Article Text
PDF
Abstract
Objectives Total hip replacement (THR) is one of the most common elective orthopaedic surgeries performed, with increasing demand among younger individuals. Few evidence-based guidelines exist on safe activity participation following THR, including whether high-intensity sport participation is safe for individuals. The purpose of this study was to identify if increased intensity of physical activity following THR was associated with increased activity-related pain and increased revision rates.
Methods Two groups undergoing THR were recruited: preoperative (cohort 1) and 5–7 years postsurgery (cohort 2); both followed for 5 years. Activity was self-reported through validated questionnaires and grouped into categories from ‘A’ (low intensity, eg, aquafit) to ‘F’ (high intensity, eg, tennis). The primary outcome was the presence of hip pain during activity (binary variable, Y/N), measured by the association between hip pain and intensity of activity (categories A–F). Secondary outcomes included activity duration, revision rate or a change in patient-reported outcome measures (PROMs).
Results 1098 individuals were included in this study (cohort 1: n=588, cohort 2: n=510). Regression analysis showed no significant interaction between activity intensity and hip pain across all time points. Approximately 20.6% of all activity (11.0% of participants) occurred in the highest intensity categories (E and F); these subjects showed no decrease in activity duration, worsening PROMs or increased revision rates compared with lower intensity activity groups (all p>0.05). When analysing by individual activities, certain activities (eg, snowboarding, squash, tennis and backpacking) were more correlated with hip pain (r>0.60), while others (eg, snorkelling, swimming, home weights, aquafit, cross-country skiing and sledding) were less likely to have hip pain (r<−0.60).
Conclusions This study showed that higher-intensity activities do not lead to decreased activity duration, worsening patient-reported outcomes or increased revision rates following THR, although certain activities may be associated with increased pain. These findings can inform patient counselling after THR.
- Exercise
- Hip
- Longitudinal Studies
- Physical activity
- Sport
Data availability statement
Data are available on reasonable request. We request investigators to contact us directly to discuss data sharing.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
It is hypothesised that higher-intensity activity may not increase symptoms and the revision rate following total hip replacement (THR), but current evidence is limited.
WHAT THIS STUDY ADDS
This is the first large-scale prospective study showing that higher-intensity activity participation does not lead to increased amounts of hip pain following primary THR. Higher-intensity activity participants also do not have a greater risk of revision, nor fractures or other injuries necessitating intervention. However, some specific activities (snowboarding, squash, tennis and backpacking) were more strongly correlated to the presence of hip pain.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Surgeons and physicians may counsel people undergoing THR that it is safe to participate in high-intensity activities following THR.
Introduction
Total hip replacement (THR) has widely been regarded as one of the most successful surgical procedures in modern medicine, revolutionising the treatment of arthritis and other conditions of the hip joint.1 2 It has become one of the most common elective surgeries performed worldwide, with an incidence of 191.5 per 100 000 population per year.3 Demand for THR in young patients is projected to grow exponentially in the coming decades, with a sevenfold higher growth rate among patients younger than 65 living in Organisation for Economic Co-operation and Development (OECD) nations from 1990 to 2011.3 This increased demand may be attributed to both an ageing population and an increase in the proportion of younger individuals undergoing the procedure.
Despite its success, individuals can develop postoperative hip pain or discomfort over time through a variety of mechanisms such as local inflammatory reactions (eg, secondary to liner wear), muscle fatigue, tendinitis/bursitis, bony impingement or—at the extreme—failure of the hip implant.4 5 Surgery to revise a hip implant is a more costly procedure associated with higher morbidity and poorer functional outcome. In 1961, John Charnley—one of the earliest THR innovators—wrote in the Lancet that it was unrealistic to expect a hip prosthesis to last 30 years or enable an individual to engage in high-impact sport.5 However, individuals have expectations of participating in higher-intensity activity on their replaced hips,6 and registry data have shown that most patients can expect their THR to last at least 25 years.7 Longevity of the newest implant materials and designs is anticipated to be even longer. Although people have higher expectations than prior cohorts with regards to physical activity following THR,8 equipoise remains in the surgical community as to the necessary long-term restrictions to prevent complications and maximise longevity.
