Introduction

Low back pain (LBP) is the main cause of years lived with disability (YLD) globally [1]. The associated economic burden is high, especially in high income countries such as the USA and Australia [2], with direct medical costs for LBP averaging $315 billion/year from 2012 to 2014 in the USA [3] and $3.4 billion from 2018 to 2019 in Australia [4].

Up to two-third of people report persistent pain at twelve months [5], following an episode of LBP [6]. One-third of those who recover will report another episode within one year of the previous episode [7]. Given the limited efficacy in commonly used treatments for LBP (e.g. paracetamol and anti-inflammatories) [8, 9], attention has shifted towards helping patients to adopt healthy lifestyles and be pain resilient [5]. Healthy and positive lifestyle behaviours such as smoking cessation, improvements in sleep quality, engagement in moderate to high physical activity, healthy BMI, and moderate to low alcohol consumption have been associated with lower risk of chronic LBP (e.g. 11% lower risk in those engaging in adequate leisure time physical activity [10]) and improvements in LBP intensity (e.g. OR 4.30, 95% CI 2.21–8.5 in those with better sleep quality) [11].

Positive emotional health factors are also thought to influence coping responses [12] and the capacity to overcome adversity [13, 14]. Positivity, optimism, and active coping have been found to be associated with less intensity of pain in people with arthritis [15], fibromyalgia [16], and knee osteoarthritis [17], whilst optimism has been associated with less experimental pain intensity in pain-free subjects (mean 46.40/100 vs. 53.59/100) [18] and may be important contributors to LBP resilience.

Resilience is recognised as a dynamic process and an outcome of adaptation and adjustment to adversity [19] resulting in either recovery or sustainability [19, 20]. The concept of resilience is appealing in LBP because it shifts the focus away from the negative aspects of the condition (e.g. high disability, low recovery rates), to positive aspects, potentially explaining why some people recover from, or continue to function well in the presence of LBP.

Familial factors and early environmental factors have been found to explain population differences in obesity [21, 22], sleep quality [23], smoking status [24], alcohol use [25], and physical activity [26], with genetic factors also accounting for 21–67% of the variability of LBP [27]. Twin study designs are instrumental to control for unobserved confounding variables such as genetics and the early familial environment [28, 29] in the relationship between risk or prognostic factors and clinical outcomes. This study, therefore, aimed to evaluate the relationship between lifestyle behaviours and emotional health factors on LBP resilience assessed through recovery and sustainability. Our secondary aim was to confirm the results with a within-pair analysis, controlling for genetic and early environmental factors.

Methods

Study design

This retrospective longitudinal cohort study included participants from the Washington State Twin Registry (WSTR). The WSTR is a community-based registry of twin pairs primarily recruited through the Washington State Department of Licensing (DOL) records. Details about the WSTR’s prior recruitment procedures and additional information are reported elsewhere [30,31,32]. Twins included in this study completed baseline surveys between 2010 and 2018 with follow-up surveys sent to participants every two to three years.

Participants

Data from twins who had completed at least two surveys were included in this study, resulting in 2334 participants (1167 complete twin pairs). Previous exposure to LBP was considered to be the stressor that may or may not lead to participants’ recovery or sustainability. As such, all 1065 participants who reported, at baseline, having LBP that lasted for at least one day in the previous three months were included in the analyses.

This resulted in 1065 participants (487 individual twin members of an incomplete pair and 578 twin members of a complete pair) who had a history of LBP. Follow-up surveys that contained the variables of interest were included, resulting in a mean time between surveys of four years. (Participant flowchart is included as Supplementary digital content.)

Variables used and data collection

Predictors of LBP resilience

We investigated two sets of predictors of LBP resilience: lifestyle factors and emotional health factors (Table 1).

Table 1 Scoring system used for computing lifestyle and emotional health scores
Full size table

Lifestyle factors included were BMI, smoking status, leisure time physical activity, sleep quality, and alcohol consumption (Table 1) and were based on previous studies that investigated the impact of lifestyle behaviours on self-rated health [33], life expectancy [34], chronic diseases [35], LBP [6], and overall health [36, 37].

Emotional health factors included were depressed mood, perceived stress, and active coping (Table 1) and were based on available data and variables that have been shown to impact on psychological resilience [15, 19] and LBP [38, 39].

