The protocol of this systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO) with the registration number CRD42022303173. This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) statement and Cochrane recommendations.22

The bibliographical search was conducted in PubMed (MEDLINE), Physiotherapy Evidence Database (PEDro), Scopus; Cochrane Library, and Web of Science (WOS) from inception to 20 November 2022. The Population, Intervention, Comparison, Outcome, and Study design (PICOS) framework was used to define the search strategy. Medical Subjects Headings (MeSH) were used as the keywords in the search strategy: back pain, low back pain, sciatica, low back ache, mechanical low back pain, hip, hip joint, pain measurement, disability evaluation, and clinical trial. The strategies used for each database are shown in Supplementary material – Table S.1. Scopus database was included as a tool for searching gray literature, and a hand search of the reference list of the included studies was performed. Searches were limited to studies in English, French, and Spanish.

The included studies met the PICOS criteria: 1) Population: adults diagnosed with LBP; 2) Intervention: hip-targeted conservative interventions applied in isolation or combined with specific low back interventions; 3) Comparison: specific low back interventions in isolation defined as the use of any physical intervention (manual therapy, massage therapy, exercise therapy, thermotherapy, electrotherapy, or other therapy) that targeted the low back specifically and not the hip joint; 4) Outcomes consisted of pain intensity and disability; 5) Study design: RCTs.

Studies were excluded if they: 1) were trials conducted with animals, cadavers, or simulators; 2) included participants with traumatic injuries or after surgical interventions; 3) the intervention was applied to the whole body; 4) the outcome variables reported were not the outcomes of interest or were not measured using a valid and reliable instrument.

After results of searches were retrieved, references were exported to Mendeley desktop, and duplicates were removed. Two reviewers independently (LC and SJ) assessed the title and abstract of each reference to determine potential eligibility. The same independent reviewers assessed potential full texts. A third author (EE) resolved the discrepancies between the two reviewers. Three authors of original papers were contacted by e-mail to clarify eligibility criteria. The inter-rater agreement was calculated using the Cohen's kappa coefficient.23

The two authors independently extracted the data from the identified studies using the standardized process adapted from the Cochrane Collaboration. Extracted information included: 1) characteristics of the study population; 2) aspects of the intervention performed; 3) outcome measures; 4) results; and 5) follow-up period. The third assessor resolved any disagreements. Data were analyzed using a qualitative synthesis and, whenever possible, using a quantitative synthesis (meta-analysis).

Two assessors assessed the quality of the studies using the Cochrane Risk of Bias tool. The Cochrane Risk of Bias tool determines the potential bias and the internal validity of the studies and classifies them as “low”, “unclear”, or “high” risk based on 7 criteria.24 This tool has shown to be reliable for evaluating the quality of studies and assessing risk of bias. Funnel plot asymmetry to assess publication bias in the meta-analyses was not conducted in this study because the meta-analyses presented did not meet the rule of at least 10 trials.

The GRADEpro GDT was used to develop a summary of the findings. This classification categorizes the evidence as “high”, “moderate”, “low”, or “very low” and allows researchers and clinicians to discern the importance of the results. The certainty of evidence for the meta-analysis was downgraded according to the presence of the following: risk of bias (downgraded by one level or two levels if more than 25% or 50% of the participants were from studies with poor or fair methodological quality: lack of allocation concealment, random allocation and/or sample size calculation, participant, and personnel blinding, blinding of outcome assessors), inconsistency of results (downgraded by one level if there was significant heterogeneity regarding outcome measurement or intervention, or if the I2 value was ≥50%, and two levels if the I2 was ≥75%),25,26 indirectness of evidence (downgraded by one level if different populations, interventions, or comparators were included), and imprecision (downgraded by one level if fewer than 100 participants were included in the comparison, and two levels when the sample sizes were ≤30 individuals).26–28 Single randomized trials were considered inconsistent and imprecise and provided “low certainty” evidence. This could be further downgraded to “very low” certainty evidence if there was also a high risk of bias.24,29

The quantitative synthesis of the results was conducted according to the outcomes considered: pain intensity and disability. When studies used different tools to assess the same outcome, the authors performed inverse variance methods.

