Associations between lifestyle-related risk factors and back pain: a systematic review and meta-analysis of Mendelian randomization studies

Abstract

Background

Mendelian randomization (MR) studies have an advantage over conventional observational studies when studying the causal effect of lifestyle-related risk factors on back pain. However, given the heterogeneous design of existing MR studies on back pain, the reported causal estimates of these effects remain equivocal, thus obscuring the true extent of the biological effects of back pain lifestyle-risk factors.

Purpose

The purpose of this study was to conduct a systematic review with multiple meta-analyses on the associations between various lifestyle factors and low back pain.

Methods

We conducted a PRISMA systematic review and specifically included MR studies to investigate the associations between lifestyle factors—specifically, BMI, insomnia, smoking, alcohol consumption, and leisure sedentary behavior—and various back pain outcomes. Each meta-analysis synthesized data from three or more studies to assess the causal impact of these exposures on distinct back pain outcomes, including chronic pain, disability, and pain severity. Quality of studies was assessed according to STROBE-MR guidelines.

Results

A total of 1576 studies were evaluated and 20 were included. Overall, the studies included were of high quality and had a low risk of bias. Our meta-analysis demonstrates the positive causal effect of BMI (OR _{IVW−random effects models}: 1.18 [1.08–1.30]), insomnia(OR _{IVW−random effects models}: 1.38 [1.10–1.74]), smoking(OR _{IVW−fixed effects models}: 1.30 [1.23–1.36]), alcohol consumption(OR _{IVW−fixed effects models}: 1.31 [1.21–1.42]) and leisure sedentary behaviors(OR _{IVW−random effects models}: 1.52 [1.02–2.25]) on back pain.

Conclusion

In light of the disparate designs and causal effect estimates presented in numerous MR studies, our meta-analysis establishes a compelling argument that lifestyle-related risk factors such as BMI, insomnia, smoking, alcohol consumption, and leisure sedentary behaviors genuinely contribute to the biological development of back pain.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-024-07727-0.

Introduction

Back pain is a prevalent and debilitating condition affecting a significant proportion of the global population. It poses considerable challenges both in terms of personal well-being and societal economic burden. Specifically, back pain is defined as pain localized between the costal margins and the inferior gluteal folds, with or without sciatica [1]. It is estimated that approximately 60–70% of adults will experience back pain at some point in their lives, contributing to substantial healthcare costs and productivity losses [2]. When these symptoms endure beyond three months, they transcend mere symptomatic discomfort to manifest as a chronic disorder with multifaceted underlying factors. The Global Burden of Disease study has shed light on the extensive impact of low back pain, consistently ranking it among the top four most burdensome nonfatal conditions worldwide [3,4].

Various lifestyle factors have been implicated in the development and exacerbation of LBP. Biomechanical stresses, arising from poor posture, improper lifting techniques, and sedentary lifestyles, exert undue strain on the spine and supporting structures, precipitating musculoskeletal imbalances and pain [5]. Similarly, lifestyle habits such as physical inactivity, smoking, and obesity have emerged as modifiable risk factors, compromising spinal integrity and exacerbating susceptibility to back pain. Occupational hazards further compound the risk, particularly in industries requiring heavy lifting, repetitive motions, or prolonged periods of sitting or standing. Workers in construction, agriculture, healthcare, and transportation face heightened vulnerability to occupational-related back injuries. Moreover, psychosocial factors, encompassing stress, anxiety, depression, and socioeconomic status, wield significant influence over the perception and management of back pain. Chronic stress and psychological distress can exacerbate symptoms, impeding recovery, while socioeconomic disparities may impede access to effective treatment modalities [5]. However, observational studies investigating these associations are often limited by confounding variables and potential biases, leading to inconsistent findings. Mendelian Randomization (MR) provides a methodological approach to address these limitations. MR uses genetic variants as instrumental variables to assess the causal relationships between modifiable risk factors and health outcomes. This technique leverages the random assortment of genes at conception, which mimics the randomization process in controlled trials, thus reducing confounding and avoiding reverse causation [6,7]. Despite the potential of MR, existing studies using this methodology to investigate the links between lifestyle factors and LBP have reported conflicting results.

In this study, we aim to conduct a systematic review with multiple meta-analyses of MR studies on the associations between various lifestyle factors and low back pain. Through this comprehensive approach, we aspire to deepen our understanding of low back pain etiology and inform targeted interventions for its prevention and management.

Methods

Literature search

This systematic review followed the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines [8,9]. A search, conducted on June 30, 2023, in PubMed and Embase databases, utilized the query “(Mendelian randomization) AND (body mass index OR overweight OR obesity or adiposity OR insomnia OR smoking OR alcohol OR sedentary OR lifestyle OR depression OR anxiety OR sleep OR physical OR exercise OR pain OR ache).” When multiple studies addressed the same outcome and ethnicity, inclusion prioritized the study with the largest number of cases or genetic variants. Moreover, as part of our systematic review process, we conducted an additional search for new literature on June 30, 2023, to ensure the most current data was included. This search confirmed that no new relevant studies meeting our inclusion criteria were published since our initial search.

There were no restrictions based on the number of cases. Detailed inclusion and exclusion criteria for all screened studies are presented in Table 1. The study protocol was registered with PROSPERO (CRD42023452452).

Table 1.

