OBJECTIVE. Although hamstring injuries are common in athletes, the distribution and location of such injuries have not been well defined. We used MR imaging to determine the frequency of injury by muscle, involvement of one or more muscles, and location of injuries within the musculotendinous unit.
SUBJECTS AND METHODS. We performed MR imaging on 15 consecutive college athletes with clinically diagnosed acute hamstring injuries. A hamstring injury was diagnosed and located on MR imaging by identifying high signal intensity within the muscle on T2-weighted images.
RESULTS. We found that 10 athletes had injuries of a single muscle with six injuries of the biceps femoris, three of the semitendinosus, and one of the semimembranosus. In an additional five athletes, we found primary injuries of the biceps femoris and secondary injuries of the semitendinosus. The injuries occurred in diverse locations within the muscles including five injuries at the proximal musculotendinous junction, two at the distal musculotendinous junction, four within the proximal half of the muscle belly, and four in the distal half. All eight intramuscular injuries were located at the musculotendinous junction within the muscle.
CONCLUSION. The biceps femoris is the most commonly injured hamstring muscle and the semitendinosus is the second most commonly injured. Although hamstring injuries often involve one muscle injured proximally, multiple muscles were involved in 33% of athletes (5/15) and the injuries were distal in 40% of athletes (6/15). All intramuscular injuries occurred at the musculotendinous junction, either at the ends of the muscle or within the muscle belly.
The hamstring muscles are the most commonly injured muscles in athletes and are usually injured during running and jumping [1, 2]. The high incidence of these injuries is emphasized by a prospective study of 180 soccer players that showed that 10% of the athletes suffered a hamstring injury during one season . Although complete hamstring avulsions have been reported [3, 4, 5], such avulsions are uncommon, and most hamstring injuries are partial tears [1, 2].
Despite the frequency of acute hamstring injuries, there have been few studies on their cross-sectional imaging [6, 7, 8, 9]. Because surgery is seldom required for muscle injuries, there is no method of studying the anatomy of such injuries before cross-sectional imaging. Cross-sectional imaging provides a non-invasive method of examining the pattern and location of injury. Previous studies on CT and MR imaging of hamstring muscle injuries have reported differing results regarding the frequency of injuries by muscle and the involvement of one or more muscles [6, 7, 8].
To our knowledge, no studies have examined the location of hamstring injuries within the musculotendinous unit. Although review articles state that hamstring injuries are most common at the proximal musculotendinous junction, the frequency of injury at other parts of the musculotendinous unit is unknown [1, 10].
To provide a better understanding of acute hamstring injuries, we studied hamstring injuries in college athletes. Our goals were to determine the frequency of injury by muscle, involvement of one or more muscles, and location of injuries within the musculotendinous unit.
Subjects and Methods
Our study was developed in consultation with the team physicians and head athletic trainer at our institution. With their support, we obtained approval to perform this study from our institution's human subjects committee. As part of the protocol, university athletes suffering a hamstring injury underwent MR imaging of the injured thigh within 5 days of their injury and compression imaging of their uninjured thigh.
We obtained axial spin-echo T1-weighted (TR range/TE range, 550-650/15-20) and fat-saturated fast spin-echo T2-weighted (TR range/TE range, 2000-4000/100-110) sagittal and axial images of the thighs. MR imaging was performed using a 1.5-T MR imaging unit (Signa; General Electric Medical Systems, Waukesha, WI) with a torso phased array coil and 7- to 10-mm sections. For T1-weighted images, we used a matrix of 256 × 192 with one excitation. For T2-weighted images, we used a matrix of 256 × 256 with two excitations. The field of view was 30 cm for the sagittal images and 40 cm for the axial images.
