Introduction

The effect of optimal graft selection on clinical outcomes and subsequent graft failure remains controversial in anterior cruciate ligament reconstruction (ACLR). Several grafts may be used in primary ACLR, including quadruple hamstring tendons, bone–patellar tendon–bone (BPTB), the quadriceps tendon (QT), allografts and synthetic grafts [1, 2]. Each of these graft options has its own advantage and disadvantages. Current evidence indicates that synthetic grafts and allografts have poor clinical outcomes and a higher complication rate compared to autografts [3,4,5]. Thus, many surgeons prefer autografts even in revision cases as the first graft choice.

The most popular grafts used in primary ACLR are the hamstring tendons and BPTB [1,2,3,4,5,6]. However, these grafts also have various disadvantages. BPTB graft increases the risk of patellar fracture, residual anterior knee pain, kneeling difficulty, patellar tendon rupture, and knee flexion contracture [2, 7]. Similarly, hamstring tendon harvesting may cause iatrogenic saphenous nerve injuries, hamstring muscle weakness, tunnel widening, a high residual laxity rate, and re-rupture, particularly with small-sized grafts less than 7 mm [2, 8, 9].

Conversely, QT autografts offer several advantages over hamstring and BPTB tendon autografts, such as a higher cross-sectional area and biomechanical properties similar to those of the native anterior cruciate ligament ACL [10,11,12]. The QT tendon harvesting technique was first described by Marshall et al. in 1979 [13]. However, this technique necessitated extensive incision, and the initial clinical results were not favorable. In the late 1990s, Fulkerson revised the technique and described a minimally invasive QT harvesting method with less donor site morbidity [14]. Subsequently, QT autografts gained popularity. It is suspected that using QT autografts will become widespread and override other options in standard practice [15,16,17]. Recent studies also demonstrated that the QT autograft is not inferior to other graft alternatives in terms of functional outcomes [1, 6, 18].

Despite these apparent advantages, QT autografts are not without complications or risks. One problem is that the tendon length might be inadequate, particularly when harvested as a free tendon autograft [19, 20]. Anatomical studies report that the QT length might be as short as 37.6 mm, far shorter than required for a proper ACLR [21]. Preoperative magnetic resonance imaging (MRI) can be used to measure QT length, but a standard MRI examination does not display the entire length of QT because the field of view (FOV) is usually only between 14 and 18 cm. Preoperative information about the QT autograft adequacy is useful in preventing intraoperative problems. A few studies previously have investigated the relationship between QT length and anthropometric measures. Although these studies reported a significant positive correlation between the person's height and QT length, no diagnostic threshold was determined [19, 20, 22].

It is hypothesized that simple anthropometric measures such as height, weight, and thigh circumference may be used to estimate free QT autograft adequacy. Therefore, the purpose of this study was to investigate the relationship between anthropometric measures and the QT length to provide a practical diagnostic guide into identify patients with inadequate free QT autograft lengths.

Materials and methods

Patients and study design

Patients who presented to the orthopedic outpatient clinic with knee symptoms and were referred to the radiology department for a knee MRI between December 2020 and February 2021 were enrolled in this prospective study. Skeletally mature patients between 18 and 45 years of age with a full range of knee motion and who required dominant-sided knee MRI were included. Patients with ACL rupture, QT pathologies, and patellofemoral disorders, patients with a history of previous knee surgery, and neuromuscular and chronic inflammatory diseases were excluded. The institutional review board approved the study protocol (approval date/issue: 09.12.2020/26-5), and the study was conducted in accordance with the Declaration of Helsinki and later amendments. All participants provided written informed consent for participation.

MRI protocol

The MRI examination was performed using a 16-channel knee coil with a 1.5-Tesla MRI device (GE Healthcare, Signa, Milwaukee, WI, USA). The imaging protocol included axial coronal and sagittal fat-saturated, proton-density-weighted (TR/TE: 3500/20, matrix: 320 × 192, FOV: 16 × 16, slice thickness: 4 mm); sagittal fast spin-echo, T1-weighted (TR/TE: 650/25, matrix: 320 × 192, FOV: 16 × 16, slice thickness: 4 mm). In addition to standard knee MRI, a high-resolution axial fast spin-echo, T1-weighted image was obtained (TR/TE: 550/20–21, matrix: 512 × 256, FOV: 32.6 × 21.2, slice thickness: 0.5 mm). The FOV (32 cm) in the additional examination included the tibial tubercle to the rectus femoris muscle belly to visualize the full length of the QT.

