Enhancing Postoperative Rehabilitation Following Knee Arthroplasty Using a New Cryotherapy Product

Outline

We performed a prospective study to evaluate P, a novel cryotherapy device, against C in patients undergoing primary total knee arthroplasty by a single surgeon (SG) using their default surgical technique. All patients were operated on in 1 hospital between March 2011 and February 2012.

During this period, all patients were managed as part of a protocol-driven ERP. Preoperative information leaflets including expected length of stay were routinely used. Patients were preassessed prior to surgery and attended on the day of their operation to a dedicated admission unit. Patients received either spinal or general anesthesia depending on their preference and clinical indications. Thigh tourniquets were in place for the entire procedure in all patients. The Triathlon Knee (Stryker, Kalamazoo, Michigan) system was used in all patients through a medial parapatellar approach. Local anesthetic infiltration using standardized volume, concentration, and technique was routinely used. A retransfusion drain was inserted prior to closure and removed on the first postoperative day. An integrated care pathway was used to record a patient’s progress throughout the admission, and postoperative rehabilitation was standardized.

Power Calculation

Following the results of a pilot study, a priori power calculation (G*Power, version 3.1.9.2; Universitat Kiel, Germany) was completed with several assumptions; our final study would have a standard power of 80% and that a p value <0.05 is significant. It revealed at least 32 patients in each group would be sufficient to show a significant difference in outcomes.

Patient Selection

Ninety consecutive patients were entered into the study. Allocation into each group was based on their hospital numbers which are randomly designated. Those with even numbers were entered into the P group, and those with odd numbers were placed in the C group. Consent was obtained from all patients, and their rights were protected at all times. Those patients with incomplete data sets or with surgical complications not as a result of the cooling product were excluded (Figure 1)

OPEN IN VIEWER

Figure 1.

CONSORT 2010 flow diagram for patient enrollment and group allocation.

Eight patients were excluded from the study. Seven did not meet the inclusion criteria of which 1 patient in C group experienced an allergic reaction to the wound dressing (not the cryotherapy system) and 6 had incomplete data sets (4 in C vs 2 in P). One patient declined to take part in the trial who would have otherwise been allocated to the P group and therefore received the local standard intervention (C). Two patients in the P group did not receive the allocated intervention.

Outcome Measures

Primary outcome measures were pain and ROM. Pain was assessed using a visual analogue scale (VAS) and ROM using a long-arm goniometer. The VAS scores and ROM were recorded before and 30 minutes after application of the cooling device on the first and second postoperative day. The patients were often discharged home on the third postoperative day.

Data Collection

Data collection was performed by physiotherapists. All data for each patient was recorded on a custom designed data collection form. The patient hospital number was recorded on the cover sheet with subsequent pages for recording pain and ROM for each postoperative day. Although all physiotherapists were appropriately trained and briefed on the methodologies of the study, these were also detailed within this tool. Data was kept in a secure location throughout the study period.

The devices were assigned on the first postoperative day. Both devices were applied in the first instance by a physiotherapist. Following this, the patient was responsible for reapplication of the device, unless they were unable to do so.

In the C group, ice and water changes were performed by physiotherapists or nursing staff. The P group patients resoaked the bandage themselves via a spray bottle when they perceived the cooling benefit to have diminished. All patients in the group were given one extra fluid bottle for this.

Ethics

All methodology and data collection were approved and conformed to the hospital trust protocols.

Statistics

The age of the groups was compared using a Student t test, whereas the ROM and VAS were compared using a Mann-Whitney U test. Normality of the data was tested using a Shapiro-Wilk test. This statistical analysis was performed using SPSS (IBM, Chicago, Illinois). A p value <0.05 was considered statistically significant.

Demographics

There was no significant difference in age or sex in either group (Table 1).

OPEN IN VIEWER

Table 1.

Comparable Age and Sex of Both Cryocuff and Physicool Groups.

	Cryocuff	Physicool
Age, mean, years	74.9	75.1
Female, n	19	18
Male, n	21	22

Range of Movement

Improvement in ROM (range) after intervention on day 1 C versus P was 12.34° (−10° to 60°) versus 20.38° (−5° to 55°), respectively. Improvement in ROM on day 2 C versus P was 8.88° (−20° to 45°) versus 16.25° (−15° to 50°), respectively. The difference in ROM between P and C was statistically significant on both day 1 (p = 0.004) and day 2 (p = 0.009) postoperatively (Figure 2).

OPEN IN VIEWER

Figure 2.

Box-plot displaying distributions of mean improvement in range of movement (ROM).

