Cohort Profile Update: The Brazilian Longitudinal Study of Adult Health Musculoskeletal (ELSA-Brasil MSK) cohort

Abstract

The Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) is a multicentric cohort set up to investigate the development and progression of chronic diseases, particularly cardiovascular diseases and diabetes, in 15 105 adults aged 35–74 years at baseline (2008-2010). Because highly prevalent musculoskeletal (MSK) disorders were not addressed by ELSA-Brasil, the ancillary study ELSA-Brasil MSK was initiated to investigate their development, progression and adverse health consequences. ELSA-Brasil MSK comprises 2901 active/retired civil servants, both sexes, aged 38 to 79 years old at inception (2012-14). In addition to subclinical and clinical diseases, sociodemographic, lifestyle and psychosocial data collected at concurrent ELSA-Brasil visits, participants are assessed for MSK diseases and symptoms (mainly chronic MSK pain, low back pain and knee/hand osteoarthritis), physical functioning, mechanical risk factors, personal beliefs and behaviour. Collaboration is welcome through research proposals sent to principal investigators at: [http://www.elsa.org.br] or by e-mail to the corresponding author. The Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) is an ongoing multicentre cohort set up to investigate the incidence and determinants of the main non-communicable diseases (NCDs), especially cardiovascular disease (CVD) and diabetes.1,2 The cohort comprises 15 105 active and retired civil servants, aged between 35 and 74 years at inception (2008–10), from teaching and research institutions located in six Brazilian states: Minas Gerais, São Paulo, Rio de Janeiro, Espírito Santo, Bahia and Rio Grande do Sul.1,2 Although ELSA-Brasil does not include the extremely poor and unemployed, the participants present reasonable variability pertaining to socioeconomic status and other demographic factors, and the prevalence of behavioural risk factors and chronic conditions in the cohort are similar to the general Brazilian population.2 ELSA-Brasil cohort surveillance includes annual telephone interviews, clinical events investigation and quadrennial face-to-face assessments. Musculoskeletal (MSK) disorders encompass a group of conditions linked with bodily pain and impaired function. Some of them rank among the top causes of years lived with disability (YLD) and/or have accounted for large increases in risk-attributable disability-adjusted life-years (DALYs) over the past decade.3,4 They are also major sources of health inequalities and health care and societal costs worldwide.5,6 Reducing the burden of MSK disorders to individuals and society requires a clearer understanding of biological, environmental, occupational, psychological and social determinants of the onset and progression of chronic pain and disability. Currently, there is limited evidence from low- and middle-income countries on relevant exposures that potentially shape response to pain and disability and their inter-relationships with age-related changes in the MSK system. Because highly prevalent MSK disorders were not addressed in the original ELSA-Brasil cohort, an ancillary cohort was initiated: the ELSA-Brasil MSK. ELSA-Brasil MSK investigates the development, progression and adverse health consequences of the most prevalent MSK disorders worldwide, particularly chronic MSK pain (CMP), including chronic low back pain (CLBP) and osteoarthritis (OA) of knees and hands. Differently from the majority of epidemiological studies on this topic, the insertion of ELSA-Brasil MSK in the larger ELSA-Brasil cohort allows the investigation of complex aetiological pathways shared between MSK disorders and major NCD. For instance, despite growing evidence of common pathways underlying the development and progression of cardiometabolic conditions and OA,7,8 low back pain9 and other MSK pain syndromes,10 limited data on subclinical diseases (e.g. subclinical atherosclerosis, pre-diabetes or other early endocrine and metabolic disruptions) and their inter-relationships with MSK disorders have hindered our understanding of the interplay between these multiple conditions across the lifespan. All civil servants enrolled at the second largest ELSA-Brasil investigation centre (IC-MG) and attending the second quadrennial face-to-face visit from the original cohort (2012–14) were invited to participate (N = 