Lumbopelvic pain (LPP) is common during pregnancy. However, there is a lack of strong systematic reviews summarizing the evidence on conservative care.
AimsTo evaluate the effectiveness of physiotherapy interventions compared to standard obstetric care (SOC) or other treatments for physical function, pain intensity, and quality of life in pregnant women with LPP.
MethodsMedline, Emcare, Embase, Emcare, Physiotherapy Evidence Database (PEDro), Web of Science, and Cochrane Central Register were searched for randomized controlled trials evaluating physiotherapy interventions for pregnancy-related LPP. Risk of bias was assessed using Cochrane Risk of Bias 2, and evidence was summarized using the GRADE framework. When appropriate, a meta-analysis was conducted to synthesize the available evidence.
ResultsThirty-one trials were included (n = 3279 screened): 23 investigated exercise, 7 manual therapy, 3 electrophysical modalities, and 2 education. Exercise was not better than SOC for the short-term pain (low certainty evidence, MD = -18.80, 95 % CI -45.97 to 8.33), but was superior at 6–12 weeks (moderate certainty evidence, MD = -25.84, 95 % CI -35.04 to -16.64). After a sensitivity analysis, exercise significantly improved physical function compared to SOC at 0–5 weeks (low certainty evidence, SMD= -0.65 95 % CI -1.25 to -0.04). Exercise was superior to SOC at 6–12 weeks (very-low certainty evidence, SMD -2.79 95 % CI -4.64 to -0.95).
ConclusionExercise interventions show low certainty evidence for improving pain and physical function in pregnancy-related LPP over medium-term periods. Higher quality research is necessary, and future studies should adhere to standardized reporting guidelines and include larger sample sizes.
Lumbopelvic pain (LPP) is the most common musculoskeletal complaint during pregnancy, significantly affecting quality of life (QoL).1 LPP may include low back pain (LBP), pelvic girdle pain (PGP), or a combination of these conditions.2 The prevalence of LPP ranges from 33 % to 50 % before 20 weeks of gestation and rises to 60 %−70 % in the later stages of pregnancy.3–5 The severity of symptoms is also shown to increase over the course of pregnancy, reaching its highest intensity between the 24th and 26th week of gestation.2,6 Risk factors for LPP during pregnancy include a history of low back and pelvic pain, strenuous work, increased body mass index (BMI), smoking, and pre-existing orthopedic conditions.2,7 Psychological risk factors include emotional distress,8 high levels of catastrophizing, work dissatisfaction,7 and fear-avoidance beliefs,9
Despite being associated with a significant health burden, LPP is often underdiagnosed and undertreated due to a lack of understanding of its etiology among healthcare providers.10 Traditionally, LPP was thought to arise from biomechanical changes during pregnancy,11–14 but current evidence suggests the role of central pain mechanisms including central sensitization.7,15,16 Central sensitization is characterized by an increased responsiveness of nociceptive neurons in the central nervous system to normal and subthreshold sensory input,17 as well as reduce the efficacy of descending inhibitory pathways.18 The possibility of a diversified etiology of LPP advocates for a biopsychosocial management approach.7 Current clinical practice guidelines (CPGs) support the use of conservative management, including physiotherapy interventions such as manual therapy, pelvic stability training, acupuncture, and education as the first line of care for pregnancy-related LPP.7,19 A systematic review and meta-analysis found that manual therapy had positive effects on LPP severity when compared to standard care, but not when compared to sham treatments.20 Moreover, another systematic review identified through meta-analysis that any land-based exercise significantly reduced LPP in pregnant individuals.1 However, both reviews and available CPGs are dated from 2017 or earlier,1,7,20 indicating that a comprehensive update has not been conducted in over seven years. In addition, the quality of evidence from reviews and the strength of recommendations from CPGs are low to very low.1,7,20 As such, there is a need for an updated review of the effectiveness of conservative interventions for LPP with a focus on guideline-recommended interventions, primarily those that fall within the scope of practice of physiotherapy. Thus, the purpose of this study is to determine if physiotherapy interventions are more effective than standard care or other interventions for improving function, pain intensity, and QoL in pregnancy-related LPP.
