A systematic review on quantifiable physical risk factors for non-specific adolescent low back pain

Abstract

PURPOSE:

The aim of this systematic review was to analyze the results of studies on quantifiable physical risk factors (beyond questionnaires) for adolescent low back pain (LBP).

METHODS:

A systematic search was conducted in Medline (OvidSP), Premedline (PubMed), EMBASE, Cochrane, CINAHL, PEDro and PsycINFO. Cross-sectional, prospective and retrospective English language studies on LBP in adolescents aged 10 to 18 years were included.

Results:

Twenty-two mostly cross-sectional studies were included. Trunk muscle endurance in particular seemed to be associated with adolescent LBP, while a possible association of trunk muscle strength and spinal flexibility was less clear.

CONCLUSION:

There is a need for prospective studies on quantifiable physical risk factors for adolescent LBP. Such studies should focus on back and abdominal muscle endurance, possibly in combination with sagittal spinal mobility, sagittal postural alignment and neurodynamics as possible modifiable risk factors for LBP.

Keywords

Adolescent low back pain physical risk factor systematic review

1. Introduction

Back pain starts early in life and its prevalence increases with age, accelerating in the early teens around the age of 12 to 15 [1, 2, 3] and reaching adult prevalence by the age of 18 [1, 4]. As for low back pain (LBP), an eight-year follow up from adolescence to adulthood showed a fourfold increase in the risk of adolescents with LBP for having LBP in adult life. Thus, it was postulated that the focus of research, prevention and treatment in this area should be changed from the adult to the young population [5].

Although there are several specific pathologies that can result in spinal pain in adolescence, a considerable number of cases of adolescent spinal pain, around one third in a recent study [6], are considered non-specific. As for risk factors for non-specific adolescent LBP, some psychosocial parameters, such as back pain of one or two parents [7, 8] as well as anxiety and depression [7, 9] were identified. Indeed, about 10% of the adolescents with LBP showed an increased probability of having sleep disorders and headaches along with the corresponding psychological problems, possibly linked to a dysregulation of the hypothalamic-pituitary-adrenal axis [10]. As for lifestyle factors, TV consumption, computer use and smoking, but not body weight or fitness level, correlated with adolescent LBP [11, 12, 13, 14]. With regard to physical parameters, controversial results were reported that differed from comparable studies in adults and, thus, suggesting that risk factors for adult LBP cannot be extrapolated to adolescent LBP [15]. So far, no physical risk factors for adolescent spinal pain could clearly be defined [11, 16, 17], and the suggestion that psychosocial factors are more important than physical factors for spinal pain in young populations [18] remains disputed. A recent study, found that although psychosomatic symptoms were most strongly associated with 1-month prevalence of adolescent spinal pain, these were followed by factors from the physical and psychosocial domains. Consequently, these authors suggested that the importance of physical risk factors for non-specific adolescent spinal pain may have been underestimated [19].

This systematic review aimed to determine whether there are quantifiable physical risk factors for LBP in adolescence. It focused on adolescents between 10 and 18 years and included prospective, cross-sectional or retrospective studies that used quantifiable measures beyond questionnaires.

2. Methods

2.1 Search strategy

A systematic literature search of studies on risk factors for back pain in adolescents up to and including September 25, 2015 in Medline (OvidSP), Premedline (PubMed), EMBASE, Cochrane, CINAHL, PEDro and PsycINFO was performed by a professional medical librarian from the local university library. The search was repeated on October 24, 2016. The two resulting additional publications from this repetitive search [19, 20] were included in the discussion section.

In a first step, the search was not restricted to physical risk factors as studies on psychosocial risk factors might have investigated physical risk factors as secondary outcomes. The literature search was also not restricted to LBP because the present study was part of a larger project investigating physical risk factors for any type of adolescent back pain. No limits were applied for the publication date of the articles. Medical subheadings (MeSH) were used as search terms. To find the most recently published studies that have not yet been linked with MeSH, keywords were also searched for in the title or abstract. The search strategies can be seen in Appendix A.

2.2 Inclusion criteria

A study was included if it was a cross-sectional, a retrospective or prospective cohort study in English and investigated back pain in adolescents from age 10 to 18. This age range was chosen because there is some evidence that puberty is a risk factor for back pain in the young [15] with pubertal development starting at 9.5 years for girls and at 10 years for boys [21]. If a study covered a wider age range, it was included only if the participants’ mean age was within the age limits of this review. In addition, to be included in this review, the studies had to use quantifiable measures and not be restricted to questionnaires only. Studies were excluded if they focused on back pain of a specified pathology such as scoliosis, Scheuermann’s disease, spondylolisthesis, disc degeneration, hypermobility, coccydynia, fibromyalgia, posttraumatic or postoperative back pain, radiographic studies., Also excluded were particular populations for example athletes, or disabled children or exclusively on lifestyle factors such as computer use, school bag weight, body weight, and sport activities. Studies that focused on possible genetic background for adolescent LBP were also excluded. Lastly, only studies that explicitly investigated back pain were included, while studies on the sacro-iliac area or on musculoskeletal pain in general were excluded.

Figure 1.

Flow diagram of the study selection process.

2.3 Study selection

In a first step, titles and abstracts of the articles were screened by two authors (TP and BW) according to the inclusion and exclusion criteria listed above. In a second step, the same two authors each screened the full text of the remaining articles for eligibility. The full texts were also retrieved if no abstract was available or the selection could not be made on the basis of the abstract. In two consensus meetings, some discrepancies regarding eligibility of the articles were resolved. A third author (KH) was consulted in case no consensus could be reached for giving the final decision. In a last step, only studies that focused on LBP were selected for the present review. Figure 1 depicts the search flow.

2.4 Quality assessment

Quality assessment was performed by two authors (SR and BW) and a consensus meeting was held after individual ratings to clarify possible disagreements. Raters’ agreement was quantified using Cohen’s kappa, calculated using IBM SPSS Statistics 21. The quality assessment was based on the “Critical appraisal form for quantitative studies” by Law et al. [22]. Because no interventions were followed, questions referring to any intervention were removed; e.g. as done in a comparable review [23]. In their place, question 14 (estimates of random variability of data) was added from the Downs and Black checklist [24]. Moreover, two questions about biases [23] and about the adequate description of the assessments [23, 24] were included (questions 4 and 7, Appendix B). All questions were either answered by YES ( $=$ 1 point) or NO ( $=$ 0 point) except from the question on biases where the scoring was reversed. As the question addressing drop-outs was only applicable to prospective studies, the total quality score was a maximum of 14 points for cross-sectional and retrospective studies and 15 points for prospective studies. The assessment form is shown in Appendix B. According to the guidelines by the Physiotherapy Evidence Database (PEDro), a study was regarded as being of moderate to high quality if it reached at least 60% of the maximum score [25], which was 8.5 points for cross-sectional and retrospective studies and 9 points for prospective studies.