2922 participants). Those who completed a minimal set of MSK health measures and provided valid data were included (N = 2901; mean age 56.0 years, SD 8.9, 52.9% women). Of these, 97.6% (N = 2830) underwent hand and knee X-rays. The timeline of the original and ancillary cohorts and a flowchart of participants included at baseline of ELSA-Brasil MSK are described in Figure 1. Table 1 lists participants’ sociodemographic and health characteristics. Flowchart of participants included at baseline of the Brazilian Longitudinal Study of Adult Health Musculoskeletal (ELSA-Brasil MSK) cohort (A) and timeline of ELSA-Brasil and ELSA-Brasil MSK (B). MSK, musculoskeletal; IC-MG, Investigation centre—Minas Gerais. *Civil servants from IC-MG who remained actively enrolled in the ongoing ELSA-Brasil cohort during its second visit of face-to-face assessments Sociodemographic, clinical and behavioural characteristics of participants and prevalence of chronic musculoskeletal pain (any), chronic low back pain and knee osteoarthritis according to participants’ sociodemographic and general health characteristics at baseline of the Brazilian Longitudinal Study of Adult Health Musculoskeletal (ELSA-Brasil MSK) cohort (N = 2901, 2012-14) Small differences in total N for each variable are due to missing values. BMI, body mass index; MSK, musculoskeletal; CLBP, chronic low back pain; OA, osteoarthritis; mNKOA, knee OA according to adapted National Institute for Health and Care Excellence guidelines (present if age ≥45 years + frequent knee pain activity-related joint pain in the same knee); RKOA, radiographic knee OA (left and/or right knee, tibiofemoral and/or patellofemoral OA); SKOA, symptomatic knee OA (ROA + frequent knee symptoms in the same knee). N = 2824. N = 2817. N = 1651. Comorbidity: hypertension, diabetes, myocardial infarction, heart failure, stroke, kidney disease, chronic obstructive pulmonary disease (COPD), cirrhosis, cancer. Cohort surveillance by annual telephone interviews started in 2015 (Figure 1B). From 2640 participants who were still alive during the second quadrennial visit from ELSA-Brasil MSK (2017-19), 2597 provided data on MSK health and 2516 repeated their hand and/or knee radiography (89.5% and 88.9% retention rate, respectively). The third quadrennial visit is planned for 2022-24 (Figure 1 and Table 2). Overview of the measures used in quadrennial cohort waves and annual telephone interviews of the Brazilian Longitudinal Study of Adult Health Musculoskeletal (ELSA-Brasil MSK) cohort Q, questionnaire; E, examination; T, test; OA, Osteoarthritis; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; WRRSI, work-related repetitive strain injury; WOMAC®, Western Ontario and McMaster Universities Osteoarthritis Index; AUSCAN®, Australian/Canadian Hand Osteoarthritis Index. Past 12-month pain history was assessed using a body diagram depicting the neck, shoulder, upper back, elbow, lower back, wrist/hand, hip/thigh, knee, ankle/foot (laterality of knee and wrist/hand pain assessed from firstwave onwards, and of shoulder/elbow/hip/thigh/ankle/foot from second wave onwards). Past 30-day and 7-day pain history was assessed for lower back and L-R (left and right) knee and wrist/hand only. Defined as pain in the past 12 months with >6 months duration (all cohort waves) and pain in the past 30 days with ≥3 months duration (second wave and lower back during telephone interviews). Assessed for lower back, L-R knee and L-R hand only, using a 0–10 numerical scale (cohort waves) or 5-point Likert scale (telephone interviews). Assessed for L-R knee and wrist/hand, and defined as pain on most days for ≥1 month in the past 12 months (first wave) or pain on most days of the past 30 days (all other assessments). Defined as pain that prevented normal daily activities (e.g. work, domestic and leisure activities). Assessed by the International Physical activity Questionnaire—IPAQ (all waves) and accelerometer (second wave onwards). Information obtained during concurrent assessments from the larger cohort includes sociodemographic measures, health-related behaviours (e.g. smoking, alcohol consumption), psychosocial data (e.g. depression, sleep patterns) and cognitive function, anthropometric measures (e.g. body mass index, bioimpedance), medication, subclinical and laboratory measurements and NCDs incidence.1,2 ELSA-MSK added major assessments on MSK disorders and potential risk