Donor-site morbidity after osteochondral autologous transplantation for osteochondritis dissecans of the capitellum: a systematic review and meta-analysis

The most important finding of the present study is that donor-site morbidity after osteochondral autologous transplantation for capitellar osteochondritis dissecans occurred in 7.8% within the knee-to-elbow group and 1.6% within the rib-to-elbow group. In the knee-to-elbow group, knee pain during daily activities (7.0%) and locking sensations (0.8%) were reported; in the rib-to-elbow group, one case was complicated by a pneumothorax. There was no significant difference in proportion of donor-site morbidity between the two harvest techniques. The findings of this systematic review emphasize the importance of associated donor-site morbidity following graft harvesting in the treatment of capitellar OCD.

The rate of donor-site morbidity in our study is comparable to the morbidity rate after knee-to-knee OATS (6%), as reported by Andrade et al. [2]. Interestingly, the authors reported a higher rate after knee-to-ankle OATS (17%) [32]. The higher rate compared to our findings may be the result of more grafts that were harvested (three versus two) [2], although the influence of the number of grafts on morbidity is controversial [1, 2, 25, 27]. Patient characteristics may have played a role as well. The vast majority of patients in our study were adolescent (15 years), high-demanding athletes, while patients with talar OCD are typically older (32 years [25]). We hypothesized that patients in our study may have been in a better physical condition before treatment, and therefore faster recovery with less donor-site effects may be expected.

In the present study, the rate of donor-site morbidity ranged from 0 to 57% in studies in whom grafts were harvested from the knee. Weigelt et al. reported morbidity in eight of 14 patients (57%): occasional pain during heavy lifting in seven patients and intermittent locking sensations in one patient [33]. The advanced age after a relative long follow-up (7 years) may be the reason for more morbidity, as well as it may be the result of large grafts (8–11 mm) that were harvested [1, 2, 25, 27]. Two included studies focused on donor-site effects in particular [9, 21]. Despite encouraging clinical results reported by Iwasaki et al., MRI evaluation revealed alterations in signal intensity in 89% of donor sites, suggesting fibrous filling of donor holes [9]. Long-term follow-up is needed to see if these changes in signal intensity are permanent and clinically meaningful. Nishimura et al. found a delay in recovery of quadriceps muscle strength [21]. This indicates that even if symptoms may resolve quickly after harvesting, the donor knee is at risk for injury due to muscle weakness within the first year. Three studies attempted to prevent postoperative bleeding by filling donor tunnels [9, 15, 21]. As included studies were limited by small case series, potential beneficial effects remain unclear. Favorable results have been reported using synthetic implants in repair of knee OCD [4, 6], although literature lacks large sample comparison studies.

Three studies that were included in this review harvested grafts from either the 5th or 6th costal-osteochondral junction to repair capitellar OCD [22, 28, 30]. Although hard scar tissue was detected in palpation, radiographs showed new subperiostal bone formation and no long-term symptoms were observed [30]. However, harvesting grafts from the rib area is a technically demanding procedure that has been described by only a few studies. Most surgeons who perform capitellar OATS are more familiar with knee anatomy rather than rib anatomy. Moreover, risk for devastating donor-site morbidity remains as harvesting may be complicated by a pneumothorax due to close proximity to the costal pleura, which may lead to a prolonged hospital stay [30]. If familiar with this technique, this may be an option in the treatment of large capitellar OCDs (>15 mm). As large lesions usually require multiple cylindrical grafts, one may want to avoid the risk for donor-site morbidity of the knee.

Using the IHE scale to evaluate the methodological quality of included studies [5, 20], the estimated potential risk of bias ranged from low to high, with the majority of studies found to be of moderate risk of bias (9 of 11). Studies scored the lowest on criteria related to outcome measures, as well as on the fact that cases were collected in a single center in each study. Therefore, the findings of this systematic review should be interpreted by taking into account some limitations. First, a major limitation is the incomplete reporting regarding outcomes related to the donor site. As physical examination or imaging was rarely thoroughly reported, studies lack objective assessment of donor-site effects. Additionally, subjective assessment of knee function by means of patient reported outcome measures was only performed in four of eight studies (Lysholm score) [9, 16, 21, 33]. In the remaining studies, no attempt was made to evaluate knee function, nor was physical examination thoroughly described [11, 15, 24, 35], as this was also the case in two rib-to-elbow studies [22, 28]. Also, in none of the included studies, either subjective or objective preoperative assessment of the donor site was reported. As data were obtained from studies that evaluated donor-site effects in varying degrees, we hypothesized that donor-site morbidity may be substantially underreported in this population. Also, due to incomplete reporting, we were unable to investigate associations between morbidity and harvest characteristics including donor-site location and the size or number of grafts. Second, included studies are limited by case series with small numbers of patients because capitellar OCD is a rare condition in the general population, and capitellar OATS is a relatively new procedure. Third, after pooling the data of included studies, we found no significant difference between the proportions of donor-site morbidity for the knee-to-elbow group and the rib-to-elbow group (p > 0.05). One may argue whether it is valid to combine study data because of between-study variance; however, we attempted to alleviate statistical heterogeneity with the use of a random effects model. This model has been used previously in systematic reviews who pooled data from case series studies [19, 34].

Future studies should comprehensively evaluate effects related to the donor site. Patient’s symptoms and physical examination should be reported for each patient. Additionally, the use of a patient reported outcome measure to assess knee function should be a routine part of clinical evaluation, both preoperatively and postoperatively. Radiographs should be performed to evaluate potential progression to osteoarthritis at 1 year and may be performed in cases in which donor fillers were used. MRI evaluation should, because of cost-effectiveness reasons, only be ordered in case of persistent symptoms. Besides the evaluation of donor-site effects, harvest characteristics should be reported in great detail, such as donor location, number of grafts, graft diameter, and depth. Also, alternative harvest methods should be investigated to have no longer the need to violate the integrity of an asymptomatic knee or rib in an adolescent athlete.

The findings of the present investigation demonstrate a considerable risk for donor-site morbidity following capitellar OATS. Although good-to-excellent results related to the elbow have been reported after capitellar OATS, surgeons should be aware of the risk for donor-site morbidity and patients should be counseled about this issue. Knowing the overall risk for donor-site morbidity is also relevant in surgical decision making. Taking this into consideration, surgeons could consider other resurfacing techniques such as allografting or autologous chondrocyte transplantation.