OBJECTIVE. The purpose of this study was twofold: to determine whether asymptomatic distance runners exhibit cortical tibial abnormalities on CT and to determine the diagnostic accuracy of CT in athletes with medial tibial stress syndrome.
MATERIALS AND METHODS. A cross-sectional study with high-resolution CT of both tibiae was performed on 41 subjects: 20 asymptomatic distance runners, 11 distance runners with unilateral or bilateral pain due to medial tibial stress syndrome (14 painful tibiae), and 10 volunteers not involved in a sport. The group was composed of 13 women and 28 men, ranging in age from 18 to 26 years. A total of 82 tibiae, 14 painful and 68 painless, were evaluated. On the basis of CT findings, tibiae were classified in three groups, and correlation between CT classification and symptoms was made.
RESULTS. Among distance runners, the presence of CT abnormalities was found in 14 (100%) of 14 painful tibiae in patients with medial tibial stress syndrome and in 8 (16.6%) of 48 painless tibiae. The difference was statistically significant (p < 0.001, Fisher's exact test). Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of CT in diagnosing medial tibial stress syndrome were 100%, 88.2%, 63.6%, 100%, and 90.2%, respectively.
CONCLUSION. High-resolution CT has high diagnostic accuracy in depicting medial tibial stress syndrome. Cortical abnormalities can also be seen in some asymptomatic distance runners.
Running is an increasingly popular form of exercise and competition in our society. Consequently, the incidence of exercise-induced lower leg pain is also increasing. Leg pain in athletes has many causes. The clinician must strive to specifically define the clinical problem to administer the appropriate treatment for the athlete's condition. Clinical conditions in the leg causing symptoms in athletes include chronic exertional compartment syndrome, tendinitis, stress fractures, fascial defects, musculotendinous junction disruption (tennis leg), popliteal artery entrapment syndrome, effort-induced venous thrombosis, nerve entrapment, and medial tibial stress syndrome [1, 2]. Even for an astute clinician, distinction among these different medical causes may be difficult given that many of their presenting features overlap . Therefore, appropriate diagnostic studies are needed to allow accurate diagnosis.
Medial tibial stress syndrome is one of many overuse lower leg injuries that may be found in athletes . Medial tibial stress syndrome is characterized by localized pain that occurs during exercise at the medial surface of the distal two thirds of the tibial shaft . Medial tibial stress syndrome, otherwise known as shin splints, is a common injury experienced by runners. It accounts for between 13.2% and 17.3% of all running injuries .
Johnell et al.  were the first to show that symptoms of medial tibial stress syndrome are correlated to bone stress reaction after taking biopsies and finding bone porosity. Using dual-energy X-ray absorptiometry, Magnusson et al.  also found abnormally decreased regional bone density in athletes with medial tibial stress syndrome. Recently, Gaeta et al.  proved that both CT and MRI provide early findings of stress injury in patients with activity-related tibial pain. In addition, they showed that CT is the method of choice in detecting osteopenia, the earliest sign of fatigue damage of the cortical bone in tibial diaphysis. However, although these studies have explained a large part of the pathophysiology for this common condition, some questions about it remain unresolved [5, 7, 8]—namely, quoting Magnusson et al. , “it is still unknown whether the symptoms result from or precede the loss of tibial bone mineral density.”
We performed a cross-sectional study to investigate whether CT abnormalities can occur in asymptomatic distance runners and to evaluate the accuracy of CT findings in distinguishing between asymptomatic and symptomatic subjects, and we also attempted to answer the question of Magnusson et al. .
Materials and Methods
Twenty consecutive distance runners (14 men, 6 women; age range, 18-25 years) agreed to participate in the study. All had a history of long-term running and had participated in an intensive training program for 2 to 6 months during which they ran 30-50 miles (48-80 km) per week. The CT examinations were performed during this training. All subjects were asymptomatic for lower extremity pain. Exclusion criteria were recent tibial pain or tenderness or a history of a tibial stress fracture or stress injury.
In addition, 11 consecutive distance runners (7 men, 4 women; age range, 20-24 years) who had a comparable history of long-term running and training regimen but complained of symptoms of medial tibial stress syndrome and 10 consecutive asymptomatic volunteers who were not involved in a sport (7 men, 3 women; age range, 20-26 years) were included in the study as age- and sex-matched control subjects. Differences in body habitus among the participants were not controlled.
