Surgical and Non-surgical Treatment in the Management of Hip Osteoarthritis: Total Hip Arthroplasty versus Progressive Resistance Training in Patients with Severe Hip Osteoarthritis (The PROHIP Trial)

Introduction
Hip osteoarthritis (OA) is a major contributor to pain and functionalimpairments worldwide, with great financial costs for society. Hip OAaffects around 1 out of 10, and it is the leading cause for total hiparthroplasty (THA). In individuals with severe hip OA, THA is considered aneffective treatment to reduce pain, improve physical function, and increasequality of life, but the procedure is associated with a small risk of seriouscomplications. On the contrary, progressive resistance training (PRT) is safeand appears to provide moderate improvements in pain, physical function,and quality of life even in individuals with severe hip OA. To date, THA hasnot been compared directly with PRT. This comparison is highly wanted asnon-surgical treatment is considered as an alternative to surgery for manymusculoskeletal disorders.

Aim and Objectives
The overall aim was to investigate the effectiveness of surgical and nonsurgical treatment in individuals with severe hip OA considered eligible forTHA. The specific study objectives were:

Study I. To explore patient, clinician, and decision maker stakeholders’perceptions on a randomised trial evaluating the effectiveness of THAcompared with PRT to inform the development of the trial protocol.

Study II. To develop and describe a study protocol for a randomised trialcomparing THA with a PRT program in individuals severe with hip OA.

Study III. To evaluate whether individuals with severe hip OA enrolled in arandomised trial comparing THA with PRT differed in self-reportedcharacteristics and outcome scores from those who only acceptedenrolment in a parallel prospective observational cohort at baseline.

Study IV. To investigate whether THA improved self-reported hip pain andfunction more than PRT in individuals with severe hip OA from baseline to6 months follow-up.

Materials and Methods
Study I. Participants were enrolled into three key stakeholder groups: patients with severe hip OA considered eligible for THA, clinicians (orthopaedic surgeons and physiotherapist), and decision makers (politicians and non-governmental organisation representatives). Focus group interviews were conducted using open-ended semi-structured interview guides, according to group status. Interviews were recorded, transcribed verbatim, and thematic analysed using an inductive approach.

Study II. A protocol for a high quality, multicentre, parallel group, randomised controlled superiority trial comparing THA with PRT in individuals with hip OA was designed and described, including considerations about study design, inclusion and exclusion criteria, enrolment procedures, randomisation and allocation concealment, blinding, parallel observational cohort, interventions, outcomes, data collection, sample size and power calculation, and statistical methods. 

Study III. Baseline data from the randomised trial (PROHIP) and parallel observational cohort (Non-PROHIP) was used. Participant characteristics and the Oxford Hip Score (OHS, range 0 [worst] to 48 [best], minimal important difference [MID]=5 points), the Hip disability and Osteoarthritis Outcome Score (HOOS) subscales (range 0 [worst] to 100 [best], MID=10 points), and University of California Los Angeles (UCLA) activity score (range 1 [inactive] to 10 [regular physical activity with high intensity], MID=2 points) were collected using electronic online questionnaires. Data were analysed using descriptive statistics, standardised differences, nonparametric tests, and univariate logistic regression.

Study IV. Participants were randomly assigned to receive either a THA followed by standard care or a 12 week supervised PRT program targeting lower extremity muscles performed twice a week followed by 12-weeks of optional non-supervised PRT. The primary outcome was the betweengroup difference in change in the OHS (range 0 [worst] to 48 [best], MID=5 points) from baseline to 6 months follow-up, analysed using a repeatedmeasures mixed effects linear model with adjustments for covariates.

