated in 14 RCTs, one non-randomized trial and one case series (Table 2b). Seven studies reported benefit [29, 30, 32, 70, 72, 73, 74], while the rest did not [23, 27, 28, 34, 75, 76, 77, 78]. Similarly, PGP was reported to improve with pelvic tilt [73] and stabilization exercises [29], but not with truncal exercises [28]. Three studies did not show the benefit of exercise with PGP [27, 77, 78].

Eleven RCTs, 2 cohort studies and 2 case reports evaluated the role of complementary therapies (Table 2c). In a cohort study, the use of pelvic belts was found to reduce SIJ mobility, thereby providing relief from LBP [79]. In one RCT, support devices (BellyBra®, Tubigrip®) were found to be effective in reducing PLBP [80]. OMT reduced LBP and the progression of back-specific dysfunction in pregnancy and postpartum in 3 studies [26, 81, 82]. Water gymnastics [35] and yoga [83] were also found to be effective in reducing pain intensity.

Two case reports [84, 85] and 6 RCTs studied the role of acupuncture [22, 32, 33, 36, 86, 87]. Two of these RCTs demonstrated analgesic benefit when compared to controls [22, 36]. Acupuncture, as an adjunct to standard treatment, was found to be better than standard treatment alone for reduction of pain and improving the ability to perform daily activities in pregnancy-related PGP [32, 33]. However, timing of acupuncture during pregnancy (20 vs. 26 weeks) or its intensity (superficial versus deep) did not alter its benefit for LBP [86, 87]. One study compared no treatment (control arm) to 3 intervention groups (exercise, acetaminophen and TENS respectively) and found TENS to be as effective as exercise and acetaminophen [88].

Three case reports described the injection of steroids into the epidural space (ESI) to relieve LBP and radicular pain [40, 41, 42]; but all patients eventually required operative intervention due to the recurrence or progression of neurological symptoms (Table 2d). Ultrasound-guided injection of local anaesthetic and steroid into the symphysis pubis joint in a pregnant patient to relieve severe symphysis pubis pain was also reported [89]. There were 2 reports of injections into the SIJ in pregnancy with good analgesic response [65, 90].

Eight case reports and 2 case series described surgical interventions for intervertebral disc herniation causing neurological deficits (sensory, motor, bladder, or bowel involvement) including discectomy (open [38, 39, 41], micro-discectomy [18, 42], or endoscopic [45]), laminectomy [46], or both [40, 43, 44] (Table 2d). Three case reports [47, 48, 50] and one literature review [91] described surgical management of vertebral fractures and vertebral hemangiomas respectively. Surgery was generally successful in relieving pain and restoring function without increased morbidity or mortality.

Two case series and 2 literature review described the use of neuraxial techniques (epidural or spinal injections and/or catheters) for labour analgesia or caesarean section in pregnant patients with a history of LBP (with/without neurological deficits), spinal deformities and previous spinal surgery or implants (such as SCS; Table 3) [17, 92, 93]. While 2 publications focussed on the efficacy of interventions and technical feasibility [92, 93], one study also studied long-term neurological outcomes [17]. Although the techniques were successful in a majority of cases, there were multiple reports of complicated spinal or epidural injections (multiple insertion attempts, bloody taps and dural taps) [92, 93]. There were also some technical (inability to place epidural catheters) and therapeutic (patchy blocks with missed dermatomes) failures.

Hebl et al. [17] retrospectively reviewed the risk of neurologic complications in 937 patients with a history of pre-existing spinal canal pathology, with or without a history of spine surgery. This included 34 cases of labour analgesia and 10 caesarean sections. The incidence of new neurological deficits or worsening of pre-existing symptoms was 1.1% (10 out of 937 patients), though this study included both pregnant and non-pregnant populations. Further, data was not reported separately for the pregnant cohort.

Young et al. [94] retrospectively reviewed 24 cases of partituents with cervical or thoracic level SCS. Of the 17 cases in which a neuraxial was attempted for either labour analgesia or operative anaesthesia, 16 cases were successful. One patient who received epidural anaesthesia developed foot drop in the post-partum period. The etiology of this was unclear but assumed to be related to prolonged second stage of labor.

Several physiological factors may contribute to low back and pelvic pain in pregnancy. This includes a rise in hormones (of relaxin, progesterone, and estrogen) possibly resulting in an increase in joint laxity [95, 96, 97]. Asymmetric SIJ laxity [66] and greater pubic symphysis mobility [98] are associated with moderate to severe posterior pelvic pain in pregnancy and puerperium. Structural imbalances created by weakened abdominal muscles (due to enlarged gravid uterus), compensatory hyperlordosis (sagittal rotation of pelvis) and shifting of centre of gravity anterior increase the load on lumbar spine and SIJ, contributing to back pain [97]. Axial loading of the spine may compress intervertebral discs, reduce disc height and cause disc protrusion culminating in radicular symptoms. Osteoporosis in pregnancy has been associated with vertebral [48] and sacral stress fractures [60]. Other rare pathologies may include osteomyelitis [99] and spinal tumours [91, 100]. A detailed account of the aetiology of low back and pelvic pain in pregnancy can be found in other publications [11, 12].

