RISK FACTORS AND CONTEMPORARY MANAGEMENT OF LOW BACK PAIN

Abstract

Low back pain is common and causes more burden in terms of years lived with disability than any other health condition globally. In most cases, the patho-anatomical cause of low back pain cannot be determined. Less commonly, specific spinal pathologies can be identified as the cause of low back pain, including conditions involving neurologic compromise, such as sciatica and lumbar spinal stenosis. Despite extensive research over the past decades, questions remain in terms of the underlying mechanisms, risk factors, and current treatment options for these conditions. The broad aim of this thesis, therefore, is to contribute to a better understanding of factors associated with low back pain onset and the safety and efficacy of contemporary management strategies. Risk factors associated with the onset of a new episode of low back pain can be divided into those involving long-term exposure (e.g., smoking) and those involving transient or brief exposure to the risk factor (e.g., a fall). A recent case-crossover study identified that commonly endorsed physical and psychosocial triggers (e.g., awkward postures, distracted during an activity) increase substantially the risk of sudden onset low back pain, with odds ratios ranging from 2.7 to 25.0. This study focussed on triggers for an acute episode of low back pain and did not consider the triggers that increased the risk of an episode of longer duration. This is an important issue as most of the costs of low back pain are associated with persistent cases. The study presented in Chapter Two includes the 12-month follow-up of this case-crossover study and examined the association between the previously identified triggers and the risk of a low back pain episode that persisted for greater than six weeks. This study was based on data from 782 patients presenting to primary care clinics for a new episode of low back pain, who were successfully followed-up. Conditional logistic regression models suggested that previously identified psychosocial and physical triggers, such as being fatigued or tired during an activity or manual tasks involving awkward postures, increased the risk of persistent episodes of low back pain, with odds ratios (OR) ranging from 2.9 (95% confidence interval [CI]: 1.3–6.4) to 11.7 (95% CI: 5.4–25.3). The results were similar to those for acute episodes of low back pain, suggesting that controlling exposures to these triggers may prevent not only the cases of low back that resolve within six weeks, but also the cases that persist, which are believed to cause the greatest burden of this condition. While a great proportion of patients with low back pain experience recovery within six weeks, recurrence of low back pain is common. However, estimates of recurrence within one year range from 26% to 84%. Part of this variability can be attributed to different definitions of episodes of low back pain used across studies. Moreover, only a few studies have used appropriate methodology to investigate predictors of recurrence. The study presented in Chapter Three determined the 1-year incidence of recurrence in participants who had recently recovered from an acute episode of low back pain, and identified predictors of future recurrences. This was an inception cohort study with 12 months follow-up. Recurrence was defined based on a 12-month recall of a new episode of pain or a new episode of care seeking with data from 469 participants. The 1-year incidence of recurrence of low back pain was 33%, and the recurrence rate for a new episode of care seeking for low back pain was 18%. Multivariable regression analysis revealed that having more than two previous episodes of low back pain increased the odds of a future recurrence by 3.2 (95% CI: 2.1–4.8). This factor was also associated with recurrent episodes of care seeking (OR: 2.9, 95% CI: 1.7–4.8). No other factors were associated with recurrence. This study contributes to the lack of research on recurrence of low back pain. Patients with low back pain seeking primary health care are often recommended paracetamol as the first line analgesic medication. This medicine is also widely used to treat osteoarthritis. However, a randomised trial published in 2014 concluded that paracetamol was ineffective for acute low back pain, and there was also conflicting evidence for its use in osteoarthritis. The systematic review with meta-analysis of randomised placebo-controlled trials presented in Chapter Four investigated the safety and efficacy of paracetamol in patients with low back pain, as well as neck pain, or osteoarthritis. Searching eight databases revealed 13 trials that met the inclusion criteria. Pain and disability scores were converted to a 0 to 100 scale, and a 9-point threshold was used to define smallest worthwhile effect. Pooling showed no effects of paracetamol on pain (mean difference [MD]: –0.5, 95% CI: –2.9 to 1.9) or disability (MD: 0.4, 95% CI: –0.9 to 1.7) for acute low back pain. No trials investigated the effects of paracetamol for patients with neck pain. Paracetamol had small and not clinically important effects for osteoarthritis in pain relief (MD: –3.7, 95% CI: –5.5 to –1.9) or disability reduction (MD: –2.9, 95% CI: –4.9 to –0.9). Patients taking paracetamol were 3.8 times (95% CI: 1.9– 7.4) more likely to have abnormal test results of liver function compared with placebo. The results of this systematic review support the reconsideration of recommendations to use paracetamol for these conditions. The study was published with an editorial and has received various prizes, including the BMJ 1st prize for the most accesses in 2015. The impact of withdrawing recommendations for paracetamol from clinical guidelines of low back pain is that the use of nonsteroidal anti-inflammatory drugs (NSAIDs), second line analgesic, is set to increase. A comprehensive review and appraisal of the literature on the efficacy and safety of NSAIDs was therefore paramount. Moreover, the effects of NSAIDs for some forms of spinal pain, such as acute low back pain and neck pain, remain uncertain. Chapter Five, therefore, presents a systematic review with meta-analysis of randomised placebo-controlled trials that aimed to determine the efficacy and safety of NSAIDs for low back pain, as well as neck pain, with or without radicular pain. Systematic searches were conducted in five large databases and 35 randomised trials were included in the review. Pain and disability outcomes were converted to a 0 to 100 scale, and a between-group difference of 10 points was used as the smallest worthwhile effect. Numbers needed to treat were also calculated providing the number of