MRI measurements

Since the QT follows an oblique course proximally and laterally, an oblique sagittal section was obtained by performing multiplanar reformation (MPR) over the high-resolution axial T1-weighted sections to determine the correct QT length. On the sagittal oblique T1-weighted images, the QT length was measured between the patella’s superior pole (patellar insertion site) and the musculotendinous junction of the rectus femoris muscle (Fig. 1). To measure the ACL length, oblique sagittal reformatted images parallel to the longitudinal axis of the ACL were created. The ACL length was measured between the midpoints of femoral and tibial attachment sites (Fig. 2).

Fig. 1
figure 1

a The course of QT on the oblique sagittal plane. b The measurement of QT length on oblique sagittal T1-weighted MRI

Full size image
Fig. 2
figure 2

a The course of ACL on the oblique sagittal plane. b The measurement of native ACL length on oblique sagittal T1-weighted MRI

Full size image

Two independent observers measured all variables once on the workstation (Advantage Windows 4.2, GE Healthcare). Both observers were radiologists with more than 10 years of experience in musculoskeletal radiology. Interobserver reliability was calculated using the interclass correlation coefficient (ICC) and 95% confidence interval (CI). ICCs of 0.81–1.00, 0.61–0.80, 0.41–0.60, 0.21–0.40, and 0.00–0.20 were interpreted as excellent, good, moderate, fair, and poor, respectively [23]. Both measurements showed excellent agreement; ICC ranged between 0.952 and 0.997 (Table 1), and the mean was used for the analysis.

Table 1 Interobserver reliability of radiological measurements on MRI
Full size table

Calculation of sufficient free QT autograft length

Previous studies advocated that at least a 15 mm tendon is required in the bone tunnels for secure graft fixation and subsequent healing [19, 24, 25]. The minimum required QT autograft length was calculated by adding the native ACL length and 30 mm for the bone tunnels (15 mm for the tibial tunnel, 15 mm for the femoral tunnel). If the measured QT length was shorter than the calculated autograft length, the subject was assigned to the insufficient free tendon autograft group. If the length was equal or longer, the subject was assigned to the sufficient group (Fig. 3).

Fig. 3
figure 3

The calculation of adequate QT autograft

Full size image

Anthropometric measurements

Anthropometric measurements included height, weight, and thigh circumference. Thigh circumference was measured at 10 cm above the superior pole of the patella while the patient was standing by the senior author using the same tape measure for standardization. Body mass index (BMI) was calculated by dividing weight (kg) by height squared (m2). The Tegner activity scale was used to grade the patients’ work and sporting activities.

Statistical analysis

Descriptive statistics for categorical variables were presented as frequencies and percentages, and continuous variables were presented as mean ± standard deviation and range. The Kolmogorov–Smirnov test was used for testing normality. Comparative analysis between independent groups was performed using the Student t test, Mann–Whitney U test, and Chi-square tests. Correlation between variables was assessed using Pearson’s and Spearman’s correlation tests. Multivariate logistic regression was used to identify independent risk factors associated with insufficient QT length. Odds ratio (OR) and 95% CI were used to quantify the associations with risk. Receiver-operating characteristic (ROC) curve analysis was used to assess the diagnostic performance of each variable. The sensitivity, specificity, cutoff value, and area under the ROC curve (AUC) were analyzed. A value of p < 0.05 was accepted as statistically significant.

Results

One hundred and eighty-four patients (92 male and 92 females, all ethnically Caucasian) with a mean age of 34.1 ± 8.0 years (range 18–45) were included in the analysis. Male subjects were younger (p = 0.012), taller (p < 0.001), heavier (p < 0.001), and more active (p < 0.001) than female subjects, but the BMI (p = 0.328) and thigh circumference (p = 0.072) were similar between sexes. The mean QT length was 69.0 ± 8.8 mm (range 48.1–90.3 mm), and the mean ACL length was 35.6 ± 2.5 mm (range 29.2–42.6 mm). The QT length and the ACL length were significantly longer in male subjects than in females (p < 0.001 for both). Patient demographics and anthropometric characteristics are shown in Table 2.