MethodsThe protocol for this systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO ID CRD42022313181). The Cochrane Handbook of Systematic Reviews of Interventions guided this systematic review.21 The statement of Preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA) was used as the guideline for reporting in this review.22
Data sources and searchesThe following electronic databases were systematically searched, from inception to the search date: Medline, Emcare, Embase, Allied and Complementary Medicine through the OVID platform, Physiotherapy Evidence Database, and Web of Science. The search also included the Cochrane Central Register of Controlled Trials in the Cochrane Library for additional unpublished or ready-to-be-published studies. Search strategies were constructed using the search terms from key articles19,20,23 and finalized after consultation with an experienced university librarian (Appendix 1). Due to the broad scope of physiotherapy practice, and to not limit our search to specific interventions, search terms for physiotherapy interventions were not included in the search strategy. Authors additionally scanned the reference list of included studies to ensure all relevant articles were retrieved.
Study selectionTypes of studiesStudies in any language were included if they were a full-length randomized controlled trial (RCT) published in a peer-reviewed journal. Non-RCTs, quasi-RCTs and other types of studies were excluded.
Types of patient populationStudies were included if participants were antepartum individuals of any age presenting with pregnancy-related LPP. Studies were excluded if participants presented with LPP of specific etiology, including history of spinal surgery, radiculopathy, inflammatory conditions (e.g., rheumatoid arthritis), or serious spinal conditions (e.g., cancer).
Types of interventionInterventions that fell within the physiotherapy scope of practice against any comparator were included. Interventions for both experimental and control conditions were grouped into the following categories: a) standard care (defined as obstetric/prenatal care or minimal intervention which may include education, exercise, ergonomic education among others, b) exercise (e.g., aerobic exercise, resistance exercise, flexibility), c) psychosocial interventions (e.g., cognitive behavioural therapy), d) electrophysical modalities (e.g., TENS, ultrasound), e) acupuncture, f) bracing, g) education, and h) manual therapy. No intervention/placebo or minimal intervention (e.g., booklet or educational material without consultation with a health professional) were included as groupings for control. Studies were included if the intervention was delivered antepartum only, however, outcome data from antepartum and postpartum were extracted. It was determined that due to the large number of studies identified, the results on acupuncture, taping or bracing will be presented in separate reviews.
Types of outcome assessmentsThe primary and secondary outcomes of this study were physical function and pain intensity, respectively. This decision was informed based on the recommendations from the CPG for Pelvic Girdle Pain in the Antepartum.7 Data was extracted for any time point antepartum and was categorized into two time points: 1) short-term 0–5 weeks post-randomization, 2) > 6 weeks after randomization. These time points represent a deviation from the protocol to better align with the time points presented within the included studies.
Covidence software was used to record and manage the selected studies.24 A training phase was implemented for authors to screen 30 titles and abstracts to ensure consistency between the screeners. Two pairs of authors (LM, MS, JM, NT) independently screened titles and abstracts and full text, with conflict resolved by a third-party member (PN). Furthermore, if a third-party member could not settle the agreement, Authors were contacted if more information about the study was needed, and studies were excluded if information provided was considered unclear, and the study was excluded if a response was not received.
Data extraction and quality assessmentThe data extraction table was piloted (3 studies) by authors (MM, LC, MS, PN) to ensure adequate data collection. Pairs of reviewers (MM, LC, MS, PN) independently extracted data and assessed risk of bias (RoB) using the standardized Cochrane Risk of Bias tool 2.0 (RoB2).33
The Consensus on Exercise Reporting Template (CERT) was also used to extract relevant data from interventions in the included study.25
The quality of the evidence was summarized using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) tool.26 Risk of bias was judged by examining methodological limitations such as lack of allocation concealment, blinding issues, incomplete outcome data, or selective reporting. Inconsistency was assessed by evaluating heterogeneity of results, including variance in point estimates, overlap of confidence intervals, and I² statistics, with downgrading applied when unexplained heterogeneity was substantial (I² > 50 %). Imprecision was considered in relation to sample size, number of events, width of confidence intervals, and whether these crossed thresholds of clinical importance or failed to meet optimal information size. Indirectness was evaluated by assessing applicability of the evidence to the review question, including differences in population, intervention, comparator, outcomes, or study setting. Publication bias was judged by considering the possibility of selective publication, funnel plot asymmetry, results of statistical tests, and whether our search strategy identified unpublished or non-indexed studies. Each domain was evaluated systematically and transparently, and evidence was downgraded by one or two levels when concerns were identified. Based on this process, the certainty of evidence for each outcome was rated as high, moderate, low, or very low.