This systematic review was performed to identify clinically useful and safe diagnostic investigations and management strategies for LBP in pregnancy and to make recommendations for neuraxial analgesia and anaesthesia during labour and delivery. The findings have been discussed below and recommendations outlined in Figure 3.

The aetiological diagnosis of LBP is mostly clinical with imaging reserved for severe or challenging cases. Most common pathologies diagnosed on imaging in pregnancy include intervertebral disc herniations, vertebral hemangiomas and fractures and sacral stress fractures. MRI is the most commonly used imaging modality for pregnant women with neurological compromise (dermatomal lower limb numbness [38], perineal numbness [39, 40], lower limb weakness [18, 39, 40, 43, 45, 46], or loss of bladder/bowel control [40, 43, 45]). Suspected disc herniation was the most frequent reason for imaging, followed by incapacitating LBP without neurological compromise [42]. Unlike most patients with LBP and radicular lower limb pain that present in the second and the third trimesters, disc herniation was also reported in the first trimester [46]. This has implications for both diagnostic imaging and interventions that are often delayed until after the first trimester.

MRI without contrast is considered the safest imaging modality for spinal pathologies that manifest during pregnancy [11, 101, 102], offering better resolution when compared to ionizing imaging modalities (fluoroscopy and CT). While there have been concerns regarding MRI-induced fetal teratogenicity, acoustic damage and heating-induced effects, recent evidence does not support this [103, 104]. However, the safety of the more recently introduced 3 Tesla MRI (which may provide better imaging because of a stronger magnetic field than the conventional 1.5 tesla MRI) has not yet been well studied [105]. There have been conflicting recommendations regarding the use of MRI in pregnancy, with some suggesting delaying its use until after the first trimester [106], while others proposing its use regardless of gestational age [107, 108]. Since the intravenous contrast agent, gadolinium, could have teratogenic effects [109], caution in its use in pregnancy is recommended [108] although it is highly improbable that toxic levels would be attained in fetal circulation [110, 111]. Discarding breast milk for 24 h after its administration can minimize the less-intensive risk to the breast-fed infant.

The use of modalities with ionizing radiation (fluoroscopy and CT) for imaging in pregnancy has been associated with the risk of spontaneous abortions, intrauterine growth restriction, neurobehavioral abnormalities, fetal malformations, or carcinogenesis [112, 113]. Such effects may depend upon the radiation dose absorbed by the fetus and the trimester of the pregnancy [114]. Although the estimated fetal absorption of radiation for a CT-scan of pelvis (20-79 mGy) is below the significant risk threshold (100 mGy) [115], a dose of 100-200 mGy can be lethal for an embryo. Despite this, the fetal risks with clinically used doses are minimal and thus radiologic examinations that may provide significant diagnostic information should not be withheld from pregnant women [115]. Our review showed that CT scans were reserved for patients with upper back pain [116] or post-partum PGP [57, 58, 65], especially when thoracic vertebral collapse or associated pathology in the hip was suspected. The reason for this is that CT is often preferred over MRI when bony pathology is suspected [117].

Although ultrasound is considered safe in pregnancy, it does not provide the information required to evaluate most spinal pathologies [118]. The indications for using diagnostic ultrasound in patients with LBP include suspected SIJ pathology [20, 119], diagnosing and monitoring pubic diastasis, and the assessment of pelvic floor dysfunction [120]. Clinical evaluation and MRI may not accurately diagnose SIJ dysfunction.

Electro-diagnostic studies (NCS and EMG) are valuable tools to supplement anatomical information obtained from imaging in pregnant women with lumbar radiculopathy secondary to nerve root compression by intervertebral disc disease, ligamentous hypertrophy or arthritic changes [121, 122, 123]These can help ascertain the correlation between MRI findings and presenting neurologic deficits, allowing an informed decision to perform a therapeutic injection or operative intervention [124]. Despite this, we found a low rate of use for electrodiagnostic studies, which may be due to patient preference (electrodiagnostic studies can be painful), urgency of the presentation or lack of awareness among health care providers.

Table 4 summarized the common pain generators in the lower back and pelvis, their common presentations, imaging or investigations of choice and their common findings. It must be borne in mind that the correlation between the causes of back pain and the radiological findings is poor [125, 126] and that findings on imaging do not always imply aetiology [127].

Multimodal management of LBP that includes physical therapy, patient education and psychological therapies, complimentary treatments, and interventional techniques (injections and surgery) is recommended for optimizing outcomes in non-pregnant patients [128]. The role of these modes of therapy in pregnant women with LBP is discussed below.