The inclusion and exclusion criteria

Inclusion criteria

1. Any studies that used MR to investigate the causal effect between any risk factor with back pian. 2. Any studies that used genetic variation as a proxy for an exposure to make causal inferences regarding the effect of the instrumental variable on the outcome of back pain.

Exclusion criteria

1. Any studies that were unrelated to risk factors on back pain. 2. Any studies that were not two-sample MR studies. 3. Any case reports, narrative reviews, letters, editorials, opinions, incomplete manuscripts, and conference abstracts. 4. Any studies in which a full English manuscript was not accessible. 5. Animal or basic research studies.

Data extraction and quality assessment

We extracted data from the included studies, as summarized in Table 2. For the population (P) component, our study includes patients or individuals from population-based cohorts, encompassing diverse demographic backgrounds and ethnicities. We examine individuals with varying degrees of back pain severity or duration to capture a comprehensive understanding of the associations between lifestyle-related risk factors and back pain outcomes. In terms of intervention (I), we focus on lifestyle-related risk factors, such as physical inactivity, smoking, obesity, occupational hazards, and psychological factors. These factors are analyzed to determine their causal relationships with back pain outcomes. Regarding comparison (C), we compare individuals exposed to specific lifestyle-related risk factors with those having low or no exposure to these factors. The choice of comparator groups may vary depending on the specific risk factor and outcome under investigation. For outcome (O), our primary focus is on the presence or severity of back pain, chronicity or persistence of symptoms, and functional disability related to back pain. Additionally, we explore secondary outcomes such as specific types of back pain, impact on quality of life, and associations with other musculoskeletal disorders. Lastly, our study design (S) involves Mendelian randomization (MR) studies investigating causal relationships between lifestyle-related risk factors and back pain outcomes. Through meta-analysis of MR studies, we synthesize findings across multiple studies to assess overall causal effects robustly.

Table 2.

Study characteristics of all 20 studies included for qualitative analysis

Study	Year	Ethnicity	Cohort	Genetic Instrument	Exposure	Sample Size	Conclusion
Elgaeva EE et al.[[12]]	2020	European	GIANT consortium	60 SNPs associated with BMI	BMI	N = 322,154	Elevated BMI causally correlates with heightened risks of both high blood pressure (BP) and cardiovascular events (LBP).
Gao et al.[[13]]	2021	European	UK Biobank	9 SNPs associated with moderate-vigorous PA	Moderate-vigorous PA	N = 337,234	The inverse correlation between PA and CBP is likely attributed to decreased physical activity levels among individuals with CBP, rather than indicating a protective effect of PA against CBP.
Zhou et al.[[14]]	2021	European	UK Biobank GIANT consortium	SNPs for BMI and obesity-related traits (WC, HC and WHR)	Traits denoted by SNPs	N(BMI) = 681,275 N(WC) = 232,101 N(HC) = 213,038 N(WHR) = 212,244	Elevated BMI has causal associations with increased risks of LBP, making weight control a valuable measure to prevent its development.
Broberg M et al.[[31]]	2021	European	UK Biobank and 23andMe	231 SNPs used in MR with insomnia	Insomnia	N = 1,331,010 or UK Biobank alone N = 453,379	Insomnia was causally associated with an increased risk of BP.
Shu et al.[[15]]	2022	European	UK Biobank and 23andMe	138 SNPs for insomnia and 38 SNPs for daytime sleepiness	Traits denoted by SNPs	Insomnia: N = 1,331,010 UKBiobank (109,402 cases and 277,131 controls) 23andMe (288,557 cases and 655,920 controls) Daytime sleepiness N = 452,071	Insomnia showed a causal association with an elevated risk of LBP, while no genetic effect of daytime sleepiness on LBP risk was identified.
Tang et al.[[16]]	2022	European	Genetics of Iron Status (GIS) consortium	3 SNPs HFE (rs1799945), HFE (rs1800562), TMPRSS6 (rs855791)	Systemic iron status	N = 48,972	Iron status is causally linked to a higher incidence of high blood pressure (BP).
Luo et al.[[17]]	2022	European	FinnGen consortium	27 insomnia, 4 sleep duration, 10 short sleep duration, 19 long sleep duration and 29 daytime sleepiness SNPs.	Traits denoted by SNPs	13,178 cases and 164,682 controls	While a causal relationship was observed between insomnia and LBP, bidirectional Mendelian randomization analyses between other sleep traits and LBP yielded negative results.
Chen et al.[[18]]	2022	European	UK Biobank	69 BMI, 42 WC, 52 HC and 29 WHR SNPs.	Traits denoted by SNPs	N(BMI) = 322,154 N(WC) = 232,101 N(HC) = 213,038 N(WHR) = 212,244	While positive causal effects of BMI on CBP were observed, no such associations were found with WC and HC.
Tang et al.[[19]]	2022	European	UK Biobank	95 SNPs for MDD	MDD	246,363 cases and 561,190 controls	Depression has a positive causal effect on CBP.
Jiang et al.[[20]]	2022	European	IEU consortium	107 SNPs for serum 25(OH)D levels	Serum 25(OH)D levels	N = 417,580	Elevated serum 25(OH)D levels, influenced genetically, were associated with a decreased risk of LBP.
Williams MK et al.[[21]]	2022	European	UK Biobank	304 education, 133 smoking, 94 AC, 16 PA, 5 LSB, 65 sleep duration, 46 MDD SNPSs.	Traits denoted by SNPs	78,935 cases and 360,896 controls	Fewer years of schooling, smoking, greater AC, and MDD increases the risk of CBP. But There was no evidence of a causal relationship between the, and LSB and CBP.
Lv et al.[[22]]	2022	European	UK Biobank	285 smoking, 68 AC and 13 CC SNPs.	Traits denoted by SNPs	N(smoking)= 1,232,091 N(AC) = 941,280 N(CC) = 375,883	Smoking is casually associated with an increased risk of LBP. In terms of AC and CC, there was no evidence to suggest a causal association on LBP.
Zhou et al.[[23]]	2022	European	CHARGE consortium	6 SNPs for Plasma omega-3 levels ELOVL2(rs3798713, rs3734398, rs2236212), C11orf10(rs174538), GCKR(rs780094), FADS1(rs174547)	Plasma omega-3 levels	N = 8866	A putative causal link between genetically increased plasma omega-3 levels and the reduced risk of low back pain.
Elgaeva EE et al.[[24]]	2022	European Icelandic Danish	UK Biobank	91 SNPs for neuroticism	Neuroticism	N = 380,060	The significant positive feedback loop between neuroticism and back pain.
Zhao et al.[[25]]	2023	European	UK Biobank	19 moderate to vigorous PA, 95 LSB (television viewing) SNPs.	Traits denoted by SNPs	N = 422,218	Leisure television viewing increased the risk of LBP, however, there is no causal associations between genetically PA phenotypes and LBP.
Suri P et al.[[26]]	2023	European	UKBiobank, FinnGen, participants from Iceland and Denmark	244 diastolic blood pressure, 230 systolic blood pressure, 72 LDL-C, 94 HDL-C, 83 total cholesterol, 65 triglycerides and type II diabetes SNPs.	Traits denoted by SNPs	119,100 cases and 909,847 controls	Diastolic blood pressure and systolic blood pressure were significantly associated with BP.
Gou et al.[[27]]	2023	European	FinnGen biobank	261 education, 50 MDD, 458 BMI, 374 WC, 93smoking, 99 AC, 113 LSB, 25 heavy physical work SNPs.	Traits denoted by SNPs	13,178 cases and 164,682 controls	Educational level was negatively associated with LBP. Depression, BMI, WC, smoking, AC, LSB and heavy physical work were positively associated with LBP.
Yao et al.[[28]]	2023	European	IEU consortium	30 insomnia, 12 anxiety and 39 MDD SNPs.	Traits denoted by SNPs	N = 1,137,057	Insomnia, anxiety, and depression are causally related to the genetic susceptibility of BP.
Huang et al.[[29]]	2023	European	UK Biobank	41 SNPs for dried fruit intake.	Dried fruit intake	N = 421,764	Greater dried fruit intake was associated with decreased risk of LBP.
Zhu et al.[[30]]	2023	European	FinnGen biobank	79 LSB (television viewing) SNPs.	LSB	25,163 cases and 248,831 controls	Genetically predicted LSB (television viewing) was a risk factor for LBP.

