Introduction

Rupture of the Anterior Cruciate Ligament (ACL) is one of the most common athletic injuries of the knee [14, 17, 61]. Lyman et al. recently estimated that the frequency of ACL reconstruction is increasing in the United States and that younger patients are at a higher risk for re-rupture of the ACL graft [41]. The consequences of an ACL rupture to the function of the involved limb are multifaceted and possibly include a decrease in joint stability, muscle weakness, meniscal damage, pain and, in the long term, development of osteoarthritis [29, 30, 46, 47, 50, 51, 54, 58, 62, 67, 73]. In an attempt to prevent these deficits in joint function, reconstruction of ACL has become one of the most common orthopaedic interventions. Although many different surgical techniques and an increasing number of graft types have been described in the literature, autograft reconstruction using Bone Patellar Tendon Bone (BPTB) or Hamstrings Tendon (HST) appears to be the most popular graft choices [4, 12, 23, 24, 26, 45, 53].

Despite a plethora of recently published comparative studies, the relative effectiveness of the different grafts used for the reconstruction of ACL remains unclear [4, 6, 7, 11, 12, 15, 23, 24, 27, 33]. Maximizing knee stability after ACL reconstruction is one of the most important criteria for the choice of graft. Post-operative stability allows the performance of rehabilitation protocols that aimed to restore normal function and thus safe and fast return to pre-injury activity level [48]. The superior post-operative stability afforded by BPTB autograft is likely to be related to enhanced healing from bone-to-bone attachments [3, 59, 72]. However, increased donor-site morbidity has been reported after harvesting BPTB autograft. Specifically, anterior knee pain, quadriceps weakness and worse results in functional tests along with an increased rate of patellar fractures have been observed [27, 40, 43, 44, 49]. Harvesting of the HST autograft may avoid some of these post-operative problems, but it is associated with hamstring muscle weakness and slower healing of the graft attachment site that may predispose patients to higher risk of re-rupture [2, 66, 70]. Thus, both of these graft choices are limited in the ability to restore knee function for people with ACL rupture, and there is consequently an ongoing debate concerning the superiority of one graft over the other. An important aspect of this debate is the outcome of lower limb muscle strength following either of these graft types.

Evaluation of muscle strength can be accomplished using functional tools (incorporating hop or twisting) or single-joint evaluation tools [2, 9, 52]. One of the most commonly used tools that is reliable in assessing single-joint muscle strength is isokinetic dynamometry [19, 56]. In comparison to other measures of strength, isokinetic dynamometry allows quantification of muscle strength deficit through the assessment of specific parameters such as work per unit, torque at specific joint angles and the widely used peak torque value [19, 22, 56, 67]. In the majority of studies that investigate muscle strength following ACL reconstruction, the strength of the operated limb is recorded as a deficit or gain in comparison to the contralateral healthy limb. Restoration of similar muscle strength between reconstructed and healthy knee is considered to be a critical factor for a safely return back to dynamic activities [48]. Thus, the restoration of muscle strength ratio between the operated and contralateral limbs for both the quadriceps and the hamstrings is crucial after an ACL reconstruction for a fast and uneventful return to pre-injury activities [48]. There is evidence that muscular recovery is closely related to pre-operative muscle strength, the time between injury and reconstruction and the pre- and post-surgery rehabilitation [8, 22, 55]. In addition, changes in the sensory system with ACL reconstruction, such as alterations in the somatosensory evoked potentials or the development of inconsistent postural synergies, may also influence muscle [18].

Although many authors have compared lower limb muscle strength in patients with ACL reconstruction after either BPTB or HST grafts, the plethora of information is difficult to interpret. Therefore, a systematic review of the literature is warranted to synthesize reported findings of the isokinetic muscle strength in studies comparing ACL reconstruction using either BPTB or HST autografts. Clarification of muscle strength recovery after ACL reconstruction using either graft type will enhance decision-making with regard to graft choice and rehabilitation.