Data synthesis and analysisStudies were statistically pooled if there was homogeneity in terms of PICO (population, intervention, comparison, and outcome). If statistical pooling was not possible due to heterogeneity of PICO elements, the findings were presented in narrative form with mean differences or risk ratios and 95 % confidence intervals (CIs) as appropriate. Quantitative data synthesis was carried out with RevMan 5.4.1 (Copenhagen, Denmark: The Nordic Cochrane Centre, The Cochrane Collaboration [www.Cochrane.org]). Pooled results were reported as mean difference (MD) or standardized mean difference (SMD) with a 95 % CI as appropriate. To synthesize results a random-effect model was used and I² was reported for statistical heterogeneity. Furthermore, we used the size of the effect to determine whether the results were clinically relevant or not, where a reduction of 25 % in the MD outcome measure or a SMD ≥ 0.5 was reported. The 0.5 SMD threshold reflects a standardized interpretation of medium effect, but should not be interpreted as equivalent to a 25 % change in the MD.
Sensitivity analysisWe conducted sensitivity analyses by excluding studies in which participants had only low back pain (LBP) or only pelvic girdle pain (PGP), in order to assess whether the location of pain influenced the overall treatment effect.
ResultsStudy selectionA total of 3279 studies were screened after duplicate removal. After full-text review, 49 studies evaluating interventions within the scope of physiotherapy were identified27–75; with 3128,30,32,33,37,39,40,42–47,51–53,56–61,63,65–71,75 eligible for inclusion in this review. A comprehensive list of excluded studies is presented in Appendix 2. As previously identified, studies evaluating taping, bracing or acupuncture, will be summarized in a separate review. Twenty-three studies investigated some type of exercise,28,30,32,33,37,40,42,45,46,51–53,56–58,60,61,63,65,66,68,69,75 7 manual therapy,39,43,44,47,59,66,71 3 on electrophysical modalities,51,70,71 2 education,63,67 and 2 multimodal therapy.43,47 A PRISMA22 diagram of the search results, including reasons for exclusion, is shown in Fig. 1.
Study characteristics
Trials were conducted in 10 countries with 4690 participants. The most represented countries were Iran (4 studies, n = 423)32,53,58,61 and Turkey (3 studies, n = 272).51,65,69 Exercise studies compared core strengthening,28,37,40,42,53,76 general exercise,32,33,45,51,52,58,60,61,63,65 Pilates,69,75 Yoga77 or Mckenzie30 to standard obstetric care, acupuncture,37 manual therapy,66 neuroemotional technique,66 TENS,51 physiotherapy routine,46 or other forms of exercise.63,68 Manual therapies were compared to standard obstetric care,39,47 TENS,71 or manual therapy combined with exercise.59 Other studies compared education delivered by phone to in-person education,67 TENS to acupuncture70 and multimodal approaches to standard care.43,44 Standard obstetric care within the included studies consisted of physician care, general education, and over-the-counter pain medication. The characteristics of included studies can be seen in Table 1.
Characteristics of included studies.
ODI – Oswestry Disability Index, NPRS – Numeric Pain Rating Scale, NRS – Numeric Rating Scale, SF-12 – Short Form Health Survey-12, GPE – Global Perception Effect, EQ-5D – European Quality of Life, WHOQOL – World Health Organization Quality of Life Scale, DRI – Disability Rating Index, PGPQ – Pelvic Girdle Pain Qestionnaire, PSFS – Patient Specific Functional Scale, PGQ – Pelvic Girdle Questionnaire, RMDQ – Roland Morris Disability Questionnaire, VAS - visual analog scale, HADS-A - Hospital Anxiety and Depression Scale-Anxiety, HADS-D - Hospital Anxiety and Depression Scale-Depression, PSQI - Pittsburgh Sleep Quality Index.
The exercise interventions varied between studies in setting, type, and dosage. The average CERT score for the studies was 7.5 (±1.8) (Appendix 3). Most exercise programs required no specific equipment and varied from 1–5 times per week.28,30,66 Study durations ranged from 3 to 12 weeks.51,58,75 None of the 23 studies reported adverse events or clear starting level for exercises (Fig. 2). Tailored interventions were rare, with only two studies providing patient-specific programs.30,63 Detailed descriptions of exercise strategies were reported in a minority of studies, making the comparison of interventions difficult (Fig. 2).