The only guidelines available for post-THR activity are based on limited data and a survey of physician preferences with considerable disagreement among clinicians regarding many activities.9–11 Some surgeons continue to recommend a highly restrictive protocol (ie, no high-impact activity), while others encourage their patients to engage in activity without limitations.12 There is a lack of scientific data to support either approach at present.
This prospective cohort study aimed to determine whether the intensity of physical activity following THR was associated with pain, subjective outcomes or early revision surgery within 10–12 years of the index procedure. The findings of the current study are intended to inform evidence-based clinical practice recommendations in the future, as both younger and more physically active individuals continue to undergo this life-changing surgical procedure, and specifically guide individuals who may want to participate in high-intensity sport following THR.
Methods
Study design
A novel prospective ‘parallel study’ design consisting of two cohorts: one recruited at the time of THR surgery (cohort 1), the other at 5–7 years postsurgery (cohort 2); both followed for 5 years. The rationale for this design is described in the study protocol (online supplemental appendix A); in short, it circumvents issues with a conventional cohort approach, including high cost, sample size and confounders. Particularly, if 2000 patients were recruited over a 5-year period, many observed revisions would be due to infection or trauma and not related to implant wear. Therefore, the parallel cohort design enables capturing a longer duration of information on physical activity, pain, revision and function.
Supplemental material
Inclusion criteria included patients aged <80 at surgery, with end-stage hip arthritis (including osteoarthritis, AVN, hip dysplasia), undergoing primary unilateral THR. Exclusion criteria included metastatic cancer, multiple sclerosis, Paget’s disease, lupus, previous trauma/fracture in the affected hip, cognitive limitations, severe inflammatory/systemic disease, DMARD use or other conditions significantly limiting mobility or function.
Eligible patients (online supplemental appendix A) were recruited from two large academic hospitals. Cohort 1 patients were identified from a list at their preoperative appointment and approached by the research assistant to obtain written informed consent. Cohort 2 patients were identified from an electronic list based on appointment scheduling data; then contacted by someone in the circle of care at an appointment or via letter and invited to meet with the research assistant for screening. Physical activity was assessed at baseline and annually (or biannually) thereafter. Outcome measures were evaluated by trained research personnel at baseline and thereafter during their clinical follow-up visit with their surgeon and over the phone.
Data collection
Baseline data on age (at the time of surgery), sex, marital status, height, weight, occupation, implant characteristics, complications (all in-hospital surgical and medical complications, postsurgery complications including those causing revision such as fracture), comorbid medical conditions and medications were collected in both cohorts (presurgery in cohort 1, at baseline enrolment 5–8 years postoperative in cohort 2). Presurgical physical activity, pain and physical function were also collected in cohort 1.
Physical activity and sports participation were evaluated with the Minnesota Leisure Time Physical Activity Questionnaire (MLTPAQ), a validated tool which captures the frequency and intensity of a wide range of physical activities.13 In brief, the MLTPAQ collects self-reported data on the duration (minutes per occasion), frequency of participation (months per year, average occasions per month) and the presence of hip pain (binary variable for hip pain with activity: yes/no, if hip pain reduced activity duration: yes/no) for 55 leisure activities.9 14
From the MLTPAQ, we grouped activities into six categories (A–F) based on their level of impact and torque on the prosthetic joint (table 1). Activity intensity was categorised into six levels: A (low-intensity simple motions, eg, aquafit, cycling) to F (high intensity complex motions, eg, downhill skiing, squash) (table 1). These classifications were based on biomechanical studies that measured the contact forces and torsion on the hip joint.15–17
Activity classification
Patient-reported outcome measures (PROMs) were also used, including the physical function subscale of The Western Ontario and McMaster Universities (WOMAC) Osteoarthritis Index, the Pain Catastrophising Scale (PCS) and the hip version of the ICOAP (Measure of Intermittent and Constant Osteoarthritis Pain) scale.18 The WOMAC is a widely used, well-studied and reliable health status measure for patients with arthritis. The tool consists of 24 questions rated on an ordinal scale of 0–4, with lower scores indicating lower levels of physical disability, divided into 3 subscales (pain: 5 items, stiffness: 2 items, physical function: 17 items). The PCS consists of 13 questions related to thoughts and feelings associated with chronic pain. Responses range from 0 to 4 (0=not at all; 4=all the time). The scores for each question are summed. A higher score indicates greater pain catastrophising. The ICOAP evaluates patients’ experiences with different kinds of pain (including aching or discomfort) in their hip, distinguishing between intermittent and constant pain. It contains 11 questions rated on an ordinal scale of 0–4, with lower scores indicating lower levels of physical disability. Data were collected in Microsoft Excel and stored securely within our institutional firewall.