Abbreviation BP, back pain; LBP, low back pain, CBP, chronic back pain; BMI, body mass index; PA, physical activity; WC, waist circumferences; HC, hip circumferences; WHR, waist-hip ratio; AC, alcohol consumption; CC, coffee consumption; LSB, leisure sedentary behavior; MDD, major depression disorder

Mendelian randomization relies on three key assumptions: Genotype must be linked to the phenotype (in this case, lifestyle-related risk factors). Genotype should not be associated with confounding factors. Genotype should affect the outcome solely through the risk factor. While the first assumption is easily assessed, confirming the second and third assumptions, collectively known as absence of pleiotropy, is challenging and often relies on investigator judgment. Although various statistical tools aim to verify these assumptions and prevent bias from pleiotropic variants, their effectiveness is often limited and may not detect bias in all scenarios. Standardized tools for assessing bias risk in Mendelian randomization studies during meta-analysis are lacking. Hence, in our evaluation of included studies, we assessed validation of the three Mendelian randomization assumptions and the methods employed for validation.

Employing the guidelines from Strengthening the Reporting of Mendelian Randomization Studies (STROBE-MR), we conducted a quality assessment to rigorously evaluate adherence to these guidelines and ensure the robustness of our meta-analysis [10,11]. After converting the quality assessment scores into percentages, we categorized the 20 studies in our analysis based on risk of bias: Scores less than 75% indicated a high risk, scores between 75% and 85% suggested a medium risk, and scores exceeding 85% were associated with a low risk (Table 3) [12–31]. In the absence of standardized tools for assessing bias in Mendelian randomization studies during meta-analysis, our focus shifted to evaluating the quality of the included studies, emphasizing the validation of the three core assumptions of Mendelian randomization. Subsequently, we closely examined the methods employed to validate these assumptions within the selected studies [32].

Table 3.