Each factor was given a score of optimal (two points), intermediate (one point), or poor (zero points) [35], with a total lifestyle behaviour and emotional health score being computed (Table 1). This created a continuous scale spanning from the lowest positive/healthy score with lowest potential for resilience to highest positive/healthy score with greatest potential for resilience [35]. The scale also allows the examination of the impact of the accumulation of positive lifestyle behaviours and emotional health factors rather than individual factors alone [33].

Outcomes of LBP resilience

LBP resilience was assessed through two outcome measures that describe resilience: recovery and sustainability. Recovery is thought to best describe a successful response to an acute struggle, and sustainability to describe a successful response to a more chronic stressor [20] and as such, both outcomes were considered for this analysis (Table 2).

Table 2 Variables used to create the outcomes of recovery and sustainability
Full size table

Recovery was defined as the absence of LBP in the previous 3 months, assessed at follow-up. (\(\overline{x }=\) 4 years, n = 906).

Sustainability was assessed through a function score based on the level of pain interference in participants’ physical or work activities. Participants had to report experiencing LBP in the previous three months, at follow-up, in combination with a function score built utilising three questions from the WSTR survey at follow-up (Table 2). Each question was scored from one to five, and scores were added to provide a total function score ranging from 3 to 15 (n = 375).

Covariates

The following variables were considered for inclusion in the analysis as covariates because they have been associated with both LBP and physical/mental health in previous research: sex [40], age [41], educational attainment [42], annual household income [42], presence of comorbid musculoskeletal conditions [27], and mental health conditions [5], in addition to the time difference between baseline and follow-up survey completion.

Statistical analyses

A descriptive statistical analysis was performed to summarise the data. Logistic regression models were built to assess the relationship between lifestyle behaviours, emotional health factors, and recovery with odds ratios and 95% confidence intervals (CI) used to quantify the association between the predictor and the outcome. Linear regression models were built to assess the relationship between lifestyle behaviours, emotional health factors, and sustainability with a correlation coefficient and 95% CI used to quantify the association between the predictors and the outcome. All analyses were adjusted for dependency of data to accommodate for the use of twin data. Univariate analyses were performed on all potential confounding variables and p-values of < 0.2 were used to determine inclusion in the multivariate model [43]. Secondary confirmatory within-pair analysis was performed when the original models revealed a statistical association at the 0.05 level, by regressing twin-pair differences of predictors on outcomes [44]. Univariate analyses were performed on all previous covariates to determine inclusion in the within-pair model. Only complete twin pairs were included in the confirmatory within-pair analysis which adjusts for genetic and shared environmental factors. Secondary sensitivity analyses were performed on different combinations of lifestyle behaviour categories and emotional health categories. No adjustments for multiple tests in the sensitivity analysis were performed to reduce the likelihood of type 2 errors [45]. Participants with missing data were excluded and a complete case analysis was performed as missing data was considered independent of the outcome.

Statistical analyses were performed using stata SE 16.1. All p-values were two-sided, and CIs were set at 95%. STROBE guidelines were used for accuracy of reporting [46].

Results

Baseline characteristics

Baseline characteristics of the 1065 participants included in the study are summarised in Table 3, whilst study sample resilience outcomes are summarised in Table 4.

Table 3 Study sample demographics
Full size table
Table 4 Study sample resilience outcomes and lifestyle behaviour and emotional health scores
Full size table

LBP resilience: recovery outcome

After adjusting for covariates, there was no relationship between the combined lifestyle behaviour score and the likelihood of people recovering from LBP at follow-up (OR 1.05, 95% CI 0.97–1.15, p = 0.218). In addition, after adjustment, there was no relationship between the cumulative emotional health score with recovery from LBP at follow-up (1.08, 95% CI 0.98–1.19, p = 0.142) (Table 5).

Table 5 Logistic regression of lifestyle behaviour score, emotional health score, and recovery
Full size table

The sensitivity analysis revealed a statistically significant relationship between combinations of fewer lifestyle behaviours and recovery. However, no relationship was found between the fewer combinations of emotional health factors and recovery (Supplementary Table 1).