Study outcomes
The primary outcome is activity-related pain (binary variable: Y/N on the MLTPAQ) at any time point up to 5 years (cohort 1) and 10–12 years (cohort 2) postsurgery. The association between hip pain (dependent variable) with respect to activity intensity (A–F) and duration (independent variables) was tested through regression models to investigate the relationship.
Secondary outcomes include reoperation (revision) rates (eg, for fracture, dislocation or aseptic loosening of the prosthesis), change in activity level over time, change in PROMs over time. Each of these outcomes was analysed with respect to participant activity level.
Sample size justification
From our preliminary studies, we estimated that 30% of THR patients engage in activity associated with hip pain (unpublished data from pilot study). From the literature at the time of study commencement, THR patients engaging in high-intensity activity ranged from 10% to 20%.19 Assuming a probability of high-intensity activity of 0.1 (10% of patients) and a frequency of hip pain of 0.3 (30% of patients), a sample of 578 individuals will yield 80% power (α=0.05) of detecting an increase in pain in the high-intensity category. Therefore, we aimed to recruit 600 subjects in each cohort. The sample size calculation was run using PASS V.12 Power Analysis and Sample Size Software (2024), NCSS.
Statistical analysis
Descriptive statistics are presented in a narrative summary fashion. With respect to baseline demographics and activity participation data, continuous measures were summarised using means and SD whereas categorical measures (eg, revision rates) were summarised using counts and proportion.
For the first study aim, logistic regression models were run for each cohort at each time point, looking at the presence of hip pain (dependent variable) in relation to activity classification, total time doing the activity and the interaction between classification and time doing the activity (independent variables). Linear mixed models were run to assess changes across all time points. The participant was treated as the random effect in the models with an assumed normal distribution. For these models, activity intensity was grouped into high intensity (E and F) versus low intensity (A and B) and medium intensity (C and D). No other covariates were included in the initial linear regression and linear mixed models. If a positive association was discovered, future models would control for covariates.
Another logistic regression model was run on the ‘hip pain’ dependent variable against each of the 55 surveyed activities on the MLTPAQ (independent variables) to assess which specific activities resulted in a statistically significant predictor of hip pain. No other variables were included in this model. This allowed a reclassification of the 55 activities into high/medium/low ‘risk’ categories based on the level of hip pain associated (logistic coefficient).
For the second aim, linear regression models were run in a similar fashion looking at PROMs (dependent variable) in relation to activity classification. Both the initial classification (A–F) and the reclassification (high/medium/low risk) were used as independent variables for these models. The relationship between activity intensity and revision rates was analysed using a χ2 test. All analyses were two-sided and were run using SAS V.9.4 (SAS Institute).
Data visualisation
We applied a combination of data analysis and data visualisation tools to the dataset to leverage and cross-validate the potential insights. For example, clustering algorithms were used to suggest key variables to graph, and bar charts revealed possible trends in the data worthy of further analysis. The final results were then organised into an interactive dashboard (Tableau) which allows viewers to meaningfully process high-dimensional data in a compact and intuitive format. The interactive component allows deeper analysis to be performed with an emphasis on verification (eg, removal of outliers). Static snapshots of the dashboard are included as figures in future sections, and the fully functional dashboard is publicly available as a supplemental resource (https://public.tableau.com/app/profile/dawit.gulta/viz/Safe_TDataVisualizationDashboard/OverviewMain).