Risk of bias
A Cochrane RoB2 assessment demonstrated that 13 studies28,30,40,43–47,67–70,75 had high RoB, 9 had some concerns32,33,51,53,58,60,61,63,65 and 9 had low RoB37,39,42,52,59,66,71,76,77 (Appendix 4). The most bias was related to missing data because of high attrition rate (22.5 % of the studies), followed by measurement of outcome (16.1 % of the studies) due to lack of blinding as per nature of the intervention.
Intervention effectExercise vs. standard careOf the 23 studies evaluating exercise, 20 compared it to standard obstetric care for pain and function.28,30,32,33,37,40,42,45,51–53,58,60,61,63,65,69,75–77 Eleven of the studies were statistically pooled using random effect models,30,33,40,42,45,53,58,61,65,69,76 while nine studies28,32,38,51,52,60,63,75,77 were reviewed narratively due to heterogeneity or insufficient data. See Appendix 5 for the results of each study. Results from the meta-analysis for pain and function can be seen in Figs. 3–6.
Pooled results of four studies showed low certainty evidence (downgraded by inconsistency and imprecision) that any exercise was not better than standard care for pain intensity at 0–5 weeks (−18.80, 95 % CI −45.97 to 8.33, n = 601) (Fig. 3), however, there is moderate certainty (downgraded by inconsistency) that exercise was better than standard care for pain intensity at 6–12 weeks (−25.84, 95 % CI −35.04 to −16.64, n = 1143) (Fig. 4).
A subgroup analysis based on the type of exercise (core /strengthening stabilization exercises or general exercises - as subjectively characterized by each study) showed low certainty evidence (downgraded by inconsistence and imprecision) that general exercise provided a greater decrease in pain in the long term as compared to standard care (MD= −35.56, 95 %CI −60.87 to −10.24, n = 257) (Fig. 4). There was moderate certainty evidence (downgraded by study limitations) that core strengthening (−18.39, 95 % CI −22.83 to −13.94, n = 565) provided a greater reduction in pain in the long-term compared to standard care (Fig. 4). In the short-term, there was low certainty evidence (downgraded by inconsistence and imprecision) that core strengthening is not better than standard care to reduce pain (Fig. 3). There was insufficient data to pool the effectiveness of general exercises for pain in the short-term.
Nine studies were included in the narrative synthesis of this review.28,32,37,51,52,60,63,75,77 Of these, six reported exclusively on short-term outcomes,32,37,51,52,75,77 one study focused on long-term outcomes,63 and two studies reported on both short- and long-term effects.28,60 The primary reason these studies could not be pooled in a meta-analysis was due to differences in intervention types (e.g., yoga, education, and supervised exercises combined with other interventions), and comparator groups (e.g., TENS and Acupuncture). Among the narratively reviewed studies, two studies (n = 78)51,75 found low certainty evidence favorable to the combination of different exercises,51 and Pilates75 in reducing pain in the short-term (downgraded by study limitation and size), and seven studies (n = 1044)28,32,37,52,60,63,77 showed low certainty evidence in favour of exercises over standard care for pain reduction in the long-term (downgraded by inconsistency and indirectness
Outcome: functionThe meta-analysis showed low certainty evidence (downgraded by inconsistency and imprecision) that exercise did not improved physical function compared to standard care at 0–5 weeks (SMD= −1.38 95 % CI −3.08 to 0.31, n = 772) and very-low certainty evidence (downgraded by study limitation, inconsistency and imprecision) for the results in favours of exercise at 6–12 weeks (SMD −2.79 95 % CI −4.64 to −0.95, n = 1111) (Figs. 5 and 6). Four of the six studies that narratively reported outcomes of physical function provided low certainty evidence (downgraded by study limitation and inconsistency) that exercise was superior to standard care in the short term (n = 389)28,32,51,52 and three studies provided low certainty evidence (downgraded by study limitation and inconsistency) of the superiority of exercise over standard care in the long-term (n = 345).28,32,52
Subgroup analysis revealed low certainty evidence (downgraded by inconsistency and imprecision) that general exercises (SMD = −0.42, 95 %CI −0.99 to 0.15, n = 267) and very-low certainty evidence (downgraded by study limitation, inconsistency and imprecision) that core strengthening (−2.02, 95 %CI −4.70 to 0.67, n = 505) were not better than standard care for physical function in the short-term (Fig, 5). Also, there was low certainty evidence (downgraded by inconsistency and imprecision) that general exercise (SMD= −1.69, 95 %CI −4.80 to 1.42, n = 205) and very-low certainty evidence (downgraded by study limitation, inconsistency and imprecision) that core strengthening, SMD= −4.18, 95 %CI −9.74 to 1.38, n = 525) did not reduce pain or improve physical function in the long-term (Fig. 6).