2.5 Data extraction

One author (TP) extracted the following information from each article: study design, number, age and gender of participants; physical risk factors that were investigated; assessments and tests that were used, and the results.

3. Results

3.1 Study selection

The selection process is illustrated in Fig. 1. The database searches resulted in a total of 6703 articles of which 2224 were duplicates. After screening the titles and abstracts for the inclusion and exclusion criteria, 48 articles were considered potentially suitable for this review. After reading the full texts, 25 studies remained that fulfilled the inclusion and exclusion criteria. In a last step, three studies that solely focused on neck pain were excluded. This resulted in 22 studies identified as suitable for this review. Three studies were prospective, 16 cross-sectional and one study was retrospective. Two studies consisted of both a cross-sectional and a prospective part. Three of the 22 studies [26, 27, 28] investigated ‘back pain’ and did not differentiate between lower and upper back, while the remaining focused explicitly on LBP.

3.2 Quality of the selected studies

After the consensus meeting, the two reviewers agreed on the scores of 19 articles. In 3 articles [29, 30, 31], there was a difference of one point in the item covering the report of clinical importance of the study results, which was deemed acceptable. The average of both ratings was used for further analysis. The resulting $\kappa$ value was 0.84 ( $p<$ 0.001). The mean quality score of all 22 studies was 10.1 $\pm$ 2.2 points (range: 5 to 13.5 points). The mean quality score for the prospective studies (including the two studies with a prospective and a cross-sectional part) was 10.6 $\pm$ 1.8 (range: 8 to 13 points) out of 15 points and 10.0 $\pm$ 2.3 (range: 5 to 13.5 points) out of 14 points for the cross-sectional (including the retrospective) studies, respectively. Four cross-sectional studies and one prospective study were below moderate quality according to the PEDro guidelines.

3.3 Quantifiable physical risk factors

The results for all outcome variables are shown in detail in Table 1 and summarized per outcome variable in Table 2. In this text section, due to the large amount of information gained, we will only present summarised information on the main results for the sake of clarity.

3.3.1 Trunk

There was concordant evidence from several cross-sectional studies that reduced endurance of trunk extensor muscles [32, 33, 34] as well as of trunk flexor muscles [31, 34] was associated with LBP. This finding was slightly challenged by one study that found both reduced and greater abdominal and back muscle endurance in girls with LBP [28] and one low-quality study that found no association between abdominal muscle endurance and adolescent LBP [33]. The evidence for trunk strength was conflicting with two studies that investigated different aspects of trunk strength and found controversial results [26, 35]. As for trunk mobility, there was evidence from one prospective study that sagittal range of motion (ROM) might be associated with adolescent LBP [35]. In contrast, when measuring either spinal mobility in flexion or extension, the results were conflicting with several studies using different clinical tests and finding markedly controversial results [31, 34, 45, 36, 37, 38]. High lumbar sagittal mobility combined with low lumbar extension strength (high mobility/strength ratio) emerged as a risk factor for LBP from one prospective study of high quality [37]. Regarding functional trunk stability, the results were controversial [38, 39]. Trunk asymmetry, in contrast, was consistently reported to be a risk factor for LBP in three studies, one of which was cross-sectional [38], one retrospective [8] and one prospective [40].

3.3.2 Lower limb

There was evidence from two studies that lower limb endurance was associated with LBP [32, 33], which was supported by a third study that found an association between lower limb power and LBP in girls [28]. Three studies reported that changes in hamstrings flexibility might be associated with more severe-LBP [34], particularly in boys [28, 41], where both, reduced [28, 41] and enhanced [28] flexibility was associated with LBP. There was, however, one low-quality cross-sectional study that found no association between hamstrings flexibility and severe LBP [42]. For the hip, one cross-sectional study presented decreased hip flexion with knees extended as a risk factor for recurrent LBP [31], while another cross-sectional study reported that decreased hip flexion and internal rotation were associated with LBP in boys [41]. Fur-

Table 1
Characteristics of the studies included and their outcome variables

Author, year	Title	Study design	Participants: Number (m/f)Age	Outcome variables	Main results	Quality score
Astfalck et al., 2010 [32]	A detailed characterization of pain, disability, physical and psychological features of a small group of adolescents with non-specific chronic low back pain	Cross-sectional	N $=$ 56 (28/28) 15.4 $\pm$ 0.5/15.7 $\pm$ 0.5 years (with LBP/without LBP)	Sitting posture (trunk angle/lumbar angle) Trunk extensor endurance (Biering Sorensen test) Thigh muscle endurance (squat)	Significant difference between LBP patients and controls in back muscle endurance and squat endurance	13
Astfalck et al., 2010 [29]	Sitting postures and trunk muscle activity in adolescents with and without nonspecific chronic low back pain.	Cross-sectional	N $=$ 56 (28/28) 15.4 $\pm$ 0.5/15.7 $\pm$ 0.5 years (with LBP/without LBP)	Subclassification by O’Sullivan (based on video) Sitting posture: comparison of spinal angles, pelvic angles, back and abdominal muscle activity (EMG) in usual and slump sitting position	Without subgrouping: no significant differences in sitting posture (usual and slump sitting) between adolescents with and without LBP. After subgrouping: lower activation of the internal oblique muscle in usual sitting position in the LBP group.	13.5
Balague et al., 2010 [35]	The association between isoinertial trunk muscle performance and low back pain in male adolescents.	Cross sectional and prospective over 2 years	N $=$ 95 (95/0) 14.0 $\pm$ 1.7 years Prospective: 11 drop-outs	Unresisted range of motion in the three anatomic axes Isometric and isoinertial trunk strength in flexion, extension, lateral flexion and rotation Fingertip-floor distance Schober test Beighton hypermobility score	Cross-sectional study: No association between trunk performance and LBP Prospective study: ‘Persistent’ LBP group: reduced trunk mobility (higher fingertip-floor distance) than the other groups ‘Never LBP’ and ‘new LBP’ group: lower baseline sagittal ROM than the other groups No association between isoinertial trunk performance and LBP	11
Bernard et al., 2007 [33]	Muscle assessment in healthy teenagers Comparison with teenagers with low back pain.	Cross-sectional	N $=$ 51 (11/40) LBP patients, N $=$ 276 (154/122) controls 14.6 $\pm$ 1.7 years	Back muscle endurance (Biering Sorensen test) Abdominal muscle endurance (Shirado’s test) Quadriceps endurance (Killy’s test) Hip extensor endurance Fingertip-floor distance Heel-cheek distance	Significantly lower endurance of back muscles, quadriceps and hip extensors in adolescents with LBP. No difference in trunk flexor endurance.	$<$ 60%/ 8.5
Burton et al., 1996 [36]	The natural history of low back pain in adolescents.	Prospective over 5 years	N $=$ 216 (approximately 50%/50%) 11.7 years	Lumbar sagittal flexibility (Flexicurve)	Mean flexion mobility significantly decreased in girls. No significant relationship between LBP and sagittal mobility	$<$ 60%/ 9.0
Cudre-Mauroux et al., 2006 [39]	Relationship between impaired functional stability and back pain in children: an exploratory cross-sectional study	Cross-sectional	N $=$ 125 (57/68) Median age $=$ 10 years	Functional stability (Matthiass Test using a new scoring system)	Significant association between Matthiass test and LBP in the last week (OR 1.77, 1.08 to 2.91) No association between Matthiass test and upper back pain in the last week.	$<$ 60%/ 8.5

