Surgical reconstruction is a cost-efficient treatment option for isolated PCL injuries

Posterior cruciate ligament (PCL) injuries are occurring in increasing frequency, and studies report that injuries to the PCL account for as much as 17% of all knee injuries [10]. Treating these injuries is related to a cost to the society. A majority of such injuries can be treated nonoperatively with a satisfactory result [6, 12, 17]. As a consequence, most patients suffering an isolated PCL injury traditionally go through a regime of rigid rehabilitation. If the injury is acknowledged within the first few weeks, the nonoperative approach usually consists of a PCL brace and a series of training exercises [6, 8]. If a nonoperative approach is unsuccessful, surgical reconstruction of the PCL is the alternative. In recent years, there has been increased focus on PCL reconstruction (PCLR). The PCL consists of two functional bundles [8]. Traditionally single-bundle (SB) reconstruction has been the preferred method of reconstruction, but double-bundle (DB) reconstruction has also been reported since 1999 [4]. DB PCLR commonly includes the use of allografts instead of or in addition to the hamstring graft as well as extra time in the operating room, which adds costs to a SB reconstruction with an autograft. For DB reconstruction, the results have not been reported to being superior to those of SB reconstructions until 2010. Describing a new technique with the use of two allografts, Spiridonov et al. [19] reported excellent outcomes both when it comes to objectively measured stability and patient-reported outcomes. Later, biomechanical studies have shown superior results with the same DB PCLR technique compared to SB technique [7, 20]. To our knowledge, there are no studies that have looked at the costs of the different treatment algorithms for treating PCL injuries. With the cost related to surgery for both SB and DB reconstructions, a gain in outcome should be required to justify the procedure over nonoperative treatment alone. As patient-reported outcome measure, the knee injury and osteoarthritis outcome score (KOOS) is most commonly used in the Scandinavian countries. If there is an improvement in the KOOS, following treatment is therefore an important question. Other important issues include whether the patients are able to return to work following surgery and the duration of the sick leave, health care utilization, and in addition cost per health care gain, measured in KOOS QoL. In a longer perspective, it is interesting to evaluate whether reconstruction has an effect on the development of osteoarthritis (OA) and whether choosing DB over SB reconstruction leads to a decrease in revision surgery rates.

This study aims to enlighten the cost of both operative and nonoperative treatments of PCL injuries and explore the cost per expected gain measured in quality of life following surgery.

The average calculated cost of nonoperative treatment was €3382. Incremental cost for SB PCLR was 154% and another 61% for DB PCLR, given that the preoperative rehabilitation programme is of the same length and intensity as that for the nonoperatively treated patients, details in Table 2. The additional cost of reconstruction on average equals the cost to society for 3 (SB)- or 6 (DB)-week absence from work in Norway. The allografts, material for extra fixation and time in surgery are the factors differentiating cost of the two surgical treatment options. In addition to the cost related to surgery, there is a cost of post-operative rehabilitation for both the SB and DB groups compared to that of the nonoperatively treated patients (Table 2). Given an expected gain in KOOS QoL of 27 points, this provides an incremental cost-effectiveness ratio (ICER) of 365. Adding the additional cost for DB reconstruction, this equals another relatively low incremental gain in QALYs. For a similar ICER score with DB over SB PCLR, this requires an incremental gain in KOOS QoL of 28 points (Table 3).

The most important finding in the present study is that it can easily be argued that the gain in QoL-measures versus cost justifies the surgical option for the studied patient group. The average reported gain in KOOS QoL is about 27 points compared to no surgical treatment, which is a clinically relevant increase. The fact that the cost of PCLR is equivalent to a very small gain in QALYs supports this. The mentioned gain in KOOS may represent the ability of returning to work or sports at the same level as prior to injury, which in both cases may be of high value to the individual as well as to society.

Surgery adds a substantial extra cost in treating PCL injuries. Current literature reports good/excellent results following surgical reconstruction both when it comes to objective measures such as stress radiographs and when it comes to patient-reported outcomes. Taken into consideration that there is an additional cost related to DB surgery, one should demand a superior result for DB reconstruction that is both statistically significant and clinically relevant. For KOOS, the minimal clinically relevant change is about 10 points for the most relevant subscales [15]. If an additional gain in 10 points in KOOS QoL or Sport/Rec are enough to justify, the extra cost of DB surgery can be debated. For DB to be as cost-effective as SB, this would require an additional gain in KOOS QoL of 28 points which is a highly unlikely outcome.

When it comes to total cost, the differences between SB and DB reconstructions are relatively small when compared to the cost to society for work absence. If only one extra person per 100 treated would be able to return to work, or if those treated in general would be able to return to work more than 3 weeks earlier, DB surgery would have the same or lower cost as SB surgery. On average, we can estimate that patients have a sick leave period of 10 weeks following PCLR. Whether the required 3 weeks of earlier return to work is a realistic goal, is uncertain at best. A measure of the value of patients returning to sports is interesting, but difficult to calculate.

Sufficient data considering surgical revisions are not yet available to confirm whether the number of revisions is reduced after DB reconstruction. The current 5-year revision rate in Norway is about 4% which is similar to reported revision rates reported for ACL reconstructions in the Scandinavian countries [1, 9]. Even if the number of revisions were reduced to 0% following DB reconstruction, this alone would not be enough to save the extra money spent on switching form primary SB to DB reconstruction. Hence, a possible reduction in revision rate is not a sufficient argument for choosing DB over SB. This is based on the assumption that an average revision is done in one stage with the same post-operative rehabilitation programme used as for primary reconstructions.