A sensitivity analysis was performed excluding the study developed by Kordi, Abolhasani53 which recruited participants with PGP only. No change was observed in the results for pain (VAS) or function (ODI) in the short or long-term after its removal.
Although this was not a planned sensitivity analysis, an observation of the forest plots and results of individual studies demonstrate that one study (n = 408)40 reported extreme improvements with MDs as high as −47.24 on the VAS in the short-term, and SMDs as high as −13.5 for the Quebec Back Pain Disability Scale in the long-term. After removing this study in a sensitivity analysis, the results for function in the short-term (Fig. 7) became statistically significant. Although Filipec and Matijević40 (n = 408) reported favourable results to exercise, their results were so distinct when compared to the rest of the studies that its addition led to very wide confidence interval turning polled results not statistically significant. Note that single studies that did not fit any subgroup were included in the plot but omitted from the meta-analysis. In this case, these studies were given a weight of zero.
Outcome: quality of life
Four studies reported on QoL, these studies were included for meta-analysis for pain and function but the data for QoL could not be pooled due to a lack of reporting of total scores within the studies (n = 205).33,53,69,76 Although Barbier, Blanc33 and Mamipour, Farazmehr76 found improvements in the physical and psychological domains, Kordi, Abolhasani53 found significant improvements only in the physical function domain and Sonmezer, Özköslü69 did not find significant improvements in QoL. However, these findings are based on very-low certainty evidence (downgraded by study limitation, inconsistency, publication (small study) bias).
Exercise vs. other interventionsAt the short-term, there is very-low certainty evidence (downgraded by publication (sample size) bias) showing that stabilization exercises are better than general exercises46 for pain (NPRS, MD= 2.25, p = 0.037) and physical function (modified ODI, MD= 13.1, p = 0.006). TENS was superior to exercise in reducing pain51 (VAS, median exercise= 6(1), median TENS = 4(1), p < 0.001) and physical function (RMDQ, median exercise=13(3), median TENS=7(2), p < 0.001) in one small study, while another study (n = 22) found no difference in pain (p = 0.12) and physical function outcomes (p = 0.5) between water and land-based exercises.68 These recommendations are based on very-low certainty evidence (downgraded by publication (sample size) bias). No adverse events of TENS application on pregnant women were reported by Keskin, Onur.51
Very-low certainty evidence (downgraded by publication (sample size) bias) showed no difference for pain (p = 0.52) and physical function (p = 0.99) when comparing exercise with home exercise (n = 78),63 and manual therapy (n = 31) or neuroemotional technique for pain (n = 35, p > 0.05) in the long-term.66 One study (n = 113)37 provided very-low certainty evidence (downgraded by publication (study size) bias) that acupuncture was superior to stabilization exercise for evening pain only (MD= −14, 95 %CI −18.1 to −3.3). Individual studies results can be found in appendix 3.
Manual therapyFour studies (n = 613)39,47,59,71 evaluated manual therapies. Osteopathic manipulation was not different than sham US for reducing pain (Quadruple VAS, MD= −0.16, p > 0.99) or improving physical function (RMDQ, MD= −0.21, p > 0.99) at the long-term.47 However, osteopathic manipulation improved physical function (RMDQ, MD= −2.3, p < 0.001) but not pain intensity (Quadruple VAS, MD= −0.38, p = 0.07) compared to standard obstetric care at the long-term.47 Sacroiliac joint mobilization was better than TENS to reduce pain (VAS, MD = −1.8, p < 0.001) and physical function (RMDQ, MD= −2, p = 0.02) at the short-term.71 Craniosacral therapy improved pain in the morning (VAS, median difference= −8, p = 0.002) and physical function (ODI, median difference= −8, p = 0.02) but there was no difference for pain in the evening (VAS, median difference= −7, p = 0.08), quality of life (EQ5D, median difference= 0.10, p = 0.07) and sick leave ( %, median difference= 5, p = 0.28) at the long-term as compared to standard care.39 Foot manipulation was not better than foot massage plus mobilization and home exercise for reducing morning pain (VAS, MD= 9, p = 0.64), evening pain (VAS, MD= 18, p = 0.28), or sick leave at the long-term59 However, all recommendations are based on very-low certainty evidence (downgraded by publication (study size) bias).