Table 1, continued
Author, year	Title	Study design	Participants: Number (m/f)Age	Outcome variables	Main results	Quality score
Dolphens et al., 2012 [44]	Sagittal standing posture and its association with spinal pain.	Cross-sectional	N $=$ 1196 (639/557) Boys: 12.6 $\pm$ 0.5 years Girls: 10.6 $\pm$ 0.5 years	Global sagittal alignment (pelvic displacement, trunk lean angle, body lean angle) and local spinopelvic parameters (e.g. number of vertebrae in the lumbar lordosis, vertebral level of apex, pelvic orientation in the sagittal plane) in habitual standing using digital images, inclinometry and accelerometry.	Significant association between LBP lifetime prevalence and pelvic displacement in boys. No association between local spinopelvic parameters and LBP.	12
Dolphens et al., 2013 [45]	Classification system of the normal variation in sagittal standing plane alignment. A study among adolescent boys.	Cross-sectional	N $=$ 639 (639/0) 12.6 $\pm$ 0.5 years	Global sagittal alignment (pelvic displacement, trunk lean angle, body lean angle) and local spinopelvic parameters (e.g. number of vertebrae in the lumbar lordosis, vertebral level of apex, pelvic orientation in the sagittal plane) in habitual standing using digital images, inclinometry and accelerometry.	Cluster analysis: 3 clusters of global alignment: neutral, sway-back, leaning-forward. Significantly higher LBP lifetime and month prevalence in boys with sway-back posture. No association between local spinopelvic parameters and LBP.	11
Dolphens et al., 2014 [46]	Classification system of the sagittal standing alignment in young adolescent girls	Cross-sectional	N $=$ 557 (0/557) 10.6 $\pm$ 0.5 years	Global sagittal alignment (pelvic displacement, trunk lean angle, body lean angle) and local spinopelvic parameters (e.g. number of vertebrae in the lumbar lordosis, vertebral level of apex, pelvic orientation in the sagittal plane) in habitual standing using digital images, inclinometry and accelerometry.	3 postural subtypes, comparable to boys. No association between posture clusters and spinal pain measures.	9
Ebrall, 1994 [30]	Some anthropometric dimensions of male adolescents with idiopathic low back pain.	Cross-sectional	N $=$ 125 (125/0) (38 controls)	13 anthropometric dimensions (e.g. standing height, sitting height, pelvic height, suprapelvic height, trunk length, leg length)	Significantly greater trunk length, sitting height, suprapelvic height and pelvic height in adolescents with LBP.	8.5
Harreby et al., 1999 [42]	Risk factors for low back pain in a cohort of 1389 Danish school children: an epidemiologic study	Cross-sectional	N $=$ 1389 (671/718) 13 to 16 years	Tightness of hamstrings (knee extension deficit when hip in 90 ${}^{\circ}$ flexion)	No association of hamstring tightness and LBP, not even in a subgroup of adolescents with severe LBP	$<$ 60%/ 8.5
Jones et al., 2005 [31]	Biological risk indicators for recurrent non-specific low back pain in adolescents.	Cross-sectional	N $=$ 56 28 with recurrent LBP (15/13) 14.9 $\pm$ 0.7 years 28 matched controls (15/13) 14.9 $\pm$ 0.7 years	Lumbar flexion mobility (modified Schober test) Lateral flexion (side bending) Hip range of motion in flexion (knee extended) Abdominal muscle endurance (sit up test) Sit and reach test	Lumbar sagittal mobility, lumbar mobility in lateral flexion and endurance of abdominal muscles significantly reduced in adolescents with recurrent LBP No association between hip flexion mobility and recurrent LBP.	10.5