EDI statement
The study recruited all consecutive patients, without any specific purposive recruitment of patients from marginalised groups. The author team was composed of a diverse set of individuals with different demographics, educational backgrounds and expertise. Among the listed authors, we have one female and four individuals of visible minorities. The team includes two board-certified orthopaedic surgeons, one orthopaedic resident, one statistician, one senior scientist, one data scientist and one master’s student. Although all are based in Toronto, Canada, one author is from a developing country. The broader safe activity participation following THR investigator team includes several women, including the protocol author (now retired). Data were not collected, nor were special accommodations made, for those from disadvantaged backgrounds.
Results
Participant characteristics
In total, 1098 subjects were included (table 2). 590 individuals were female (53.7%), mean age at surgery was 68.5 (SD: 9.8) and mean body mass index (BMI) at surgery was 29.4 (5.9). Cohort 1 had 99.1% follow-up at 5 years, and cohort 2 had 99.3% follow-up at 11 years postoperative. Significant differences between cohorts were observed with regards to American Society of Anesthesiologists (ASA) class (a common surrogate for comorbidities), use of cemented stems, liner material, length of stay and discharge destination (p<0.05), but not revision rates. While the mean BMI was similar between cohorts, the BMI categories showed significant differences, with more underweight individuals in cohort 2. Age at surgery was not different between cohorts.
Baseline demographics
Primary outcome
The primary aim was to determine whether higher activity intensity was associated with greater hip pain. Across visits, 21.5% of individuals reported hip pain with activity, and 12.5% of all individuals reported that they reduced participation in activities due to their hip pain. The logistic model relating hip pain (dependent variable) to activity intensity (category A–F, independent variable) showed no significant association between activity intensity and the presence of hip pain (table 3). This was tested using logistic regression models assessing the hip pain variable in relation to the activity classification, total time doing the activity and the interaction between classification and time spent doing activity. Another linear mixed model followed this analysis by adding in a component testing interactions by duration of activity. These models did not show significant changes over time. Grouping activity category into high intensity (E and F) versus low intensity (A and B) and medium intensity (C and D) did not change our findings.
Results from a logistic model showing no correlation between hip pain and activity intensity (category A–F) and duration at each time point
Activity-specific results
Mean duration of activity spent in each category did not change in either cohort from baseline (preoperative) to any postoperative time point. Time spent doing low-intensity activities (categories A and B) was approximately 8.9 hours/month per person (20.6% of all recorded activities (43.1 hours/month)), compared with medium-intensity (categories C and D) at 25.3 hours/month (58.7%) and high-intensity (categories E and F) at 8.9 hours/month (20.6%). The percentage of individuals with their maximum activity in each category is shown in table 4. 11.0% of individuals reported participating in category E and F activities across time points, with no statistically significant change in the percentage across time points. In summary, individuals in higher-intensity activity groups did not reduce their activity intensity over time.
Percentage of individuals with maximum activity duration in each activity intensity category at each time point
Reclassification
Our initial classification of MLTPAQ activities (categories A–F based on biomechanical loading) showed a reasonable correlation to hip pain. Logistic regression between hip pain (dependent variable) and activity (independent variable) was modelled (online supplemental eFigure 2). This analysis allowed us to reclassify activities (on the MLTPAQ) into low, medium and high-risk groups, based on the logistic coefficient (table 5 and on the dashboard (https://public.tableau.com/app/profile/dawit.gulta/viz/Safe_TDataVisualizationDashboard/OverviewMain).
Reclassification of activities based on the probability of hip pain
This logistic regression model identified snowboarding, squash, tennis and backpacking as having a high risk of hip pain (correlation threshold r>0.60). In contrast, snorkelling, swimming, home weights, aquafit, cross-country skiing and sledding were associated with an absence of hip pain (r<−0.60).
Revisions
A secondary aim was to determine the association between activity intensity and revision rates. 30 revisions occurred overall in the study (online supplemental eTable 2); 22 revisions were noted in cohort 1 (8 due to periprosthetic femur fracture, 6 due to infection, 3 due to aseptic loosening) and 8 in cohort 2 (3 due to infection, 2 due to aseptic loosening, 2 due to recurrent dislocations). There was no statistically significant relationship between maximum activity level and revisions (p=0.48).
