Resistance Training Falls Short in Severe Hip Arthritis

Patients with severe osteoarthritis (OA) of the hip did better after 6 months when they had fast-tracked arthroplasty versus 12 weeks of supervised resistance training in a randomized trial.

Average Oxford Hip Score values rose 15.9 points among 53 patients assigned to surgical treatment, compared with an improvement of 4.5 points for the 56 participants in the resistance training group (difference 11.4 out of a possible 60-point total score, 95% CI 8.9-14.0), according to Thomas Frydendal, PhD, of Vejle Hospital in Vejle, Denmark, and colleagues.

Five patients randomized to surgery hadn’t actually received it by month 6, the group reported in the New England Journal of Medicine, whereas 12 of those assigned to resistance training had symptoms worsen to the point of needing immediate surgery.

“[T]otal hip replacement resulted in clinically important, superior reductions in patient-reported hip pain and improvements in function at 6 months as compared with resistance training,” Frydendal and colleagues concluded.

That’s not to say, however, that exercise therapy should have no role for patients like those enrolled in the trial, called PROHIP (Progressive Resistance Training Versus Total Hip Arthroplasty in Patients With Hip Osteoarthritis).

“Although our results favor treatment with total hip replacement for most outcomes,” they wrote, “they do not oppose the use of resistance training as initial treatment. In this regard, nearly three of four patients had not received exercise therapy before enrollment, and almost one in four who had been assigned to resistance training had not undergone surgery after 24 months … Even for patients who ultimately undergo total hip replacement, previous studies support faster postoperative recovery in those who undergo supervised resistance training than in those who do not.”

Mean age for PROHIP participants was 68, and Oxford Hip Score values averaged 25 at baseline; subscales of the Hip Disability Osteoarthritis Outcome Score (HDOOS) averaged around 50 for pain, overall symptoms, and function in daily living. For patients with this degree of severity, total arthroplasty is typically recommended. Yet other studies have suggested that resistance training can reduce pain and improve function, “even in patients who are scheduled to undergo total hip replacement,” Frydendal’s group noted. In OA generally, exercise is often urged at least as a prelude to surgery. Thus, the researchers said a randomized trial was warranted to compare the two approaches directly.

The trial was conducted at four orthopedic specialty units in different regions of Denmark. Patients older than 50 judged eligible for total hip arthroplasty were enrolled. Half of those enrolled were men. Resistance training consisted of twice-weekly, 1-hour sessions with a physiotherapist, with four hip-related machine exercises. Resistance loads were gradually increased while the number of repetitions per exercise was reduced.

Changes in mean HDOOS values paralleled those for the Oxford scale, with increases from baseline of 33-44 points in its various subscales with arthroplasty, versus increases of 9-15 points with resistance training (all PP=0.009).

Participants were also followed as long as 24 months after randomization with self-evaluations of pain, overall symptoms, function, quality of life, and “global perceived effect.” At months 12 and 24, these all favored surgery, significantly so in most measures.

Numbers of serious adverse events through month 6 were about the same in the two groups, at six versus five. For those assigned to arthroplasty, most were related to the procedure (i.e., prosthetic joint infection, hip dislocation, or revision surgery). However, one patient initially assigned to resistance training proceeded to arthroplasty and then suffered a hip dislocation — Frydendal and colleagues categorized this event with the resistance training group, even though it was related to surgery. None of the other adverse events in those assigned to resistance training appeared directly related to it, except possibly for one serious case of atrial fibrillation.

Limitations to the study included its open-label design and the 86% of eligible patients who declined to participate when asked. Most said they wanted immediate arthroplasty, the authors observed.

“Since treatment preference may be associated with outcome, our trial may be prone to selection bias because the enrolled patients may differ from the general population of persons with hip osteoarthritis,” they wrote.