Back care training programs educate pregnant women on the basic anatomy, ergonomics, correct postures, relaxation and pain management strategies [11]. Although our review shows conflicting evidence of their role, it promotes awareness on the benefits of being active and of self-care, potentially reducing the severity of LBP and consequent disability. The National Institute for Health and Clinical Excellence in the United Kingdom (UK) recommends that “exercising in water, massage therapy and group or individual back care classes” may help in reducing LBP during pregnancy [129]. Bio-psychosocial factors (such as anxiety and catastrophisation) can prevent the conditioning of back muscles and worsening LBP [58, 120], both during and after pregnancy [131], and should be considered when formulating preventive strategies and in rehabilitation.

The evidence for exercise in management of LBP in pregnancy is equivocal. The results of 2 recent systematic reviews differ from each other, probably due to the inclusion of post-partum women in one [132] and not in the other [133]. Although exercise does not appear to benefit patients with PGP [23, 32, 33], it is often recommended during pregnancy because it is a non-invasive, low-risk, inexpensive intervention that may have benefits in addition to providing relief from back pain [129, 134].

Pelvic belts and other support devices may limit the laxity at SIJ, provide symptomatic relief, and are safe [135]. However, their use is not supported in a recent systemic review of lumbar supports [136]. While there is a paucity of RCTs utilizing yoga for alleviating LBP in pregnancy, meta-analysis from non-pregnant populations with LBP found moderate-to-strong evidence of its effectiveness [137] with the caveats that pain-related catastrophising and fear of movement may decrease its use [138, 139]. Though 2 RCTs support the use of acupuncture for managing LBP and PGP in pregnancy [22, 36], evidence for acupuncture in managing LBP is uncertain, mainly due to methodological flaws and significant heterogeneity among included studies [140, 141, 142]. Based on one RCT, TENS appears to be as effective as exercise therapy or acetaminophen at reducing both pain and disability from pregnancy-related back pain.

We found weak evidence for the analgesic and surgery-delaying effect of ESI in pregnant patients with LBP, which is consistent with observations in non-pregnant patients [128, 143]. Although a single dose of epidural steroid appears to be of low risk to the fetus, it is recommended that ESI be reserved for patients with new onset of signs or severe symptoms of lumbar nerve root compression [14]. Unknown variables about ESI in pregnant patients include frequency of its use, efficacy, and duration of therapeutic effect. It might, however, be reasonable to use higher volume (5-10 mL) ESI based on clinical diagnosis, whereas recurrence or worsening of neurological symptoms may prompt imaging and a transforaminal ESI using lower volumes (1.5-3 mL) that is performed with the objective of the medication reaching the dorsal root ganglion of the spinal nerve root [144].

Lumbar spine surgery under general anaesthesia [38, 39, 40, 41] is indicated during pregnancy for disc herniation, tumours and vertebral fractures associated with significant neurological compromise. Although surgery on the lumbosacral spine in pregnant women poses hazards to the mother and fetus because of requirement for prone or lateral positioning and exposure to anaesthetic medications, it should not be delayed if essential for preserving maternal life and function [145]. Preoperative discussion regarding risks with the pregnant patient and a team approach involving obstetricians, anaesthesiologists, neonatologists and spine surgeons is advocated for optimizing outcomes [44]. Co-interventions include perioperative fetal monitoring after 23 weeks of gestation [41], administration of corticosteroids to accelerate fetal lung maturity before 34 weeks of gestation and fetal surveillance in the immediate postoperative period [146]. Delivery is rarely indicated to allow spinal surgery at the same time [38, 39, 41, 43]. Prone position may be utilized to access the spine during the first trimester or in post-partum [39, 41], while lateral decubitus that reduces aortocaval compression, is preferable in second and third trimesters [45].

We found that neuraxial techniques (epidural or spinal) in pregnant patients with history of LBP, spinal deformities or previous spinal surgery are associated with higher rates of complications and failures. It is noteworthy that the risk of neurological complications reported by Hebl et al. [17] is considerably higher than that found as a result of a UK national audit (1.1/100 vs. 2.0-4.2/100,000), although the evidence synthesized by Hebl et al. [17]included both pregnant and non-pregnant patients [147]. These findings may lead to anaesthesiologists not offering neuraxial techniques to pregnant patients with LBP because of the potential for worsening of pre-existing neuraxial deficits. This is unfortunate because neuraxial analgesia and anaesthesia have a better efficacy and safety profile than systemic analgesia and general anaesthesia [148, 149]. Although selecting a different level than the level of pathology to perform the neuraxial techniques may appear reasonable, there is not much evidence to support this theory. In their review, Hebl et al. [17]found that 7 out of 10 patients who reported new or worsening neurological deficits had a neuraxial block at a different level than the level of the pathology. These results cannot be confidently interpreted because the palpatory method of spinal level localization utilized in this review is unreliable [150]. Despite this, avoiding the level of an obvious or known pathology by use of an ultrasound-guided technique may be considered potentially protective until further evidence is available. The use of ultrasound for guiding neuraxial procedures helps in better identification of spinal levels, allows accurate estimation of midline and epidural depth, improves success rate and ease of performance and may reduce the risk of traumatic procedures [151] this is applicable to patients with previous spine surgery [152]. We envision expansion in the use of ultrasound-guided neuraxial interventions in parturients and further research is in progress [153].

