1 Introduction

Total hip (THA) and total knee arthroplasties (TKA) are considered the treatments of choice for patients with end stage hip and knee osteoarthritis that no longer benefit from conservative treatment options.¹ Despite the overall effectiveness of both procedures,² a subgroup of patients experience unsatisfactory results.³

Patient satisfaction depends on the extent to which their preoperative expectations were met after surgery.⁴ According to a study by Conner-Spady et al.,⁵ the most frequent expectations in patients undergoing TKA or THA were pain relief and improvement in mobility, walking, physical activities, and daily activities. The same authors reported that physical activities had the lowest percentage of met expectations 12 months post-surgery. Satisfied patients had a significantly greater percentage of met expectations in comparison to dissatisfied patients.

Patient related outcome measures (PROMs) are important to assess improvement in health status out of the patients’ perspective. Different PROM instruments were developed in the last decades.^6,7 The West Ontario and MacMaster University Osteoarthritis Index (WOMAC) is a valid, reliable and responsive patient-centered self-reported health status questionnaire that was developed in the early 1980s as a disease-specific measure for hip and knee osteoarthritis patients.^8,9 In our institution the WOMAC score is administrated routinely preoperatively and at 3 and 12 months after primary THA or TKA.

Clinicians can only interpret PROM score changes at an individual patient level when benchmarks of response are available. Bellamy at al. reported on minimal clinically important improvements (MCII) for WOMAC scores in patients with hip and knee osteoarthritis.¹⁰ The patients were evaluated before and four weeks after treatment with nonsteroidal antiflamatory drugs. Their absolute MCII estimate (95% CI) for the WOMAC total score was 7 (4–10) scores points. Clement et al. reported a minimum clinically important difference (MCID) in WOMAC total score of 10 (6.1–12.9) score points for patients 12 months after TKA.¹¹

Little is known whether gender and the affected joint (hip or knee) influence WOMAC score changes after THA or TKA. The primary aim of this study was to investigate whether WOMAC total and sub-score changes after THA or TKA are gender and joint specific. Furthermore, we investigated whether the WOMAC scores are significantly different among patients with different clinically complexity levels (PCCL) and whether there were significant predictors of the 12 months WOMAC total index score.

2 Material and methods

3 Results

The preoperative data of 1539 (855 THA) were analyzed. The patients in the sample were on average 69.4 ± 10.8 years old. There were no significant differences between the patients in the TKA and THA groups concerning their mean age (p = 0.2) and length of hospital stay (LOS) (p = 0.5). The distribution of male and female patients was not significantly different between TKA and THA groups (p = 0.9). The number of patients with higher PCCL was higher in the THA group (p = 0.008). Further demographic data details are presented in Table 1.

The GLMM models for the dependent variables WOMAC total, pain, stiffness and function scores were all statistically significant (p < 0.001). FU (p < 0.001), gender (p < 0.001), joint (p < 0.001), and the interaction “FU by joint” (p < 0.001) had significant effect on the WOMAC total, pain, stiffness and function scores. The interaction “gender by joint” was only significant for WOMAC total and function scores (p = 0.02).

The model-estimated mean WOMAC total and sub-scores with 95% CI for THA and TKA patients by FU are shown in Table 2. For THA and TKA patients the estimated mean WOMAC total score and the sub-scores for the dimensions pain, stiffness and function improved significantly between preoperative and 3 months and between 3 and 12 months for a p < 0.001 (Fig. 1a–d).

Close

Fig. 1 Estimated means with 95% CI for (a) WOMAC total, (b) pain, (c) stiffness and (d) function scores for the patients in the THA and TKA groups across the follow-up times.

Export to PPT

THA and TKA patients had statistically significant different estimated mean WOMAC total, stiffness and function scores at each follow FU (Table 3). While preoperatively THA patients had significantly worse WOMAC total, stiffness and function scores, at 3 months FU the difference was inverted with THA patients having significantly better WOMAC total, stiffness and function scores (Fig. 1a, c, d). These significant differences remained until 12 months FU. Preoperatively, THA and TKA patients did not differ concerning their pain. At 3 and 12 months FU patients after TKA had significantly higher pain scores in comparison to THA patients (Fig. 1b).

The estimated mean WOMAC scores for male and female patients in both groups across the FU times are presented in Table 4. Within the THA group, male patients had significantly lower WOMAC total score preoperatively (p < 0.05) (Fig. 2a). There were no differences between both genders at 3 and 12 months FU. In the TKA group male patients had significant lower WOMAC total scores preoperatively and at 12 months FU (p < 0.05) (Fig. 2b). Regardless the group and the FU time, the overall estimated mean WOMAC total score for male and female patients were 18.4 [17.6–19.2] and 20.7 [19.9–21.5], respectively. Male patients had significantly lower estimated mean WOMAC total scores (p < 0.001).

