Knee Surgery, Sports Traumatology, Arthroscopy

, Volume 22, Issue 9, pp 2094–2101

Iliotibial band autograft versus bone-patella-tendon-bone autograft, a possible alternative for ACL reconstruction: a 15-year prospective randomized controlled trial


    • Arthroscopic Centre Amager, Amager HospitalCopenhagen University Hospital
  • Kristian Thorborg
    • Arthroscopic Centre Amager, Amager HospitalCopenhagen University Hospital
  • Lars Konradsen
    • Section of Sports SurgeryCopenhagen University Hospital
  • Uffe Jørgensen
    • Parken Private Hospital
  • Per Hölmich
    • Arthroscopic Centre Amager, Amager HospitalCopenhagen University Hospital

DOI: 10.1007/s00167-013-2630-9

Cite this article as:
Stensbirk, F., Thorborg, K., Konradsen, L. et al. Knee Surg Sports Traumatol Arthrosc (2014) 22: 2094. doi:10.1007/s00167-013-2630-9



The long-term results after using the iliotibial band autograft (ITB) in anterior cruciate ligament (ACL) reconstruction are not fully known. If equal in quality to conventional methods, the ITB graft could be a useful alternative as a primary graft, in revision surgery or multi-ligament reconstruction. The purpose is to assess whether the ITB autograft is a long-term reliable alternative to the bone-patella-tendon-bone (BPTB) autograft, using a prospective randomized controlled trial design.


From 1995 to 1996, sixty patients scheduled for primary ACL reconstruction were included in a prospective randomized controlled trial. Three senior knee surgeons, experienced in both types of ACL surgery, performed all the operations. A standardized and supervised rehabilitation programme was used for both groups for 6 months. Thirty patients received the ITB reconstruction, and 30 received the BPTB reconstruction. Forty-nine participated at follow-up in 2010 (82 %). Primary outcome was the failure rate after ACL reconstruction. Secondary outcomes were knee injury osteoarthritis outcome score (KOOS) [pain, symptoms, Sport/Rec, quality of life (QOL), daily living function], Tegner activity scale, anterior knee pain-score, Lysholm score, Rolimeter laxity, extension deficit, single hop and crossover hop for distance.


At 15-year follow-up, no significant difference existed between the groups. Graft failure occurred in 4 ITB subjects (16 %) and 3 BPTB subjects (13 %). KOOS (Sport/Rec) for the ITB group was 75 and 73 for the BPTB group. The KOOS (QOL) was 72 and 68 for the ITB group and BPTB group, respectively.


Similar graft failure rates and KOOS were found when comparing ITB- and BPTB-operated individuals, at 15-year follow-up. The ITB graft had equal long-term clinical results compared to the BPTB graft and is recommended as a reliable alternative autograft for ACL reconstruction.

Level of evidence

Therapeutic studies, Level I.


ACL reconstructionPatellar ligamentIliotibial bandKOOSLong term


A number of different autografts are available for anterior cruciate ligament (ACL) reconstruction, the most popular being the bone-patella-tendon-bone (BPTB) graft and the hamstring tendons. Both short- and the long-term results using these two grafts for ACL reconstruction seem to be equal [24, 30, 34].

The iliotibial band (ITB) autograft was popular both as a primary reconstruction graft and as an augmentation technique in Scandinavia in the 1980s and beginning of the 1990s. The graft offered the possibility of a combined intra-articular and lateral tenodesis without disrupting the knee extensor or flexor mechanism. Results for this graft have been reported in the literature as a primary reconstruction [3, 4, 17, 38] or as an augmentation procedure [6, 9, 12, 23, 37]. In these cohort studies, the graft performance was comparable to other methods of ACL reconstruction. Furthermore, the ITB has been commonly proposed for ACL reconstruction in skeletally immature athletes due to physeal-sparing benefits [13]. The procedure, however, was more or less abandoned as the BPTB, and later hamstring reconstruction methods became predominant, and therefore, follow-up studies concerning ACL reconstructions with alternative autografts like the ITB are scarce. Why the procedure was abandoned is unknown, as our experience was that clinical results with this graft seemed comparable to the results of other graft reconstructions at that time. It could be due to surgical fashion in conjunction with the cosmetic complaints primarily due to the defect in the ITB creating a muscle herniation on the lateral side of the femur.