Quality assessment based on STROBE-MR guidelines

Author	Abstract	Introduction		Methods						Results				Discussion				Other information
	Title and Abstract	Background	Objectives	Study Design and Data Sources	Assumptions	Statistical Methods: Main Analysis	Assessment of assumptions	Sensitivity and Additional Analysis	Software and Pre- Registration	Descriptive Data	Main Results	Assessment of assumptions	Sensitivity and Additional Analysis	Key Results	Limitations	Interpretation	Generalizability	Funding	Data and data sharing	Conflicts of interest	Total Score (out of 20)	Score(%)
Elgaeva EE[[12]]	1	1	1	1	1	1	1	1	0.5	1	1	1	1	1	1	0.5	1	0	1	1	18	90
Gao[[13]]	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	1	1	19	95
Zhou[[14]]	1	1	1	1	1	1	1	0.5	1	1	1	1	1	1	1	1	1	0.5	1	1	19	95
Broberg M[[31]]	1	1	1	1	1	1	0.5	0.5	1	1	1	1	1	1	1	0.5	1	1	1	1	18.5	92.5
Shu[[15]]	1	1	1	1	1	1	0.5	1	1	1	1	0.5	1	1	1	1	0	1	1	1	18	90
Tang[[16]]	1	1	1	1	1	1	1	1	0.5	1	1	1	1	1	1	1	1	0.5	1	1	19	95
Luo[[17]]	1	1	1	1	1	1	0.5	1	1	1	1	1	0.5	1	1	1	1	0	1	1	18	90
Chen[[18]]	1	1	1	1	1	1	1	1	1	1	1	0.5	1	1	1	1	1	0	1	1	18.5	92.5
Tang[[19]]	1	1	1	1	1	1	0.5	1	0.5	1	1	1	1	1	1	1	1	0.5	1	1	18.5	92.5
Elgaeva EE[[24]]	1	1	1	1	1	1	1	0.5	1	1	1	1	1	1	1	0.5	0	1	1	1	18	90
Williams MK[[21]]	1	1	1	1	1	1	0.5	0.5	1	1	1	1	1	1	1	1	1	0	1	1	18	90
Lv [[22]]	1	1	1	1	1	1	0.5	1	0.5	1	1	0.5	1	1	1	1	1	0.5	1	1	18	90
Jiang[[20]]	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	1	1	1	19	95
Zhou[[23]]	1	1	1	1	1	1	1	1	1	1	1	1	0.5	1	1	1	1	0.5	1	1	19	95
Zhao[[25]]	1	1	1	1	1	1	0.5	1	0.5	1	1	1	1	1	1	1	1	0.5	1	1	18.5	92.5
Suri P[[26]]	1	1	1	1	1	1	0.5	1	1	1	1	0.5	1	1	1	1	1	0.5	1	1	18.5	92.5
Gou[[27]]	1	1	1	1	1	1	1	0.5	0.5	1	1	1	1	1	1	1	1	1	1	1	19	95
Yao[[28]]	1	1	1	1	1	1	0.5	1	1	1	1	1	1	1	1	0.5	1	0.5	1	1	18.5	92.5
Huang[[29]]	1	1	1	1	1	1	1	1	0.5	1	1	1	1	1	1	1	1	1	1	1	19.5	97.5
Zhu[[30]]	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0.5	1	1	1	1	19.5	97.5

Each item is scored between 0 and 1 for each criterion to yield a total score. Upon conversion of the quality assessment score to a percentage, scores of < 75%, 75–85% and > 85% were considered to indicate high, medium and low risk of bias, respectively

Statistical analysis

We conducted a meta-analysis by pooling data from studies that met the criterion of having at least three separate investigations evaluating the causal effect between the factor or genetic instrument and back pain. Meta-analyses were performed using the ‘meta’ package in RStudio (Version 2023.06.1 + 524, RStudio, PBC, Boston, MA, USA). In our analysis, we combined data from studies that utilized the same risk factor and Mendelian randomization (MR) technique. The primary outcomes focused on the association between genetic risk and the risk of experiencing back pain, reported as odds ratios (ORs) with corresponding 95% confidence intervals. To assess heterogeneity, we employed the I² statistic, interpreting a value between 25% and 50% as mild heterogeneity, 50–75% as moderate, and exceeding 75% as severe heterogeneity among the included studies. This statistical measure allowed us to gauge the variability in results and assess overall consistency [33]. A two-sided p-value of < 0.05 was considered statistically significant. Data extraction was performed by one investigator (GJB) and checked by another investigator (YKT).

Results

Study characteristics

The database searches returned a total of 4659 results, of which 3122 duplicates were removed to yield 1537 unique records (Fig. 1).

Fig. 1.

Preferred Reporting Items of Systematic Review and Meta-analyses (PRISMA) flow diagram

Meta-analysis of causal risk factors for back pain

A total of 4 studies for BMI [12,14,18,27], 3 studies for waist circumference [14,18,27], 3 studies for insomnia [15,17,28,31], 4 studies for major depression disorder [19,21,27,28], 3 studies for smoking [21,22,27], 3 studies for alcohol consumption [21,22,27], 4 studies for leisure sedentary behavior [21,25,27,30] and 3 studies for physical activity [13,21,25] were selected for quantitative analysis, on the basis that there were common risk factors between studies to meta-analyze.

Obesity-related traits

We assessed the causal effect between Body Mass Index (BMI) and Waist Circumferences (WC) on all back pain outcomes using values obtained by the inverse variance weighted (IVW) MR method. This suggested that there was a positive causal effect between BMI and all back pain outcomes under both fixed effects (OR: 1.03 [1.02–1.04]) and random effects models (OR: 1.18 [1.08–1.30]) (Fig. 2). However, there was not a causal effect between WC and all back pain outcomes under random effects (OR: 1.22 [0.99–1.49]), but a causal effect under fixed effects models (OR: 1.03 [1.01–1.04]) (Fig. 3).

Fig. 2.

Forest plot of studies that evaluated the causal effect between BMI and all back pain outcomes using values obtained by the IVW MR method

Fig. 3.