Materials and methods

A thorough search of the databases MEDLINE, Cinahal and EMBASE for articles that compared muscle strength using isokinetic dynamometry between patients that had undergone ACL reconstruction with either BPTB or HST autograft was completed in September 2009. Full text articles published in English were searched using variations and combinations of the following terms: anterior cruciate ligament reconstruction, knee reconstruction, dynamometry, strength, weakness, and torque.

To be included in this review, articles must have:

  • compared two groups of patients that had undergone ACL reconstruction: one of the groups must have received BPTB autograft and one HST autograft;

  • evaluated knee flexor and extensor isokinetic muscle strength between 4 and 24 months after ACL reconstruction surgery;

  • been published in English language.

The following criteria were used to exclude articles from the systematic review. Studies were not included if:

  • studies did not include original data;

  • any participants had undergone revision of ACL;

  • participants had undergone multiple-ligament reconstruction.

Studies of different methodological design were included in this systematic review and are subject to different biases. Therefore, multiple tools were used to assess the quality of included studies. Randomized Control Trials (RCT) were assessed for quality using the PEDRO scale [20] which assesses the quality of studies based on 11 criteria. All other study designs were assessed using the tool described by Downs and Black [21]. The assessment of methodological quality was completed by 2 reviewers independently. Disagreement was resolved by discussion with a 3rd reviewer.

Extraction of data

Two independent reviewers read all of the articles in the final yield and systematically extracted pre-defined relevant data. Demographic details of participants were extracted from all articles in addition to the descriptive variables of isokinetic strength assessment at all speeds.

A meta-analysis was conducted on the findings of isokinetic evaluations at testing speeds of 60°/s and 180°/s, an average of 12 months after ACL reconstruction surgery. To be included in the meta-analysis, the mean and measures of variability must have been reported. Wherever the outcomes were not presented in a form suitable for direct inclusion in the meta-analysis, the corresponding authors were contacted by email in an attempt to obtain the data required for meta-analysis (numbers of participants, mean scores and SDs).

Statistical analysis

Muscle strength of the operated limb was extracted when reported either as a percentage of the uninvolved limb (i.e. Limb Symmetry Index) or as a percentage deficit of the uninvolved limb (100 × deficit of injured leg/deficit of uninjured leg). Mean differences and 95% confidence intervals were calculated from the extracted data. Random-effects models were used to pool data. Review Manager 5 (Version: 5.0.24) software was used for the calculation of effect sizes.

Results

A total of 1,532 published studies were identified in the original search of databases. Following the application of inclusion and exclusion criteria, a final yield of 14 studies were included in this systematic review as presented in the flow chart (Appendix). Of the 14 included studies, eight were RCT and six non-Randomized Control Trials (non-RCT).

The study design and the characteristics of each study included in this review are presented in Table 1.

Table 1 Characteristics of the included studies
Full size table

Quality assessment of the RCTs and the non- RCTs is presented in Tables 2 and 3. Inadequate randomization may allow the introduction of bias; however, only 3 of the 8 RCTs reported the process of patient randomization. Although blinding of the patient and surgeon is not always possible in this field of research, only 2 studies reported that assessors were blinded to the group allocation of patients.

Table 2 Results from the methodological assessment of the eight RCTs using the Pedro scale
Full size table
Table 3 Results from the methodological assessment of the six non-RCTs using the Downs and Black scale
Full size table

Muscle strength outcomes

The muscle strength outcomes that were reported from all studies are presented in Tables 4 (for RCTs) and 5 (for non-RCTs).

Table 4 Muscle strength outcomes of the included RCT, at the time between 4 and 24 months
Full size table
Table 5 Muscle strength outcomes of the included non-RCT, at the time between 4 and 6 months
Full size table

Six studies [5, 7, 11, 12, 16, 69] found no significant difference between BPTB and HST for isokinetic muscle strength for knee extensors or knee flexors at follow-up times between 4 and 24 months after reconstruction.

Four studies [10, 26, 35, 42] found significant extensor muscle strength deficit in the operated limb in the BPTB group compared to the HST group at different follow-up times between 4 and 24 months. In addition, six studies [10, 13, 26, 31, 42, 71] found significant deficits of the flexor muscles in the operated limb in HST group compared to the BPTB group at different follow-up times between 4 and 24 months.