LBP resilience: sustainability outcome

After adjusting for covariates, there was a positive association between sustainability (i.e. physical function in the presence of pain) and lifestyle behaviours (β 0.20, 95% CI 0.04–0.36, p = 0.013) (Table 6). This twins-as-individuals association was supported by the within-pair analysis (β 1.79, 95% CI 0.05–3.53, p = 0.043).

Table 6 Linear regression of lifestyle behaviour score, emotional health score, and sustainability
Full size table

After adjusting for covariates, there was a positive association between sustainability and emotional health factors (β 0.22, 95% CI 0.00–0.43, p = 0.049) (Table 6), with the association being confirmed by the within-pair analysis (β 0.52, 95% CI 0.09–0.96, p = 0.021).

The sensitivity analysis revealed a relationship between most of the combinations of four lifestyle behaviour categories and sustainability, with the exception of sleep/physical activity/smoking and alcohol. Of the alternative combinations analysed, the strongest relationship was found for sleep/physical activity/BMI/and alcohol consumption (β 0.28, 95% CI 0.08 to 0.48, p = 0.005) (Supplementary Table 2). Only one combination of emotional health factors (perceived stress/depressive thoughts) showed a relationship with sustainability (β 0.47, 95% CI 0.19–0.76, p = 0.001).

Discussion

Findings of this study revealed that greater levels of sustainability in LBP (i.e. maintenance of high levels of function) are associated with an increased number of healthy lifestyle behaviours (adequate sleep and physical activity, optimal BMI, non-smoking, and minimal alcohol consumption) as well as with healthy emotional factors (lack of depressive thoughts, low perceived stress, and resilient coping). This relationship was confirmed when familial and genetic confounders were controlled for in the within-pair analyses. This study also found that there was no relationship between the lifestyle behaviour score or the emotional health score with recovery from LBP.

Resilience describes the process of successful adaptation [19], with physical resilience referring to optimising function [47], or resisting functional decline [48] following a stressor. It embodies factors that impact on recovering from episodes of pain or enable maintenance of function [12], consisting of resilience resources (stable personal characteristics), and resilience mechanisms (thoughts or behaviours that support recovery or function) [12]. Our findings suggest that these positive lifestyle behaviours and positive emotional health factors may be important resilience mechanisms for promoting high levels of function in the presence of LBP.

People’s lifestyles may have a significant impact on their physical health, with engagement in sufficient physical activity, minimal alcohol consumption, non-smoking habits, and healthy diet, associated with a decreased risk of chronic disease [35], increased functional health [49], and reduced incidence of LBP [50]. Sleep was also included in our lifestyle behaviour score given the protective impact of higher levels of sleep quality on the development of LBP [51, 52]. The findings of this study indicate that the healthier lifestyle behaviours people adopt, the more likely they are to maintain high levels of function despite having LBP, suggesting that people with LBP may benefit from accumulating resources that support the maintenance of function.

Positive affect, extraversion [17], and active coping have been shown to be associated with improvements in pain and disability associated with arthritis [53] as well as physical activity in people with osteoporosis [54]. Positive psychological factors have also been shown to correlate with physical functioning and pain tolerance in people with fibromyalgia [16] and knee osteoarthritis [17]. Our emotional health score, which combined the concepts of absence of depressive thoughts, low levels of perceived stress, and high levels of resilient coping, extends this research [16, 17, 53, 54], suggesting the more positive emotional health factors people have, the more likely they are to maintain high levels of function when faced with LBP.

In the current study no relationship was found between the positive lifestyle behaviour score or the positive emotional health score and recovery from LBP. Whilst this may be a limitation of the data that was available (e.g. absence of LBP as a measure of recovery), it may also suggest that for people who experience LBP, maintenance of function despite pain may be more important than the absence of pain. In fact, people with LBP tend not to describe recovery as the absence of pain [55]. Given that LBP is thought to be a long-lasting condition [5], sustainability may therefore be the most effective way to describe how people cope and adapt, potentially informing how LBP should be effectively managed.