Table 2 Demographic and anthropometric characteristics of the patients
Full size table

There was a weak positive correlation between QT length and height (p < 0.001), weight (p < 0.001), and activity score (p = 0.007), but there was no correlation with the other variables. Similarly, ACL length was positively correlated with height (p < 0.001), weight (p < 0.001), and activity score (p = 0.007). However, QT and ACL lengths did not correlate with each other (p = 0.118). The correlation coefficients are listed in Table 3.

Table 3 Correlation between QT length and ACL length with demographic, anthropometric and radiologic variables
Full size table

According to the adequacy calculation, 23 males and 39 females (total: 62, 33.7%) had inadequate QT length for a free QT autograft. Females had insufficient QT length more frequently than males (p = 0.009). The sufficient QT length was significantly longer, ranging from 62.1 to 90.3 mm (mean 73.5 ± 6.4 mm), compared to the insufficient length, ranging from 48.1 to 68.3 mm (60.0 ± 5.0 mm) (p < 0.001). Subjects with sufficient QT length were taller (p < 0.001), but other variables showed no significant difference (Table 4). If the QT graft was harvested with a bone block (adding 15 mm), only 6 patients (3 males, 3 females, 3.3%) would remain who still have insufficient graft size.

Table 4 Comparison of subjects with sufficient and insufficient free QT length
Full size table

Age, sex, height, weight, body mass index, thigh circumference, activity score, and ACL length were set as explanatory variables to create a regression model that predicts QT length. Testing for multicollinearity with variance inflation factor (VIF) among explanatory variables showed that weight (VIF = 103.9), height (VIF = 39.0), and BMI (VIF = 89.5) were significantly correlated; thus, weight (highest VIF) was excluded from the regression model to increase the precision of the estimate coefficients and the statistical power. After excluding weight, both tolerance and VIF values were acceptable (Tolerance > 0.20, VIF < 4) for the rest of the variables.

A stepwise backward logistic regression analysis revealed that height was the only independent variable that predicted adequacy of QT graft (r2 = 0.051, p = 0.009) (Table 5). Age, sex, weight, thigh circumference, and the activity score failed to show an association. A calibration plot was used to assess the accuracy of the predictive model, and it indicated that the model fit the quadratic data well (Fig. 4). ROC analysis showed that height could not detect a subject with an inadequate QT length (AUC: 0.384, 95% CI 0.300–0.468). The ROC curve diagram is presented in Fig. 5. A power calculation was performed to eliminate type II error in the logistic regression, and a post hoc power analysis using G*Power software (Version 3.1.9.6) showed a regression model power of 94.1%.

Table 5 Results of backward stepwise (WALD) logistic regression
Full size table
Fig. 4
figure 4

Calibration plot showing the predicted probabilities for adequate QT length and height

Full size image
Fig. 5
figure 5

ROC curve for height. State variable: inadequate QT

Full size image

Discussion

This study attempted to predict the adequacy of free QT autograft length using simple anthropometric measures, namely age, sex, weight, BMI, thigh circumference, and activity score. The findings of this study showed that the height of the subject positively correlated with QT length, and regression analysis showed that it was an independent predictor of graft length adequacy. However, the discriminatory capacity was too weak to use as a diagnostic tool to identify patients with inadequate free QT autograft. The weak diagnostic capacity may be related to calculating the required tendon length using the subject’s native ACL length. The QT length increases as the height of the subjects increases; however, the native ACL length also increases with height. Thus, a short QT might still be adequate for a particular short subject with a short native ACL. Conversely, even if the free QT autograft in a tall person is long, inadequacy may arise because the native ACL is also long in the same person.