Close

Fig. 2 Estimated mean for WOMAC total scores with 95% CI for male and female patients by follow-up time (FU) in the (a) THA and (b) TKA groups.

Export to PPT

There were significant differences between the mean WOMAC total scores of patients with different PCCL classifications at each FU (p < 0.005) (Fig. 3). Patients with higher PCCL had higher (worse) WOMAC total scores.

Close

Fig. 3 Comparison of WOMAC total scores among patients with different clinical complexity levels (PCCL) by follow-up time. The one-way ANOVAs were performed at each FU on log-transformed WOMAC total score data. Since the data was not normally distributed, we present boxplots. Statistically significant differences (*) between the pairs of means for a p < 0.05.

Export to PPT

The preoperative WOMAC score (ß = 0.20; p = 0.001), preoperative pain (ß = 0.14; p = 0.02), patients age (ß = 0.06; p = 0.05) and PCCL (ß = 0.13; p < 0.001) significantly predicted the mean WOMAC total score at 12 months [F(4) = 31.4; p < 0.001; R² = 0.13; Adj. R² = 0.13]. Lower preoperative WOMAC total scores, lower preoperative WOMAC pain scores, younger age and lower preoperative PCCL significantly predicted better WOMAC scores 12 months after surgery.

4 Discussion

In the present study we investigated whether changes in the WOMAC total and sub-scores were joint (hip and knee) and gender-specific in patients with osteoarthritis who underwent THA or TKA. Furthermore, we analyzed whether the WOMAC scores among patients with different PCCL differed within the FU times. Finally, we tested whether patient-specific factors could predict the WOMAC total score measured 12 months after surgery. The WOMAC score data was collected preoperatively and at 3 and 12 months after surgery.

The WOMAC total score and its sub-scores pain, stiffness and function improved significantly for THA and TKA patients between all FU times (p < 0.001) (Fig. 1a–d). Preoperatively the patients who submitted for THA had significantly worse WOMAC total scores and perceived significantly more stiffness and less function in comparison to TKA candidates (p < 0.001). In line with previous research,¹⁵ patients after TKA perceived more pain, more stiffness and less function at 3 and 12 months FU in comparison to THA patients. Regardless of the group and the FU-time, male patients had overall lower mean WOMAC total scores (p < 0.001). Patients with higher clinically complexity levels had worse WOMAC scores at each follow up (Fig. 3). The preoperative WOMAC total and pain scores, patients’ age and PCCL significantly predicted the mean WOMAC total score 12 months after surgery. The patterns of change in the WOMAC total and sub-scores were similar in both groups (THA vs. TKA), however they were joint-specific. These are the main results of the present study.

In a similar study it was reported that THA patients improve quicker and achieved better outcome scores than TKA patients.¹⁵ Our results reinforce their findings.

The question whether the statistically significant differences found are also clinically relevant, should be discussed too. In the past a number of authors investigated the Minimum Clinically Important Difference (MCID) for the WOMAC total score and its sub-scores.^{10,11,16–18} In most cases the MCID was defined with the use of anchor-based methods, which allow an association between the measured WOMAC score and an anchor-question that reflects the patient's perception of the changes in their health-related quality of life (HRQOL) or their perceived benefits from the treatment. As reported previously, problems persist when reporting methods and results of studies, in which the WOMAC questionnaire was used.¹⁹ In two studies the MCID for the WOMAC scores after THA¹⁸ and TKA¹⁶ were investigated. Unfortunately, the authors of both studies converted the WOMAC total and sub-scores into a scale from 0 to 100. Therefore, the MCID they have detected cannot be used unless the WOMAC data is converted into the same scale. For a matter of usability, benchmarks for MCID that allow a clinician to interpret the improvements of a single patient should be presented in the same scale as the one used by the questionnaire. This was done in a study by Clement et al.^11,20 In their work MCID was defined as the difference in the mean WOMAC score change from preoperative to 12 months FU, between the patients who have rated themselves as having perceived “no improvement” and those who reported “little improvement” according to the anchor-question. After adjusting for preoperative confounding variables, the authors estimated a mean MCID difference [95% CI] of 10 [6.1–12.9] score points for WOMAC total, 11 [7.2–15.6] for pain, 8 [3.2–11.8] for stiffness and 9 [5.6–12.7] for function scores. As referred above, these MCIDs were estimated for changes between preoperative and 12 months FU. Although the answer to whether to use the estimated average MCID or the lower or upper bound of its 95% CI was not answered yet,²¹ we have decided to use the lower bound of the 95% CI to search for clinically significant differences among our results. Since we are looking for the MCID, we think the lowest possible value, given by the lower bound of the 95% CI, should be used as threshold for MCID. Accordingly, the changes between preoperative and 12 months in our study were all clinically significant in both groups, since the mean changes are greater than the above-reported lower bounds of the 95% CI for the different WOMAC scores (Table 2). With the use of the same MCID threshold, the detected statistically significant improvements between 3 and 12 months FU would be clinically irrelevant in both groups, because the improvements are smaller than the lower bounds of the 95% CI. The authors did not consider it appropriate to use a MCID that was estimated to detect changes between preoperative and 12 months FU to detect MCID between 3 and 12 months FU. Our results show that perception of improvement by the patients after joint arthroplasty is not linear. The improvement rates were higher for all sub-dimensions of the WOMAC score during the first three months after surgery than between the third and the 12th month. For this reason, MCID thresholds calculated based on patients perceived clinical changes that took place within the time window of 12 months shouldn't be used to detect clinical changes that took place in a different time window or changes between two different groups at a certain time point. To the best of our knowledge there is no study indicating WOMAC score thresholds for MCID between the 3rd and 12th month after THA and TKA. The WOMAC score changes in this time period were much smaller than between preoperative and third month, but they were statistically significant. In order to investigate whether improvements in WOMAC scores after the 3rd month are clinically relevant further research is necessary.