In situations where additional graft material is needed, alternative reliable graft options are sought for. This is especially the case in relation to revision ligament surgery and multi-ligament reconstruction, but also even as a primary ACL reconstruction, alternatives to the BPTB/hamstring would be helpful. It is unclear whether the ITB autograft represents a longstanding alternative.

To assess the durability of ACL reconstructions performed with an ITB autograft, we performed a long-term (15 years) follow-up of a prospective randomized controlled trial comparing primary ACLreconstructions using a BPTB autograft to an ITB autograft. The main hypothesis was that the ITB-reconstructed knees would show similar re-rupture rates as those reconstructed with the BPTB graft.

Materials and methods

During 1995–1996, 60 patients scheduled for primary ACL reconstruction were recruited in a prospective randomized controlled trial comparing reconstruction with an ITB autograft versus a BPTB autograft. The study was approved by the local ethical committee and performed in accordance with current ethical standards (declaration of Helsinki). Furthermore, all persons gave their informed consent prior to their inclusion in the study.

Inclusion criteria were as follows: age ≥18; MRI or arthroscopically verified ACL rupture; informed consent; symptomatic post-traumatic functional rotational instability (episodes of giving way and/or inability to perform at the previous level of activity) despite relevant rehabilitation; no other uni- or contra-lateral lower extremity disability; ACL rupture but no other ligament instability diagnosed preoperatively.

Exclusion criteria were as follows: combined instability (rotatory instability, multi-ligament instability) recognized during testing preoperatively. Meniscal or cartilage injuries were not considered a reason for exclusion.

If the subject was considered eligible after testing under anaesthesia and arthroscopy, randomization to either BPTB or ITB reconstruction was performed in the operation theatre using closed envelopes.

Operative techniques

The three operative senior knee surgeons were very experienced in both types of surgery technique that were standard in the mid 1990s. While the operative BPTB technique is well known, the ITB technique, presumably less known, will be described in detail.

Iliotibial band reconstruction (Fig. 1a, b)

The surgery was a modification of the Hey-Grooves procedure [15]. A mean circular ITB graft diameter of 8 mm was obtained using a 4–6-cm-wide and 18-cm-long strip of the anteroseptal part of the ITB (including part of the vastus lateralis fascia). The graft was harvested using two longitudinal incisions on the lateral side of the thigh. The first started 1 cm distally to the proximal part of the lateral femoral condyle in the midline of the ITB ascending 3–5 cm depending on the patient (Fig. 1a). The subcutaneous tissue over the ITB was freed from the ITB.
Fig. 1

Harvesting procedure for the ITB autograft. a Proximal and distal incisions on the lateral side of the femur. Iliotibial tract isolated subcutaneously. Scissor ready to perform fasciotomy. b Part of iliotibial band rolled into a cylindrical graft and the remaining part of the iliotibial band

A fasciotomy was performed with a smiley knife as far anterior as possible to obtain an entire anterior part of the ITB and part of the fascia, continuing 18 cm proximally from the lateral femoral condyle (Fig. 1a). Through the fasciotomy 4 cm proximally of the femoral condyle, a pean was directed posterior to identify the septum, just anterior to which the second fasciotomy was performed also extending 18 cm proximally. At the proximal point, the fasciotomies were connected through a transverse cut. This resulted in an 18-cm-long, distally attached graft that could be rolled into the cylindrical graft. The “rolled position” was secured using a number of absorbable sutures (Fig. 1b).

Intra-articularly, a notchplasty of approximately 20 mm was performed using Odensten’s notch measuring instrument (Stryker). A one tunnel transtibial technique was used, and the femoral tunnel was thus drilled through the tibial tunnel. The direction of the femoral tunnel was determined by a wish to exit the femoral tunnel at the centre of the tuberculum at the end of the linea aspera, on the lateral side of the femur, which is the near isometric point for the lateral tenodesis. The graft was passed from the outside through the femoral tunnel and pulled out through the tibial tunnel. The graft was pretensioned and then fixed distally on the tibia in 30 degrees of flexion. Distal fixation was achieved using two Richards staples in a double-wrapback procedure, in which the first staple fixed the descending graft, which was then looped back, and the second staple fixed both the descending and the ascending part of the graft. The goal of this belt-buckle configuration was to prevent slippage.

BPTB reconstruction

The graft was harvested through a straight incision technique. The femoral tunnel was drilled through a transtibial approach, and the graft (diameter 10 mm) was fixed with metal interference screws [33].