Sufficient data were provided in only four of the RCTs [11, 16, 26, 42] and three of the non-RCTs [13, 69, 71] to conduct a meta-analysis on findings 12 months after ACLR.

Figures 1 and 2 show forest plots that summarize quadriceps and hamstring strength for patients at a speed of 60°/s. There were 3 articles where muscle strength of the operated limb was reported as a percentage of the uninvolved limb. For patients with HST graft, quadriceps strength was an average of 9% stronger and hamstrings strength was 8% weaker than patients with BPTB graft. Two articles reported muscle strength of the operated limb as percentage deficit of the uninvolved limb. Similarly, patients with HST graft showed a 3% lower deficit in quadriceps strength and 9% greater deficit in hamstrings strength than patients with BPTB.

Fig. 1
figure 1

Forest plots for isokinetic extensor muscle strength at 60°/s at 12 months. BPTB bone patellar tendon bone, HST hamstring, SD standard deviation, CI confidence interval

Full size image
Fig. 2
figure 2

Forest plots for isokinetic flexor muscle strength at 60°/s at 12 months. BPTB bone patellar tendon bone, HST hamstring, SD standard deviation, CI confidence interval

Full size image

Figures 3 and 4 show forest plots that summarize quadriceps and hamstring strength for patients at a speed of 180°/s. There were 2 articles where muscle strength of the operated limb was reported as a percentage of the uninvolved limb. For patients with HST graft, quadriceps strength was an average of 7% stronger and hamstrings strength was 9% weaker than patients with BPTB graft. Two articles reported muscle strength of the operated limb as percentage deficit of the uninvolved limb. Similarly, patients with HST graft showed a 1% lower deficit in quadriceps strength and 20% greater deficit in hamstrings strength than patients with BPTB.

Fig. 3
figure 3

Forest plots for isokinetic extensor muscle strength at 180°/s at 12 months. BPTB bone patellar tendon bone, HST hamstring, SD standard deviation, CI confidence interval

Full size image
Fig. 4
figure 4

Forest plots for isokinetic flexor muscle strength at 180°/s at 12 months. BPTB bone patellar tendon bone, HST hamstring, SD standard deviation, CI confidence interval

Full size image

Discussion

The most important finding of the present study was the apparent trend for muscle weakness that is specific to the graft donor site following ACL reconstruction. The meta-analysis performed showed that extensor muscle strength deficit exists in ACL reconstructed knees using BPTB autograft and that flexor muscle strength deficit exists in ACL reconstructed knees using HST autografts, 12 months post-operatively.

Not all studies reported muscle weakness in one group of patients or the other. Six studies [5, 7, 11, 12, 16, 69] did not find significant differences in extensor or flexor muscle strength between BPTB and HST groups, at any testing speed (60°/s, 120°/s, 180°/s, 240°/s, 300°/s). In contrast, eight studies found differences between groups. Significant quadriceps muscle strength deficit in BPTB group was observed in four studies [10, 26, 35, 42] and six studies found significant hamstrings muscle deficits in HST group [10, 13, 26, 31, 42, 71]. All of the studies evaluated patients between 4 and 24 months after surgery and muscle weakness was found to persist throughout this period. These findings are in agreement with other reviews [19, 57] that have concluded that the graft site affects muscle strength.

There is an obvious trend for quadriceps deficit at BPTB group compared to HST group and a trend for hamstrings deficit in HST group compared to BPTB group at 12 months post-operative. The results of the meta-analysis showed that difference between BPTB and HST group for extensor muscle strength was nearly 10% at the speed of 60°/s and 180°/s and that for flexor muscle strength was 20% at 180°/s. It is clinically accepted that anything less than a 10% difference between limbs is considered inconsequential [39]. Although the difference in quadriceps strength between sides was not greater than 10%, the difference in hamstring strength exceeded this clinical limit. It is difficult to know what the implications for this asymmetry between limbs are, given that most research has focused on investigating asymmetrical quadriceps weakness. Further research is, therefore, needed to establish whether such a large hamstring weakness in the operated limb of patients with HST graft has any clinical relevance.