Technical feasibility and efficacy of neuraxial anaesthesia particularly after spinal surgery are other valid concerns for healthcare providers. Epidural anaesthesia can be challenging in such patients due to abnormal or missing landmarks, and due to the presence of surgical implants. It is associated with a higher risk of unintentional dural puncture, false loss of resistance, multiple puncture attempts and consequent patchy or failed blocks [92, 93, 154, 155, 156]. In patients with previous spine surgery, a “single-shot” spinal anaesthetic may be a more reliable and efficacious approach [17]. Combined spinal-epidural [157] and continuous spinal anaesthesia [158] have also been successfully utilized in patients with previous spinal surgery but are less common choices due to perceived technical difficulty. Thankfully, the presence of spinal implants does not preclude the possibility or impact the success of neuraxial anaesthesia or analgesia in patients with a spinal cord stimulator [94]. This is because most such patients have implants at either cervical or lower thoracic levels. However, necessary details pertaining to this level, and preferably implant-related spine imaging, must be reviewed to avoid inadvertent damage to the epidurally placed leads.

Timely identification and planning of care by a multidisciplinary team for patients with severe LBP, spine pathology, or history of spinal surgery or implants in the antenatal clinic can assist appropriate planning of labour analgesia or surgical management for delivery. It is also important to note that delivery itself can cause neurological damage resulting in pain and disability post-partum [159]. Additionally, there are concerns that repeated Valsalva manoeuvers for vaginal delivery may further compress the cauda equina in a patient with disc herniation and cause worsening of neurological deficit [41]. Therefore, in cases with pre-existing neurological deficits, decisions regarding the management of labour and delivery must be made by a multidisciplinary team. The use of neuraxial anaesthesia, especially labour analgesia, may also potentially prevent the prompt recognition of worsening neurological deficits. Finally, while it is not anticipated that the risk for neurological injury in patients with LBP risk may be increased, a documentation of the patient's LBP may help differentiate this from other causes of post-partum back pain.

This systematic review employed a comprehensive search strategy covering all aspects in the management of LBP in pregnancy with evidence-based recommendations. However, due to the absence of high-level evidence in all areas, a significant proportion of evidence came from cohort studies, case series and case reports, and this increases the possibility of bias. The RCTs included in this review were of moderate quality and included several single blind studies that may influence the outcomes, particularly in interventional studies. Evidence regarding the use of neuraxial analgesia and anaesthesia is largely from non-pregnant patients. Unique anatomical and physiological changes of pregnancy suggest that inferences based on studies on non-pregnant patients may have limited applicability in pregnancy.

LBP is common in pregnancy, but there are several knowledge gaps regarding optimal modalities for diagnosis, treatment, and anaesthetic management of this condition. The evidence presented in this review suggests that though sinister causes of LBP are uncommon, “red-flags” such as radiating leg pain, neurological deficits (numbness or weakness), bladder and/or bowel symptoms, or fever should be excluded in all patients. While the evaluation of LBP is mostly clinical, MRI is the diagnostic investigation of choice during pregnancy, with recourse to imaging involving ionizing radiation, ultrasound and electro-diagnostic studies in some instances. The inclusion of back care training in antenatal educational programs that include consideration of bio-psychosocial factors may promote activity and help reduce morbidity in pregnancy. There is limited evidence for the role of exercise and acupuncture and further research is needed in establishing the role of complementary and alternative therapies in the management of LBP in pregnancy. ESI for pregnant patients with radicular pain may confer analgesic benefit, while operative intervention is indicated for the most severe cases of LBP, particularly with associated neurological compromise (Fig. 3). Performance of neuraxial blocks in pregnant patients with LBP, neurological deficits and previous spine surgery or implants is technically feasible in most cases, but until more evidence is available regarding neurological complications in pregnant patients with neurological deficits or previous spinal surgery, management should be individualised.

We would like to thank Ms. Marina Engelasakis, Librarian, University Health Network for her assistance in designing the literature search strategy.

The authors have no conflicts of interests to disclose.

A.B. and H.S. designed the study, reviewed the literature, analysed the data, and prepared the manuscript. R.D. provided obstetrical and methodological input and helped with editing the manuscript.

Not applicable.

The authors have no funding to declare.