participants treated with NSAIDs who would achieve a clinically important pain reduction compared with placebo. Pooling revealed that for every six participants (95% CI: 4 to 10) treated with NSAIDs, only one would benefit from it, considering a between-group difference of 10 points (i.e., compared with placebo) for clinical importance in the short-term. Moreover, only in three of the 14 analyses looking at different types of spinal pain, outcomes, or time points were the pooled treatment effects marginally above our threshold for clinical importance. Additionally, taking NSAIDs increased the risk of developing gastrointestinal adverse events by 2.5 times (95% CI: 1.2–5.2). The initial management of low back pain usually focuses on conservative treatments, including analgesic medications. When conservative treatments are unsuccessful, surgery may be considered. Sciatica is a common indication for spine surgery, but at present the clinical course of this condition following surgery remains largely unknown. Therefore, the systematic review with meta-analysis of cohort studies presented in Chapter Six investigated the clinical course of pain and disability in patients who had surgery for sciatica. The searches were conducted in three large databases and 40 publications (39 cohort studies) were included. Pain and disability scores were converted to a common 0 to 100 scale and modelled as a function of time. Generalised estimating equations revealed that the pooled mean leg pain intensity before surgery was 75.2 (95% CI: 68.1 to 82.4) and the mean disability was 55.1 (95% CI: 52.3 to 58.0). Pooled mean leg pain (15.3, 95% CI: 8.5 to 22.1) and disability (15.5, 95% CI: 13.3 to 17.6) reduced substantially after three months. At five years, patients still reported moderate levels of leg pain (21.0, 95% CI: 12.5 to 29.5) and disability (13.1, 95% CI: 10.6 to 15.5). These findings suggest that patients with sciatica experience rapid improvements in the first three months after surgery, but are not likely to experience full recovery (i.e., absence of pain or disability) in the long-term. Lumbar spinal stenosis is the fastest-growing indication for spine surgery among older people. However, surgeons usually rely on their own preferences to decide on the best surgical technique for their patient. The systematic review and meta-analysis presented in Chapter Seven investigated the efficacy of surgery for lumbar spinal stenosis, and the effectiveness of various surgical options for this condition. The searches conducted on seven databases revealed limited evidence, as no surgical placebo-controlled trials were found. The 24 randomised trials included in the review compared various surgical options for lumbar spinal stenosis. Pain and disability scores were converted to a 0 to 100 scale. Pooling suggested that fusion offered no additional benefits over decompression surgery alone on pain (MD: –0.3, 95% CI: –7.3 to 6.7) or disability (MD: 3.3, 95% CI: –6.1 to 12.6). The interspinous process spacers alone were not more effective than conventional decompression in pain relief (MD: –0.6, 95% CI: –8.1 to 7.0) or disability reduction (MD: 1.3, 95% CI: –4.5 to 7.0), but showed small effects when compared with decompression plus fusion on disability (MD: 5.7, 95% CI: 1.3 to 10.0). This review was originally published in PLoS ONE in 2015, but has since then been updated and published in the Cochrane Database of Systematic Reviews, presented in this thesis as an appendix. The updated results provide current evidence on the surgical options for lumbar spinal stenosis, and could be used to guide clinical decision-making in this contentious area. Even though the effects of surgical procedures for patients with lumbar spinal stenosis remain unclear, the rates of fusion procedures have increased in the United States in recent times. It is unknown, however, whether these trends are happening elsewhere. Moreover, further information on complications could better inform surgeons, referring physicians, and patients about risks of surgical procedures. The population-based health record linkage study presented in Chapter Eight determined the trends in hospital admission and surgery for lumbar spinal stenosis in Australia, and investigated associated complications and health care use. The Centre for Health Record Linkage was used to link data of admissions, discharges, and transfers records from all public and private hospitals in New South Wales between 2003 and 2013. In one decade, the age-standardised rate of hospital admissions for lumbar spinal stenosis increased from 34.8 to 39.3 per 100,000 people. In 2013, the total costs for lumbar spinal stenosis were AU $46.1 million. Decompression rates increased from 19.0 to 22.1 per 100,000 people during 2003–2013, while the rates of simple fusion doubled, from 1.3 to 2.8 per 100,000 people. The most significant increase, however, occurred for complex fusion, from 0.6 to 2.4 per 100,000 people – a 4-fold increase in the same 10-year period. Mean hospital costs with decompression surgery were AU $12,168, while simple and complex fusion cost AU $30,811 and AU $32,350, respectively. Complex fusion procedures increased the odds of major complications by 4.1 (95% CI: 1.7–10.1) compared with decompression alone. This study confirms that in Australia the number of complex fusion procedures is increasing at a much faster rate than any other surgical procedure for lumbar spinal stenosis, though it is associated with increased risk of major complications and resource use. Overall, the studies presented in this thesis provide a substantial contribution to the understanding of the mechanisms and risk factors of low back pain. The identification of transient risk factors for persistent low back pain could help develop better preventive strategies. Although a great proportion of patients experience recovery within six weeks, it is now clear that a third is expected to have a recurrence, with multiple previous episodes being the only significant predictor of future recurrences. This thesis also contributes to a better understanding of current management strategies for low back pain. Paracetamol is ineffective for acute low back pain, but NSAIDs provide small effects in pain relief and disability reduction. Recommendations in clinical practice guidelines on pharmacological interventions should be reviewed. Although patients refractory to conservative treatments are frequently referred to surgery, the postoperative clinical course of sciatica is not as favourable as previously thought. Furthermore, despite the lack of evidence on surgical options for lumbar spinal stenosis, fusion surgery is increasing at an alarming rate in Australia.