The protocol was approved by the medical ethics committee of our university hospital. Informed consent was obtained from each patient and volunteer after the nature and the aim of the examination had been fully explained.
The diagnosis of medial tibial stress syndrome in the group of symptomatic runners was made by a review of clinical findings, physical examination, and a detailed history by three experienced sports medicine physicians. Other causes of chronic leg pain were excluded. In all patients, pain was induced by exercise. In some patients, palpation of the tibia produced discomfort. There was no history of plantar paresthesia or other symptoms, such as muscle cramping and swelling, that are indicative of other causes of exercise-induced leg pain. Sonography was performed in all the patients to evaluate the muscles and tendons. Color Doppler examination was selectively performed to exclude the presence of vascular disease. Electromyography with nerve conduction was selectively used to evaluate patients with suspected radicular nerve conditions. In six patients, measurement of compartment pressure was performed, and the results excluded the presence of chronic compartment syndrome, which is the most common cause of exercise-induced leg pain other than tibial stress injury . The results of CT were not included to avoid bias .
All CT examinations were performed using a 16-MDCT scanner (Somatom Sensation 16, Siemens Medical Solutions). Both lower legs were positioned together and imaged from the middle to distal tibial diaphysis. The scanning parameters were as follows: peak tube voltage, 120 kVp; effective tube current-time product, 120 mAseff; detector collimation, 16 × 0.75 mm; rotation time, 1 sec; and table feed per rotation, 12 mm. In each subject, lead protection was used for regions of the body not included in the examination field to reduce radiation dose .
The images were stored digitally. At the conclusion of the examinations, the image sets obtained for each patient were displayed on a computer monitor and then evaluated by the observers. Each examination was reviewed at the time of acquisition by a senior radiologist to ensure that images were of diagnostic quality and to give a preliminary interpretation of the findings. In addition, an initial review was performed with knowledge of all available clinical and imaging findings to make a formal report and to optimize patient care.
CT Image Analysis
All CT examinations were evaluated by two experienced musculoskeletal radiologists in consensus. These radiologists had not been involved in the execution of the examinations. The reviewers were aware that the patients were being evaluated for a possible tibial lesion, but they were unaware of the other clinical findings.
Axial CT images were displayed with two window and level settings to assess bone structures and soft tissues. To improve evaluation of the craniocaudal extensions of abnormalities, a sagittal reconstruction for each tibia was obtained along the medial tibial cortex. The reviewers were asked to classify the tibial cortex findings according to the description of Gaeta et al.  as follows: type 0, no abnormality (Fig. 1); type 1, small, scattered area of slightly reduced cortical attenuation without clear findings of osteopenia (Figs. 2A and 2B); or type 2, cortical osteopenia without or with cavitations, striations, or both cavitations and striations (Figs. 3, 4A, and 4B). For the evaluation of cortical osteopenia on CT images, the signal difference-to-noise ratio of reduced bone attenuation was measured, according to the description of Gaeta et al. , as follows: a region of interest (ROI) was placed within the zone of reduced bone attenuation and in the nearest normal-appearing part of the bone cortex. The area of the ROI within the bone cortex varied between 0.05 and 0.1 cm2. The size of the ROI within the area of reduced bone attenuation and the size of the ROI within normal bone cortex were identical. The difference between the mean bone attenuation then was divided by the SD of the signal intensity measured outside the tibia (noise, ROI = 1 cm2). An attenuation reduction of 10% or more was considered to be osteopenia.
Data Analysis and Statistical Evaluation
A radiologist not involved in the execution of the examinations and in the evaluation of the images analyzed the collected data. Type 0 and 1 were considered to be normal cortex, and type 2 was considered to be abnormal cortex. The prevalence of all CT abnormalities was calculated, and the data were compared with the presence of pain. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated.
The two-tailed Fisher's exact test, with a p value of < 0.05 considered to indicate a statistically significant difference with a 95% CI, was used to test the difference between groups.
Because CT images of both tibiae of each subject were obtained, a total of 82 tibiae were included in this study: 20 tibiae of volunteers who do not participate in a sport; 40 tibiae of asymptomatic runners; and 22 tibiae of symptomatic runners, of which 14 tibiae were painful (three runners had bilateral tibial pain) and eight were painless. Therefore, 68 painless tibiae and 14 painful tibiae were evaluated.
In 10 of the asymptomatic volunteers not involved in a sport, 19 tibiae were classified as type 0 and one tibia as type 1.
