1 Introduction

Anterior cruciate ligament (ACL) tears are common injuries that are frequently treated with ACL reconstruction (ACLR) surgery.¹ ACLR has generally been shown to be an effective surgery with most patients attaining significant improvements in patient-reported outcomes.^2–5 Graft selection for ACLR continues to be a focus of study with patient and surgeon preference being main drivers of graft preference in ACLR.^6–8 Other factors taken into consideration include risk of re-rupture, associated symptoms (e.g. anterior knee pain, harvest site pain), and cost.⁹ Although a multitude of graft types tend to provide similar outcomes,^10,11 graft selection can have substantial economic implications, with allograft ACLR typically costing more than autograft ACLR.¹² Other factors shown to affect cost are surgical setting (outpatient vs inpatient), single-vs double-bundle technique, concomitant procedures, and operative time.^12–15

As a high-volume procedure in orthopaedics, ACLRs have been under scrutiny in value-based research. Value-based care focuses on optimizing the ratio of patient outcomes to costs of care incurred.¹⁶ Because operative time has been shown to affect the cost of ACLR, modifying this variable while maintaining patient outcomes can affect the overall value of the procedure. Previous literature has shown that operative time can be modified by patient characteristics, technique, facility, surgeon experience, presence of trainees, concomitant procedures, patient age, and graft type (i.e., allograft vs autograft).^13,17–20 Operative times associated with the most commonly used graft options have yet to be fully evaluated. In addition to generating more direct costs associated with operating room (OR) and related personnel utilization, prolonged OR time can lead to adverse effects such as higher rates of complications, some of which can lead to prolonged inpatient stays that further increase total costs of care.^21–24 Although previous literature has compared healthcare costs and utilization between allograft ACLR and autograft ACLR, few studies have reported on these for the variety of autograft sources such as bone-tendon-bone (BTB), hamstring (HS), and quadriceps tendon (QT).

While it is widely known that allograft ACLR is faster, autograft ACLR operative time has not been clearly determined. The purpose of this study was to evaluate operative time for autograft ACLR in the context of patient treatment and surgeon variables that influence operative time. The primary outcome evaluated in this study was operative time stratified by graft type (allografts, BTB autograft, HS autograft, QT autograft). The hypothesis was that utilization of BTB and QT autograft in ACLR would be associated with increased operative times compared to HS autograft use.

2 Materials & methods

This was a retrospective cohort study conducted after obtaining approval from the Institutional Review Board. All ACL-related procedures conducted from January 1, 2018 to July 31, 2022 in any of four ambulatory surgical centers (ASCs) within a single metropolitan healthcare system in the United States were identified using Current Procedural Terminology (CPT). Patients aged 16 or younger, revision procedures, procedures that did not involve ACLR (e.g., ACL debridement, tibial spine fixation), procedures conducted by surgeons with less than 10 cases annually, and bridge-enhanced ACL repair (BEAR) ACLRs were excluded. ACLRs with concomitant procedures such as additional ligament reconstructions or tendon repairs, hardware removal, chondrocyte or cartilage transplantation, and unicompartmental knee arthroplasty among others were also excluded due to high operative time variability or low volume.

The institutional electronic medical record (EMR) was queried for baseline characteristics. Patient characteristics included age, sex, body mass index (BMI), and American Society of Anesthesiologists (ASA) score. Surgical variables included date, location, surgeon, presence of surgical assistant, laterality, concomitant procedures, and graft type. Surgeon characteristics included fellowship training, years in practice, annual case volume, and graft utilization prevalence. Fellowship training, year of fellowship completion, and years employed in the study institution were extracted from web resources such as institutional websites and online curriculum vitaes. Years in practice was defined as time from fellowship completion until surgical date. Annual case volume was calculated as the ratio of total ACLRs conducted during the study period to the years employed at the study institution. Graft utilization was assessed by determining the percentage of ACLRs for each surgeon that used a particular graft compared to each surgeon's total number of ACLRs during the study period.

The primary outcome was operative time, which was recorded as time elapsed in minutes from incision to start of closure. The main variable of interest was graft type, and thus patients were classified according to graft type utilized during ACLR: (1) allograft, (2) BTB autograft, (3) HS autograft, and (4) QT autograft. All allograft types were analyzed together and included BTB, hamstring, tibialis anterior, and tibialis posterior allografts. Surgeons were also classified according to fellowship training (sports fellowship-trained vs not), years in practice ( < 10 years vs > 10 years), case volume ( < 25 cases per year vs > 25 cases per year), and graft utilization (mostly uses allograft vs BTB autograft vs HS autograft vs QT autograft).