Pain

The difference in VAS was recorded pre- and postintervention. We report our results in terms of a difference in VAS. Therefore, the greater the decrease between the pre- and postintervention VAS, the better the improvement in pain. The difference in VAS after intervention on day 1 C versus P was −1.2 (−4 to 0) versus −1.73 (−5 to 1), respectively. The difference in VAS on day 2 C versus P was −0.8 (−4 to 1) versus −1.73 (−4 to 1), respectively. The greater improvement in VAS in the P group compared to the C group was statistically significant on both day 1 (p = 0.013) and day 2 (p = 0.001) postoperatively (Figure 3).

OPEN IN VIEWER

Figure 3.

Box-plot displaying distributions of the mean visual analogue scale (VAS) improvement after interventions on postoperative day 1 and 2. *Median and 25th percentile values were both −2. **Median and 25th percentile values were both −1.

Cost

The cost of a Physicool presoaked bandage and sufficient additional cooling fluid is £15. The cost of a Cryocuff is £37, excluding the cost of water and ice production, which is difficult to quantify. The cooling reservoirs cost £80 each and are replaced as required. Patients are not able to take them home with them due to the cost. Cost savings of approximately £25 per patient were estimated accounting for the cost of disposables, replacement cooling reservoirs, and ice. This equates to a potential saving of £8750 per year based on a unit which performs 350 total knee replacements annually.

Limitations

We accept that this was not a randomized blinded study, and therefore results are susceptible to bias. In view of the nature of the cryotherapy devices, patient blinding is not feasible. It may, however, have been possible to blind assessors, although due to resource and time limitations this was not included in our methodology.

While there was no control group, it was felt that sufficient evidence exists to support the efficacy of cryotherapy versus placebo for the outcome measures we examined. Previous work has shown cryotherapy to be superior to compression bandages and external ice at reducing pain and improving early ROM. ^{16 –18}

Furthermore, we did not examine analgesic requirements between the 2 groups, although previous meta-analysis has shown a reduction in analgesia requirement with cryotherapy. ¹⁹

The groups were matched on age and sex alone. Our pilot study suggested no difference in ROM and VAS on day 1 prior to administration of any form of cryotherapy, as this would aid in forming homogenous groups. Consequently, we did not record preoperative ROM and VAS. Our study showed no statistical difference in day 1 ROM (p = 0.79) and VAS (p = 0.102) between the patients in either group before cryotherapy was applied, which was in keeping with our findings in our pilot study. This helps to justify our decision, as we felt both groups had a similar point of reference. Patients were also not matched regarding anaesthetic type (general vs spinal). Although several studies have shown that spinal anesthesia provides improved analgesia post TKA, ²⁰ studies have also confirmed that local anesthetic periarticular injections provide pain relief for up to 48 hours postoperatively, ²¹ which all of our patients had performed. We do appreciate that matching using preoperative ROM, VAS, and anesthetic type would have increased the strength of our cohorts, and therefore we will do so in further studies of a similar nature.

Pain scores and ROM data were only recorded for the first 2 postoperative days. The mean length of stay of total knee arthroplasty patients in our department is 4.2 days, with a significant number of patients being discharged on the third postoperative day. As no provision for measurement following patient discharge was made, we felt a shorter follow-up was preferable to multiple incomplete data sets.

We used VAS, as it provides a simple and validated scoring system. ²² Intraobserver variability was controlled by having the same physiotherapist recording pre- and postcryotherapy measurements with a goniometer; however, no mechanisms were put in place to control interobserver variability.

For those who had been excluded from the study, 1 individual in the C group experienced an allergic reaction to the Tegaderm dressing (3M, St Paul, Minnesota). Seven patients did not undergo postoperative measurements of either ROM or VAS due to limitations on the ward, and these incomplete data sets were excluded. The 2 patients who had not received their allocated intervention complained of the cooling fluid-soaked bandage causing dampness on bed sheets and clothing, and some found the odor of the cooling fluid unpleasant. They were therefore given the standard C system and excluded from data analysis. It was noted that application of the P device can be challenging for those with poor manual dexterity.

Implications

This study clearly demonstrates the superiority of the P cryotherapy system over that of the C method for patients undergoing total knee arthroplasty when considering pain scores, ROM, and cost in the early postoperative period.

In addition to the acute postoperative phase of knee arthroplasty, cryotherapy has been used successfully following several arthroscopic procedures, particularly cruciate ligament reconstruction. ²³ It has been used to reduce both the recovery time from traumatic injuries and in the treatment of chronic sporting injuries. It would seem likely that the advantage of Physicool over Cryocuff would be similar in these situations, so this study could prompt the exploration of the use of the Physicool system in a wide range of operative and nonoperative scenarios.