The apparent trend for muscle strength weakness related to the donor site may be explained by previous research. It seems that harvesting the patellar tendon autograft during the ACL reconstruction may alter the length–tension relationship of the extensor mechanism [32] and consequently contribute to extensor muscle strength deficit. It is also described that muscle function might be altered due to the attenuation of the gamma loop function caused by the initial ACL injury and that is not restored after the ACL reconstruction. The mechanoreceptors located within the ACL play an important role in enhancing the activity of gamma motor neurons (contributing, to a normal muscle function) [36, 38, 63]; however, this mechanism is not restored with ACL reconstruction, and may, therefore, also play a role in the extensor muscle weakness seen after harvesting the BPTB graft. Furthermore, knee pain and effusion have been documented up to 12 months following ACL reconstruction and could alter the neural control of the quadriceps [37, 65, 68].

Strength deficits in the knee flexor muscles may be more easily explained. There is evidence that tendon fibers can regenerate following harvesting of the hamstring tendon to become similar to healthy and non-harvested fibers [25, 28]. However, Hioki et al. [34] found an atrophy of hamstrings’ muscle fibers as well as hypertrophy of the semimembranosus and biceps muscles, after harvesting the hamstrings tendon. Moreover, they demonstrated that after harvesting the hamstrings tendon the semitendinosus muscle assumes different shapes and movements and that each pattern was related to different knee flexor strength. It is not clear how these changes in morphology affect muscle and knee function.

Regardless of the physiological explanations for muscle weakness, it is clear that restoration of muscle strength must focus on increasing muscle strength following ACL reconstruction to maximize functional outcomes. In particular, it appears that patients with different graft types may be susceptible to muscle strength that is specific to graft type. These findings suggest that rehabilitation that addresses muscle weakness specific to graft type may enhance strength outcomes after ACL reconstruction.

The findings of muscle weakness related to graft donor site were not consistent throughout all of the studies included in this review. There were some methodological differences between these studies that may explain this inconsistency. The method of randomization was not reported or was insufficient for the most of the RCTs. Only three [11, 26, 42] used a specific random allocation, which verifies that allocation was concealed. This allows for a bias that potentially could alter the findings of these studies. Although almost all RCTs assessed patients with the same activity level, three did not report the sex of the patients despite the plethora of information that gender influences outcome after ACL reconstruction. Therefore, the generalizability of the findings reported in these studies may be limited [1, 64]. Although blinding is one of the most important factors to limit bias in a RCT, no patients or therapist and only 2 studies reported that assessors were blinded to patient group allocation. Only in the trials of Aglietti et al. [5] and Maletis et al. [42], the assessors were blinded. Again, the potential for bias in the findings of those studies that did not blind assessors needs to be considered. The studies that were not RCTs were subjected to different biases. Because patients in these studies were not randomized to receive either a BPTB or HST graft, it is important that both groups be similar at baseline on factors that may confound muscle strength findings. However, 3 studies did not adequately describe that groups were similar on important demographic characteristics such as height and weight. These limitations need to be considered when interpreting the findings of this review. Future work that compares the muscle strength outcomes between patients with either BPTB or HST ACL reconstruction needs to consider these factors when designing future research.

There are some limitations that need to be considered when interpreting the findings of this review. The meta-analysis was limited to only half of the studies included in the review because of disparity in the parameters of isokinetic testing (e.g., the speed of testing, and the time since surgery). Nevertheless, studies that did not evaluate muscle strength according to the strict criteria were still included in the systematic review and contribute significantly to the information that details recovery of muscle strength following ACL reconstruction.

Conclusions

Although not all studies reported muscle strength differences between patients with either BPTB or HST graft ACL reconstruction, there was an obvious trend toward greater muscle weakness that was dependent on the graft donor site. Rehabilitation that is specific to this difference in muscle strength between graft types is needed.

Furthermore, more high quality studies need to be conducted assessing the muscle strength recovery after the reconstruction of the torn ACL, in order to reveal a potential superiority of a graft type over the other graft options.