Separate analyses were performed for pain intensity and disability. When studies reported pain data on a scale other than 0 to 10 (e.g., 0 to 100) we transformed the data into a 0 to 10 scale.30 Time points of assessment considered were short-term (or post intervention), medium-term (6 months), and long-term (12 months). Mean, standard deviation (SD), and sample size at each time point were extracted for each group. Outcomes were analyzed based on the post-intervention means and SDs by calculating the mean difference (MD) when studies used the same scale or standardized mean difference (SMD) when studies used different scales, with 95% CIs. Significance was set at a P value <0.05. The minimum clinically important change (MCID) on pain intensity was stated as 1.5.31 For disability, the between-group effects size was used to classify the effect estimates as small (SMD at least 0.2 but less than 0.5), medium (SMD from 0.5 to less than 0.8), or large (SMD 0.8 or greater).32

Data were combined for meta-analysis using a minimum of two trials assessed as clinically homogeneous. Trials were considered clinically homogeneous if there was a common intervention and outcome. When a three-arm study was included, the data from the control group were divided.33 Random-effect meta-analysis was performed when the combination of intervention effects could incorporate an assumption that the studies are not all estimating the same intervention effect.34

A researcher analysed data using RevMan 5.4 software.

Eight studies were included in the qualitative and quantitative synthesis. One study was excluded due to lack of information about the comparison group.35 The description of the selection process is shown in the PRISMA flowchart diagram (Fig. 1). The agreement between reviewers was calculated by kappa with a value of 0.9.

Fig. 1.

PRISMA flowchart.

(0.29MB).

A total of 8 RCTs were included comprising 508 patients with LBP. The sample size ranged from 24 to 80 patients.

Most studies included patients with LBP for at least 6 weeks,36 or 3 months.37–40 All the studies included patients with LBP older than 18.36–43 Concerning the LBP diagnosis, 3 studies included patients with non-specific LBP,36,38,39 one study included patients with mechanical LBP,41 and 3 studies did not specify the LBP diagnosis.40,42,43 The sociodemographic and clinical characteristics of the participants in each study are shown in Table 1.

The hip-targeted intervention group (hip-targeted intervention) in each trial consisted of different hip interventions. Seven studies added hip-targeted interventions to specific low back interventions,36–39,41–43 and only one applied hip treatment in isolation.40 Four studies combined hip strengthening with specific low back exercise,36–38,42 one applied hip abductor strengthening in isolation,40 one study combined hip joint stretching with specific low back exercise,38 and 3 studies applied hip joint mobilization and strengthening plus specific low back intervention.39,41,43

Specific low back intervention groups (control) were mainly based on lumbar stabilization exercises in all 8 studies.36–43Table 1 shows the treatments applied in the hip and low-back groups.

Information on the duration of the sessions, the number of sessions per week, and the total number of sessions are presented in the Supplementary material - Table S.2. The sessions duration ranged from 20 to 60 min,37–39,42,43 the most common frequency was 2–3 sessions per week,37–40,42,43 and most of the studies performed a total number of sessions between 18 and 24.37,38,40,42 Three studies did not reported the frequency, duration, and/or the total number of sessions.36,41,43

The outcomes considered in this meta-analysis were pain intensity and disability. Seven studies assessed pain intensity. Five studies used the visual analogue scale (VAS),36–40 and 2 used the numeric pain rating score (NPRS).41,43 Disability was measured for all the studies. Six studies used the Oswestry Disability Index (ODI),36–38,41–43 and 2 used the Roland Morris Questionnaire.39,40

All the studies assessed the outcome variables at baseline and after the intervention (short-term).36–43 Two studies assessed the medium- and long-term results at 6 months and at 12 months of follow-up, respectively.39,43

All RCTs included in this review showed a high risk of selection bias and reporting. Most of the studies correctly performed the random sequence generation, but the concealment allocation was only performed in 3. No study blinded the participants or therapist (this item is expected in conservative non-pharmacological interventions). The Cochrane risk-of-bias tool results are shown in Fig. 2.

Fig. 2.

Risk of bias Cochrane tool.

(0.6MB).