Statistical analysis was conducted using Intellectus Statistics (Clearwater, FL) and Microsoft Excel (Redmond, WA). Descriptive statistics included counts, frequencies, means, standard deviations, medians, ranges, and interquartile ranges as appropriate. Preliminary analysis included Spearman and Pearson correlation tests for continuous and ordinal variables in relation to operative time as well as independent sample t-tests, Mann-Whitney U tests, and analysis of variance (ANOVA) for categorical variables when appropriate. Based on these results, analysis of covariance (ANCOVA) was then conducted to evaluate the relationship between operative time and graft type while controlling for (1) age, (2) number of menisci repaired, (3) fellowship training, (4) years in practice, and (5) annual case volume. Independent sample t-tests and Mann-Whitney U tests were conducted to compare operative time according to surgeon categorical variables, while Spearman and Pearson correlation tests assessed the relationship of years of experience and case volume with operative time. Significance was set at p = 0.05 (two tails).

3 Results

Of 2567 procedures identified, 1813 primary ACLRs performed among 15 surgeons were included in the final analysis (Fig. 1). Most patients were male (51.2%) with an average age of 32 ± 12. Average operative time was 98.9 ± 33.0 min. The most utilized graft type was BTB autograft (42.6%) followed by HS autograft (32.3%) and allograft (21.4%). Only 68 cases (3.8%) used a QT autograft (Table 1).

Close

Fig. 1 Flow diagram for patient population. UKA = Unicompartmental Knee Arthroplasty. BEAR = Bridge-enhanced ACL Repair.

Export to PPT

67.8% of ACLRs involved concomitant procedures. Meniscal operations occurred in 62.3% of cases, with the majority occurring in the lateral meniscus. Meniscectomies, meniscal repairs, and minor meniscal repair procedures (trephination, saucerization, abrasion) occurred in 32.0%, 36.6%, and 4.3% of cases, respectively. Only 5.9% of cases involved bilateral meniscal repairs. Additional procedures are detailed in Table 2.

All included surgeons are described in Table 3. Fourteen of 15 surgeons were sports fellowship-trained. Average years in practice after completing fellowship were 13 years. Average ACLR volume was 121 cases (32 cases per year). The only surgeon that did not undergo sports fellowship training was responsible for 88 total cases (19.1 cases per year) and had been in practice for 38 years. Seven of the 15 mostly used BTB autograft. Only one surgeon accounting for 13 total ACLRs and just one year in practice after completing sports fellowship predominantly used a QT autograft (9 of 13 cases). One surgeon used BTB autografts and HS autografts with equal prevalence (Table 3).

Average operative time was shortest for ACLRs using allograft (70.9 min) followed by BTB autograft (103.4 min), HS autograft (109.7 min), and QT autograft (113.3 min) (p < 0.001, Fig. 2). After adjusting for age, meniscal repairs, and surgeon variables, marginal means (Table 4) revealed that there was no difference in operative time among the autografts (p = 0.342). Allograft ACLR was observed to be significantly faster by 27–33 min compared to all autograft types (p < 0.001).

Close

Fig. 2 Primary ACLR operative time (minutes) by graft. BTB = Bone-tendon-bone; Auto = Autograft; Quad = Quadriceps Tendon.

Export to PPT

Operative time and percentage of cases involving meniscal repairs differed according to surgeon categorical variables. While sports surgeons exhibited average higher operative time compared to the non-sports surgeon (99.4 vs 88.9 min, p = 0.016), they tended to more frequently repair menisci (37.9% vs 10.2% cases with meniscal repairs, p < 0.001). Similarly, surgeons with 10 or more years in practice had a shorter average operative time compared to those with less than 10 years (97.0 vs 101.3 min, p = 0.046) while performing less meniscal repairs (30.4% vs 44.7%, p < 0.001). Operative time similarly demonstrated a weak negative correlation with years of experience (r = −0.22). Surgeons performing more than 25 cases per year exhibited higher average operative time compared to those performing less than 25 ACLRs per year (99.8 vs 95.6 min, p = 0.012) while also performing more meniscal repairs (38.0% vs 31.4%, p = 0.016). Rising case volume however weakly correlated with decreased operative time (r = −0.19).