Post-operative treatment and rehabilitation

In both groups, the knee was immobilized in a brace locked at 30 degrees of flexion for 1 week, in order to protect the graft. The same standardized rehabilitation programme, following the internationally accepted progression criteria, was conducted. The rehabilitation programme was supervised by a hospital physiotherapist for 5–6 months in both groups. After 3 months, specific quadriceps training was started, and after 6 months, cutting and pivoting sports activities were allowed [10, 27].

Follow-up assessment

Three independent observers not involved in the primary treatment of the patients conducted follow-up. The two observers performing the tests were blinded to which leg had been operated and to the method of surgery. Bilaterally athletic tape was applied by a third person. The patients were instructed not to reveal information to the two blinded observers about which leg had undergone surgery, and what kind of graft that had been used.

Primary outcome measure

Graft failure rate

Graft failure during the 15 years since the operation confirmed by a specialist in orthopaedics or re-operation/arthroscopic findings. Failure rate is calculated as failures in each group divided with total number of patients in each group.

Secondary outcome measures

Knee injury osteoarthritis outcome score (KOOS)

The KOOS is a disease-specific self-administered questionnaire with 42 questions, in five subscales (symptoms, pain, daily living function (ADL), sports and recreational activities (Sport/Rec) and quality of life (QOL) scored on a 0 (extensive problems) to 100 (no problems) scale). The score has been validated for assessment of patients with ACL reconstruction [28], and the subscales Sports/Rec and QOL are of special interest because of their better capability to measure knee problems [7].

Tegner activity scale

The Tegner activity scale is a one-item instrument that assesses activity levels for sports and occupational activities. The Tegner activity scale has been tested and proven valid and reliable for evaluating level of activity years after injury to the knee [5]. Tegner score was registered preoperative, for maximal activity level postoperative and present.

Lysholm score

The Lysholm score is a physician-administered questionnaire that assesses knee function with 8 questions (limping, need of cane, locking cases, pain, instability, swelling, walk on stairs and squatting) scored on a <65 (poor function) to 100 (excellent function) scale. It has been tested and proven both valid and reliable in evaluating knee stability years after injury [5].

Anterior knee pain (AKP) questionnaire

The revised AKP questionnaire specified for ACL-injured patients was used to evaluate possible anterior knee pain in these patients [14]. It consists of 8 subscales (pain, occurrence of pain, sitting with flexed knee >30 min, walking upstairs, walking downstairs, squatting, kneeling and arretations-catching). Outcome is a score max 50 points equivalent to no AKP. This questionnaire has proven reliable to evaluate anterior knee pain after ACL reconstruction [14].

Rolimeter test

The Rolimeter is a validated and reliable device [11] designed to measure anterior displacement. This measurement was performed on both knees 3 times each. Average anterior displacement was calculated, with the difference between knees reported in millimetres.

Knee extension deficit measured by the heel height method

Measuring heel height gives valid and reliable measure of a possible knee extension deficit [32]. The test is performed with the patient lying prone on an examining table with full passive knee extension and the lower legs (including the patella) over the edge. The difference in heel height was measured with a ruler and a measuring tape, and reported in centimetres.

Single-legged hop for distance

Patients were told to hop as far as possible, taking of and landing on the same foot. They were allowed to swing their arms during the jump. Patients were instructed to perform a controlled, balanced landing and maintain this position for 2–3 s. One test jump and three approved jumps were performed on each leg. Average jump distance was calculated for each leg and compared to the opposite side.

Crossover hop for distance

The crossover hop for distance was measured under the same circumstances as the single-legged hop, but the patients had to perform three jumps in a row, each crossing a measuring line, still finishing with a controlled balanced landing maintaining this position for 2–3 s. Patients performed one test jump and three approved jumps for each leg. Average jumping length was calculated and compared to the opposite side, with the difference reported in centimetres. Both hop tests have proven valid and reliable by Reid et al. [26] for subjects with ACL reconstructed knees.

Statistical analysis

Values are expressed as means with corresponding SD unless other is given. The statistical calculations were performed in SPSS statistics. For comparisons, both parametric and nonparametric tests were used. Re-rupture rates were compared using the χ2 test. For the normally distributed data, the two sample t test was performed. The remaining comparisons were performed using the nonparametric Mann–Whitney U test.


