The results of meniscal allograft transplantation surgery: what is success?

The menisci act primarily as load bearers and secondary stabilisers within the knee joint, and are chondroprotective. Loss of meniscal tissue through trauma or degeneration increases the joint contact forces by up to 350%, exacerbating articular cartilage degeneration [1]. Meniscal tears are particularly common in sportspersons [2, 3] with the typical mechanism of injury involving twisting on a flexed knee or loaded hyperflexion [4]. Symptomatic meniscal tears may require repair, or, if irreparable, a partial meniscectomy [5], significantly increasing the risk of future chondral failure and, ultimately, secondary osteoarthritis [6].

Meniscal Allograft Transplantation (MAT) was first reported in 1989, and is a useful potential option for symptomatic patients where there is loss or discontinuity of the peripheral meniscus and where the degenerative damage is not yet severe enough or the patient is too young for artificial joint replacement surgery to be a suitable option [7‐9]. However, MAT is a technically demanding procedure, performed by only a small number of surgeons, and graft availability and cost can be problematic [10]. Patients eligible for MAT are typically young and high-demand, and they are a highly-complex population who have often already endured multiple previous surgical procedures, and many have expectations to return to high activity levels [1]. Although a challenge, restoring function in this patient population is a priority within orthopaedics today [11].

Whilst outcome data is limited by low volumes and short follow-up, there is evidence that MAT can relieve pain, optimising knee function and comfort in selected patients [7]. A pilot randomised-controlled trial in 2018 confirmed MAT had better outcomes at 1 year compared to physiotherapy [12]. Authors have reported positive patient outcome measures following MAT using the KOOS, IKDC, Lysholm, Tegner, and SF-12 scores [13‐16]. It remains unclear what proportion of any improvement might be due to MAT alone or from potential concomitant procedures [17], and it is also unclear whether MAT has any long-term chondroprotective effect [18]. Whilst many studies have found that the addition of concomitant procedures at the same time as MAT do not affect outcomes, a recent study showed that compared to isolated MAT, PROMs were improved by the addition of a high tibial osteotomy to offload the affected compartment [19]. Certainly, MAT has been shown to have a higher-failure rate in ACL-deficient knees [20].

Surgical failure is defined variably in studies, but is most often defined as the patient either requiring removal of the allograft, allograft revision or conversion to joint replacement (arthroplasty) [13]. Clinical failure is often defined as a Lysholm Score of < 65 and has been shown not to correlate well with surgical failure or poor patient-reported outcomes [13].

There are differing definitions of failure within the existing literature, with surgical failure not necessarily correlating to clinical failure. The aim of this study was to examine the outcomes of MAT in one surgeon’s patient cohort (using the same graft preparation and fixation techniques), using a variety of surgical and patient-reported outcome measures, with a view to exploring the question of what ‘success’ might actually mean. As with most studies of MAT, many patients required additional concomitant procedures alongside MAT, complicating the assessment of outcomes. It was hypothesized that patients classified as ‘surgical failures’ or ‘clinical failures’ would have significantly lower outcome measures compared to those who were not.

Ethical and Risk Assessment Approval was obtained from the University of Edinburgh’s Moray House School of Education, and use of the surgeon’s data was approved by The London Bridge Hospital Clinical Governance Committee, and information governance was maintained throughout the study.

Following informed consent, a retrospective analysis of patients aged 18 to 50 years who had undergone MAT by a single surgeon between 2008 and 2014 was undertaken. The indications for MAT were symptomatic patients with complete or subtotal loss of a meniscus causing pain and reduced function, with evidence of post-meniscectomy degenerative damage where artificial joint replacement surgery was not yet indicated. Each patient had previously been assessed for suitability for MAT by history, clinical examination, imaging and arthroscopic findings, with the principles of treatment fully explained. Alignment and stability were assessed before MAT, to check whether concomitant procedures were also required. Table 1 shows the list of the 31 concomitant procedures performed, with anterior cruciate ligament (ACL) reconstruction being the most common.

Sixty patients were identified and sent questionnaires, with zero excluded.

Of the total of 60 MAT procedures performed on 60 patients by this surgeon during the period 2008–2014, 43 (72%) patients responded, with a mean follow-up of 3.4 (±1.6) years. Seventeen patients (28%) did not respond. Table 2 outlines the baseline demographics of the patients.

Data were not normally distributed for all the PROMs except the IKDC and VAS scores. Therefore Table 3 shows the median and interquartile PROM scores for the whole cohort and for the surgical failure and clinical failure groups. For 3 patients, their responses to the IKDC were < 90% complete, rendering them invalid, and these were therefore excluded. All KOOS scores were valid.