In the 20 asymptomatic runners, 22 tibiae were classified as type 0, 13 as type 1, and 5 as type 2. Eight painless tibiae in symptomatic runners were classified as follows: type 0, n = 2; type 1, n = 3; and type 2, n = 3. All 14 painful tibiae were classified as type 2. In summary, 43 tibiae were classified as type 0, 17 as type 1, and 22 as type 2.
Forty-three of 43 tibiae classified as type 0 and 17 of 17 tibiae classified as type 1 were painless. On the other hand, 14 of 22 tibiae classified as type 2 were painful.
Among distance runners, CT abnormalities (type 2 CT findings) were found in 14 (100%) of 14 painful tibiae and in 8 (16.6%) of 48 painless tibiae. The difference was statistically significant (p < 0.001).
Sensitivity, specificity, PPV, NPV, and accuracy of CT in predicting the presence of pain were 100%, 88.2%, 63.6%, 100%, and 90.2%, respectively.
Stress injuries comprise a spectrum of soft-tissue and osseous abnormalities that occur in response to abnormal repetitive stress applied to normal bone [11, 12]. Repetitive submaximal stress creates a region of accelerated bone remodeling that may progress to a stress fracture if the stress continues. Symptomatic exercise-induced stress reactions involving the tibia are common in athletes and may account for up to 75% of exertional leg pain [13-15]. The most frequent location for tibial stress reactions is the tibial diaphysis .
Medial tibial stress syndrome, a term originally selected by Drez, according to Mubarak et al. , is one of many overuse injuries. It denotes the symptom complex otherwise known as shin splints, which are defined as long-lasting pain occurring at the medial surface and distal two thirds of the tibia that increases during exercise. Until recently, inflammation of the periosteum due to excessive traction was considered the most likely cause of medial tibial stress syndrome . Recent studies have supported the view that medial tibial stress syndrome is not an inflammatory process of the periosteum but instead a bone stress reaction that has become painful.
Bone is a dynamic tissue that requires stress for normal development. Stresses related to daily activities stimulate the remodeling process. Increased osteoclastic resorption is the initial response to abnormal stresses. If increased stress persists, an imbalance between bone resorption and bone replacement leads to weakening of the bone. Accelerated intracortical remodeling causes microscopic cracks, osteopenia, and formation of resorption cavities that may join in larger lesions . Stresses in cancellous bone may initially result in microfractures. If the inciting activity is not decreased, the accumulation of microdamage may result in stress fracture of cortical or trabecular bone [17-28]. Therefore, osseous stress injury is a pathophysiologic continuum from accelerated remodeling to frank stress fracture. Diagnosis of early stress injuries permits appropriate therapy and may prevent these lesions from progressing to stress fracture.
Magnusson et al.  measured bone mineral density in 18 adult male athletes with long-standing medial tibial stress syndrome and compared the measurements with those of 16 age- and sex-matched control subjects and with those of 18 athletes without medial tibial stress syndrome who had a comparable training regimen. They found that athletes with medial tibial stress syndrome had lower bone mineral density at the affected region compared with control subjects and athletes without medial tibial stress syndrome. In addition, bone mineral density was decreased on the unaffected side in subjects with unilateral symptoms, indicating that osteopenia may precede medial tibial stress syndrome. The article by Magnusson et al. confirms previous reports that symptoms of medial tibial stress syndrome are correlated to bone stress reaction, as shown by some authors, after taking biopsies and finding bone porosity [6, 30].
Recently, Gaeta et al.  showed that MRI is the best single technique to use to reveal tibial stress injuries at an early stage. However, in the subset of stress injuries of the tibial cortex, CT is the most sensitive technique in diagnosing early abnormalities—namely, osteopenia. They described three types of cortical lesion: osteopenia, resorption cavity, and striation.
The results of our work show that CT findings have high accuracy for the diagnosis of medial tibial stress syndrome. However, we detected clear findings of osteopenia in about 17% of painless tibiae; therefore, the presence of cortical abnormalities was not a completely reliable marker for tibial pain. Remarkably, CT confirmed the observation obtained by bone mineral density analysis that reduced tibial bone density may be seen in some asymptomatic long-distance runners .
