Concomitant Osteotomy Reduces Risk of Reoperation Following Cartilage Restoration Procedures of the Knee: A Matched Cohort Analysis

Abstract

Objective

The objective of this study is to compare the (1) reoperation rates, (2) 30-day complication rates, and (3) cost differences between patients undergoing isolated autologous chondrocyte implantation (ACI) or osteochondral allograft transplantation (OCA) procedures alone versus patients with concomitant osteotomy.

Study Design

Retrospective cohort study, level III.

Design

Patients who underwent knee ACI (Current Procedural Terminology [CPT] 27412) or OCA (CPT 27415) with minimum 2-year follow-up were queried from a national insurance database. Resulting cohorts of patients that underwent ACI and OCA were then divided into patients who underwent isolated cartilage restoration procedure and patients who underwent concomitant osteotomy (CPT 27457, 27450, 27418). Reoperation was defined by ipsilateral knee procedure after the index surgery. The 30-day postoperative complication rates were assessed using ICD-9-CM codes. The cost per patient was calculated.

Results

A total of 1,113 patients (402 ACI, 67 ACI + osteotomy, 552 OCA, 92 OCA + osteotomy) were included (mean follow-up of 39.0 months). Reoperation rate was significantly higher after isolated ACI or OCA compared to ACI or OCA plus concomitant osteotomy (ACI 68.7% vs. ACI + osteotomy 23.9%; OCA 34.8% vs. OCA + osteotomy 16.3%). Overall complication rates were similar between isolated ACI (3.0%) and ACI + osteotomy (4.5%) groups and OCA (2.5%) and OCA + osteotomy (3.3%) groups. Payments were significantly higher in the osteotomy groups at day of surgery and 9 months compared to isolated ACI or OCA, but costs were similar by 2 years postoperatively.

Conclusions

Concomitant osteotomy at the time of index ACI or OCA procedure significantly reduces the risk of reoperation with a similar rate of complications and similar overall costs compared with isolated ACI or OCA.

Introduction

Chondral and osteochondral defects of the knee are common causes of musculoskeletal morbidity and have a prevalence as high as 66%.^1-4 Because of the avascular nature of articular cartilage, chondral defects have a poor capacity to regenerate or self-repair and, if left untreated, may lead to early osteoarthritis.^5-9 The number of cartilage restoration and repair procedures has increased in recent years.^10-15

Various cartilage restoration procedures have demonstrated good results in smaller chondral defects (<2 cm²), including chondroplasty, marrow stimulation techniques, and osteochondral autograft transplantation.^12,13 However, in the management of larger cartilage defects, the use of autologous chondrocyte implantation (ACI) and osteochondral allograft transplantation (OCA) are generally recommended.^11,14,15 Both ACI and OCA have favorable long-term outcomes, with up to 87.5% and 91.0% graft survivorship reported at 10 years, respectively.^15-24

Malalignment and overload of the affected compartment can play a large role in the management of chondral defects. Osteotomies, including tibial tubercle osteotomy (TTO), high tibial osteotomy (HTO), and distal femoral osteotomy (DFO) can significantly unload the affected compartment of the knee, normalize the mechanics, and reduce contact pressures on the implanted graft, ultimately contributing to superior graft survivorship and improved clinical outcomes.^25-28 Previous studies examining concomitant osteotomy at the time of cartilage restoration procedure consist mainly of smaller case series, but have demonstrated positive, lasting results with improved clinical outcomes and graft survivorship.^29-36

Nonetheless, the extent to which concomitant osteotomy provides an improvement in objective clinical outcomes after cartilage restoration procedure is unclear as the current literature making this direct comparison is limited. The purpose of this study was to compare (1) the reoperation rates, (2) 30-day complication rates, and (3) cost differences between subjects undergoing isolated ACI or OCA procedures versus subjects with concomitant osteotomy in a large insurance database. The authors hypothesized that there would be no significant differences in (1) reoperation rates, (2) 30-day complications rates, (3) cost between subjects with and without concomitant osteotomy.

Materials and Methods

The MarketScan Commercial Claims and Encounters database (Truven Health Analytics, Ann Arbor, MI) from January 1, 2007 to December 31, 2016 was used for this study. This database is a collection of commercial inpatient, outpatient, and pharmaceutical claims of more than 75 million employees, retirees, and dependents representing a substantial portion of the U.S. population covered by employer-sponsored health insurance. MarketScan contains 53 million patient inpatient records, 40 million with employer sponsored insurance, 3.7 million with Medicare Part B and 6.8 million on Medicaid for a total of over 28 billion patient records. Only outpatient records in MarketScan were queried for this study. Due to MarketScan’s sourcing from large employers, its data boasts superior longitudinal tracking of patients. The MarketScan database contains International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) and 10th revision, Clinical Modification (ICD-10-CM), Current Procedural Terminology (CPT), and Diagnosis Related Group (DRG) codes as well as National Drug Codes (DEA).