The complication rate was 21%. Four patients were classified as surgical failures and nine were classed as clinical failures, with three patients having both surgical and clinical failure. The mean times to surgical and clinical failures were 1.8 and 2.6 years respectively. Statistical comparison tests did not show any significant differences for any of the PROMs between the two groups. However, Effect Sizes for KOOS and Lysholm scores were large. The Effect Sizes for IKDC, post-operative Tegner scores and SF-12 mental subscale were medium-large and SF-12 (whole test) was small-medium. There was no effect between the two surgical failure groups for pre-operative Tegner Activity Scale. For clinical failure, there were significant differences (p < 0.001) between groups for KOOS and Lysholm scores, for the IKDC, the VAS and the post-operative Tegner values (all p = 0.002), for the SF-12 (p = 0.041) and SF-12 physical subscale (p = 0.024) but not for pre-operative Tegner. Effect Sizes for all PROMs were large, except for pre-operative Tegner score, which was small-medium. The Wilcoxon test showed that there was a significant difference (p < 0.001) between pre- and post-operative Tegner scores for the whole sample, and this also had a large effect size.

Six patients stated that they would not have the procedure again, with two of these being clinical failures, one being a surgical failure and clinical failure, but three of these did not come under either of the definite failure categories. Five patients were unsure whether they would repeat the procedure again, with the remaining 32 stating they would undergo it again. Table 4 shows the PROM scores grouped by patients’ statements about whether or not they would undergo the procedure again (Yes / No / Not Sure). There were no statistical differences between the groups, although the KOOS (p = 0.091), SF-12 physical subscale (p = 0.085) and Lysholm scores (p = 0.056) approached significance. Effect sizes between the groups were large for KOOS and Lysholm values, medium-large for the SF-12 physical subscale and small-medium for all other PROMs.

Full results for the BMLRs are shown in Tables 5 and 6. The regression found none of the five risk factors was significantly related to the chance of surgical failure, although gender (being female) was nearly significant (p = 0.064). In the regression for clinical failure, again, none of the risk factors was significant, although gender (p = 0.077) and number of concomitant procedures (p = 0.067) were nearly significantly linked to the chance of clinical failure.

Pre-operative and post-operative patient sporting activity, using a combination of findings from experimental studies and clinical investigations to define intensity of impact [28], is shown in Table 7. Thirty two (74%) patients returned to fewer, lower intensity sporting activities following MAT, with 34 patients (79%) expecting this change. Of those who did not return to the same level of sport, 73% stated that they would undergo the procedure again. At review, 39.2% were satisfied, 20.9% were neither satisfied nor dissatisfied, and 40% were dissatisfied with their level of sporting activity achieved following surgery.

This study sought to record the PROMs of patients who had undergone MAT, and to investigate whether there were differences reported in those patients with ‘surgical failure’ or ‘clinical failure’ versus those categorised as being ‘successful’.

The results for Lysholm, KOOS and Tegner scores compared favourably or similarly to other reported series in the published literature [14‐16], although our IKDC scores were lower [15]. Naimark et al. reported similar outcomes in patients who underwent arthroscopic partial meniscectomy [29]. The Tegner scores were significantly reduced following surgery, probably due to the specific post-operative restrictions that were recommended by the operating surgeon.

It should be noted that there are other techniques to replace menisci, such as the Actifit and the Collagen Meniscal Implant. Interestingly, the studies evaluating the effect of these treatments also do not have a consensus on accepted definitions of failures [30]. However, these implants are only suitable specifically for the treatment of partial meniscal loss where the is still continuity of the peripheral meniscal rim, as they require a meniscal rim for attachment [31]. This therefore makes comparison to MAT difficult. De Coninck et al. reported more favourable outcomes in IKDC with the Actifit compared to our study, but our study showed higher results in KOOS and Lysholm scores [32]. The two studies had comparable mean ages, but De Coninck excluded patients who had severe cartilage damage, whilst in our study 44% had an Outerbridge Grade III or IV, and this may explain the higher IKDC scores in their study. Furthermore, the KOOS and IKDC scores both contain questions regarding activities such as jumping, twisting and pivoting; activities which our participants were advised not to undertake, and so may bias results, as other studies also report higher scores in these questionnaires [33, 34].

While the number of patients undergoing MAT remains small, finding the right tool to evaluate patients is problematic. New scoring systems may provide improved evaluation of MAT.

The Western Ontario Meniscal Evaluation Tool (WOMET) may be appropriate. It is a specific questionnaire aimed to detect the impact of meniscal tears on health-related quality of life [35]. When first designing this present study, it was decided not to include WOMET as there were doubts at that time with regard to WOMET’s reliability, construct validity and responsiveness [36]. It would also have caused questionnaire fatigue by adding a raft of further questions which may have reduced questionnaire completion rates. Whilst Sgroi et al. reported superior measurement of knee function and quality-of-life impairment using WOMET compared to the KOOS in patients treated for meniscal tears [37], to the authors’ knowledge, it had not been used in studies purely investigating MAT. Whether WOMET is indeed a more useful tool for the evaluation of patients who have undergone MAT, however, is a topic in itself for further research. Until the most appropriate and validated scoring system is established for the evaluation of patients who have undergone MAT, this present study suggests that success cannot be adequately defined.

