Background. Low back-related leg pain (LBLP) is clinically diagnosed as referred leg pain or sciatica. Within the spectrum of LBLP there may be unrecognised subgroups of patients. This study aimed to identify and describe clusters of LBLP patients using latent class analysis (LCA). Methods. The study population were 609 LBLP primary care consulters. Variables from clinical assessment were included in the LCA. Characteristics of the statistically identified clusters were described and compared to the clinically defined groups of LBLP patients. Results. A five cluster solution was optimal. Cluster one (n=104) had mild leg pain severity, no clinical signs suggestive of sciatica and more anxiety. Cluster two (n=122), three (n=188) and four (n=69) represented mild, moderate and severe sciatica in terms of response to clinical assessment items, pain severity and impact on function. Cluster five (n=126) was more difficult to define based on response to clinical assessment items (below knee pain and possible neural tension) and had a similar severe profile to cluster four in terms of high pain, disability, psychosocial factors, work impact and risk of poor outcome; but had longer duration pain and more comorbidities. Cluster three consistently mirrored the profile of the overall group of patients with a clinical diagnosis of sciatica. Cluster one mirrored the referred leg pain group. Conclusion. This is the first study that used LCA to classify LBLP patients, including sciatica. These clusters could represent more homogenous groups that may require different treatment approaches. Further work will describe the clinical course and longer term outcomes of these clusters. No conflicts of interest. S Stynes is supported by an NIHR/CNO Clinical Doctoral Research Fellowship (CDRF-2010-055). Dr Konstantinou is supported by a HEFCE/NIHR Senior Clinical Lectureship. Professor Hay is a NIHR Senior Investigator

Background and objectives. Low back pain (LBP) is a major health challenge globally. Research has identified common trajectories of pain over time. We aimed to investigate whether trajectories described in one primary care cohort can be confirmed in another, and to determine the prognostic value of factors collected 5 years prior to the identification of the trajectory. Methods and results. The study was carried out on 281 patients who had consulted primary care for LBP, at that point completed a baseline questionnaire, and then returned a questionnaire at 5-years follow-up plus at least 3 (of 6) subsequent monthly questionnaires. Baseline factors were measured using validated tools. Pain intensity scores from the 5-year follow-up and monthly questionnaires were used to cluster participants into 4 previously derived pain trajectories (no or occasional mild, persistent mild, fluctuating, persistent severe), using latent class analysis. Posterior probabilities of belonging to each cluster were estimated for each participant. The posterior probabilities for the assigned clusters were very high (>0.90) for each cluster except for the smallest ‘fluctuating’ cluster (0.74). Lower social class (OR 2.9; 95% CI 1.2, 7.0), higher pain intensity (1.6 per unit; 1.2, 2.2), and pain duration greater than 3 years (2.7; 1.0, 7.3), were significantly associated with a more severe trajectory 5-years later, as were higher physical disability, emotional impact of pain, and perception pain will last a long time. Conclusion. LBP trajectories identified previously appear generalizable. These allow better understanding of the long-term course of LBP and effective management tailored to individual trajectories needs to be identified. Conflicts of interest. None. Funding. This work was supported by the following grants: Arthritis Research UK [13413], the Wellcome Trust [083572] and the Medical Research Council Prognosis Research Strategy (PROGRESS) Partnership [G0902393/99558]. Time from NEF was supported by an NIHR Research Professorship (NIHR-RP-011-015). NEF is an NIHR Senior Investigator. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health

Background and purpose. Trajectories of change over time can illustrate the course of pain. Back pain trajectories have previously been identified among adults. Understanding different patterns of back pain among adolescents could shed light on how persistent back pain develops. We aimed to identify and compare trajectories of back pain among adolescents in the general population. Methods and results. This prospective cohort study followed 1,336 adolescents (initially aged 11 years) every three months for three years, using self-completion questionnaires. Latent class growth analysis was used to cluster subjects into distinct trajectories of back pain. Six clusters were identified. Most subjects (78%) had very low likelihood of back pain throughout follow-up. The second largest cluster (10%) had relatively low and falling probability of back pain. Two clusters (4% each) showed increasing probability of back pain. One cluster had a peak likelihood of pain around age 13 (2%). The smallest cluster (1%) had very high probability of back pain throughout follow-up. The more painful trajectories had significantly fewer males, higher levels of depression and somatization, and lower life satisfaction at baseline, but differed little in physical activity or BMI. Conclusions. Over a fifth of adolescents had trajectories indicating back pain, but a minority had persistent pain problems. Poorer baseline psychological status and female gender were linked with more frequently experienced pain. Trajectories indicating rising or persistent pain might be illustrating the beginning of a long-term pain problem. Headache, facial pain and stomach pain trajectories were also identified in this sample; findings were similar to back pain