Forest plot of studies that evaluated the causal effect between WC and all back pain outcomes using values obtained by the IVW MR method

Discussion

The systematic review and meta-analysis revealed positive causal effects for BMI, insomnia, smoking, alcohol consumption, and leisure sedentary behaviors on various back pain outcomes. However, waist circumference, major depressive disorder, and physical activity lacked conclusive evidence for causality in all back pain outcomes. These findings align with our qualitative analysis for BMI [12,14,18,27], insomnia [15,17,28,31], and smoking [21,22,27]. Waist circumference (WC) concurred with studies by Zhou et al. [14] and Chen et al. [18], but differed from Gou et al. [27], who reported positive causality. For major depression, our qualitative analysis differed from Tang et al. [19], Williams MK et al. [21], Gou et al. [27], and Yao et al. [28], who reported positive causality. Findings for physical activity aligned with Williams MK et al. [21] and Zhao et al. [25], but differed from Gao et al. [13], who reported negative causality. Leisure sedentary behaviors matched Zhao et al. [25], Gou et al. [27], and Zhu et al. [30], but differed from Williams MK et al. [21], who reported no causality. Alcohol consumption results matched Williams MK et al. [21] and Gou et al. [27], but differed from Lv et al. [22], who reported no causality. Overall, the included studies were of high quality with a low risk of bias.

Obesity-related traits

Obesity poses an increased risk of back pain through various factors, such as heightened load on weight-bearing joints, inflammation, and psychological impairment [34–36]. Individuals with both a painful medical condition and obesity may experience more intense pain compared to non-obese counterparts [37]. For instance, Pudalov LR et al. found that patients with obesity had higher scores on the Pain Disability Index [37]. BMI serves as an indicator for assessing obesity, while WC directly relates to the weight borne by various joints in the waist and back [38,39]. Consequently, a growing body of literature identifies increasing BMI and WC as significant predictors of back pain. Our results support the findings of studies included in our systematic review, indicating a direct causal effect of BMI on back pain [12,14,18,27]. For instance, Elgaeva EE et al. reported a causal association of BMI on back pain, with a 4.65 kg/m² increase in BMI conferring 1.15 times the odds of back pain. A recent meta-analysis by Zhang et al. also estimated an OR of 1.15 (95% CI 1.08–1.21) for incident back pain in overweight versus normal-weight individuals [40]. In our pooled analysis, an OR of 1.18 (95% CI 1.08–1.30) underscores the direct causal effect of BMI on back pain, consistent with evidence from clinical and Mendelian randomization studies.

However, our subgroup analysis of WC did not reveal a causal effect between WC and all back pain outcomes under random effects models (OR: 1.13 [0.97–1.31]). This suggests that the increased likelihood of back pain in obese individuals may not solely result from the accumulation of abdominal fat and increased pressure on weight-bearing joints in the back. Discrepancies with the meta-analysis conducted by Gou et al. [27] may stem from the exclusion of genetic instruments associated with BMI in their study. Our results indicate no causal effect of WC on back pain, emphasizing the importance for obese patients to focus on overall weight loss rather than targeting localized weight loss in the waist and back, as localized efforts may potentially contribute to the occurrence of back pain [41].

Insomnia and major depressive disorder (MDD)

The relationship between insomnia and chronic pain is bidirectional, with insomnia acting as a trigger for chronic pain, and conversely, chronic pain exacerbating insomnia [42]. Previous prospective cohort studies, such as one by Agmon et al. [43], documented a higher risk of developing back pain in adults with insomnia. Ho et al. also found that individuals with insomnia were twice as likely to experience chronic lower back pain compared to those without insomnia [44,45]. Conversely, there are reports showing a positive correlation between daytime sleepiness and the symptoms and frequency of back pain [46], and improvements in sleep problems have been associated with a reduction in lower back pain [47]. A systematic review by Van Looveren et al. [48] over the past decade on sleep and chronic spinal pain revealed a weak to moderate association between sleep parameters and chronic spinal pain, with an escalation in sleep disruptions as pain severity increases. In recent years, sleep traits have become a focus in MR studies. A study by Broberg et al. demonstrated a positive causal relationship between genetic susceptibility to insomnia and lower back pain [31]. Our meta-analysis of MR studies related to insomnia and back pain [15,17,28,31] employed various MR analysis methods, providing robust evidence for the relationship between insomnia and back pain, indicating that insomnia may increase the risk of developing back pain. Notably, treating insomnia significantly improved pain symptoms in patients with lower back pain, as observed in a randomized, double-blind, placebo-controlled trial [49]. Addressing insomnia in patients with chronic spinal pain appears to be a valuable complement to back pain treatment for optimal therapeutic outcomes.