EducationOne study67 (n = 110) compared the effectiveness of two approaches for education delivery. The results of this study provided very-low certainty evidence (downgraded by publication (study size) bias) that telephone education is better than in-person education for QoL SF36 physical (MD= 3.1, p = 0.04) and role domains (MD= 15.7, p = 0.01) but not for the remaining SF36 domains (p > 0.05), pain (VAS, MD= −2.1, p = 1.7) and function (ODI, MD= −6.5, p = 0.1) at the short-term.
Electrophysical modalitiesOne small study (n = 45)70 provided very-low certainty evidence (GRADE, downgraded by publication (study size) bias) that there is no difference between TENS and Acupuncture for pain (NRS, MD= 0.3, 95 %CI −1.06 to 1.07), function (ODI, MD= −3.4, 95 %CI −8.5 to 1.9), and quality of life (EQ5D mean, MD= 0.1, 95 %CI −0.01 to 0.27) in the long-term.
Multimodal therapiesTwo studies (n = 182)43,44 used multimodal approaches, which were based on the delivery of at least two distinct therapies to patients allocated in the intervention group. There is very-low certainty evidence of conflicting results on the effectiveness of multimodal interventions compared with standard obstetric care. One study44 showed that combining manual therapy, exercise, advice and standard care is better than standard care alone for pain (NRS, MD= −2.7, p < 0.001), physical function (QDQ, MD= −1.4, p < 0.001) and global perceived effect (GPE, MD= −1.7, p < 0.001) at the long-term. On the other hand, another study43 showed no difference for pain (NRS, MD= −1.6, p = 0.86), physical function (ODI, MD= −0.9, p = 0.8), QoL (EQ5D, MD= −3.3, p = 0.56), and sick leave (−15 %, p = 0.36) when comparing the combination of manipulation, mobilization, soft tissue treatment, exercise and education with standard obstetric care at the long-term.
DiscussionThis review analyzed 31 studies to assess the effectiveness of physiotherapy interventions, including exercise, manual therapy, TENS, education, and multimodal approaches for pregnancy related LPP. Most studies had a high risk of bias, and exercise protocols were often poorly described. Overall, certainty of evidence ranged from very low to moderate, with exercise showing better outcomes than standard care for pain, physical function, and QoL in some but not all time points. However, our subgroup analysis by exercise type showed that general exercises and core strengthening are not better than standard care to improve function in the short or long-term; likely due to the smaller sample sizes of the subgroup analysis leading to wider confidence intervals. It is worth noting that when results were statistically significant for pain and physical function, the effect sizes were large to moderate and are likely clinically significant.
An important finding of this review is that positive outcomes from exercise were observed even when interventions were very short and within a very short follow-up time. The true mechanisms of action of exercise are unclear, but its benefits are likely to span beyond biological effects. Recent literature suggests that exercise may have a significant role in addressing the biopsychosocial nature of LPP,19 such as reducing fear of movement and stress.56 The mechanism in which exercise acts to improve outcomes may be due to its ability to regulate the stress system and improve self-efficacy and coping strategies, and thus, allowing short-term improvements in pain intensity. While self-efficacy theory has not been explicitly applied to and observed within exercise interventions for pregnancy-related LPP, recent literature has found an association between LPP and low self-efficacy.78 This finding and the findings from the current review challenge the traditional belief that pregnancy-related LPP is solely an issue of biomechanics and instability, supporting recent literature regarding the use of biopsychosocial intervention for pregnancy-related LPP.56
Although our results demonstrate that exercise was superior to standard care in improving outcomes among individuals with pregnancy-related LPP, there was considerable variability in the effect sizes across trials leading to high statistical heterogeneity and imprecision. This heterogeneity may have arisen from differences in exercise protocol, as well as clinical presentation (e.g., lumbar versus pelvic pain). There are differences in the clinical presentations of pelvic pain and low back pain which likely require different management strategies. Because there was only one study pooled with pelvic pain, our sensitivity analysis was not able to demonstrate potential differences in outcomes between the group. However, future studies including individuals with lumbopelvic pain should provide disaggregate results for the two conditions.