The World Health Organisation recognises the importance of empowering people to manage their health and well-being through lifestyle behaviour change [56]. Considering 75% of those with LBP suffer other comorbid conditions such as heart and lung disease and diabetes [57], adopting these positive lifestyle behaviours and emotional health factors may impact not only on their LBP but also their general health and quality of life [58]. Whilst adopting the five positive lifestyle behaviours analysed may be difficult for those with LBP, the findings suggest these changes are likely to be important for maintaining high levels of physical and work-related function. Given the number of sensitivity analyses conducted and the potential for some of the tests to be underpowered, these results should be viewed with caution. However, the results also showed that combinations of four lifestyle behaviours had a greater positive relationship with sustainability than fewer factors, supporting the main finding that greater numbers of positive lifestyle behaviours might be important in people’s capacity to maintain high levels of function in LBP.

Strengths and limitations

This study has several strengths. The use of a sample of twins allowed the control of unobserved confounding variables such as genetics and the early familial environment [28, 29], in addition to the adjustment of observed multiple covariates in the analyses. The longitudinal nature of this study also provides important insights into people’s long-term adaptation and coping with LBP. Follow-up surveys asked participants how they felt on the average day and at most, in the past month to minimise recall bias. Further to this, as resilience changes with time, it is expected that the resilience outcomes relate to how the participants felt at the time of completing the surveys and not in the period between baseline and follow-up.

This study also has some limitations. The lack of representation of participants from lower socio-economic groups and less educated people impacts on the generalisability of the results. The study participants were largely within intermediate to optimal levels of individual lifestyle behaviours which may also impact on the generalisability of the results. In addition, the requirements of the twin analysis meant the within-pair analysis involved small numbers of twin pairs. Despite this, the twin analysis confirmed the robustness of the results obtained in the whole cohort, individual analysis. Whilst the sensitivity analysis was pre-planned, the results should be viewed with caution and future research should include multiple timepoints to more thoroughly assess the impact of positive factors on LBP resilience. Future research should attempt to confirm the results of this study through the implementation of inception cohort studies that use LBP specific outcome measures (e.g. LBP disability) and more frequent follow-ups to reduce possible risk of bias and to address the generalisability. Further analysis of the twin data—that is, the estimation of between-pair, as well as within-pair, associations between lifestyle behaviours and LBP resilience would also allow important conclusions to be drawn about the role of heritability (i.e. genetics), as well as both shared and non-shared environmental factors on LBP resilience.

Clinical implications

The findings of this study provide important clinical implications for targeting lifestyle behaviours and emotional health factors that can impact the sustainability of people who present with LBP. Using the results of the regression model for sustainability (β 0.20, 95% CI 0.04–0.36, p = 0.013), if people with LBP cease smoking, improve their sleep from less than six hrs/night to eight hrs/night, increase physical activity to more than 20 min/day, consume minimal to no alcohol, and improve their BMI from 30 to 25, they may improve their function score on average by two points (out of 15). Using the function scale, this could reduce the impact of pain on their daily lives from a category of “all of the time” to a category of “some of the time”, or from a category of “some of the time” to a category of “none of the time”, with this impact potentially being regarded as clinically meaningful. Moreover, using the results of the regression model for sustainability (β 0.22, 95% CI 0.00–0.43, p = 0.049), improving emotional health from poor to optimal through managing depressive thoughts and feelings as well as decreasing stress levels and improving resilient coping may also reduce the impact of pain on their daily lives from a category of “a little of the time” to a category of “none of the time”, or from a category of “most of the time” to a category of “some of the time”. This study has focussed on positive lifestyle factors and positive emotional health factors that impact on recovery and function with LBP. More importantly, these factors are modifiable and the findings highlight the potential benefits of managing lifestyle behaviours and the emotional well-being of people who present with ongoing LBP as opposed to promoting the use of traditional, medicalised approaches that often offer little to no value to patients with LBP [59].

Conclusion

To our knowledge, this study was the first to investigate the associations between combinations of positive, rather than negative, factors and LBP resilience, assessed through the outcomes of recovery and sustainability. People who adopt optimal lifestyle behaviours such as engaging in moderate physical activity, adequate sleep, non-smoking, minimal alcohol consumption, and maintaining a healthy BMI, as well as positive emotional factors, are more likely to be resilient and have higher levels of function despite their LBP. This has important implications for both clinical guidelines and practice, where encouragement to adopt simple, positive lifestyle behaviours and emotional health factors may have a significant impact on LBP management.