MDD is characterized by alterations in neurobiological pathways involving neurotransmitters such as serotonin, norepinephrine, and dopamine, which play a crucial role in mood regulation and pain perception [50]. Dysregulation of these pathways in individuals with MDD may lead to heightened sensitivity to pain stimuli, amplifying the perception of back pain [50]. Chronic stress associated with MDD can activate the hypothalamic-pituitary-adrenal (HPA) axis, resulting in increased secretion of cortisol and inflammatory mediators, which can exacerbate inflammation and pain in the musculoskeletal system. MDD and back pain often share common risk factors, such as sedentary lifestyle, obesity, and psychosocial stressors. Individuals with MDD may be more prone to adopting unhealthy behaviors, such as physical inactivity and poor posture, which can contribute to musculoskeletal imbalances and increase the risk of back pain [51]. Moreover, psychosocial factors such as perceived stress, anxiety, and depression can interact synergistically with biological mechanisms to exacerbate pain perception and disability in individuals with comorbid MDD and back pain. Previous MR studies have highlighted the role of genetic factors in the risk of chronic pain. Mclntosh AM et al. [52] reported a genetic correlation between chronic pain and major depressive disorder (MDD), a more severe mental illness associated with insomnia. Treatment drugs for insomnia and MDD often share similarities, leading to the consideration of a strong correlation between MDD and back pain [53,54]. Consistent with several previous MR studies [19,21,27,28], our study using the Inverse-Variance Weighted (IVW) analysis method revealed a positive causal effect between MDD and all back pain outcomes under both fixed effects (OR: 1.06 [1.04–1.07]) and random effects models (OR: 1.26 [1.08–1.46]). However, subgroup analysis using the MR-Egger method and the Weighted Median method, based on random effects models (OR: 1.06 [0.98–1.15]), did not provide support for a positive causal relationship between MDD and back pain. This study questions previous MR study conclusions, emphasizing the need for additional relevant MR studies to confirm this relationship.

Smoking and alcohol consumption

Initially, researchers believed there was an association between smoking and back pain, particularly among heavy physical workers [55]. However, the association between smoking and back pain could be confounded due to heavy physical workers both smoking more and experiencing increased physical exposures at work. Conducting relevant research on this relationship becomes challenging, as establishing a causal association between smoking and back pain through RCTs is ethically restrained. In a meta-analysis of observational studies, Shiri R et al. [56] found that both current and former smokers have a higher prevalence and incidence of low back pain than never smokers. In our study, we utilized MR studies to investigate the relationship between smoking and back pain. The pooled analysis of MR studies indicated a positive causal effect of smoking on back pain. Smoking may contribute to back pain through various mechanisms. First, smoking can lead to blood vessel constriction and reduced blood flow, impacting the blood supply to the back muscles [57]. Second, chemicals in tobacco can harm the cartilage and tissues around the spine, causing chronic inflammation [58] that may damage the spine and surrounding tissues [59]. Third, smoking may reduce bone mineral density, increasing the risk of fractures and vertebral injuries [60]. Finally, smoking can potentially impact lung function, leading to muscle fatigue, especially in the back muscles [61].

AC is a prevalent lifestyle-related risk factor associated with various diseases. Previous studies have shown that AC can cause low BMD, increasing the risk of vertebral fractures [62]. Additionally, it affects the nervous system, leading to muscle fatigue and increasing the burden on back muscles, potentially resulting in pain. Despite controversy surrounding whether AC can directly cause back pain [63–65], insufficient evidence from epidemiological studies and a lack of well-designed, AC-specific studies focusing on back pain have contributed to the existing debate. Similar to smoking, establishing a causal association between AC and back pain through RCTs is ethically constrained. Therefore, our analysis relies on existing MR studies, and the pooled analysis of MR studies [21,22,27] found that AC has a positive causal effect on back pain. This study provides valuable insights for guiding the clinical treatment of back pain.

Sedentary behaviors and physical activity

There is increasing evidence indicating that leisure sedentary behaviors (LSB) and physical activity (PA) are associated with various musculoskeletal disorders [66]. However, the causality between LSB/PA and musculoskeletal health remains unknown. LSB, whether in work or leisure time, are linked to a moderate increase in the risk of back pain (BP) in adults, children, and adolescents [67]. The pooled analysis in this meta-analysis also suggests that LSB is a risk factor for LBP. This implies that, whether from clinical or MR studies, reducing LSB remains a primary means of preventing and alleviating back pain.

Despite the long history of exercise being considered a treatment for back pain [68], our study did not demonstrate a negative causal relationship between physical activity and back pain. Williams MK et al. [21] suggested that a possible explanation for this might be that the linearity assumption contrasts with the view that moderate activity is beneficial for CBP, but very high or low activity levels may not be beneficial. Support for this perspective is also found in a single meta-analysis [69]. Moreover, confounding variables such as body weight, occupational factors, psychosocial stressors, and genetic predispositions may also influence the relationship between physical activity and back pain risk. Individuals with lower levels of physical activity may also exhibit other health behaviors or lifestyle factors that mitigate their risk of back pain, such as maintaining a healthy body weight, practicing ergonomic work habits, or utilizing coping strategies for stress management. Therefore, disentangling the effects of physical activity from these confounding factors is essential for accurately assessing its impact on back pain risk. In conclusion, the seemingly counterintuitive finding that lower levels of physical activity are not strongly associated with back pain risk underscores the complexity of this relationship. By exploring the nuances of physical activity measurement, the role of physical conditioning, potential confounding factors, and the need for longitudinal and interventional studies, we can gain a deeper understanding of how physical activity influences back pain risk and inform more effective preventive strategies and interventions.

Strengths and limitation

The article adopts a comprehensive approach to investigate the causal relationships between various lifestyle-related risk factors and back pain, covering factors such as BMI, insomnia, smoking, alcohol consumption, and sedentary behaviors. This comprehensive assessment aims to provide a thorough understanding of the potential impact of these factors. Additionally, our study adheres to established guidelines, including PRISMA for systematic reviews and the STROBE-MR guidelines for quality assessment, ensuring a rigorous and transparent methodology. The use of STROBE-MR guidelines for quality assessment adds strength to the study by evaluating the adherence of included studies to these guidelines, enhancing the credibility of the meta-analysis. The article provides clear and detailed reporting of the methods, statistical analyses, and findings, facilitating reader understanding of the research process and results.