Table 1, continued
Author, year	Title	Study design	Participants: Number (m/f)Age	Outcome variables	Main results	Quality score
Kaspiris et al., 2010 [8]	Non-specific low back pain during childhood. A retrospective epidemiological study of risk factors.	Retrospective	N $=$ 692 7.5 to 14 years	Coexisting orthopedic disorders (trunk asymmetry, kyphosis, lordosis, leg-length discrepancy, platypodia, valgus knee)	Significant correlation between LBP and coexisting pediatric orthopedic conditions (no information per each condition given)	$<$ 60%/ 8.5
Merati et al., 2004 [26]	Trunk muscular strength in pre-pubertal children with and without back pain.	Cross-sectional	N $=$ 144 (77/67) 11.9 $\pm$ 0.3 years	Isometric trunk extensor and flexor muscle strength Isokinetic trunk extensor and flexor muscle strength	Peak torque flexor/Peak torque extensor ratio higher in boys with back pain at 90 ${}^{\circ}$ /s angular velocity Authors conclusions: Isometric and isokinetic trunk muscle strength of little importance in back pain occurrence in children	12
Nissinen et al., 1994 [40]	Anthropometric measurements and the incidence of low back pain in a cohort of pubertal children.	Prospective over 2 years	N $=$ 859 (451/408) 11.8 years at baseline	Sagittal spine profile (spinal pantograph) Leg-length inequality Sitting height Trunk asymmetry (forward bending test)	Significant association between sitting height, trunk asymmetry and LBP incidence Authors conclusions: Role of anthropometric measures modest.	10
O’Sullivan et al., 2011 [27]	Association of biopsychosocial factors with degree of slump in sitting posture and self-report of back pain in adolescents: a cross-sectional study.	Cross-sectional	N $=$ 1596 14.1 $\pm$ 0.2 years	Normal sagittal sitting posture (lateral photographs) Back muscle endurance (Biering Sorensen test)	Weak positive association between LBP made worse by sitting and higher degree of slump position Weak negative association between LBP not made worse by sitting and higher degree of slump position	11
Perry et al., 2009 [28]	Fitness, motor competence and body composition are weakly associated with adolescent back pain.	Cross-sectional	N $=$ 1608 (825/783) 14.1 $\pm$ 0.2 years	Back muscle endurance (sustained back extension test) Abdominal muscle endurance (abdominal curls) Limb muscle performance (standing long jump) Hamstrings flexibility (unilateral sit and reach test) Motor competence (McCarron Assessment of neuromuscular development)	Boys: Positive association between both, reduced and enhanced flexibility and a diagnosis of LBP. Girls: Positive association between greater abdominal endurance and LBP in the last month. Positive association between both, reduced and enhanced back muscle endurance and a diagnosis of LBP. Authors conclusions: Small contribution of these factors to logistic regression models.	13

Table 1, continued
Author, year	Title	Study design	Participants: Number (m/f)Age	Outcome variables	Main results	Quality score
Salminen et al., 1992 [34]	Spinal mobility and trunk muscle strength in 15-year-old schoolchildren with and without low-back pain.	Cross-sectional	N $=$ 38 (17/21) with recurrent or continual LBP N $=$ 38 asymptomatic controls 15 years	Lumbar flexion and extension (flexicurve, Schober test) Fingertip to floor distance Lumbar sagittal mobility (side bending) Flexibility of hip flexor muscles and hamstrings (passive straight leg raise test) Dynamic strength and endurance of abdominal muscles (curl-up test) Endurance of back muscles (Biering Sorensen test)	Adolescents with recurrent/continuous LBP: Decreased lumbar extension mobility, decreased hamstrings flexibility, increased flexion mobility (measured by Schober test, not significant when measured with flexicurve), reduced endurance of abdominal and back muscles.	9
Sjölie and Ljunggren, 2001 [37]	The significance of high lumbar mobility and low lumbar strength for current and future low back pain in adolescents.	Cross-sectional and prospective over one year	N $=$ 88 (50/38) 14.7 $\pm$ 0.6 years Prospective: N $=$ 86	Lumbar flexion and extension mobility (modified Schober test) Lumbar extension strength (modification of the Biering Sorensen test),	Cross-sectional and prospective: LBP associated with low lumbar extension strength and high lumbar mobility-extension strength ratio (high lumbar mobility=sum of flexion and extension mobility, low extension strength) Associations significant for the whole group and for the girls, but not for the boys.	13
Sjölie, 2004 [41]	Low-back pain in adolescents is associated with poor hip mobility and high body mass index.	Cross-sectional	N $=$ 88 (50/38) 14.7 $\pm$ 0.6 years	Hip mobility (active abduction, flexion, extension, internal and external rotation), Flexibility of hamstrings (active knee extension test)	Significant association of reduced hip flexion, internal rotation and hamstrings flexibility with LBP in boys	10
Smith et al., 2008 [43]	Classification of sagittal thoraco-lumbo-pelvic alignment of the adolescent spine in standing and its relationship to low back pain.	Prospective over 3 years	N $=$ 766 14.0 $\pm$ 0.2 years	Sagittal thoraco-lumbo-pelvic alignment (lateral standing photographs), 3 angles: Sway angle (C7-Trochanter-Malleolus lat.) Lumbar angle (Trochanter-SIAS-Th12) Trunk angle (Trochanter-Th12-C7)	Adolescents with non-neutral postures: higher odds for all measures of back pain (7 of 15 analyses significant)	11
Wirth et al., 2013 [38]	Spine Day 2012: spinal pain in Swiss school children – epidemiology and risk factors	Cross-sectional	N $=$ 434 (211/223) 10.4 $\pm$ 2.8 years	Trunk stability (Matthiass test) Trunk asymmetry (forward bending test) Spinal mobililty (fingertip-floor distance) Coordination (single leg stance)	Trunk asymmetry increased odds for LBP	9

LBP $=$ low back pain; N $=$ number; SIAS $=$ spina iliaca anterior superior.

Table 2

Summary of the results per outcome variable

Outcome variable	Changes in adolescents with LBP	No changes in adolescents with LBP
Trunk
Endurance of trunk extensors	Reduced in recurrent/continuous LBP [34] Reduced in LBP [32, 33] Reduced/enhanced in girls with LBP [28]
Endurance of trunk flexors	Reduced in recurrent/continuous LBP [31, 34] Reduced/enhanced in girls with LBP [28]	[33]
Trunk strength (isometric, isokinetic, isoinertial)	Increased isokinetic flexion/extension ratio in boys with back pain [26]	No change in isometric and isoinertial strength in flexion/extension, lateralflexion, rotation [35]
Spinal mobility in flexion	Reduced flexion and lateral flexion mobility in LBP [31]Reduced trunk flexion mobility in ‘persistent’ LBP [35] Increased flexion mobility in recurrent LBP [34]	[36–38]
Spinal mobility in extension	Reduced extension in recurrent/continuous LBP [34]	[36,37]
Spinal mobility, sagittal ROM	Reduced in ‘never’ and ‘new’ LBP [35]
Mobility/strength ratio	Low extension strength and high sagittal mobility (increased mobility/strength ratio) in LBP [37]
Functional stability (Matthias test)	Reduced in LBP [39]	[38]
Trunk asymmetry (forward bending)	Associated with LBP [8, 38, 40]
Lower limb
Hip range of motion	Decreased hip flexion with extended knee in recurrent LBP [31] Decreased hip flexion and internal rotation in boys [41]
Hip extensors	Weakness in LBP [33]
Lower limb strength	Quadriceps weakness in LBP [33] Reduced squat endurance in LBP [32] Reduced limb power in girls with LBP [28]
Hamstring flexibility	Reduced flexibility in recurrent/continuous LBP [34] Reduced flexibility in boys with LBP [41] Reduced/enhanced flexibility in boys with LBP [28]	No change in severe LBP [42]
Anthropometric measures and posture
Anthropometric measures	Greater sitting height, pelvic height, suprapelvic height in LBP [30]Sitting height positively associated with LBP incidence [40]
Sitting posture	Weak positive association between greater slump degree and LBP [27]	Decreased thoracic kyphosis and increased lumbar lordosis in LBP only when adolescents were subclassified [29]
Standing posture, global alignment	Non-neutral posture higher odds for LBP [43] Increased pelvic displacement angle and sway back posture in boys [44, 45]	No association in girls [46]
Local spinopelvic parameters		No change in vertebral level of the lumbar apex, pelvic tilt, sacral inclination [44]
Other
Motor competence		No association between neuromuscular development and LBP [28]