Sixty patients were included in this study during 1995–1996. Ten could not be contacted due to emigration (n = 4) or insufficient contact data (n = 6) (Fig. 2). One patient did not wish to participate (Fig. 2). Thus, 49 (82 %) patients with 24 BPTB and 25 ITB reconstructions were eligible for 15-year follow-up (Fig. 2). The mean age at operation was 28 years (range 14–45 years), and the mean age at follow-up was 43 years (range 29–59). There were 32 men and 17 women (Table 1). Median time from injury to operation was 39 months in the ITB group (range 7–167) compared to 29.5 months in the BPTB group (range 1–144).
Fig. 2

Flow diagram for the participants since the beginning in 1995

Table 1

Basic data of the 49 patients included in the follow-up study


Iliotibial band


No. of patients



Mean age at follow-up (range)

41 (29–52)

45 (33–59)

Mean age at operation (range)

26 (14–37)

30 (18–44)

Gender no







Weight at follow-up (SD)

80.4 (9.4)

81.7 (12.5)

Height at follow-up (SD)

178.8 (7.5)

178.9 (10.2)

Median preinjury Tegner (range)

6 (3–10)

6 (3–10)

Meniscal injury



Meniscal repair



Meniscal surgery



Median time from injury to surgery in months (range)

39 (7-169)

29.5 (1-144)

Primary outcome measures

Four of the 25 ITB reconstructions (16 %) failed compared to 3 BPTB reconstructions of 24 (13 %). No significant difference in re-rupture rate could be identified (Table 2).
Table 2

Outcome results at follow-up





P value

Failure rate



3/24 (12.5 %)




4/25 (16 %)


Median (range) Tegner at follow-up



5.5 (2–9)




4 (2–9)


Median (range) Tegner max



6 (3–10)





6 (3–10)





86 (12)




84 (13)





42 (6)




41 (7)


Functional test



2 (2)


AP knee laxity (mm)



2 (2)


Single hop for



0.92 (0.17)


Distance (cm)



0.93 (0.09)


Crossover hop



0.87 (0.13)


For distance (cm)



0.94 (0.04)





−2 (2)


Deficit (cm)



−2 (2)


Results are given as mean values with SD in parenthesis unless other is specified. Nonsignificant P-values (n.s.)

Secondary outcome measures

The KOOS profile is shown in Fig. 3. No significant differences could be detected for the KOOS subscales QOL, Sports/Rec, pain, symptoms, ADL (n.s.).
Fig. 3

Knee injury osteoarthritis outcome score (KOOS) profile for the bone-patella-tendon-bone reconstructed individuals versus the iliotibial band reconstructed individuals at follow-up

Thirty-eight patients (18 ITB; 20 BPTB) completed the Tegner activity scale, 34 (15 ITB; 19 BPTB) completed the Lysholm score and 35 (16 ITB; 19 BPTB) completed the AKP questionnaire at follow-up. There were no significant differences between groups when comparing Tegner score for maximal activity level after reconstruction, present Tegner level of activity, Lysholm score and AKP score (Table 2). Moreover, we found no significant differences in any of the functional tests (heel height, single hop for distance and crossover hop for distance) (Table 2).


The most important findings of this study are that there is no difference in failure rates between ITB and BPTB reconstructed individuals. KOOS for the two groups displayed no significant differences, and we did not find any differences in anterior knee pain at the 15-year follow-up.

In the present study, four failures in the ITB group versus 3 in the BPTB group were identified. This is consistent with the 2-year follow-up of the same cohort [18] and with a retrospective 2-year study comparing the ITB with BPTB [25].

The ITB graft failure rate (4/25–16 %) is comparable to the only other long-term follow-up case-series [38] evaluating the ITB autograft and its failure rate. In this 13-year follow-up including 27 subjects, 3 failures were reported (11 %). The BPTB graft failure rate in the present study was 3/24 (13 %). This is also consistent with several other long-term studies evaluating the BPTB graft, reporting failure rates of 8 % [16], 13 % [31] and 12 % [30].

No study evaluating ITB reconstruction versus hamstring reconstruction has been published. However, several studies have compared the BPTB and hamstring autograft, and found very similar outcomes between these surgical procedures [19, 21, 22, 24, 30]. It therefore seems reasonable to assume that the re-rupture rates after ITB, BPTB and hamstring graft procedures are comparable. Larger randomized clinical trials are needed to verify this.

Knee injury osteoarthritis outcome score (KOOS) for the two groups in the present study displayed no significant differences (Fig. 3). When comparing the present data with a comparable long-term study evaluating KOOS in men 14 years after ACL reconstruction (primarily BPTB surgery) [36], our subjects generally displayed higher numerical scores (10–11 points) in the most relevant subscales for ACL reconstruction (Sport/Rec and QOL) [7], implying that the ITB surgery has equal patient-reported outcome to BPTB reconstructions.