Patients who underwent open autologous chondrocyte implantation or open osteochondral allograft transplantation of the knee in the outpatient setting with a minimum of 2-year follow-up were queried in the database using CPT codes 27412 and 27415, respectively. Records of the index surgery were reviewed for analysis. Patients who had both procedures were excluded, thus making the cohorts mutually exclusive. Patients without a CPT modifier for laterality (left or right) were excluded. Additionally, patients undergoing concomitant ligament reconstruction procedures were excluded. The resulting cohorts of patients that underwent ACI and OCA were then divided into those patients without osteotomy and those patients with concomitant osteotomy using CPT codes 27457, 27450, or 27418. The cohorts without osteotomy were then matched to the patients that underwent concomitant osteotomy using a propensity-score match (PSM) for each primary procedure (ACI, OCA). A caliper of 0.01 was utilized in the match, and replacement of patients in the algorithm was not allowed. All baseline ten characteristics were input into the algorithm, leading to the matched covariates having no statistically significant differences.

Patient-level variables, including age at time of surgery, sex, and comorbidities (obesity, tobacco use, myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, rheumatic disease, peptic ulcer disease, liver disease, diabetes, hemiplegia, renal disease, malignancy, and AIDS/HIV) were recorded based on claims data ( Table 1 ).

Table 1.

Patient Demographics.

Variable	ACI (n = 402)	ACI + Osteotomy (n = 67)	P	OCA (n = 552)	OCA + Osteotomy (n = 92)	P
Age (years), mean (SD)	32.5 (10.3)	32.9 (9.6)	0.7673	31.5 (10.8)	31.2 (10.3)	0.7843
Female, n (%)	244 (60.7)	40 (59.7)	0.8774	259 (46.9)	43 (46.7)	0.9743
Obesity (BMI ≥ 30 kg/m2), n (%)	7 (1.8)	1 (1.5)	0.8842	4 (0.7)	1 (1.1)	0.7139
Hyperlipidemia (%)	33 (8.2)	7 (10.5)	0.5436	37 (6.7)	6 (6.5)	0.9486
Diabetes, n (%)	9 (2.2)	3 (4.5)	0.2826	6 (1.1)	1 (1.1)	1.0000
Hypertension, n (%)	30 (7.5)	4 (6.0)	0.6627	40 (7.3)	7 (7.6)	0.9016
Tobacco use, n (%)	25 (6.2)	4 (6.0)	0.9376	34 (6.2)	5 (5.4)	0.7873
CCI, mean (SD)	0.12 (0.5)	0.18 (0.5)	0.3825	0.1 (0.4)	0.1 (0.3)	0.7965
Follow-up (months), mean (SD)	40.2 (16.8)	33 (22.4)	0.1091	40.2 (16.8)	33 (22.4)	0.1091

ACI = autologous chondrocyte implantation; OCA = osteochondral allograft transplantation; BMI = body mass index; CCI = Charlson Comorbidity Index.

The primary outcome measure was rate of ipsilateral reoperation. Reoperation was defined by ipsilateral ACI (CPT 27412), OCA (CPT 27415, 29867), osteochondral autograft transplantation (CPT 29866), synovectomy (CPT 29875, 29876), chondroplasty (CPT 29877), microfracture (CPT 29879), manipulation under anesthesia (CPT 27570), lysis of adhesions (CPT 29884), arthroscopic loose body removal (CPT 29874), debridement for infection (CPT 27301, 27303, 27310, 29871), osteotomy (CPT 27457, 27450, 27418), unicompartmental knee arthroplasty (CPT 27446), and total knee arthroplasty (CPT 27447) that occurred after the index surgery date. Ipsilateral reoperation was queried using CPT codes and modifiers for laterality to ensure that the procedures were reoperations and not contralateral procedures. The time period for potential reoperation was up to nine years postoperatively, depending on when the index ACI or OCA was performed.

The 30-day postoperative complication rates were also assessed for both cohorts using ICD-9-CM codes for bursitis, anemia, knee dislocation, infection, nerve injury, neurologic complications, cardiac complication, heterotopic ossification, sepsis, wound complication, deep vein thrombosis, pulmonary embolus, hematoma, pulmonary complication, urinary tract infection, and delirium.