Prospective studies are shown in bold. LBP $=$ low back pain; ROM $=$ range of motion.

thermore, hip extensor weakness was found to be a risk factor in one low quality cross-sectional study [33].

3.3.3 Anthropometric measures and posture

Sitting height emerged as a risk factor for adolescent LBP from one cross-sectional [30] and one prospective [40] study, although its importance was minor [40]. For sitting posture, contradictory results were reported. A weak association between greater slump degree and LBP that was observed in one cross-sectional study [27] was contradicted by another study that found differences in sitting posture only when the adolescents were sub-classified [29]. Posture alignment in stance was concordantly stated to be a risk factor for LBP by three studies. Non-neutral alignment emerged from one prospective study [43] and an increased pelvic displacement angle [44] and sway back posture [45] were demonstrated in two cross-sectional studies as risk factors for LBP in adolescence for boys, but not in girls [46]. No association of local spinopelvic parameters, such as vertebral level of the lumbar apex, pelvic tilt or sacral inclination [44], and of general motor competence, assessed by a test for neuromuscular development [28], with adolescent LBP were found.

4. Discussion

The aim of this systematic review was to summarize the current knowledge about quantifiable physical risk factors for LBP in adolescents between 10 and 18 years. Prospective, cross-sectional or retrospective studies that used quantifiable measures beyond questionnaires were included. There is wide consensus in the literature that psychosocial factors are associated with LBP in adolescents [10, 18, 47, 48]. The role of mechanical factors was reported to be minor, although many studies restricted mechanical factors to body weight, school bag weight or physical activity [18]. However, there is some evidence that the importance of physical risk factors might have been underestimated [19]. Indeed, this review revealed that there is conclusive evidence that some quantifiable physical parameters might put adolescents at risk for LBP. This accounts particularly for back muscle endurance concordantly reduced in adolescents with LBP, which was in line with a recent review by Lardon et al. [49]. There was the exception of one study that reported both reduced and enhanced back muscle endurance to be associated with LBP, but only in girls [28]. This study reported, however, small associations of all investigated factors to LBP and was the only one that did not use the Biering-Sorensen test as assessment for back muscle endurance. Similarly, reduced endurance of the abdominal muscles was, with the exception of one low-quality study, associated with LBP in three cross-sectional studies, two of which focused on adolescents with recurrent or continuous LBP. However, as no prospective study was found in this context, no conclusion about causality can be drawn. The relevance of trunk muscle strength became less clear. Only one cross-sectional study found some association between LBP and the ratio of flexion and extension peak torque in boys [26]. Thus, in accordance with the review of Lardon et al. [49], who reported no association between back muscle strength and adolescent LBP, the measurement of trunk muscle strength seems to be of minor importance. Accordingly, the results regarding spinal mobility were rather diverse as were the measurement methods that were used (Schober test, fingertip-floor distance, flexicurve). In a recent prospective study by Aartun et al. [20], both methods, the fingertip-floor distance as well as the Schober test (with dichotomized results), could not predict future LBP. Interestingly, one prospective study of high quality reported a combined measure of strength and mobility, more precisely a high lumbar mobility-extension strength ratio to be predictive for adolescent LBP [37]. However, in fact, the authors conducted a modification of the Sorensen test, which means that the proposed measure is rather a combination of back muscle endurance and spinal mobility. Thus, it can be hypothesized that the decrease in back muscle endurance is a risk for adolescent LBP, which becomes more accentuated when spinal mobility is enhanced. The results of the study by Salminen et al. [34] seem to further support this hypothesis as they found decreased endurance of back muscles and abdominal muscles along with increased flexion mobility in the adolescents with recurrent/continuous LBP. Unfortunately, they did not combine these measures. Trunk asymmetry, assessed by forward bending, was consistently associated with LBP in this review, but did not have a predictive value in the study by Aartun et al. [20]. This test is broadly used in scoliosis screening, and was reported even in this context to have limited positive predictive and sensitivity values [50]. Thus, its application in the field on non-specific adolescent LBP seems of minor value.

Lower limb performance was associated with LBP in three cross-sectional studies. Accordingly, knee extensor strength was significantly smaller in adult runners with chronic LBP when compared to control runners [51]. Furthermore, adult patients with chronic LBP who reported psychologic distress and a high pain level showed reduced quadriceps muscle torque, probably as a result of increased inhibition of muscle activity [52]. Given these findings, a prospective study in adolescents that includes quadriceps strength might be of interest. The situation in terms of hip mobility and hamstrings flexibility seems more complex as the measuring methods differed substantially. Jones et al. [31] tested hip flexion with the knee extended, which is comparable to the passive straight leg test used by Salminen et al. [34] for assessing hamstrings flexibility. Both studies dealt with a subpopulation of adolescents with more severe LBP and found a positive association between reduced flexibility and recurrent LBP. However, it was shown via ultrasonography that the straight leg raise score is not determined by hamstring elongation [53], but rather by altered neurodynamics such as sliding capacity of the nerve roots and the spinal cord in the thoracolumbar region [54]. Thus, it might rather be impaired neural sliding than hamstrings tightness that could be a risk factor for adolescent LBP. The limited importance of hamstrings flexibility is underlined in the present review of two studies by Sjölie [41] and Harreby et al. [42] that measured hamstrings flexibility via knee extension deficit in 90 degrees of hip flexion and found contradicting results. Furthermore, in the study done by Perry et al. the use of sit-and-reach tests for assessing hamstrings flexibility was not recommended [55] or only with reservations [56].