Each MR image was analyzed to determine which muscle or tendon was injured and to locate the injury within the musculotendinous unit. A muscle or tendon was considered injured if there was intramuscular or intratendinous high signal intensity on the T2-weighted images. If more than one muscle was injured, the muscle with the greatest length and cross-sectional area of signal abnormality was considered the primary injury site. Areas of increased signal intensity smaller than the primary injury were considered secondary injuries. Within each muscle, the injury could be located at the muscle origin or insertion, at the proximal or distal tendon, at the proximal or distal musculotendinous junction, or within the muscle itself (Fig. 1). If the injury was within the muscle, we noted whether the abnormal signal was adjacent to or separated from the intramuscular tendon. Injuries within the biceps or semitendinosus muscles were classified as proximal or distal on the basis of their relationship to the image that first showed the origin of the short head of the biceps femoris. The short head begins approximately at the midpoint of the femoral shaft and served as a reproducible marker for the midpoint of the thigh .
We evaluated other signs of muscle injury including hematoma, recognized as a focal fluid collection with high signal intensity on T1- and T2-weighted images; perifascial edema; and hemorrhage surrounding the muscle.
From April 1996 to December 1997, we recruited 15 injured athletes to participate in our study (one woman and 14 men; age range, 18-22 years). Our study group included four football players, one hockey player, and 10 track athletes. Seven of the 10 track athletes competed in the long jump or the triple jump. MR imaging was performed 1-5 days (average, 3 days) after the injury.
Hamstring muscle injuries were found in all athletes as detailed in Table 1 (Fig. 2). The long head of the biceps femoris was the most commonly injured muscle, with six isolated injuries and five multiple muscle injuries in which it was the primary muscle of involvement. The semitendinosus was the second most commonly injured muscle, with three isolated injuries and five secondary injuries. The semimembranosus was the least injured muscle, with one injury.
TABLE 1 Location of Acute Hamstring Injuries in 15 College Athletes
Primary with secondary semitendinosus
All six isolated injuries of the biceps femoris were within the muscle at the proximal or distal intramuscular musculotendinous junction (Figs. 3A, 3B, 4, 5A, 5B, 5C). The close relationship of these injuries to the intramuscular musculotendinous junction was emphasized by a characteristic featherlike appearance of high signal intensity surrounding the tendon and dissecting the attached muscle fibers (Fig. 4). This pattern was seen in three of the six injuries. In the multiple muscle injuries, the primary site of injury was the biceps femoris at the proximal musculotendinous junction in four athletes and within the proximal muscle in one. These five patients also had secondary injuries of the semitendinosus (Fig. 6A, 6B, 6C).
Of the three isolated semitendinosus injuries, one was at the proximal intramuscular musculotendinous junction and two were at the distal musculotendinous junction (Fig. 7A, 7B). All the isolated semitendinosus injuries occurred in track and field jumpers. The isolated semimembranosus injury occurred at the proximal musculotendinous junction (Fig. 8).
All injuries were partial tears; we found no evidence of tendon avulsions, isolated tendon injuries, injuries of the short head of the biceps femoris, or intramuscular injuries distal to the musculotendinous junction. Hematomas were uncommon with only one 2-cm round collection (Fig. 5A, 5B, 5C) and two crescentic subfascial collections. Perifascial fluid was a common finding seen in 13 (87%) of the 15 athletes (Figs. 6A, 6B, 6C, 7A, 7B, 8). Gross muscle fiber disruption was noted in four patients.
The hamstring muscles consist of the semimembranosus, semitendinosus, and biceps femoris. They originate from an incompletely separated tendon mass on the lateral and proximal aspect of the ischial tuberosity . In our study of 15 college athletes, we found six isolated and five primary biceps injuries, three isolated and five secondary semitendinosus injuries, and one isolated semimembranosus injury.
The predominance of biceps injuries is consistent with the clinical perception that the biceps femoris is the most commonly injured hamstring muscle . Our findings are consistent with previous CT and MR studies of hamstring injuries in which the biceps femoris was the most commonly injured muscle [6, 7, 8]. Using CT, Garrett et al.  found seven injuries of the biceps femoris, one of the semimembranosus, and two of the semitendinosus. Using CT or MR imaging, Speer et al.  found 11 injuries of the biceps femoris, four of the semimembranosus, and two of the semitendinosus muscles. Pomeranz and Heidt  found six injuries of the biceps femoris, five of the semimembranosus, and one of the semitendinosus muscles.