Bellamy et al. presented population-based normative values for the WOMAC score.²² The authors evaluated the WOMAC questionnaire of 7300 subjects and reported that in general, pain, stiffness and physical function subscales and subsequently the WOMAC total score increased with age and there were also, to a minor extent, gender differences. Our results are in line with their findings. Age was a significant predictor of 12 months WOMAC total score. A one-year increase in patients’ age resulted in a slightly higher 12 months WOMAC scores. In our sample we have also detected gender differences. Regardless the group (THA or TKA), male patients had lower estimated overall mean WOMAC scores throughout the FU-times.

Between January 2016 and December 2018, 2907 patients underwent TKA or THA at the clinic. Thereof, the preoperative WOMAC data of 53% (n = 1539) were available in the database of the clinic. There is no information on the reasons why the WOMAC data of 47% of the patients was missing. We presume organizational reasons or missing written consent of the patients behind this high number of missing data. In line with other studies in which follow-up data was acquired by means of postal survey, there were also dropouts at 3 and 12 months FU-times. The number of cases with preoperative WOMAC data (n = 1539) decreased by 42% (n = 905) at 3 months FU and by further 4% (n = 891) at 12 moths-FU. This is a study limitation. The GLMM allowed us to use all available data at each FU-time.

As mentioned above, WOMAC data at 3 and 12 months FU was acquired with questionnaires that were sent by post to the patients. The patients did not come to the clinic for FU visits. For this reason, clinical data of the patients at 3 and 12 months FU is not available. It would have been interesting to compare the patient reported outcomes with the outcomes out of the clinicians’ perception.

MCID thresholds are time-window specific. Further research should investigate MCID thresholds in shorter time windows after THA and TKA to detect the point in time after surgery where changes are no longer clinically relevant since they are not perceived by the patients anymore.

5 Conclusions

The mean WOMAC score changes after THA and TKA arthroplasties are joint-specific. Patients after TKA perceived less improvement in all dimensions. The results of this study can be used to adjust patients’ expectations before surgery. The detected differences between preoperative status and 12 months were all under the minimum clinically important difference (MCID).

Funding

Not applicable. None of the authors or their institutions received funding

Ethics approval

According to the Ethics Commission of the Federal State of Hamburg, Germany, retrospective database-based studies do not require ethical approval and patient informed consent whenever the data were acquired, saved and treated anonymously. This applies to the present study. The study was conducted in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments.

Consent to participate

Not applicable.

Consent for Publication

This manuscript does not contain any individual person’s data. All data exposed in this manuscript was anonymized.

Availability of data and material

All SPSS outputs can be provided by the corresponding author upon reasonable request.

Code availability

The SPSS syntax used to run the GLMM models can be provided by the corresponding author upon reasonable request.

Authors’ contributions

CJM and FL contributed to the study conception and design. CJM prepared the data for analysis. HOP gave assistance with the SPSS syntax to run the GLMM. CJM analyzed the data. CJM wrote the first draft of the manuscript. HOP, KB, JL and FL reviewed the manuscript. All authors approved and read the final manuscript.