**3. Results**

Sixteen studies described post-operative therapy for MCP joint replacement in enough detail to understand the treatment schedule. Four hundred and twenty-seven patients in these studies had rheumatoid arthritis, 19 had osteoarthritis and one had psoriatic arthritis. There were four randomised trials (one about post-operative therapy), three prospective cohort studies, three prospective case series (two about post-operative therapy), one case study about post-operative therapy, and the remaining were retrospective case series. Missing data was obtained from two authors, to assist in reviewing the rigour of the studies, but many authors could not be contacted.

The randomised controlled trial study found to specifically compare post-operative regimes for metacarpophalangeal arthroplasty (8) randomised patients into postoperative therapy groups that both included dynamic splinting, but the treatment group also included continuous passive motion. These researchers found no difference between treatment groups. Thomsen, Boeckstyns and Leth-Espensen(2003) (9) and retrospectively reviewed consectutive patients who had either dynamic MCP extension splinting, or had static splinting that was removed for exercises post-operatively. They found that residual extension lag was significantly less (*p* = 0.002) in the dynamicallysplinted group, concluding that postoperative dynamic splinting was useful. Groth, Watkins and Paynter, (1996)(10) retrospectively compared patients who had dynamic flexion with those who had dynamic extension splinting, and found that those who had post-operative dynamic flexion splinting had greater post-operative MCP flexion. Burr, Pratt and Smith (2002)(11), Burr and Pratt (1999)(12) focussed their research on postoperative therapy, but neither study had a comparison treatment group. No further studies compared post-operative treatment regimes, therefore the results of the remaining studies can only be appreciated as a combination of surgery, implant and post-operative therapy.

One outcome common to nearly all postoperative patients in every study was the relief of pain once the diseased joint had been removed by surgery. Negative outcomes such as wound infection, implant loosening and migration were reported, in small proportions. Compliance with splinting and therapy was not discussed. Sixteen studies described the outcome of different implants and postoperative therapy regimes for MCP arthroplasty.

Features common to many regimens (Table 1) were postoperative avoidance of any hand activity for the first three to six weeks and long-term avoidance of ulnar forces on the fingers. Nearly all regimens began between the second and seventh postoperative day.

Regimens could be divided into two main categories with regard to splinting and exercise. Static splint regimens involved removal of the splint for active MCP range of motion exercises, and dynamic splint regimens involved active-assisted MCP extension and active MCP flexion exercises within the splint.


566 Recent Advances in Arthroplasty

Studies were appraised as described by the Cochrane Collaboration (7) for sources of methodological bias that could decrease the internal validity of a study. The types of methodological bias were in patient selection, equality of treatment, attrition of patients, and detection of all relevant outcomes. If the study could not be fully appraised from the

Sixteen studies described post-operative therapy for MCP joint replacement in enough detail to understand the treatment schedule. Four hundred and twenty-seven patients in these studies had rheumatoid arthritis, 19 had osteoarthritis and one had psoriatic arthritis. There were four randomised trials (one about post-operative therapy), three prospective cohort studies, three prospective case series (two about post-operative therapy), one case study about post-operative therapy, and the remaining were retrospective case series. Missing data was obtained from two authors, to assist in reviewing the rigour of the studies, but

The randomised controlled trial study found to specifically compare post-operative regimes for metacarpophalangeal arthroplasty (8) randomised patients into postoperative therapy groups that both included dynamic splinting, but the treatment group also included continuous passive motion. These researchers found no difference between treatment groups. Thomsen, Boeckstyns and Leth-Espensen(2003) (9) and retrospectively reviewed consectutive patients who had either dynamic MCP extension splinting, or had static splinting that was removed for exercises post-operatively. They found that residual extension lag was significantly less (*p* = 0.002) in the dynamicallysplinted group, concluding that postoperative dynamic splinting was useful. Groth, Watkins and Paynter, (1996)(10) retrospectively compared patients who had dynamic flexion with those who had dynamic extension splinting, and found that those who had post-operative dynamic flexion splinting had greater post-operative MCP flexion. Burr, Pratt and Smith (2002)(11), Burr and Pratt (1999)(12) focussed their research on postoperative therapy, but neither study had a comparison treatment group. No further studies compared post-operative treatment regimes, therefore the results of the remaining studies can only be appreciated as a combination of surgery, implant and