Our finding that the semitendinosus was the second most commonly injured muscle agrees with one study and differs from two [6, 7, 8]. Differences in the frequency of injuries might reflect the statistical variation within a small series or the different injury patterns occurring in various sports. For instance, in our study we noted that all the athletes with an isolated injury to the semitendinosus were track and field jumpers. We speculate that the correlation of knee extension with hip flexion might predispose track athletes to semitendinosus injuries.
Although the frequency of injury within specific hamstring muscles has been previously reported [6, 7, 8], the location of injuries within the musculotendinous unit has not been studied. Clinically, the most common site of a biceps femoris injury is at the proximal musculotendinous junction so that the maximum pain is about 10 cm below the gluteal fold .
In their anatomic dissections, Garrett et al.  found that the hamstring muscles have complex relationships at the musculotendinous junction. Rather than having a narrow zone of transition from the muscle to the tendon at each end of the muscle belly, the tendons extend within the muscle belly for each of the hamstring muscles. Each hamstring muscle has a tendon extending completely or almost completely down the length of the muscle . In an animal model of muscle-stretch injury, Garrett et al.  noted that all injuries occur at the musculotendinous junction either at the ends of the muscle or within the muscle belly.
Our clinical study is in agreement with the animal study performed by Garrett et al. ; in our study, all the acute hamstring injuries were located at the musculotendinous junction (Table 1). Our study is the first to emphasize that injuries of the musculotendinous junction can be intramuscular or located at the proximal or distal junction between the tendon and the muscle belly.
The location of the injuries at the intramuscular musculotendinous junction is defined by the featherlike pattern of edema (Fig. 4C). This pattern was noted in previous MR imaging studies of muscle injury [9, 13, 14]. A featherlike pattern of muscle edema appears to be characteristic of muscle strain.
In our study, patients often injured multiple hamstring muscles. In fact, five (33%) of the 15 athletes had injuries of more than one muscle. Only one previous study has noted that more than one muscle may be injured . Of eight athletes with diagnoses of acute hamstring injuries on CT, Garrett et al.  found isolated injuries in six athletes and combined injuries of the biceps femoris and semitendinosus in two. Because all hamstring muscles are subject to overstretching or forceful contractions, more than one muscle is often injured.
We were able to identify injured muscles because MR imaging provides a detailed delineation of muscle anatomy and is very sensitive to edema. We found that T1- and T2-weighted axial images were necessary to locate injuries. The T2-weighted images permitted easy identification of edema, but the loss of tissue plane definition with fat suppression made it difficult to determine the relationship to the intramuscular tendon or whether one or more muscles were involved without comparison with T1-weighted images at the same level.
In three patients, we noted the presence of a small hematoma at the injury site. Although, to our knowledge, such focal blood collections have not been previously reported on imaging of acute hamstring tears, hematomas are not unexpected when muscle fibers are torn. The high frequency (87% of athletes) of perifascial fluid was consistent with what was previously reported as a “rim” of blood and edema . Fluid collects in a perifascial location because each muscle is surrounded by a thick perimysium (also called the fascia of the muscle) [9, 11].
In summary, the biceps femoris was the most commonly injured muscle in acute hamstring injuries in college athletes. The semitendinosus muscle was the second most commonly injured. In our study, five of the 15 athletes injured more than one muscle. These five patients sustained primary injuries of the biceps femoris and secondary injuries of the semitendinosus. Six (40%) of the 15 injuries were distally located. All the injuries were of the musculotendinous junction at diverse locations including the proximal musculotendinous junction, the intramuscular musculotendinous junction, and the distal musculotendinous junction.
We thank Dennis Helwig and all the university athletic trainers and team physicians.
Address correspondence to A. A. De Smet.
Kujala UM, Orava S, Järvinen M. Hamstring injuries: current trends in treatment and prevention. Sports Med 1997; 23:397-404