Anthropometric measures and sitting posture seem to play, at most, a minor role in the context of adolescent LBP, while alignment in stance seems more important. In the prospective study by Smith et al. [43], any deviation from the neutral posture (sway, flat, hyperlordotic) was a risk for back pain. The three cross-sectional studies by Dolphens et al. [44, 45, 46] reported sway-back posture to be associated with LBP in boys, while the investigated local spinopelvic parameters, such as apex of lumbar lordosis and pelvic orientation in the sagittal plane did not show a relationship. In contrast, in a recent study of this group, higher lumbar lordotic apex and an increase in pelvis retroversion were the two physical factors with the highest odds for LBP [19]. This corresponds well with findings in adults with chronic LBP, where a greater proportion of patients with a low sacral slope and a long but small lumbar lordosis were reported when compared to healthy controls [57].

4.1 Limitations

Narrowing this review to quantifiable physical risk factors led to some limitations. By limiting the age of the adolescents from 10 to 18 years, we might have missed some interesting studies on slightly older adolescents. Indeed, a recent prospective study in adolescents aged 19 years at baseline, studied proprioceptive control using muscle vibration and found that an ankle-steered control strategy was a risk factor for developing recurrent mild LBP [58]. Furthermore, we excluded studies on physical activity as this was mostly asked via questionnaires. However, some studies measured aerobic capacity as did the study by Perry et al. This study reported an association between greater aerobic capacity and LBP in the last month [28]. Such an association however disappeared when the model was adjusted for back muscle endurance [59]. Thus, a recent review hypothesized that the aerobic capacity could be a proxy measure for back muscle endurance and both measures could be two manifestations of body build with a genetic background [49]. Indeed, the exclusion of genetic studies is another limitation of the present review. Although there are not many such studies, they focus, in a broader sense, also on quantifiable physical risk factors. For example, Skouen et al. found in 1004 adolescents that genetic variants in an adrenergic candidate gene (beta-2 adrenergic receptor) were associated with chronic comorbid neck and low back pain [60]. This might further support the hypothesis of a dysregulated hypothalamic-pituitary-adrenal axis being involved in chronic pain in adolescents [10]. A last limitation that should be mentioned is the focus on correlations between quantifiable physical risk factors and LBP. Correlations do not infer causation. Therefore, it might well be that in longitudinal studies investigating treatment effects through modification of physical risk factors will not result in changes in LBP as seen in adults with LBP [61]. Nonetheless, this review is a first step in identifying modifiable physical risk factors that should be further studied, possibly in prospective studies.

5. Conclusions

This review demonstrated 1) correlations between LBP and back and abdominal muscle endurance, and to a lesser extent, with sagittal spinal mobility, sagittal postural alignment and neurodynamics; and 2) a need for prospective studies of high quality that assess pos- sible causality between physical risk factors and adolescent LBP. Based on the results of this study, it is recommended that such studies focus on back and abdominal muscle endurance, possibly in combination with sagittal spinal mobility, sagittal postural alignment and neurodynamics as possible modifiable risk factors for the development of LBP in adolescence.

Footnotes

Acknowledgments

We thank Martina Gosteli, information specialist at the Main Library of the University Zurich, for conducting the systematic literature search.

Conflict of interest

The authors have no conflict of interest to report.

Appendix

Appendix A: Search strategies

Appendix B: Quality assessment form

Adapted from the “Critical appraisal form for quantitative studies” by Law et al. [22] using the checklist by Downs and Black [24] and the form by Prins et al. [23]:

STUDY PURPOSE	Was the purpose stated clearly?	Yes/No
LITERATURE	Was relevant background literature reviewed?	Yes/No
DESIGN	Was the design appropriate for the study question?	Yes/No
	Were there any biases (random/nonsystematic error or measurement bias/systematic error) that may have influenced the results (apart from reliability and validity of the outcomes)?	Yes/No
SAMPLE	Was the sample described in detail?	Yes/No
	Was sample size justified?	Yes/No
OUTCOMES	Were the methods of outcome measurement described sufficiently?	Yes/No
	Were the outcome measures reliable?	Yes/No
	Were the outcome measures valid?	Yes/No
RESULTS	Results were reported in terms of statistical significance?	Yes/No
	Were the analysis method(s) appropriate?	Yes/No
	Clinical importance was reported?	Yes/No
	Drop-outs were reported?	Yes/No/Not applicable
DATA VARIABILITY	Does the study provide estimates of the random variability in the data for the main outcomes?	Yes/No
CONCLUSIONS AND IMPLICATIONS	Conclusions were appropriate given study methods and results?	Yes/No

TOTAL: Maximal 14 points for cross-sectional studies and 15 points for prospective studies.

References

Jeffries

, Milanese

, Grimmer-Somers

. Epidemiology of adolescent spinal pain: a systematic overview of the research literature. Spine 2007; 32: 2630-7.

Kjaer

, Wedderkopp

, Korsholm

, Leboeuf-Yde

. Prevalence and tracking of back pain from childhood to adolescence. BMC Musculoskeletal Disorders 2011; 12: 98.

Roth-Isigkeit

, Schwarzenberger

, Baumeier

, Meier

, Lindig

, Schmucker

. Risk factors for back pain in children and adolescents. Schmerz 2005; 19: 535-43.

Leboeuf

. At what age does low back pain become a problem? Spine 1998; 23: 228-234.

Hestbaek

, Leboeuf-Yde

, Kyvik

, Manniche

. The course of low back pain from adolescence to adulthood: eight-year follow-up of 9600 twins. Spine 2006; 31: 468-72.

Gennari

, Themar-Noel

, Panuel

, Bensamoun

, Deslandre

, Linglart

, et al. Adolescent spinal pain: The pediatric orthopedist’s point of view. Orthop Traumatol Surg Res 2015; 101: S247-50.

Balague

, Skovron

, Nordin

, Dutoit

, Pol

, Waldburger

. Low back pain in schoolchildren. A study of familial and psychological factors. Spine 1995; 20: 1265-70.

Kaspiris

, Grivas

, Zafiropoulou

, Vasiliadis

, Tsadira

. Nonspecific low back pain during childhood: a retrospective epidemiological study of risk factors. Journal of Clinical Rheumatology: Practical Reports on Rheumatic & Musculoskeletal Diseases 2010; 16: 55-60.