One outcome common to nearly all postoperative patients in every study was the relief of pain once the diseased joint had been removed by surgery. Negative outcomes such as wound infection, implant loosening and migration were reported, in small proportions. Compliance with splinting and therapy was not discussed. Sixteen studies described the outcome of different implants and postoperative therapy regimes for MCP arthroplasty. Features common to many regimens (Table 1) were postoperative avoidance of any hand activity for the first three to six weeks and long-term avoidance of ulnar forces on the fingers. Nearly all regimens began between the second and seventh postoperative day. Regimens could be divided into two main categories with regard to splinting and exercise. Static splint regimens involved removal of the splint for active MCP range of motion exercises, and dynamic splint regimens involved active-assisted MCP extension and active

publication, information was sought by writing to the authors.

**3. Results** 

many authors could not be contacted.

post-operative therapy.

MCP flexion exercises within the splint.




ADL=Activities of Daily Living, MCP= metacarpophalangeal, IP= interphalangeal joints; RD= radial deviation; RCT= randomized

Table 1. Studies reporting post-operative therapy for metacarpophalangeal arthroplasty, their design, treatment protocols, and

controlled trial; ROM= range of motion, RA=rheumatoid arthritis OA=osteoarthritis

outcomes.


controlled trial; ROM= range of motion, RA=rheumatoid arthritis OA=osteoarthritis 

Table 1. Studies reporting post-operative therapy for metacarpophalangeal arthroplasty, their design, treatment protocols, and outcomes.

quality, or quantity of the treatment being studied, which was suspected in the continuous passive motion (CPM) study by Ring et al. Ring et al. describe the application of CPM in detail, except passive forces are described as "low" and treatment quantity is described as "as tolerated." As a result, the reader remains unsure of what amount of passive force is

Detection bias is determined if the timing of assessment, the outcome assessment used, or knowledge of the assessor of the patient's previous state could miss any relevant aspect of the outcome. This may have occurred in the study by Groth et al.,(10) in which some preoperative data were unavailable and patients were assessed at different postoperative time frames. Detection and comparison of outcomes between studies are only possible when the same outcome measures are used in a standardized manner. The researchers in this review all measured range of motion, but at different time frames (Table 1). Those who measured pain, cosmesis, and function applied different assessments at different time frames. The challenge of outcome measurement in rheumatology has led to the formation of focus groups such as OMERACT (Outcome Measures in Rheumatoid Arthritis Clinical Trials), who have made recommendations for outcome measures to be used in drug trials. OMERACT recommendations are not fully relevant to hand therapy research; however, the process of forming a focus group, and the development of assessment guidelines that allow comparison between homogeneous patients, is possible

Attrition bias is determined if the loss of patients in the study is significant or varies between the treatment and control groups. This is common in long-term studies involving patients with rheumatoid arthritis, and was experienced by Groth et al.,(10) who were unable to obtain long-term follow-up of the patient group who received their extension protocol. Long-term follow-up is an issue with rheumatoid populations. These patients undergo numerous surgical and drug interventions, while their disease progresses and fluctuates, making the long-term effects of the MCP surgery and therapy difficult to define. Once more, large numbers of patients in each treatment group would be required to decrease the effects of attrition bias and to dilute the effects of subsequent

The difficulties of past studies guide the planning of future studies. Although the issues of low patient numbers, variable preoperative status, additional surgical and drug interventions, and chronic disease cannot be altered, study designs can. Large randomized trials may not be possible; however, samples of patients, paired according to preoperative status, may be allocated to different treatment protocols. Standardized measurement of pain, cosmesis, impairment, disability, and impact on the patient, made at similar postoperative time frames, would further assist in determining treatment

This review suggests that *all* regimens contribute toward an increase in MCP motion and an increase in hand function, but despite the efforts of patients and clinicians, hand therapists remain unaware of the most effective postoperative protocol for MCP arthroplasty or the suitability of each regimen for specific implants and soft-tissue

ineffective, as well as what quantity of treatment per day is ineffective.

(28).

interventions.

efficacy.

**5. Conclusion** 