This study also found that the QT length was insufficient for a free QT autograft in approximately one-third of subjects. Almost half of the female subjects (42%), and a quarter of the male subjects (25%) had inadequate QT length. However, the inadequacy rate dropped to 3.3% if the tendon was harvested with a 15 mm bone block. Thus, surgeons who prefer free QT autografts should keep in mind that the tendon length may be short, particularly in female patients. The bone block harvesting technique described by Blauth increases the QT autograft length by 15 mm and an adequate graft can be achieved in almost all patients [26]. However, a notable complication of this technique is patellar fracture [27]. Fu et al. reported 3.5% intraoperative and 8.8% postoperative patellar fractures in a series of 67 patients who underwent QT autograft with bone block [28]. Apart from the bone block technique, the length of the QT autograft might also be extended by the elevation of periosteal flap (2 cm strip) from the surface of the patella. The periosteal part of the graft is folded in the middle and it provides approximately 1 cm of additional length [29].

The findings of the current study conflict with relevant studies in the literature. Xerogeanes et al. measured the QT length using three-dimensional MRI in 60 patients and found a moderate correlation between QT length and height [30]. They reported that 90% of patients had grafts longer than 70 mm, and 96% of patients had grafts > 60 mm. However, males and females were on average taller in this study than that in ours. They reported that the free QT graft was not below 60 mm in any subject taller than 167 cm. In contrast, 24 out of 66 patients with inadequate free QT grafts were taller than 167 cm in our study. Ukogewa et al. measured the QT length in 51 patients who underwent total knee prosthesis and found a positive correlation between height and QT length [20]. The QT was longer than 80 mm in 48 out of 51 patients, and it was less than 75 mm in only three female patients shorter than 160 cm. However, in both studies, the ACL length was not considered. Similarly, many authors suggested that a 65–70 mm tendon length is sufficient for an all-inside ACLR [31, 32]. However, we believe that these dogmatic values do not have meaning without knowing the native ACL size. The ACL length is subject to variations according to gender and height of the subject. In a recent systematic review, the mean ACL length ranged between 26.0 and 38.2 mm in 16 different studies [33]. In our study, the native ACL length was between 28.5 and 42.6 mm; subsequently a sufficient tendon length ranged between 58.5 mm and 72.6 mm.

Only one study supports our findings. Yamasaki et al. calculated the required tendon length by leaving a 30 mm margin for the femoral and tibial tunnels and reported that 37% of the patients had inadequate free QT [19]. Similarly, in our study, the tendon length was inadequate in 33% of the patients. The Yamasaki study was conducted in an Asian population, and the authors suggested that the tendon length might be shorter due to ethnic differences. Two cadaver studies have been conducted in Asian populations. In both, the average tendon length ranged from 60 to 65 mm [21, 34]. However, in studies conducted in Caucasian races, the tendon lengths were reported to be longer (70–99 mm) [10, 20, 29, 35, 36]. These findings suggest that the QT lengths may differ by ethnicity.

This study has both strengths and limitations. First, MRI was used for the measurements, which may not be an accurate technique compared to direct dissection measurements. However, a high correlation was detected in a previous study between the MRI measurements and the harvested QT length with a 0.1 mm precision [19]. The mean of the two ratings derived by the experienced radiologists was used to reduce measurement errors. The FOV was also wider than a standard knee MRI to visualize the entire QT length. Moreover, the QT and ACL lengths were measured through MPR since they follow an oblique course, and standard sagittal sections might underestimate the true longitudinal dimensions [37]. The current study focused on the QT length and did not evaluate the thickness and volume, which are also critical issues for the adequacy. But it is well known that the QT thickness and strength are quite sufficient and larger than other autograft options, as shown in several previous studies [22, 38,39,40,41,42]. To our knowledge, this study had the largest number of patients, and the high statistical power strengthened the results.

In conclusion, free QT autograft may be inadequate for one-third of the patients, but a QT autograft with a bone block is almost always sufficient. Inadequate free QT autografts are more common in women than men. Although the QT length correlated with height, it cannot be used as an accurate diagnostic tool to identify subjects with an inadequate QT autograft. Measuring the ACL and QT lengths by MRI might eliminate the problems during the surgery, if a free QT graft is chosen, especially in female patients.