The discrepancy in the reported operating time from the respective clinics can probably mainly be explained by the surgeons’ experience. Of the hospitals performing PCLR, eleven have reported only 1 PCLR per year. More experienced surgeons can probably reduce the duration of surgery even more, although the difference in time between single- and double-bundle surgeries should be fairly consistent as it only adds a relatively short standardized procedure. If this extra time should be considered, an extra cost is also a matter of discussion. The cost per minute in the OR is here calculated based on average numbers per minute from start in the morning until the day ends. It does not take into consideration how many patients are treated per day. Only 15 min extra time for the surgical procedure possibly implies that the same number of patients will be treated per day (8 am–4 pm in a typical clinic) for both SB and DB surgeries. This is because normally only one or two PCLRs are performed the same day and therefore combined with other relevant knee procedures in the relevant OR on the actual day. Then, we are left with the cost of the allografts and extra fixation material as the only additional cost for DB over SB.

With DB reconstruction, there is an increased volume of graft which theoretically transfers to more restrain to stress and improved stability. This is supported by biomechanical studies [20]. Some clinical studies report no difference in results or an advantage of DB reconstruction [14, 21]. For the high-performance athlete planning return to sport at preinjury level, DB reconstruction is a good treatment option as there are no studies implying inferior results with DB reconstruction. When it comes to other patient categories, both approaches should be considered. Although it can be argued that the extra cost with a DB technique is not so high that it should influence the method of choice given that one yields a superior result.

The cost to society and employers for time away from work following surgery comes in addition when compared to conservative treatment. If the PCL-injured patients are to be treated surgically with either method, one should expect both short-term and long-term results that justify this cost. There is of today no available literature comparing nonoperative treatment to operative treatment. There is one study available showing both statistically significant and clinically relevant improvement in KOOS scores 2 years following PCLR for an isolated PCL injury [11]. All these patients have gone through a period of rehabilitation prior to surgery. They have comparable increases in KOOS similar to that of patients treated for an anterior cruciate ligament injury. However, their KOOS scores are not even close to that of a control population with no prior knee injuries. The PCL patients ended up with a KOOS Sport/Rec score of 46 and a QoL score of 52 post-operatively. In a control population, the scores are 91 for Sport/Rec and 90 for QoL. It can be debated whether a clinically relevant increase in KOOS score is sufficient to justify operating on these patients or whether one should demand better outcomes. In addition, we do not yet know how good the results are after a longer period than 2 years. On the other hand, one could argue that the KOOS is not a linear scale and that an increase in the QoL subscale from 30 to 50 is an important increase as this translates to a better function in daily life activities and maybe a sufficient improvement for return to work for those that have moderately physical demanding work. This again translates to a benefit to society.

As there is room for improvement in post-operative results, one can argue that a central part of this is the surgical procedure. The only way to improve is to develop the reconstruction procedures and continue the research in the area.

A limitation to the data is the numbers representing total cost for surgical treatment. This is the cost in the authors’ hospital which should be comparable to other Norwegian and maybe even European hospitals, but there will be variations from country to country. These variations will also depend on how the costs are calculated. The cost to society will vary according to the organization of each country’s health services including who pays the hospital bill and how the expenses for work absence are calculated.

Another limitation may be the choice of procedure for DB surgery. This procedure has not been widely performed over years and not yet compared to other procedures for DB reconstruction. It is also an issue with allografts as they are not permitted for use in all countries. Another limitation is regarding the patients who are treated surgically without first going through a rehabilitation period of several months. For these patients who are treated more or less acutely, the first rehab cost is saved and leaves surgery as the only additional cost assuming the cost of preoperative and post-operative rehabilitation is equivalent. This economically favours early over later surgery. It could also be argued that twice a week with a physiotherapist is insufficient for an optimal rehabilitation. This has the potential of increasing the cost of preoperative and post-operative rehabilitation accordingly. QALY is a common parameter in evaluating the health outcome of new treatments. As data necessary to calculate QALYs are not available, this can be considered as a limitation to the current study. We will still argue that cost per gain in knee-related quality of life (KOOS QoL) is as good alternative measure to QALY for this patient group. Depending on which questionnaire or scoring tool that is applied for calculating QALY, one will get variation in the results. This again makes comparing studies looking at the same patient population less relevant if not the same questionnaire is for the QALY calculation.

In a clinical setting, the current study supports surgery as a treatment option in a health economic perspective as the cost is equivalent to a low gain in QALY. Both SB and DB reconstructions should be considered, keeping in mind that DB is less cost efficient than SB and as such has the potential of increasing cost with an uncertain incremental gain in QALY. In a future perspective, randomized controlled trials are needed to compare the outcomes of SB to DB PCLR including nonoperative treatment as a control group.

There is a substantial extra cost to treating PCL injuries if it comes to surgical reconstruction. As the results following nonoperative treatment are reported to be good to excellent, it is difficult to justify the cost of surgery without first trying a nonoperative approach for isolated PCL injuries. If surgery is indicated, the method that most likely restores the biomechanics of the knee should be preferred almost irrespective of costs aspects.