The surgical failure rate of 9% was similar to that reported in other comparable studies, which have reported a failure rate of 10.4% [13]. The clinical failure rate of 21% was greater than that reported by Lee et al. [16] (8.5%) and Zaffagnini et al. [14], (11%). The complication rate was 21%; similar to the 21.3% rate reported in the meta-analysis by El Attar et al. [7]. A consensus regarding the current definition of clinical and surgical failure is required, since two (50%) patients with surgical failures and six (66.7%) with clinical failures said that they would undergo the MAT procedure again or were unsure. It is concerning that previous authors have solely used these endpoints to report the efficacy of MAT.

This study, therefore, suggests that the term ‘patient satisfaction’ may define success better than ‘surgical failure’ or ‘clinical failure’. Patient satisfaction, in terms of activity levels and current symptoms, do not in this study correlate with surgical outcomes. Patient satisfaction may be best incorporated into a questionnaire using the WOMET score, but again, this requires further research and discussion with specific patient focus groups.

The BMLR analysis did not find any risk factors that were significantly associated with failure, although gender (surgical and clinical failure) and number of concomitant procedures (clinical failure) were nearly significant. However, given the heterogeneity of concomitant procedures, future studies could investigate the effect of these different procedures on the success of MAT. The odds ratios showed that except for lateral (vs medial) MAT in surgical failure, all risk factors increased the odds of surgical or clinical failure. However, many of these odds ratios were close to unity, reinforcing their lack of statistical significance.

This study’s analyses support the findings of Parkinson et al. [13] and Ahn et al. [38] that potential risk factors of age, gender and number of concomitant procedures were not related to surgical failure. However, our study disagrees with those studies, in that Outerbridge classification and lateral vs medial were not found to be significant factors in our analysis. Whilst some studies have found that articular cartilage damage at the time of surgery and medial (vs. lateral) allografts are associated with a higher risk of surgical failure [13], this study supports the findings of other papers that conclude that cartilage damage does not make a difference [38‐40] and that medial allografts are not at a higher risk of surgical failure [40]. Importantly, however, it should be noted that where articular cartilage damage was found in the relevant compartment of a patient’s knee at the time of MAT surgery, the articular cartilage pathology was also treated at the same time; with radiofrequency coblation chondroplasty for unstable or rough partial thickness chondral damage, with microfracture for small (<2cm²) full-thickness defects or with articular cartilage grafting (using Chondrotissue grafting) for larger (>2cm²) defects. Therefore, it could be surmised that appropriate concomitant treatment of articular cartilage lesions possibly lessens or even eliminates the otherwise increased risk of failure after MAT that might otherwise be associated with increasing severity of articular cartilage damage.

One reason for the lack of significance found in some of our analyses might be the small sample size in the present study, combined with the number of risk factors examined, which could have led to a lack of statistical power. The methods in the present study were better controlled compared to some previous research: only fresh-frozen meniscal allografts (which are biologically and mechanically superior to other forms of preservation [41]) were used, and these were all sourced from a single provider, and a single one-surgeon technique was used. Previous authors have either not reported their allograft sources [38, 42] or have used multiple providers with different preservation techniques [13]. Additionally, the generally good outcomes following MAT in the present study may be attributable to the surgical technique; suture fixation of the allograft to the capsule combined with trans-tibial bone tunnel suture fixation is the current preferred technique [14, 43, 44]. Techniques using bone plug fixation risk articular cartilage damage and have been associated with a higher failure rate [45, 46].

There was no difference in the SF-12 mental sub-scores in patients who said that they would undergo surgery again versus those who would not, although the physical sub-score was approaching significance. Saltzman et al. [47] and Rue et al. [48] also previously reported that the physical component improved significantly following MAT but that the mental score did not.

It was observed that following MAT, patients returned to fewer and lower levels of sport. Seventeen patients were satisfied with the level of sport achieved following MAT, but 17 others were dissatisfied. Nearly 80% of patients did expect there to be a difference in the level of physical activity that they could achieve after MAT, although there were no differences between the satisfaction levels between those who expected there to be a difference and those who did not. It is unclear whether the protective post-operative regime or the MAT procedure itself explains the dissatisfaction about not being able to return to sport.

This retrospective study, with 3-year follow-up of 43 patients undergoing MAT, reported outcome scores that are consistent with success, with scores of 82.1, 62.1 and 88 for KOOS, IKDC and Lysholm, respectively. Those patients classified as clinical failures had significantly lower outcome scores, although the four patients with surgical failure did not, possibly due to the small sample. It was not possible to identify significant pre-operative risk factors associated with surgical or clinical failure, although gender (being female) and the number of concomitant procedures were nearly significant and had high odds ratios. The use of the terms ‘surgical failure’, ‘clinical failure’ and ‘patient satisfaction’ may need further consideration, since up to 60% of patients defined as clinical / surgical failures said that they would undergo the procedure again. Either newer outcome measures, such as the WOMET, or a bespoke MAT-specific measure could help to re-define success / failure and establish a more sensitive marker of graft survival and the anticipated chondroprotective effects of MAT.