Our study has some weaknesses and limitations. First, the use of a cross-sectional study design limits interpretation of the results. A major weakness of our study is the lack of clinical follow-up to determine whether the presence of bone stress injuries in asymptomatic runners can predict later development of symptoms. It is intriguing to hypothesize that asymptomatic osteopenia may precede the development of medial tibial stress syndrome. Recently, the presence of tibial abnormalities in asymptomatic distance runners has been shown on MRI in 33% of asymptomatic distance runners . The presence of these changes was not found to be a predictor of future symptoms. However, in that work , only soft-tissue abnormalities were shown on MRI. Therefore, it is not possible to extend these results to patients studied on CT because bone changes could have a different predictive value than soft-tissue stress reactions.
Several bone scintigraphic studies have shown that it is not rare to identify asymptomatic stress lesions [11, 15, 32, 33]. In a series, the frequency of detection was as high as 46% . The authors of that series have seen that some of these patients may develop pain at the site of asymptomatic uptake in the weeks and months after a bone scan examination . These scintigraphic data support the hypothesis that asymptomatic cortical abnormalities that can be shown by CT are areas of stress-related bone remodeling detected earlier on the pathophysiologic continuum and that they may be a potential predictor of medial tibial stress syndrome. Additional studies are necessary to clarify this point definitively.
Moreover, the lack of follow-up does not permit us to ensure that the resolution of symptoms was related to the resolution of the CT findings. However, we believe that this is a reasonable hypothesis. This belief is supported by the work of Magnusson et al.  who, using bone mineral density analyses, have seen that low tibial bone density in athletes with medial tibial stress syndrome normalizes after recovery from symptoms.
Another weakness of our work is the lack of histopathologic specimens that would be needed to correlate imaging with pathologic findings. This limitation cannot be avoided because, in the clinical setting of tibial stress injuries, it is not easy to obtain bone biopsy. However, the existence of previous studies on the physiopathology of the stress lesion permits us to hypothesize with good reliability that the pathologic abnormalities may correlate with and explain the imaging findings. For example, the appearance of resorption cavities and striations within the cortex may correlate with osteoclastic proliferation. These abnormalities are similar to those visible in metabolic osteoporosis .
Despite these limitations, we think that this study has both clinical and research significance. Clinically, our results indicate that CT has a high sensitivity in detecting abnormalities of the tibial cortex both in patients with medial tibial stress syndrome and in distance runners with presymptomatic bone remodeling.
Recently, Gaeta et al.  have shown that MRI is the better single method to use for the diagnosis of tibial stress injuries. However, 4 (8%) of 50 symptomatic patients with normal findings on MRI examination had cortical abnormalities shown only on CT. Consequently, we believe that a practical diagnostic imaging approach to athletes with lower leg pain should start with MRI examination as a first step. If MRI is not diagnostic, CT should be considered as an additional investigation with which to detect cortical abnormalities when a strong clinical suspicion of medial tibial stress syndrome exists.
We think that this work also has research significance. High-resolution CT has a potential role as a tool with which to understand the in situ response of bone to physiologic forces. Because these responses are difficult to model in the laboratory using tissue or animal models, there is a growing interest in using noninvasive imaging techniques to study these processes in humans.
An interesting point that has to be addressed to better understand pathophysiology of bone stress reactions is the relationship between bone changes and pain. Although bones have some nerves, it is not clear whether the periosteum and the associated vascular changes induced by overuse generate the pain and the cortical porosity is a secondary feature. Recently, Magnusson et al.  stated that “it is still unknown whether the symptoms result from or precede the loss of tibial bone mineral density.” Our results showed that bone osteoporosis and resorption cavities can be found in asymptomatic athletes. Obviously, in these subjects, pain cannot be the cause of bone remodeling. On the other hand, no patient with medial tibial stress syndrome had normal tibial cortex. We think this is evidence that bone remodeling always precedes pain and therefore is highly probable that it is the cause of symptoms.
In conclusion, we have shown that CT has a high accuracy in revealing stress-induced cortical bone remodeling, both in symptomatic and asymptomatic athletes. Although agreement is growing that MRI is the best technique for the assessment of stress injuries to bone, we think that CT can be a useful imaging tool both in clinical and in research approaches to this pathophysiologic entity.
Magnusson HI, Henrik GA, Karlsson C, Nyquist F, Karlsson MK. Low regional tibial bone density in athletes with medial tibial stress syndrome normalizes after recovery from symptoms. Am J Sports Med 2003; 31:596-600
Gaeta M, Minutoli F, Scribano E, et al. CT and MRI findings in athletes with early tibial stress injuries: comparison with bone scintigraphy and emphasis on cortical abnormalities. Radiology 2005; 235:553-561