Graft selection was associated with surgeon variables (Table 4). Specifically, average surgeon case volume was slightly lower for allograft cases and QT cases versus BTB and hamstring, while years of experience for QT cases was substantially lower compared to other grafts. For surgeons that mostly utilized allografts or BTB autografts, their average operative time was lower for these respective grafts compared to those surgeons that used other grafts more prevalently (57 vs 77 min for allograft, 99 vs 118 min for BTB autograft; p < 0.001) (Fig. 3). In contrast, surgeons that more prevalently used HS and QT autografts exhibited longer operative times for these grafts compared to those that did not frequently utilize them (114 vs 97 min for HS, 138 vs 110 min for QT; p < 0.004). For the surgeon that equally used BTB autograft (n = 17) and HS autograft (n = 16), the average time for BTB autograft ACLR was higher compared to HS (150 vs 128 min, p = 0.020).

Close

Fig. 3 ACLR operative time based on surgeon and preferred graft type. BTB = Bone-tendon-bone; Auto = Autograft; Quad = Quadriceps Tendon. Split = BTB and hamstring auto with equal preference.

Export to PPT

4 Discussion

Time in the OR is a precious commodity typically shared among numerous surgeons. Being able to effectively use allotted OR time is dependent upon on the type of procedure as well as the surgeon carrying it out. In ACLR, graft selection as well as the involved surgeon have implications regarding operative time. The overall average operative time for ACLRs in this study was about 99 min, with there being no difference in average operative time among BTB, HS, or QT autograft. ACLRs involving allografts were faster by an average of 30 min compared to autograft ACLRs. Despite demonstrating a statistical difference, there were no notable differences in primary ACLR operative time based on surgeon experience or volume. However, there seems to be substantial variability in graft utilization and associated operative times among surgeons. Surgeons who more frequently used allograft and BTB autografts were about 20 min faster with these grafts compared to surgeons who typically used other grafts. While surgeons who regularly used HS and QT autografts were about 17 and 28 min slower, respectively, than surgeons that did not. These results indicate that even though HS and QT autografts are becoming more prevalent, BTB autograft ACLR does not take additional operative time, and in some cases where it is used the procedure is shorter compared to other autografts. As orthopaedics transitions towards adopting value-based care, surgeons should be cognizant of the impact that their decision-making has on patient outcomes as well as healthcare costs and utilization.

Although prior works have identified patient factors that dictate graft usage,^6–8 this study notably did not find any association between patient factors and operative time with the exception of age (r = −0.34). This has been observed in prior works and is likely explained by the preference for allograft use with older patients (average age 45 for allograft ACLR versus 28 for autograft in this study).^6,18 BMI has been associated with ACLR operative time, with one study identifying nearly a 1 min increase for every BMI point increase. The present study found no correlation between these variables (r = 0.01, p = 0.556). This is likely because the prior work was an Amercican College of Surgeons National Surgical Quality Improvement Program (ASC-NSCIP) study including only patients aged 18 or younger that likely all received autografts (unable to discern with NSQIP). The current study included allograft usage and patients that were older, thus likely limiting this relationship to the younger population. It is important to recognize that primary ACLR operative time can vary substantially based on non-surgeon factors.

The literature regarding how graft type influences outcomes is discordant, with some studies suggesting no difference.^{4,10,11,25,26} Graft choice is thus ultimately driven by patient and surgeon preference^7,8,27,28 making it yet another modifiable factor that can affect operative time and, consequently, outcomes and costs. Several studies have commented on the operative time and cost differences between autograft and allograft usage.^{12,13,29–35} The majority indicate that while allograft usage can save from 11 to 33 min of operative time,^13,30–32 this does not offset the cost associated with the actual allograft.^{13,26,29–32} While the majority of these studies did not account for confounders, this study adjusted for factors that contribute to operative time and similarly found that allograft ACLR saves an average of 27–33 min compared to autograft ACLR. Additionally, based on data from prior works,²⁷ the cost savings would equal an average of $245 to $300 per allograft ACLR, which would not offset the average $2300 cost of the allograft at the study institutions. In conjunction with the similar outcomes achieved regardless of graft, this data suggests that allograft use may not be more cost-effective than autograft use.

Prolonged operative times for ACLR have been found to lead to adverse outcomes including higher rates of complications and longer hospital stays.^21–24,36 One study of 1640 ACLRs identified increased operative time as a risk factor for infection post-ACLR with an incidence rate ratio of 0.179 (p = 0.011).³⁶ A recent study including 14,159 procedures from the ACS-NSQIP database found that every 15-min incremental increase in operative duration was associated with increased risk in complications such as deep vein thrombosis (RR 1.12, p = 0.042), surgical site infections (RR 1.21, p = 0.001), and sepsis (RR 1.66, p < 0.001). It also led to higher healthcare utilization such as increased readmission rates (RR 1.23, p = 0.001) and an extended length of stay (RR 1.18, p = 0.008).²¹ Other recent ACS-NSQIP studies involving machine learning with up to 21,636 elective ACLRs identified prolonged operative time as a risk factor for overnight stay versus same-day discharge,^22,23 with one specifically concluding that an operative time greater than 200 min was associated with overnight stay.²³ Another study of 11,060 patients identified that an operative time greater than 90 min led to a higher risk of 30-day readmission.²⁴ Considering the directly proportional relationship between costs and operative time as well as the subsequent costs associated with longer stays and readmissions, these studies seem to highlight the importance of surgeons limiting unnecessary operative time for ACLR as a promoter of value-centric care.