Results
Study I. Four focus group interviews with a total of 14 patients, one focus group with 4 clinicians (2 orthopaedic surgeons and 2 physiotherapists), and one focus group with 4 decision makers (3 politicians and 1 nongovernmental organisation representative) were conducted. Two main themes were developed. ‘Treatment expectations and beliefs impact management choices’ covered three sub-themes: Treatment without surgery is unlikely to lead to recovery; Clinician authority affects the management narrative; The ‘surgery versus exercise’ debate. ‘Factors influencing clinical trial integrity and feasibility’ highlighted three subthemes: Who is considered eligible for surgery?; Facilitators and barriers for surgery and exercise in a clinical trial context; Improvements in hip pain and hip function are the most important outcomes. 

Study II. The THA intervention consisted of a standard fast-track multimodal surgical program, which included patient information, optimised pain management, and early mobilisation. The surgical procedure was performed in accordance with the standard posterior approach, and postoperative rehabilitation followed hospital-specific procedures. The PRT programme comprised 10-min warm-up on stationary bicycle followed by four exercises performed unilaterally in machines or cable pulleys with full as full range-of-motion as possible in three sets separated by 60 seconds of rest. Progression of training load followed a linear model of periodization with an initial relative load of 12 repetition maximum (RM) in week 1-2, 10RM in week 3-6, and 8RM in week 7-12. Absolute training load was adjusted on set-by-set basis using muscular contraction to volitional failure and between sessions guided by hip pain.

Study III. In total, 402 participants were included in the analysis with 109 in PROHIP and 293 in Non-PROHIP. The PROHIP group had a better mean (±standard deviation) OHS (25.1±5.9 vs. 22.6±6.9, group difference 2.5 points [95% CI 1.1 to 4.0]) than the Non-PROHIP group. The PROHIP group also had better mean scores in all HOOS subscales (group differences were 4.9 points [95% CI 1.4 to 8.3] in HOOS pain; 5.6 points [95% CI 1.9 to 9.3] in HOOS symptoms; 7.4 points [95% CI 3.5 to 11.3] in HOOS function in activities of daily living; 8.1 points [95% CI 4.0 to 12.2] in HOOS function in sport and recreation; and 7.7 points [95% CI 4.4 to 11.0] in HOOS qualityof-life), while there was no difference between the groups in the UCLA activity score (group difference 0.2 points [95% CI -0.2 to 0.6]). For the selfreported characteristics, the PROHIP group had a higher body mass index (group difference 1.4 kg/m2 [95% CI 0.5 to 2.3]) and a lower proportion had previously received THA (group difference -13.5% point [95% CI -20.9 to -6.0]), total knee arthroplasty (TKA) (group difference -6.4% point [95% CI -10.4 to -2.3]) or supervised exercise (group difference -10.8% point [95% CI -21.0 to -0.7]) compared with Non-PROHIP group.

Study IV. Among the 109 randomised participants (mean age, 67.6 years; 54 [49.5%] females), 103 (94.5%) completed the randomised trial. At the 6 month follow-up, the adjusted mean change (±standard error) in the OHS was 15.9±1.0 points in the THA group and 4.5±1.0 points in the PRT group (group difference 11.4 points [95% CI 8.9 to 14.0; P<0.001]). 

Conclusion
Patient, clinician, and decision maker stakeholders had treatment expectations and beliefs that could lead to selection bias during enrolment procedures, treatment crossovers, and reduced generalisability of the randomised trial. To improve methodological rigorousness of the trial protocol, a parallel observational cohort was included, an enrolment procedure using generic guidance delivered by an independent clinician to facilitate communication of clinical equipoise was developed, and change in self-reported hip pain and function was selected as the primary outcome. A high quality randomised controlled trial comparing THA with PRT was designed, described, initiated, and completed with enrolment of 109 participants. At baseline, participants enrolled in the randomised trial had less self-reported hip pain and impairments in function, higher body mass index and fewer had previously received THA, TKA or supervised exercise than those enrolled in the parallel prospective observational cohort. Among individuals aged 50 years or above with severe hip OA considered eligible for surgery, THA yielded clinically significant superior improvements in hip pain and function compared with PRT at 6 months follow-up.