The cost of the procedures per patient was calculated using the Diagnosis-Related Group (DRG) system. This included the entire reimbursement by the insurer on the day of surgery. The cumulative cost of the procedure per patient was also calculated at 9 months and 2 years postoperatively, including the cost of reoperations.

Statistical comparisons of reoperations, postoperative complications, and payments were made between each matched cohort (osteotomy vs. no osteotomy) for ACI and OCA procedures separately. The significance of differences in means of continuous variables between the 2 groups was determined by the Student t test. The differences in categorical variables between groups was determined by the chi-square test. Percentages in this study were a representation of proportion of the cohort. All P values were reported with significance set at P < 0.05.

As this study includes only analysis of secondary de-identified data, it was not considered human subject research and received exemption from the institutional review board approval at our institution.

Results

A total of 1,113 patients (402 ACI, 67 ACI + osteotomy, 552 OCA, 92 OCA + osteotomy) were included in the study with a mean follow-up of 39.0 months. Demographics and baseline comorbidities were similar between all groups ( Table 1 ).

The overall reoperation rate was significantly higher in the isolated cartilage restoration procedure groups (ACI or OCA), compared with the cartilage restoration procedure plus concomitant osteotomy (ACI 68.7% vs. ACI + osteotomy 23.9%, P < 0.0001; OCA 34.8% vs. OCA + osteotomy 16.3%, P = 0.0004) ( Table 2 ).

Table 2.

Reoperations.

Variable	ACI (n = 402)	ACI + Osteotomy (n = 67)	P	OCA (n = 552)	OCA + Osteotomy (n = 92)	P
Reoperation rate, n (%)	276 (68.7)	16 (23.9)	<0.0001*	192 (34.8)	15 (16.3)	0.0004*
Knee arthroplasty
UKA	1 (0.3)	0 (0.0)	0.6828	1 (0.2)	0 (0.0)	0.6829
TKA	0 (0.0)	0 (0.0)	—	0 (0.0)	0 (0.0)	—

ACI = autologous chondrocyte implantation; OCA = osteochondral allograft transplantation; UKA = unicompartmental knee arthroplasty; TKA = total knee arthroplasty.

^*Statistically significant.

There was no significant difference in the conversion to arthroplasty between the isolated cartilage procedure and the cartilage procedure plus concomitant osteotomy groups. Only a single patient in each of the isolated ACI and OCA groups underwent unicompartmental knee arthroplasty (ACI 0.3% vs. ACI + osteotomy 0.0%, P = 0.68; OCA 0.2% vs. OCA + osteotomy 0.0%, P = 0.68); no patients in any group converted to total knee arthroplasty during the follow-up period ( Table 2 ). Overall complication rates were similar between the ACI (3.0%) and ACI + osteotomy (4.5%) groups (P = 0.52) and the OCA (2.5%) and OCA + osteotomy (3.3%) groups (P = 0.69) ( Table 3 ).

Table 3.

Thirty-Day Complications.

Variable	ACI (n = 402)	ACI + Osteotomy (n = 67)	P	OCA (n = 552)	OCA + Osteotomy (n = 92)	P
Any complication, n (%)	12 (3.0)	3 (4.5)	0.5203	14 (2.5)	3 (3.3)	0.6881
Deficiency anemia, n (%)	2 (0.5)	0 (0.0)	0.5629	2 (0.4)	1 (1.1)	0.3447
Knee dislocation, n (%)	1 (0.3)	0 (0.0)	0.6828	3 (0.5)	1 (1.1)	0.539
Infection, n (%)	3 (0.8)	1 (1.5)	0.5386	1 (0.2)	0 (0.0)	0.6829
Sepsis, n (%)	0 (0.0)	1 (1.5)	0.0142*	0 (0.0)	0 (0.0)	1.000
Heterotopic ossification, n (%)	4 (1.0)	2 (3.0)	0.1796	6 (1.1)	2 (2.2)	0.3835
DVT, n (%)	2 (0.5)	0 (0.0)	0.5629	1 (0.2)	0 (0.0)	0.6829
Hematoma, n (%)	0 (0.0)	0 (0.0)	—	1 (0.2)	0 (0.0)	0.6829

ACI = autologous chondrocyte implantation; OCA = osteochondral allograft transplantation; DVT= deep vein thrombosis.

^*Statistically significant.