The relationship of surgeon training, experience, and volume with operative time and graft choice has seldom been evaluated.^18,37 One study found that surgeons who performed greater than 50 ACLRs per year and greater than 300 total career ACLRs were on average 43 min faster compared to surgeons who performed fewer than 10 procedures per year and fewer than 30 total career ACLRs while producing equivalent outcomes.³⁷ Only 107 patients were enrolled in this study. A more recent study of 1180 ACLRs among 135 surgeons using the NSQIP database found correlations between operative time and surgeon number of arthroscopy cases per month (r = −0.20), surgeon performing more than 20 knee arthroscopy cases per month (r = −0.12), and surgeon years in practice (r = −0.27). Multivariable regression analysis found similar results for years in practice, although case volume was not subjected to the analysis. The regression also found that sports fellowship training led to lower operative time, but this did not reach significance (p = 0.197).¹⁸ The present study evaluated 1813 primary ACLRs that were manually chart reviewed and similarly identified weak correlations between operative time and surgeon volume (r = −0.19) as well as between operative time and years of practice (r = −0.22). In contrast to the aforementioned NSQIP study, this study was also able to evaluate graft type effect on operative time.

There were several strengths and limitations to this study. This study was able to control for concomitant procedures as well as surgeon experience and case volume when comparing operative time among graft types, which has typically not been accounted for in prior works. This is also the first study to describe and compare operative time for the most common types of autografts including the QT graft, which has been growing in popularity recently.^38,39 The most notable limitation in this study is the lack of outcome data. Although this data was not available and pertinent to the study purpose, it would provide knowledge regarding the relationship between operative time and value for ACLR. One limitation of this study is the grouping of the differing types of allografts into one category, which may confound the results due to the various tissues and types of graft preparation used in allografts. This study evaluated solely the amount of time to perform the procedure and did not include incision closure. This has the potential to affect the study results as the locations and sizes of incisions used for the various techniques may change the overall time spent on the case. Another limitation of this study is the fact that, due to large variability, it was unable to account for technical details that may also influence operative time (e.g., drilling technique, graft preparation, implants). However, these factors have not been identified in the literature as contributing to operative time. Future research should evaluate the effect of these variables on operative time and cost. The study also had a low number of QT graft ACLRs, which likely reflects the newfound interest in the QT autograft. Finally, this study was conducted among four institutions of a healthcare system in a single metropolitan area, limiting the generalizability due to the unique time-related confounders that may be present within this healthcare system and region. Future multicenter studies could more effectively evaluate this subject.

5 Conclusion

When controlling for concomitant procedures as well as surgeon experience and volume, no operative time differences were found among BTB, HS, or QT autograft ACLRs. Allograft ACLR took approximately 30 min less to perform compared to autograft ACLR. Graft utilization varies considerably among surgeons. Value in ACLR has yet to be fully understood.

Source of funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Patient consent

Patients consent was not considered to be necessary for this study as this was a retrospective review.

Institutional ethical committee approval

All procedures were conducted in accordance with the approval of the authors’ IRB. This study was assigned approval number A18-284.

CRediT authorship contribution statement

All authors included on this manuscript have made significant contributions to this manuscript. Fernando A. Huyke-Hernandez contributed to conceptualization, methodology, validation, formal analysis, data curation, writing of the original draft and reviewing/editing and visualization. Stephen A. Doxey contributed to conceptualization, methodology, validation, data curation, writing of the original draft and reviewing/editing and visualization. Arthur J. Only contributed to conceptualization, validation, data curation, writing of the original draft and reviewing/editing the draft. Nizar Mikhael contributed to investigation, data curation, writing of the original draft and reviewing/editing the draft. Andrew Sibley contributed to investigation, data curation, writing of the original draft and reviewing/editing the draft. Christopher Y. Kweon contributed to conceptualization, methodology, writing of the original draft and reviewing/editing, visualization, supervision and project administration. Brian P. Cunningham contributed to conceptualization, methodology, validation, data curation, writing of the original draft and reviewing/editing visualization, supervision and project administration.