factors (Table 2). Participants are assessed for doctor-diagnosed MSK diseases (OA, rheumatoid arthritis, systemic lupus erythematosus, gout, fibromyalgia and work-related repetitive strain injury), as well as 12-month, 30-day and 7-day history of any pain, discomfort and/or stiffness (herein named pain, and not necessarily related to a previous doctor diagnosis) (Table 2). A body diagram assists the identification of pain at the following sites11: neck, shoulder, upper back, elbow, lower back, wrist/hand, hip/thigh, knee, ankle/foot. Annual telephone interviews obtain data on symptoms in the lower back, knees and hands. Information regarding trauma-related pain onset, self-reported disability (pain that prevented normal daily activities), pain intensity (0-10 numerical scale) and care-seeking behaviour are also collected. Data on pain duration will allow chronicity to be defined according to the two most common definitions used in epidemiological research: (i) pain in the past 12 months of >6 months duration; and (ii) pain in the past 30 days of >3 months duration.12 This has a number of advantages, such as increasing comparability, allowing analyses that match research standards for studies on a given MSK symptom or disease, and addressing research questions that might benefit from different approaches to pain measurement. For knee and hand pain, assessments are complemented by the identification of frequent pain, defined as pain on most days for ≥1 month in the past 12 months (baseline assessment) or pain on most days for the past 30 days (follow-up telephone and 4-yearly face-to-face interviews), and by the Australian and Canadian Hand Osteoarthritis Index® (AUSCAN®) and Western Ontario and McMaster Universities Osteoarthritis Index® (WOMAC®) questionnaires. AUSCAN® was applied to those with hand symptoms and WOMAC® to those with knee symptoms up to participant 1469, and everyone thereafter (irrespective of knee symptoms). Bilateral non-fluoroscopic digital radiography of left and right knees [posteronterior (PA) fixed-flexion, using a novel knee positioning device,13 and lateral view] and hands (PA) are performed by radiographers, under radiologist supervision.14 The radiographic images are interpreted, blind to participants’ characteristics, according to Kellgren & Lawrence,15 Osteoarthritis Research Society International (OARSI)-revised16 and the ELSA-Brasil MSK Atlases of osteoarthritic lesions, the latter developed from participants’ images used during training and calibration procedures. Radiographic readings follow a two-step protocol: (i) preliminary screening for ‘possible knee OA’ by two independent radiographers; and (ii) diagnosis of radiographic knee OA (RKOA) by an experienced radiologist after the revision of all images rated as ‘possible knee OA’ by at least one radiographer. The screening protocol was tested by having a random selection of 108 images rated by the two screeners as ‘no possible knee OA’ evaluated by the experienced radiologist, with no misclassification being found. RKOA is defined as tibiofemoral (TF) OA (Kellgren & Lawrence classification grade ≥2)17 and/or patellofemoral (PF) OA (definitive osteophyte in the PF joint or definitive PF joint space narrowing plus any bone abnormality).18,19 Symptomatic knee OA (SKOA) is defined as RKOA plus frequent knee pain in the same knee. Additional data collected allow the classification of OA according to modified National Institute for Health and Care Excellence (NICE) criteria20 (mNKOA), which comprise the presence of frequent knee pain + age ≥45 years + activity-related knee pain in the same knee (identified by WOMAC® scores >1 on questions regarding pain during walking and/or climbing stairs) (Supplementary Table S1, available as Supplementary data at IJE online). The NICE criteria do not include knee radiography. Efforts have been made to ensure the quality and comparability of radiographic readings. For instance, training and calibration sessions were performed with a highly experienced, academically-based bone and joint radiologist (external reader), who was involved in the Framingham Osteoarthritis Study, Beijing Osteoarthritis Study and Multicentre Osteoarthritis Study. Inter-rater reliability between the external reader and two ELSA-Brasil MSK researchers [RWT (rheumatologist) and LCS (radiologist)], and intra-rater reliability of the ELSA-Brasil MSK radiologist analysis revealed good