Pain intensity was measured in 7 studies after the intervention,36–41,43 and in 2 studies at 6 and 12 months of follow-up.39,43

Very low certainty evidence showed that hip strengthening plus specific low back exercise provide statistically significant but clinically unimportant improvement in pain intensity compared to specific low back intervention (MD = −0.66; 95% CI −0.86, −0.48; I2: 0%; 3 studies, 153 participants) (Fig. 3A).

Fig. 3A.

Forest plot of pain intensity for hip-targeted interventions in isolation or in combination with LBP interventions versus specific low back treatment in the short-term.

(0.4MB).

Very low certainty evidence showed that hip joint stretching plus specific low back exercise provide statistically significant but clinically unimportant improvement in pain intensity compared to specific low back intervention (MD = −0.55; 95% CI −1.02, −0.08; 1 study, 32 participants) (Fig. 3A).

Very low certainty evidence showed that hip abductor strengthening provide statistically significant but clinically unimportant improvement in pain intensity compared to specific low back intervention (MD = −0.93; 95% CI −1.67, −0.19; 1 study, 24 participants) (Fig. 3A).

Very low certainty evidence showed that hip joint mobilization and hip strengthening plus specific low back intervention provide no better improvement in pain intensity compared to specific low back intervention (MD = −0.15; 95% CI −0.59, 0.30; I2: 0%; 3 studies, 230 participants) (Fig. 3A).

These studies measured pain intensity in the medium-term and long-term.39,43 Very low certainty evidence showed that hip joint mobilization and hip strengthening plus specific low back intervention provide no better improvement in pain intensity compared to specific low back intervention at medium-term (MD: −0.32; 95% CI: −0.96, 0.31; I2: 0%; 2 studies; 146 participants) (Fig. 3B), or long-term (MD: −0.10; 95% CI: −0.65, 0.45; I2: 0%; 2 studies; 146 participants) (Fig. 3C).

Fig. 3B.

Forest plot of pain intensity for hip-targeted interventions combined with LBP interventions versus specific low back treatment in the medium-term.

(0.14MB).

Fig. 3C.

Forest plot of pain intensity for hip-targeted interventions combined with LBP interventions versus specific low back treatment in the long-term.

(0.14MB).

Disability was measured in all the studies after the intervention,36–43 and in 2 studies at 6 and 12 months of follow-up.39,43

Very low certainty evidence showed that hip strengthening plus specific low back exercise provide a large improvement in disability compared to specific low back intervention (SMD = −0.81; 95% CI −1.53, −0.10; I2: 80%; 4 studies, 220 participants) (Fig. 3D).

Fig. 3D.

Forest plot of disability for hip-targeted interventions in isolation or in combination with LBP interventions versus specific low back treatment in the short-term.

(0.4MB).

Very low certainty evidence showed that hip joint stretching plus specific low back exercise provide a large improvement in disability compared to specific low back intervention (SMD = −1.03; 95% CI −1.82, −0.25; 1 study, 32 participants) (Fig. 3D).

Very low certainty evidence showed that hip abductor strengthening provide no improvement in disability compared to specific low back intervention (SMD = −0.77; 95% CI −1.60, 0.07; 1 study, 24 participants) (Fig. 3D).

Very low certainty evidence showed that hip joint mobilization and hip strengthening plus specific low back intervention provide no improvement in disability compared to specific low back intervention (SMD = 0.06; 95% CI −0.30, 0.42; I2: 46%; 3 studies, 230 participants) (Fig. 3D).

These studies measured disability in the medium-term and long-term.39,43 Very low certainty evidence showed that hip joint mobilization and hip strengthening plus specific low back intervention provide no improvement in disability compared to specific low back intervention in the medium term (SMD = −0.01; 95% CI; −0.33, 0.31; I2: 0%; 2 studies; 146 participants) (Fig. 3E), or long-term (SMD = 0.05; 95% CI; −0.27, 0.38; I2: 0%; 2 studies; 146 participants) (Fig. 3F).

Fig. 3E.

Forest plot of disability for hip-targeted interventions combined with LBP interventions versus specific low back treatment in the medium-term.

(0.15MB).

Fig. 3F.

Forest plot of disability for hip-targeted interventions combined with LBP interventions versus specific low back treatment in the long-term.

(0.15MB).

According to GRADE, the overall certainty of evidence for all comparisons were rated as very low for pain intensity and disability (Table 2).