Patient-reported outcome measures
Another aim was to determine if higher activity intensity affects subjective outcomes. As expected, all PROMs were significantly improved from the preoperative to first postoperative visit (T1) (p<0.001) in cohort 1 (table 6) and did not change across postoperative time points (p>0.05). Analysis was undertaken to determine if there were statistical differences between the intensity of activity (independent variable) and a change in PROM scores (dependent variable) across time points using linear regression models. Participation in higher-intensity activities did not result in worsened PROMs over time (table 7). The only statistically significant finding was that higher activity individuals had a greater change in ICOAP score (F=3.4, p=0.009) from preoperative to T1, but this finding did not persist at any other time point.
Patient-reported outcome measures across time points
Analysis of activity intensity versus PROM scores over time
Additional analysis was undertaken using the reclassified activities (high/medium/low risk) to determine if participation in high-risk activities resulted in decreased PROMs over time (online supplemental eTable 1). Once again, participation in higher risk activities did not result in a statistically significant decrease in PROMs over time. In summary, higher-intensity levels of activity participation did not result in a clinically significant decrease in PROMs over time.
Data visualisation
The results were visualised using an interactive online dashboard (Tableau, Salesforce, USA) (https://public.tableau.com/app/profile/dawit.gulta/viz/Safe_TDataVisualizationDashboard/OverviewMain). Tableau is a visual analytics platform that is used across many sectors to help people see, understand and act on data. Our dashboard shows the interplay of variables including age, sex, BMI, PROMs and activity participation in individuals before and after THA. Sample screenshots of the dashboard are presented below (online supplemental eFigure 1). This dashboard allows readers to interact with our data, looking at specific subgroups and visualise the analyses presented above (tables 4 and 6 and online supplemental eTable 1).
Discussion
This study aimed to examine the relationship between participation in high-intensity activities and concomitant hip pain, subjective outcomes and early revision. The most significant finding in this study is that participation in activities of higher intensities is not independently associated with the presence of hip pain at any postoperative time point. Participation in biomechanically categorised ‘high-risk’ activities at higher frequencies and durations was not associated with the cessation of activity, hip pain, worse patient-reported outcomes, or higher rates of revision. However, our reclassification of individual activities based on hip pain showed that some high-intensity activities (eg, squash, tennis, backpacking and snowboarding) elicited hip pain more often, while other commonly recommended activities (eg, swimming, aquafit and cross-country skiing) were protective.
We found that more than one in five (21.5%) individuals experience some degree of hip pain with activity following THR, although we were not able to demonstrate a relationship with biomechanical activity intensity (categories A–F). Current literature shows that individuals return to physical activity and sports at high rates following THR, although some fall short of their pre-surgical expectations20—hip pain may be one explanatory variable for this. Whereas literature evaluating specific sports following THR with modern implants has not yet found any specific activities to be associated with early failure requiring revision,21 22 these studies have not similarly considered the patient experience when returning to these activities. Communicating that returning to activity may be safe but occasionally uncomfortable or painful should be an important element of the preoperative counselling process to better inform patient expectations.
The improvement in PROMs between the preoperative assessment and at 1 year following surgery in cohort 1 did not wane over the ensuing 10-year follow-up (in cohort 2) and were not related to the intensity or duration of activity undertaken by individuals in either cohort. However, only 11.0% of all THR patients participated in the highest-impact activities (categories E and F) following surgery, which is consistent with other published data,6 but may ultimately limit insight into this subgroup. Although age-related lifestyle changes may explain much of this activity profile, it is also possible that the low frequencies may be due to the unfounded fears of decreased implant longevity with high-impact activity. Alternatively, as discussed above, it may be because of pain or functional limitations experienced by individuals when returning to these activities, thereby resulting in self-imposed restrictions.