However, our analysis has several limitations. Firstly, due to the nature of a pooled analysis of individual MR studies, a patient-level analysis cannot be conducted. The article does not explicitly address the potential for publication bias, a common concern in meta-analyses, which may lead to an overestimation of effect sizes if studies with significant findings are more likely to be published. Secondly, although we screened for several risk factors associated with back pain in MR studies, it remains unknown which specific risk factor is most strongly associated with back pain. Thirdly, our study evaluates several lifestyle-related risk factors as a whole, precluding a subgroup analysis of specific genetic mutations and their influence on the outcomes of interest. Additionally, while our article assesses heterogeneity using the I² statistic, it lacks a detailed discussion of potential sources of heterogeneity across the included studies. Finally, although MR methods offer advantages over traditional meta-analytic approaches and can provide evidence for associations between lifestyle-related risk factors and back pain, they still rely on modeling experiments and assumptions, necessitating experimental and clinical verification.

Clinical implications and future research

The findings of our article carry significant clinical implications for healthcare professionals, researchers, and policymakers. Firstly, our study underscores the importance of specific lifestyle-related risk factors, including BMI, insomnia, smoking, alcohol consumption, and sedentary behaviors, as potentially modifiable risk factors for back pain. Healthcare providers can leverage this information to identify individuals at higher risk and tailor interventions accordingly. Secondly, given the established causal relationship between BMI and back pain, healthcare professionals should prioritize emphasizing the significance of weight management and healthy lifestyles, particularly for patients with a history of back pain or those at risk. Weight loss programs and dietary interventions may prove beneficial in reducing both the incidence and severity of back pain. Furthermore, the observed association between insomnia and back pain highlights the need for healthcare providers to integrate sleep management into back pain treatment strategies. Evaluating and addressing sleep problems in patients with back pain, coupled with interventions to improve sleep quality, may positively impact pain management outcomes. The identified positive causal effect of smoking on back pain underscores the imperative for smoking cessation programs. Healthcare providers should actively encourage patients who smoke to quit, not only for overall health benefits but also to mitigate the risk of developing or exacerbating back pain. While the relationship between alcohol consumption and back pain is less straightforward, healthcare professionals should be cognizant of this potential association. Patients reporting back pain should be queried about their alcohol consumption habits, and those with excessive alcohol intake may benefit from counseling or intervention. Our study also suggests that leisure sedentary behaviors are associated with a heightened risk of back pain. Encouraging patients to reduce sedentary time and increase physical activity emerges as a viable approach in both back pain prevention and management. The findings underscore the variability in the impact of these lifestyle factors on back pain among individuals. Healthcare providers should adopt individualized treatment plans, considering a patient’s specific risk factors and lifestyle choices.

In conclusion, our article underscores the need for further research in key areas, particularly the relationship between MDD and back pain. These findings serve as a starting point for more in-depth investigations into these intricate relationships. Future research in the field of back pain and its causal factors, as illuminated by this article, should prioritize several important avenues to enhance our understanding and guide clinical practice. Longitudinal studies tracking individuals over time can yield valuable insights into the enduring effects of lifestyle-related risk factors on the development and progression of back pain, contributing to a more definitive establishment of causality. Additionally, conducting RCTs or intervention studies to assess the effectiveness of lifestyle modifications, such as weight loss programs, smoking cessation, and sleep interventions, in reducing the incidence and severity of back pain. Further exploration of genetic variations that may predispose individuals to back pain and interact with lifestyle factors is essential, offering a more nuanced understanding of the genetic component of back pain risk. Investigating the outcomes of multidisciplinary pain management programs that integrate medical, physical, psychological, and lifestyle interventions for individuals with chronic back pain can provide valuable insights into comprehensive treatment approaches. Lastly, the exploration of potential biomarkers for early detection, diagnosis, and monitoring of back pain and its associated risk factors can significantly advance our ability to manage and intervene in this complex condition. These proposed avenues for future research aim to contribute to a more holistic understanding of the multifaceted nature of back pain and inform evidence-based interventions in clinical settings.

Conclusions

The conclusion of this study is that BMI, insomnia, smoking, alcohol consumption, and leisure sedentary behaviors have a significant causal influence on the development of back pain. With established biological underpinnings supporting these associations, future research should focus on elucidating the precise mechanisms by which modifying these risk factors may alter the incidence and progression of back pain.

We then meta-analyzed studies that used the other methods to include all back pain outcomes to assess whether results would support or rebut those from IVW data. This again showed a positive causal effect between BMI and all back pain outcomes under both fixed (OR: 1.03 [1.02–1.05]) and random effects models (OR: 1.13 [1.06–1.20]) (Fig. 4). But there was not a causal effect between WC and all back pain outcomes under random effects models (OR: 1.13 [0.97–1.31]) (Fig. 5). This suggests that they corroborated the positive causal effect of BMI on all back pain, but the causal effect of WC on all back pain was instability

Fig. 4.

Forest plot of studies that evaluated the causal effect between BMI and all back pain outcomes using values obtained by the other methods

Fig. 5.