The total day of surgery payments were significantly higher in the ACI + osteotomy group compared to ACI alone ($41,171 vs. $15,122, P < 0.0001) ( Table 4 ). This trend narrowed but remained significant at 9 months ($52,034 vs. $41,970, P = 0.01). However, the overall total 2-year payments were similar between the 2 groups (ACI + osteotomy $60,851 vs. ACI $55,373, P = 0.22). Similarly, the OCA + osteotomy group had significantly higher day of surgery ($24,452 vs. $15,122, P < 0.0001) and 9-month total payments ($32,539 vs. $24,386, P < 0.0001) compared with OCA alone. The overall total 2-year payments were similar between the 2 groups (OCA + osteotomy $35,661 vs. OCA $31,917, P = 0.13).

Table 4.

Procedure Payments.

Variable	ACI (n = 402)	ACI + Osteotomy (n = 67)	P	OCA (n = 552)	OCA + Osteotomy (n = 92)	P
Day of surgery payments	$15,122	$41,171	<0.0001*	$14,455	$24,452	<0.0001*
Total payments
9 months	$41,970	$52,034	0.0104*	$24,386	$32,539	<0.0001*
2 years	$55,373	$60,851	0.2214	$31,917	$35,661	0.1286

ACI = autologous chondrocyte implantation; OCA = osteochondral allograft transplantation.

^*Statistically significant.

Discussion

Isolated cartilage restoration procedures, specifically autologous chondrocyte implantation and osteochondral allograft transplantation, have significantly higher reoperation rates, similar complication rates and similar 2-year costs compared with the same cartilage restoration procedure with a concomitant osteotomy. Thus, the addition of an osteotomy at the time of the index cartilage restoration procedure was protective against early reoperation, while not causing a significant increase in complications or overall costs.

Concomitant osteotomy at the time of cartilage restoration procedure was identified as an independent predictor of a reduced rate of reoperation in both the ACI and OCA groups in the present study. These findings are supported by the current literature. The utilization of an osteotomy when necessary to correct malalignment and unload the affected compartment is an important aspect of the surgical management of cartilage defects and may help to avoid preventable graft failures.^24,37-41 Osteotomy can significantly unload the affected compartment of the knee, normalize the mechanics, and reduce contact pressures on the implanted graft, ultimately contributing to superior graft survivorship and improved clinical outcomes.^25-28

Multiple studies have reported improved clinical outcomes and graft survivorship when concomitant osteotomy is utilized with ACI.^31,42-44 Trinh et al. ⁴³ found that the 23% of patients who underwent ACI plus concomitant patellofemoral realignment osteotomy had significantly greater improvements in clinical outcome scores compared with isolated ACI. They found no significant difference in reoperation rate between isolated ACI and ACI plus osteotomy once hardware removal was excluded. While in our study only 14% of ACI cases had a simultaneous osteotomy, in the case series by Zardakis et al., ⁴⁴ 91% of U.S. Army Servicemembers who underwent patellofemoral second-generation ACI had a concomitant TTO. They reported good results after ACI and ACI plus TTO with a successful return to work rate of 78% in a demanding military population at a mean follow-up of over 4 years.

In the case of osteochondral allograft transplantation, residual ipsilateral malalignment has been shown to be an independent risk factor for graft failure.^35,40 León et al. ³⁵ reported graft survivorship of nearly 60% at 25 years in patients who underwent femoral condyle OCA with concomitant osteotomy. Persistent postoperative malalignment resulted in more graft failures (29% vs. 4%) and was a risk factor for graft failure with a hazard ratio of 6.55 at long-term follow-up. Krych et al. ⁴⁰ found that 56% of failed cartilage surgeries, including OCA, were associated with residual limb malalignment. In their systematic review of OCA, Assenmacher et al. ⁴⁵ found osteotomies to be the three most common concomitant procedures: patellofemoral realignment 39%, DFO 25%, and HTO 15%.

Patients undergoing isolated ACI as the index procedure had a significantly higher reoperation rate (68.7%) compared with those undergoing ACI plus concomitant osteotomy (23.9%) at final follow-up. Additionally, patients undergoing isolated OCA as the index procedure had a significantly higher reoperation rate (34.8%) compared with those undergoing OCA plus concomitant osteotomy (16.3%) at final follow-up. Our study notably has higher reoperation rates than those described in the literature for both ACI and OCA, which have a reported survivorship free from reoperation ranging from 71% to 88% and 76% to 91% at up to 10 years postoperatively, respectively.^15-24

The higher reoperation rate in this study compared to the current literature is likely multi-factorial. Previous studies used multiple definitions for graft failure, including reoperation, revision surgery, conversion to knee arthroplasty, and postoperative failure to improve from baseline per patient reported outcomes. In the current study, we are able to make a direct comparison of reoperation rates between groups by using the same definition of reoperation for ACI, ACI + osteotomy, OCA, and OCA + osteotomy. Thus, providing this study with an internal control.