Manual therapy techniques, such as osteopathic manipulation and sacroiliac joint mobilization, as a stand-alone intervention, showed very-low certainty evidence of varied effects on pain and function outcomes depending on the follow-up period and comparator. Previous research also found limited evidence for the recommendation of manual therapies to improve pain and function related to LBP and PGP in pregnancy.20,79,80 This limited evidence is mainly due to the reduced number of studies, small sample size and poor description of the intervention.81 It is worth noting that in some cases, pregnancy may be considered a contra-indication to spine manipulation due to the increased risk of venous thromboembolism and ligament laxity experienced by women in this period.82,83 Although the occurrence of adverse events resulted from the application of manual therapies during pregnancy is not well-described,81 the available evidence show that adverse events are scarce.84
An interesting finding of this review is that exercise provided similar effects to manual therapy (although results were very low certainty). Recent evidence for non-pregnancy related non-specific low back pain suggests no difference between exercise and manual therapy for reduction of pain and physical function in the short term.85 Similar to the results of this review, there were only a small number of studies comparing the interventions, leading to poor certainty of the evidence. Manual therapy is considered a passive intervention and may not address the multifactorial nature of low back pain. Thus, future high-quality studies should evaluate the effectiveness of manual therapy in the long-term.
Results regarding the effectiveness of electrophysical agents and multimodal therapies were conflicted, underscoring the need for further investigation and standardization of treatment protocols. An intriguing result was that one of the included trials found that TENS may be favoured over exercise in the short-term.51 However, this study had a very small sample size and some concerns for risk of bias, suggesting that further studies are needed before conclusions can be made about these findings. Current data on the effectiveness of TENS for the treatment of pregnancy-related LPP is limited, and thus, more research should be conducted on this topic.
The poor reporting of the interventions included in this study is evidenced by the CERT scores. This result reinforces the findings of a recent systematic review showing that only 33.3 % of the CERT items were reported in studies providing exercises for LBP.86 While only a few studies measured exercise adherence,45,52,60,66,75 this has the potential to play a significant role in study outcomes and results. Lower adherence is consistently reported within exercise groups in LPP research,87 and in pregnant individuals this may be due to a lack of time, motivation, or understanding of the intervention. Furthermore, feelings of nausea and fatigue have also been identified as barriers to exercise participation by women who are pregnant.87,88 A recent review highlights that taking into account patient preferences and goals in addition to tailored education is essential to promote exercise adherence during antepartum.89
The strengths of this review include a protocol developed and registered a priori as well as the use of the PRIMSA-P guidelines to facilitate transparent reporting and the CERT guidelines to evaluate the quality of reporting interventions. A significant strength was the use of the Cochrane RoB2 and GRADE tools to evaluate the RoB and summarize the quality of the evidence. Furthermore, this review included trials published in all countries and languages and the application of a robust search strategy developed through consultation with a librarian.
Limitations of this review include the inclusion criteria that were generalized to pregnancy-related LPP, and did not differentiate between LBP and PGP in the analysis of included trials. This limits the specificity of results and does not directly account for studies that included women with only one of these pain disorders. Furthermore, standard care was not specifically defined when creating the inclusion criteria, leading to variability in the comparators of results. The included trials also had large variability in the intervention protocols, including different types, frequency, and duration of exercise. Additionally, most of the included trials described the intervention in poor detail, which made it challenging to compare interventions and make specific recommendations for exercise guidelines. This further limit the generalizability of results, as additional studies are required to determine the optimal parameters for exercise interventions among individuals with pregnancy-related LPP.
ConclusionThis systematic review provides valuable insights into the effectiveness of physiotherapy interventions for pregnancy-related LPP. While exercise interventions show promise in reducing pain and particularly in improving function compared to standard obstetric care, the overall certainty of evidence remains low, emphasizing the need for further high-quality research. Future trials for individuals with pregnancy-related LPP should include a robust methodology that follows CERT reporting guidelines and includes larger sample sizes to strengthen the quality and specificity of physiotherapy and exercise recommendations.
FundingThere was no funding for this study.
The authors have no competing interests.
The authors would like to acknowledge Jessica Manuel, Lauren Murray, Nicole Taylor, and Callie Berlet for their help on this review.
articles