to an excellent agreement (Supplementary Table S2, available as Supplementary data at IJE online). Training and calibration using a similar protocol are currently under way for hand radiography. At baseline, physical functioning was assessed by the Five Times Seat-to-Stand test21 and handgrip test using the Jamar hydraulic dynamometer (Sammons Preston, Bolingbrook, IL).22 From the second visit onwards, assessments have been complemented with gait velocity and static balance tests from the Short Physical Performance Battery.23 Participants are asked about their history of hand/knee trauma/surgery and their engagement in joint-loading activities (repetitive knee-bending exposure), including climbing ≥10 flights of stairs/day, kneeling/squatting/deep knee-bending ≥30 min/day and lifting or moving objects/people weighing ≥8 kg. The General Self-Efficacy Scale and the Multidimensional Health Locus of Control-A scale were applied at baseline assessments.24 Self-efficacy and locus of control are psychological constructs derived from social learning theories and relate to personal beliefs about one’s competency to deal with life stressors, and the location of where one’s control over health resides, respectively.25,26 Leisure-time physical activity has been assessed by the International Physical Activity Questionnaire at every quadrennial visit, and by the accelerometer-based device ActiGraph wGT3X-BT (ActiGraph, LLC, Pensacola, FL) from the second visit onwards. At baseline of ELSA-Brasil MSK, OA was the most prevalent self-reported doctor-diagnosed MSK disease (18.3%), followed by work-related repetitive strain injury (7.4%), fibromyalgia (6.5%), gout (3.4%) and rheumatoid arthritis (2.8%). Figure 2(A–C) describes the prevalence of any pain, CMP, and chronic disabling musculoskeletal pain according to site (A), number of sites (B), and number of body regions affected (C). Nearly all participants (94.9%) reported pain in the past 12 months. The 12-month prevalences of CMP and chronic disabling MSK pain were 55.1% and 25.4%, respectively. The neck was the most prevalent site of any pain (49.8%), whereas chronic and chronic disabling pain were more frequently reported in the knee (22.5% and 7.9%, respectively) and lower back (18.6% and 6.9%, respectively). Twelve-month prevalence of any pain, chronic musculoskeletal pain and chronic disabling musculoskeletal pain according to site (A), number of sites (B), and number of body regions (C) at baseline of the Brazilian Longitudinal Study of Adult Health Musculoskeletal (ELSA-Brasil MSK) cohort (2012–14). A total of nine pain sites were evaluated: neck, shoulders, upper back, elbows, lower back, wrists/hands, hips/thigh, knees, ankles/feet. Body regions were divided as follows: (i) upper limbs: shoulders, elbows and wrists/hands; (ii) lower limbs: hips/thigh, knees and ankles/feet; (iii) axial skeleton: neck, upper back and lower back. Participants with chronic/chronic disabling musculoskeletal pain in at least one site/region, but who had pain in two or three or more sites/regions, were included in the categories of 2/3+ sites/regions of pain. Small differences in total N for each variable are due to missing values. *N = 2901; **N = 2889; ***N = 2895 The 12-month prevalences of multisite pain (present in three or more sites from nine sites depicted in the body diagram) and generalized pain (present simultaneously in three body regions: upper + lower limbs + axial skeleton) were: 54.5% and 34.0% for multisite and generalized pain of any duration; 41.0% and 27.7% for chronic multisite and generalized pain; 21.3% and 15.8% for chronic disabling multisite and generalized pain, respectively. There was a higher frequency of chronicity and disability as the number of pain sites or regions increased (Figure 2B and C). The baseline prevalence of radiographic knee OA was 18.1%, but only 7,8% presented symptoms (SKOA). Among those aged >45 years, 28% showed mNKOA (Table 1). Figure 3 shows the overlap between 1605 participants aged 45+ years with data on all three OA criteria: nearly 40% of them fulfilled at least one knee OA criterion, and 10.4% all three criteria. Mechanical and frequent knee pain without signs of RKOA was in of had RKOA in one knee and mNKOA in the radiographic OA in one knee and early OA in the other (Figure Among all individuals with RKOA = had PF RKOA (Figure for SKOA, knee OA was more frequently in each knee (Supplementary Table and Supplementary Figure S1, available as Supplementary data at IJE online). of different osteoarthritis criteria in the Brazilian Longitudinal Study of Adult Health Musculoskeletal (ELSA-Brasil MSK) (A) diagram of three different osteoarthritis criteria. diagram of knee sites of radiographic *N = number of participants with data on all three criteria. **N = number of participants with right and tibiofemoral and radiographic mNKOA, modified knee OA defined according to National Institute for Health and Care Excellence (present if age ≥45 years + activity-related joint pain according to questions on knee pain during walking and/or climbing + frequent knee pain at the same knee); RKOA, radiographic knee OA (left and/or right knee, tibiofemoral and/or patellofemoral OA); SKOA, symptomatic radiographic knee OA + frequent knee symptoms at the same knee); Western Ontario and McMaster Universities Osteoarthritis Table 1 describes the baseline prevalence of CMP (any), chronic low back pain, RKOA, and mNKOA according to participants’ sociodemographic and health characteristics. disorders were more common among and retired prevalence was higher among participants with multiple for RKOA, using baseline data from ELSA-Brasil MSK between risk factors and showed of multisite and generalized CMP for individuals with bodily or with of increasing at higher duration of to and pain a between symptoms and chronic knee pain, which was increased in pain was by daily disability. an between life and chronic knee pain, which was also in the presence of knee disability. described an between by the with CMP at any with of multisite and generalized from ELSA-Brasil MSK researchers the of the General Self-Efficacy Multidimensional Health Locus of and WOMAC® and have these to be for in participants of the The knee radiographic protocol and positioning device used during OA assessments in the cohort were also to and highly measurements of joint space In a previous analysis of the of knee OA in a (N = RKOA and showed excellent but lower mNKOA and the of an experienced was used as the the burden of MSK the prevalence and of MSK pain disorders are due to of available studies (e.g. of symptoms and with and related to and data from developed Because of these and health beliefs and common social and life have been ELSA-Brasil MSK will a data on the burden and of most common MSK disorders in a low- or middle-income a demographic and include the of and measures of functioning and physical and the early identification of and age-related changes about of the cohort participants were aged years at the baseline of ELSA-Brasil MSK, whereas other OA cohorts comprise the three OA criteria in with participants’ sociodemographic, clinical and psychosocial characteristics, allow the investigation of different OA the of participants with and without MSK pain and knee OA allows analyses on the determinants of the incidence as well as of these including knee OA and its adverse ELSA-Brasil MSK, ELSA-Brasil, was as an cohort to health in a large and to participant retention over of Although this the external of that are by the and the of ELSA-Brasil MSK still to a good the and socioeconomic of the Brazilian as these are to the of and risk measures obtained in the Because of and that osteoarthritic changes to disease such as and bone and other potentially relevant physical measures (e.g. and or (e.g. and Osteoarthritis were not used in ELSA-Brasil MSK. Collaboration in data analysis and are welcome through research proposals sent to principal the ELSA-Brasil [http://www.elsa.org.br] or e-mail the corresponding author. ELSA-Brasil was by the National for in Research de and the of the six participant research ELSA-Brasil MSK was by the and research of was performed following the standards of the of and its and all participants Supplementary data are available at IJE the study and its including quality and the knee radiographic the the and the of the All the ELSA-Brasil has been by the Brazilian of Health and and the Brazilian of and de and de ELSA-Brasil MSK has been by de de de de do de Minas Gerais, and ELSA-Brasil MSK baseline investigation was was a by is a research from The the and participants of ELSA-Brasil and ELSA-Brasil MSK for their and de and for with and The also and from the of of and from the of at and for their and with training and quality control of radiographic readings.

Topics & Concepts