It is important to note that the majority of the peer-reviewed literature on this topic is from high-volume experts in knee reconstruction and potentially are more familiar with the procedures being studied. Therefore, the current literature may underestimate the true reoperation rate in the general population of orthopedic surgeons. ⁴⁶

Another possible explanation for the elevated reoperation rate in the ACI group is the inclusion of multiple generations of ACI. Clinical outcomes between the different generations are similar, but there is a significantly higher rate of graft hypertrophy in ACI using the periosteal patch (22%) compared with the membrane (6%).^47-50 Based on the dates of our collection period, it is possible that many of the ACI procedures were performed using the periosteal patch, which may explain the overall higher rates of reoperation in both the ACI and ACI + osteotomy groups compared to the OCA groups.

While our study is not able to discern in which compartment of the knee the cartilage restoration procedure took place, it is conceivable that the majority of ACI procedures were performed in the patellofemoral joint. This would be more compatible with the elevated failure rates of 60% previously reported in the literature that occur in patellofemoral ACI at long-term follow-up. ⁵¹

The total day of surgery payments was significantly higher in both osteotomy groups compared with the isolated cartilage procedure groups. This initial increase in cost is expected as an osteotomy is an additional procedure that increases operating room time and adds implant costs. This trend continued, although to a lesser extent, as the osteotomy groups remained significantly higher in cost at the 9-month mark. Nevertheless, the gap between the osteotomy groups and the isolated cartilage procedure group begins to narrow by the 9-month mark, possibly due to the significant difference in reoperation rates between the groups. Furthermore, the cost difference between isolated cartilage procedure and cartilage procedure plus concomitant osteotomy converged and were not significantly different by 2 years postoperatively. Thus, despite the initial increased cost incurred by adding an osteotomy at the time of index cartilage procedure, addition of an osteotomy remains cost-effective compared with isolated cartilage procedure by 2 years postoperatively. Importantly, initial staging procedures for OCA or cartilage biopsy for ACI were not included in the cost analysis as the actual cartilage restoration procedures served as the index surgery date. Therefore, the total cost of both cartilage restoration procedures with and without osteotomy would likely be even higher than reported in the present study, but the cost differences would likely remain the same. Despite high costs of ACI and OCA, both cartilage restoration procedures remain cost-effective options for the majority of patients up to 10 years postoperatively. ⁵²

There are some limitations to this study inherent to all studies that use large databases. A major limitation in a database study looking at realignment is the lack of pre- and postoperative alignment measurements. We also are not able to discern in which compartment of the knee the cartilage procedure took place. Furthermore, the analysis depends the accuracy of the ICD and CPT codes reported, leading to potential inaccuracies, miscoding, or noncoding by physicians as a source of error. Apart from conversion to arthroplasty, the inherent flaws of using CPT coding in a large database study limit our ability to confidently conclude exactly which revision surgery each patient underwent. Additionally, patients can change insurances and therefore, leave the database. However, this was controlled for through inclusion of subjects with minimum 2-year follow-up. It is also possible that the database is not a true representation of the health care population in the United States. However, through the inclusion of outpatient records from patients with federal and commercial insurance, the current study is more inclusive. Additionally, the database only included records during a nine-year period (January 1, 2007 to December 31, 2016), so procedures (index or reoperation) that occurred outside that timeframe may have been missed. The current study also only identified complications that occurred within 30 days from the index procedure to increase the likelihood that the complications identified were related to the postoperative course following ACI or OCA and not another medical condition that arose. Thus, late complications that occurred outside the 30-day window were not captured in the present study. Cost analysis is based on the average, per patient total insurer payout per diagnosis at the time of surgery and follow-up time points including reoperations. This may not be representative of all institutions. Finally, there are multiple unknown confounding variables such as no direct physical contact, patient reported outcomes, or medical records access to determine diagnosis, prior procedures, chondral defect size, chondral defect location, alignment, treatment, and all concomitant procedures.

Cartilage defects often occur in the setting of joint malalignment or overload. When addressing cartilage defects with a cartilage restoration procedure, particularly involving the patellofemoral joint, it is important to identify and address concomitant pathologies at the time of index procedure. This study demonstrates that the addition of an osteotomy at the time of index ACI or OCA procedure significantly reduced reoperation rates without increasing complications or overall costs.