The method of intra-articular autologous ITB reconstruction used in the present study differs from the original method described by Hey-Grooves (1920), by using an 4–6-cm-wide graft preserving the lateral femoral attachment combining extraarticular lateral tenodesis with an intra-articular isometrically placed graft. This theoretically helps avoiding excess pivoting, and the extraarticular tenodesis has in cadaver models demonstrated a capability to decrease stress on the intra-articular graft with up to 43 % [9]. Furthermore, a recent study by Delcroix et al. [8] investigating histologic structure, tensile strength and stress-relaxation properties of the ITB supports the use of this graft for ligament reconstruction.

Harvesting the ITB graft leaves the knee extensor and flexor mechanism intact, but results in a lateral femoral quadriceps hernia that may cause cosmetic complaints, and may cause post-operative haematoma due to bleeding from small vessels along the ITB graft harvest site [17]. Two studies investigating the effect of the ITB on knee biomechanics during a simulated pivot shift test [35, 39] indicate that the ITB aids in increasing knee stability in response to rotatory load. According to this, it may seem problematic to utilize the ITB as a donor site. As described previously, we use the anteroseptal part of the ITB, leaving the strong postero-septal part in situ with its insertion on the lateral femoral condyle, and the protective function and dynamic tensioning of the ITB should remain intact. For the BPTB graft, it is well documented that the harvesting procedure interferes with the extensor mechanism of the knee. Known donor site morbidity related to this procedure includes patella fracture, patella tendon tendinitis and anterior knee pain both in general and during kneeling [2, 16, 20, 29, 31]. In the present study, anterior knee pain was evaluated with the AKP questionnaire, and we did not find any differences in anterior knee pain at the 15-year follow-up. The lack of standardized reporting methods of AKP makes it difficult to compare between studies.

Harvesting the hamstring graft [19, 22] interferes with the flexor mechanism of the knee. This leads to reduced hamstring muscle power and reduced hamstring to quadriceps ratio in up to 3 years post-operative [1]. Zebis et al. [40] found reduced preactivity of the semitendinosus combined with elevated preactivity of the vastus lateralis during a side-cutting manoeuvre, as a significant risk factor for future noncontact ACL injury. Using the semitendinosus tendon as a graft for ACL reconstruction therefore might prove detrimental, as this muscle is involved in protecting the ACL.

The surgical techniques used in the present study have throughout the previous 15 years been refined. Today better alternatives for fixating the grafts distally have been developed. Additionally, a more gentle technique to harvest the graft has been developed for the ITB reconstruction, using less material and shorter incisions, leading to reduced problems with muscle herniation and cosmetic complaints. However, the ITB reconstruction has nearly been abandoned since the emergence of the hamstring reconstruction in the mid 1990s, without any studies supporting the choice of the hamstring graft.

The number of participants originally included in this study represents a limitation. When considering the primary outcome measure “failure rate”, a post hoc analysis shows that we would have needed to identify 10 or more failures in the ITB group to show a statistically significant difference compared to the BPTB group, assuming the BPTB failure rate incidence is around 13 % as found in the present study and in the existing literature [16, 30, 31]. Therefore, studies including larger numbers of participants are needed when analysing re-rupture rates, as it seems that the differences, if any, are small. It should also be kept in mind that this is not a study only comparing two different types of grafts, but also a study comparing two different surgical techniques. In the present study, there is no certainty that placement of the femoral and tibial tunnels was the same in the two groups or the fixation methods equally adequate. Gender distribution and activity level would be possible confounders but no differences were found between the groups related to these issues. (Table 1).

This study has clinical implications as it suggests that the ITB autograft could be a relevant choice for revision ligament surgery, multi-ligament reconstruction and as a primary ACL reconstruction method.


Using a randomized clinical trial design, we found no differences in re-rupture rates between ITB and BPTB reconstructed patients at 15-year follow-up. Thus, it seems that the ITB reconstruction can be recommended as an attractive and realistic alternative to the conventional methods, and could also be considered in relation to revision ligament surgery and multi-ligament reconstruction where additional graft material is needed.


The authors extend special thanks to Lise Ellerø, Natasha Kiss and Camilla Quist Larnkjær for collecting the data at follow-up. Thanks also extend to Tom Nicolaisen for reading the manuscript and providing useful comments.

Conflict of interest

The authors declare that they have no conflict of interest.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013