Finally, our finding that revision risk is not elevated by activity intensity is consistent with recent studies, which have found no deleterious effect of sports participation on revision due to loosening of the implant,10 and a low medium-term prosthetic failure rate among tennis players receiving THR.23 Older literature demonstrating higher rates of implant failure,24 with high-impact/high-intensity activity is no longer relevant in the era of modern implant materials and less invasive surgical approaches.25 26 More recent literature—although more supportive of our findings—has been predominantly retrospective or registry/database driven, which has inherent limitations (eg, focus on ‘hard outcomes’ such as failure resulting in revision of implant) and has not generally included function or pain as outcome measures.27 28 These gaps and inconsistencies in evidence have resulted in the variability of recommendations, ranging from those found in consensus-based guidelines to the experience-based opinions of institutions or surgeons.29 Our study presents objective and high-quality evidence in support of softening restrictions and creating more permissive activity guidelines.30
Clinical implications
The clinical implication of this study is that there is no evidence to support arbitrary restriction of activity levels following THR; rather the decision should involve a discussion between the individual and their surgeon with regards to the low rates of implant failure/revision, the higher potential for pain with certain activities (eg, snowboarding, squash, tennis and backpacking), and uncertainties in the available evidence with respect to such higher impact activities. This is of utmost importance given that individuals undergoing THR are increasingly younger, more active and have higher expectations than in previous generations.
Strengths and limitations
Our study is the largest prospective cohort study to date evaluating the influence of activity on a variety of THR outcomes (including pain and function) with the use of modern implants and bearing technology. The data collected were detailed with regards to type and time spent on each activity by each participant; we also used a novel study design which enabled the collection of 10-year outcomes in half the time by employing two temporally separate cohorts recruited from the practices of multiple surgeons distributed over two separate institutions.
Limitations of our study included the use of patient self-reported data, which is prone to recall or other reporting biases, despite validation of the MLTPAQ.14 However, given the subject area and the variety of activities undertaken by individuals, it would be challenging (time/resource-intensive) to record daily activity in a more objective manner, especially in larger cohort studies. The presence of confounding variables not captured in our models must be noted given the observational design. The relationship between activity and pain may not be causative and could be susceptible to correlation in both directions (eg, hip pain causes cessation of high-intensity activities). Our analysis was also restricted to clinical outcomes; we were not able to analyse radiographs for evidence of loosening or fracture, and other signs of implant failure. Finally, revision rates in our cohort were very low (as expected), most occurred early in cohort 1, and the majority (19/30) were due to infection or leg length discrepancy. As such, although we followed individuals up to 10 years postoperatively, longer follow-up times may be needed for any relationship between revision rates and activity level to emerge.
Conclusions
In conclusion, surgeons may counsel individuals that most activities are safe following THR with respect to implant failure rates. Important information that can be shared with individuals is that while certain activities (eg, snowboarding, squash, tennis and backpacking) may be associated with hip pain, they are not associated with revision surgery up to 10 years.
Supplemental material
Data availability statement
Data are available on reasonable request. We request investigators to contact us directly to discuss data sharing.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by Sunnybrook Hospital Research and Ethics Board #2784. Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We sincerely thank the SAFE-T investigators, research manager, coordinators and assistants, in particular Monica Kunz, for their efforts in study ideation, protocol development, patient recruitment, data collection and storage. We also thank the Data Visualisation team at York University for their innovative approach to representing this complex dataset. We dedicate this paper to Dr. Iris Weller who had the vision to initiate this important research.
References
Footnotes
X @ajaymedic
Collaborators SAFE-T investigators: Harman Chaudhry MD MSc, Stephen Chen PhD, Jeffrey Gollish MD, Monique Gignac PhD, Dawit Gulta, Gillian Hawker MD PhD, Richard Jenkinson MD, Alex Kiss PhD, Hans Kreder MD MPH, Michael Ade Conte MD (c), Steven Macdonald MD, Ajay Shah MD, and Cari Whyne PhD.
Contributors AS, HC and CW contributed to drafting the manuscript, interpreting the data and revising the manuscript critically for intellectual content. AK, SC and DG provided statistical analysis and manuscript writing. HK: responsible for supervising the project, securing funding and ensuring ethical compliance. Guarantor: HC is the guarantor of this work and accepts full responsibility for the accuracy, integrity and submission of the manuscript in its final form.
Funding This study was funded by the Canadian Institute of Health Research Musculoskeletal Health and Arthritis (grant number 84315).
Disclaimer The authors report no competing sources of funding or conflicts of interest relevant to the statements of this manuscript. Patients were not involved in the design, or conduct, or reporting, or dissemination plans of our research.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.