Stanford

, Chambers

, Biesanz

, Chen

. The frequency, trajectories and predictors of adolescent recurrent pain: a population-based approach. Pain 2008; 138: 11-21.

10.

Beales

, Smith

, O’Sullivan

, Straker

. Low back pain and comorbidity clusters at 17 years of age: a cross-sectional examination of health-related quality of life and specific low back pain impacts. J Adolesc Health 2012; 50: 509-16.

11.

Cardon

, Balague

. Low back pain prevention’s effects in schoolchildren. What is the evidence? Eur Spine J 2004; 13: 663-79.

12.

Harreby

, Nygaard

, Jessen

, Larsen

, Storr-Paulsen

, Lindahl

, et al. Risk factors for low back pain among 1.389 pupils in the 8th and 9th grade. An epidemiologic study. Ugeskrift for Laeger 2001; 163: 282-6.

13.

Hestbaek

, Leboeuf-Yde

, Kyvik

. Are lifestyle-factors in adolescence predictors for adult low back pain? A cross-sectional and prospective study of young twins. BMC Musculoskeletal Disorders 2006; 7: 27.

14.

Troussier

, Davoine

, de Gaudemaris

, Fauconnier

, Phelip

. Back pain in school children. A study among 1178 pupils. Scand J Rehabil Med 1994; 26: 143-6.

15.

Lardon

, Leboeuf-Yde

, Le Scanff

, Wedderkopp

. Is puberty a risk factor for back pain in the young? a systematic critical literature review. Chiropr Man Therap 2014; 22: 27.

16.

Balague

, Troussier

, Salminen

. Non-specific low back pain in children and adolescents: risk factors. European Spine Journal: Official Publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society 1999; 8: 429-38.

17.

Burton

, Balague

, Cardon

, Eriksen

, Henrotin

, Lahad

, et al. Chapter 2. European guidelines for prevention in low back pain: November 2004. Eur Spine J 2006; 15(Suppl 2): S136-68.

18.

Watson

, Papageorgiou

, Jones

, Taylor

, Symmons

, Silman

, et al. Low back pain in schoolchildren: the role of mechanical and psychosocial factors. Arch Dis Child 2003; 88: 12-7.

19.

Dolphens

, Vansteelandt

, Cagnie

, Vleeming

, Nijs

, Vanderstraeten

, et al. Multivariable modeling of factors associated with spinal pain in young adolescence. Eur Spine J 2016 Sep; 25(9): 2809-2821.

20.

Aartun

, Hartvigsen

, Hestbaek

. Validity of Commonly Used Clinical Tests to Diagnose and Screen for Spinal Pain in Adolescents: A School-Based Cohort Study in 1300 Danes Aged 11-15 Years. J Manipulative Physiol Ther 2016; 39: 76-87.

21.

Gasser

, Molinari

, Largo

. A comparison of pubertal maturity and growth. Ann Hum Biol 2013; 40: 341-7.

22.

Law

, Stewart

, Pollock

, Letts

, Bosch

, Westmoreland

. Critical review form – quantitative studies. McMaster University: Occupational Therapy Evidence-Based Practice Research Group 1998.

23.

Prins

, Crous

, Louw

. A systematic review of posture and psychosocial factors as contributors to upper quadrant musculoskeletal pain in children and adolescents. Physiother Theory Pract 2008; 24: 221-42.

24.

Downs

, Black

. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J Epidemiol Community Health 1998; 52: 377-84.

25.

PEDro. PEDro Physiotherapy Evidence Database. Available from: http://www.pedro.org.au/english/downloads/pedro-statistics/.

26.

Merati

, Negrini

, Carabalona

, Margonato

, Veicsteinas

. Trunk muscular strength in pre-pubertal children with and without back pain. Pediatr Rehabil 2004; 7: 97-103.

27.

O’Sullivan

, Smith

, Beales

, Straker

. Association of biopsychosocial factors with degree of slump in sitting posture and self-report of back pain in adolescents: a cross-sectional study. Phys Ther 2011; 91: 470-83.

28.

Perry

, Straker

, O’Sullivan

, Smith

, Hands

. Fitness, motor competence, and body composition are weakly associated with adolescent back pain. J Orthop Sports Phys Ther 2009; 39: 439-49.

29.

Astfalck

, O’Sullivan

, Straker

, Smith

, Burnett

, Caneiro

, et al. Sitting postures and trunk muscle activity in adolescents with and without nonspecific chronic low back pain: an analysis based on subclassification. Spine (Phila Pa 1976) 2010; 35: 1387-95.

30.

Ebrall

. Some anthropometric dimensions of male adolescents with idiopathic low back pain. J Manipulative Physiol Ther 1994; 17: 296-301.

31.

Jones

, Stratton

, Reilly

, Unnithan

. Biological risk indicators for recurrent non-specific low back pain in adolescents. Br J Sports Med 2005; 39: 137-40.

32.

Astfalck

, O’Sullivan

, Straker

, Smith

. A detailed characterisation of pain, disability, physical and psychological features of a small group of adolescents with non-specific chronic low back pain. Man Ther 2010; 15: 240-7.

33.

Bernard

, Bard

, Pujol

, Combey

, Boussard

, Begue

, et al. Muscle assessment in healthy teenagers, Comparison with teenagers with low back pain. Ann Readapt Med Phys 2008; 51: 263-83.

34.

Salminen

, Maki

, Oksanen

, Pentti

. Spinal mobility and trunk muscle strength in 15-year-old schoolchildren with and without low-back pain. Spine (Phila Pa 1976) 1992; 17: 405-11.

35.

Balague

, Bibbo

, Melot

, Szpalski

, Gunzburg

, Keller

. The association between isoinertial trunk muscle performance and low back pain in male adolescents. Eur Spine J 2010; 19: 624-32.

36.

Burton

, Clarke

, McClune

, Tillotson

. The natural history of low back pain in adolescents. Spine (Phila Pa 1976) 1996; 21: 2323-8.

37.

Sjolie

, Ljunggren

. The significance of high lumbar mobility and low lumbar strength for current and future low back pain in adolescents. Spine (Phila Pa 1976) 2001; 26: 2629-36.

38.

Wirth

, Knecht

, Humphreys

. Spine Day 2012: spinal pain in Swiss school children-epidemiology and risk factors. BMC Pediatr 2013; 13: 159.

39.