Forest plot of studies that evaluated the causal effect between WC and all back pain outcomes using values obtained by the other methods

Insomnia and major depression disorder (MDD)

According to the IVW MR method, there was a positive causal effect between insomnia and all back pain outcomes under both fixed effects (OR: 1.21 [1.17–1.26]) and random effects models (OR: 1.38 [1.10–1.74]) (Fig. 6), and there was also a positive causal effect between MDD and all back pain outcomes under both fixed effects (OR: 1.06 [1.04–1.07]) and random effects models (OR: 1.26 [1.08–1.46]) (Fig. 7).

Fig. 6.

Forest plot of studies that evaluated the causal effect between insomnia and all back pain outcomes using values obtained by the IVW methods

Fig. 7.

Forest plot of studies that evaluated the causal effect between MDD and all back pain outcomes using values obtained by the IVW methods

We then meta-analyzed studies that used the other methods to include all back pain outcomes to assess whether results would support or rebut those from IVW data. This again showed a positive causal effect between insomnia and all back pain outcomes under both fixed (OR: 1.14 [1.10–1.19]) and random effects models (OR: 1.16 [1.09–1.23]) (Fig. 8). They corroborated the positive causal effect of insomnia on all back pain. But there was not a causal effect between MDD and all back pain outcomes under random effects models (OR: 1.06 [0.98–1.15]) (Fig. 9). This suggests that they corroborated the positive causal effect of insomnia on all back pain, but the causal effect of MDD on all back pain was instability

Fig. 8.

Forest plot of studies that evaluated the causal effect between insomnia and all back pain outcomes using values obtained by the other methods

Fig. 9.

Forest plot of studies that evaluated the causal effect between MDD and all back pain outcomes using values obtained by the other methods

Smoking and alcohol consumption (AC)

We assessed the causal effect between smoking and AC on all back pain outcomes using values obtained by the IVW MR method. This suggested that there was a positive causal effect between smoking and all back pain outcomes under both fixed effects and random effects models (OR: 1.30 [1.23–1.36]) (Fig. 10). And there was also a causal positive effect between AC and all back pain outcomes under both fixed effects and random effects models (OR: 1.31 [1.21–1.42]) (Fig. 11).

Fig. 10.

Forest plot of studies that evaluated the causal effect between smoking and all back pain outcomes using values obtained by the IVW methods

Fig. 11.

Forest plot of studies that evaluated the causal effect between AC and all back pain outcomes using values obtained by the IVW methods

We then meta-analyzed studies that used the other methods to include all back pain outcomes to assess whether results would support or rebut those from IVW data. This again showed a positive causal effect between smoking and all back pain outcomes under both fixed and random effects models (OR: 1.14 [1.10–1.19]) (Fig. 12), and also showed a positive causal effect between AC and all back pain outcomes under both fixed (OR: 1.34 [1.19–1.50]) and random effects models (OR: 1.33 [1.18–1.50]) (Fig. 13). They corroborated the positive causal effect of smoking and AC on all back pain

Fig. 12.

Forest plot of studies that evaluated the causal effect between smoking and all back pain outcomes using values obtained by the other methods

Fig. 13.

Forest plot of studies that evaluated the causal effect between AC and all back pain outcomes using values obtained by the other methods

Leisure sedentary behavior (LSB) and physical activity (PA)

We assessed the causal effect between LSB and PA on all back pain outcomes using values obtained by the IVW MR method. This suggested that there was a positive causal effect between LSB and all back pain outcomes under both fixed effects (OR: 1.37 [1.23–1.53]) and random effects models (OR: 1.52 [1.02–2.25]) (Fig. 14), and there was not a negative causal effect between PA and all back pain outcomes under both fixed effects and random effects models (OR: 0.97 [0.87–1.07) (Fig. 15).

Fig. 14.

Forest plot of studies that evaluated the causal effect between LSB and all back pain outcomes using values obtained by the IVW methods

Fig. 15.

Forest plot of studies that evaluated the causal effect between PA and all back pain outcomes using values obtained by the IVW methods

We then meta-analyzed studies that used the other methods to include all back pain outcomes to assess whether results would support or rebut those from IVW data. This again showed a positive causal effect between LSB and all back pain outcomes under both fixed (OR: 1.77 [1.49–2.11]) and random effects models (OR: 1.85 [1.46–2.33]) (Fig. 16). And This again showed that there was not a negative causal effect between PA and all back pain outcomes under both fixed and random effects models (OR: 0.094 [0.83–1.08]) (Fig. 17)

Fig. 16.

Forest plot of studies that evaluated the causal effect between LSB and all back pain outcomes using values obtained by the other methods

Fig. 17.

Forest plot of studies that evaluated the causal effect between PA and all back pain outcomes using values obtained by the other methods

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

JBG and KTY designed the systematic review. JBG and TL drafted the protocol, and Ge Gao and KTY revised the manuscript. GJB and TL will independently screen the potential studies, extract data, assess the risk of bias and finish data synthesis. GJB and KTY will arbitrate any disagreements during the review. All authors approved the publication of the protocol. JBG is the first author, and HHL are corresponding author.

Funding

Not applicable.

Data availability

All data generated or analyzed during this study are included in this published article or are available from the corresponding author on reasonable request.

Declarations

Ethical approval

This systematic review and meta-analysis followed the PRISMA guidelines. It involved synthesizing data from previously published studies, without new data collection from human participants. As a result, the ethical approval was not applicable.

Consent for publication

Not applicable.

Consent to participate

Since this study only involved the analysis of previously published data, consent to participate were not applicable.

Competing interests

The authors declare no competing interests.