Cudre-Mauroux

, Kocher

, Bonfils

, Pirlet

, Meichtry

, Hilfiker

. Relationship between impaired functional stability and back pain in children: an exploratory cross-sectional study. Swiss Med Wkly 2006; 136: 721-5.

40.

Nissinen

, Heliovaara

, Seitsamo

, Alaranta

, Poussa

. Anthropometric measurements and the incidence of low back pain in a cohort of pubertal children. Spine (Phila Pa 1976) 1994; 19: 1367-70.

41.

Sjolie

. Low-back pain in adolescents is associated with poor hip mobility and high body mass index. Scand J Med Sci Sports 2004; 14: 168-75.

42.

Harreby

, Nygaard

, Jessen

, Larsen

, Storr-Paulsen

, Lindahl

, et al. Risk factors for low back pain in a cohort of 1389 Danish school children: an epidemiologic study. Eur Spine J 1999; 8: 444-50.

43.

Smith

, O’Sullivan

, Straker

. Classification of sagittal thoraco-lumbo-pelvic alignment of the adolescent spine in standing and its relationship to low back pain. Spine (Phila Pa 1976) 2008; 33: 2101-7.

44.

Dolphens

, Cagnie

, Coorevits

, Vanderstraeten

, Cardon

, D’Hooge

, et al. Sagittal standing posture and its association with spinal pain: a school-based epidemiological study of 1196 Flemish adolescents before age at peak height velocity. Spine (Phila Pa 1976) 2012; 37: 1657-66.

45.

Dolphens

, Cagnie

, Coorevits

, Vleeming

, Danneels

. Classification system of the normal variation in sagittal standing plane alignment: a study among young adolescent boys. Spine (Phila Pa 1976) 2013; 38: E1003-12.

46.

Dolphens

, Cagnie

, Coorevits

, Vleeming

, Vanderstraeten

, Danneels

. Classification system of the sagittal standing alignment in young adolescent girls. Eur Spine J 2014; 23: 216-25.

47.

Mikkonen

, Heikkala

, Paananen

, Remes

, Taimela

, Auvinen

, et al. Accumulation of psychosocial and lifestyle factors and risk of low back pain in adolescence: a cohort study. Eur Spine J 2016; 25: 635-42.

48.

Kamper

, Henschke

, Hestbaek

, Dunn

, Williams

. Musculoskeletal pain in children and adolescents. Braz J Phys Ther 2016.

49.

Lardon

, Leboeuf-Yde

, Le Scanff

. Is back pain during childhood or adolescence associated with muscle strength, muscle endurance or aerobic capacity: three systematic literature reviews with one meta-analysis. Chiropr Man Therap 2015; 23: 21.

50.

Deurloo

, Verkerk

. To screen or not to screen for adolescent idiopathic scoliosis? A review of the literature. Public Health 2015; 129: 1267-72.

51.

Cai

, Kong

. Low back and lower-limb muscle performance in male and female recreational runners with chronic low back pain. J Orthop Sports Phys Ther 2015; 45: 436-43.

52.

Verbunt

, Seelen

, Vlaeyen

, Bousema

, van der Heijden

, Heuts

, et al. Pain-related factors contributing to muscle inhibition in patients with chronic low back pain: an experimental investigation based on superimposed electrical stimulation. Clin J Pain 2005; 21: 232-40.

53.

Kellis

, Ellinoudis

, Kofotolis

. Hamstring Elongation Quantified Using Ultrasonography During the Straight Leg Raise Test in Individuals With Low Back Pain. PM R 2015; 7: 576-83.

54.

Rade

, Kononen

, Vanninen

, Marttila

, Shacklock

, Kankaanpaa

, et al. 2014 young investigator award winner: In vivo magnetic resonance imaging measurement of spinal cord displacement in the thoracolumbar region of asymptomatic subjects: part 1: straight leg raise test. Spine (Phila Pa 1976) 2014; 39: 1288-93.

55.

Muyor

, Zemkova

, Stefanikova

, Kotyra

. Concurrent validity of clinical tests for measuring hamstring flexibility in school age children. Int J Sports Med 2014; 35: 664-9.

56.

Mayorga-Vega

, Merino-Marban

, Viciana

. Criterion-Related Validity of Sit-and-Reach Tests for Estimating Hamstring and Lumbar Extensibility: a Meta-Analysis. J Sports Sci Med 2014; 13: 1-14.

57.

Chaleat-Valayer

, Mac-Thiong

, Paquet

, Berthonnaud

, Siani

, Roussouly

. Sagittal spino-pelvic alignment in chronic low back pain. Eur Spine J 2011; 20(Suppl 5): 634-40.

58.

Claeys

, Dankaerts

, Janssens

, Pijnenburg

, Goossens

, Brumagne

. Young individuals with a more ankle-steered proprioceptive control strategy may develop mild non-specific low back pain. J Electromyogr Kinesiol 2015; 25: 329-38.

59.

Bo Andersen

, Wedderkopp

, Leboeuf-Yde

. Association between back pain and physical fitness in adolescents. Spine (Phila Pa 1976) 2006; 31: 1740-4.

60.

Skouen

, Smith

, Warrington

, O’ Sullivan

, McKenzie

, Pennell

, et al. Genetic variation in the beta-2 adrenergic receptor is associated with chronic musculoskeletal complaints in adolescents. Eur J Pain 2012; 16: 1232-42.

61.

Steiger

, Wirth

, de Bruin

, Mannion

. Is a positive clinical outcome after exercise therapy for chronic non-specific low back pain contingent upon a corresponding improvement in the targeted aspect(s) of performance? A systematic review. Eur Spine J 2012; 21: 575-98.

A systematic review on quantifiable physical risk factors for non-specific adolescent low back pain

Abstract

PURPOSE:

METHODS:

Results:

CONCLUSION:

Keywords

1. Introduction

2. Methods

2.1 Search strategy

2.2 Inclusion criteria

2.4 Quality assessment

2.5 Data extraction

3. Results

3.1 Study selection

3.2 Quality of the selected studies

3.3 Quantifiable physical risk factors

3.3.1 Trunk

3.3.2 Lower limb

Table 1 Characteristics of the studies included and their outcome variables

4. Discussion

4.1 Limitations

5. Conclusions

Footnotes

Acknowledgments

Conflict of interest

Appendix

Appendix A: Search strategies

Appendix B: Quality assessment form

References

Table 1
Characteristics of the studies included and their outcome variables