Article Text
Abstract
Background The majority of sporting injuries are clinically diagnosed using history and physical examination as the cornerstone. There are no studies supporting the reliability of making a clinical diagnosis of medial tibial stress syndrome (MTSS).
Aim Our aim was to assess if MTSS can be diagnosed reliably, using history and physical examination. We also investigated if clinicians were able to reliably identify concurrent lower leg injuries.
Methods A clinical reliability study was performed at multiple sports medicine sites in The Netherlands. Athletes with non-traumatic lower leg pain were assessed for having MTSS by two clinicians, who were blinded to each others’ diagnoses. We calculated the prevalence, percentage of agreement, observed percentage of positive agreement (Ppos), observed percentage of negative agreement (Pneg) and Kappa-statistic with 95%CI.
Results Forty-nine athletes participated in this study, of whom 46 completed both assessments. The prevalence of MTSS was 74%. The percentage of agreement was 96%, with Ppos and Pneg of 97% and 92%, respectively. The inter-rater reliability was almost perfect; k=0.89 (95% CI 0.74 to 1.00), p<0.000001. Of the 34 athletes with MTSS, 11 (32%) had a concurrent lower leg injury, which was reliably noted by our clinicians, k=0.73, 95% CI 0.48 to 0.98, p<0.0001.
Conclusion Our findings show that MTSS can be reliably diagnosed clinically using history and physical examination, in clinical practice and research settings. We also found that concurrent lower leg injuries are common in athletes with MTSS.
- medial tibial stress syndrome
- diagnosis
- history
- physical examination
- reliability
Statistics from Altmetric.com
Introduction
Medial tibial stress syndrome (MTSS) is defined as exercise-induced pain along the posteromedial tibial border, and recognisable pain is provoked on palpation of this posteromedial tibial border over a length of ≥5 consecutive centimetres.1 MTSS is a common overuse sports injury,2 3 with incidence rates from 4% to 19% in athletic populations.4
MTSS is diagnosed clinically using history and physical examination. Various imaging techniques have been studied for their ability to identify athletes with and without MTSS. These studies used the clinical diagnosis as the gold standard and examined if imaging compared with this.5 6 In this case, imaging’s accuracy will always be less than the clinical diagnosis. Studies into imaging of other sports injuries, such as patellofemoral pain syndrome and groin pain, have also shown to lack discriminatory ability between symptomatic and asymptomatic athletes.7 8 In MTSS, the underlying pathology is equivocal, with both bony overload or periosteal inflammation being reported.4 9 There seems to be a need for a shift in the diagnostic paradigm for sports injuries where the pathogenesis is unclear; from making a diagnosis based on imaging or histological findings towards a diagnosis based on clinical findings.10
While history and clinical examination are the cornerstones of the diagnostic process in MTSS, the reliability of this approach has never been examined. Making a reliable clinical diagnosis forms a good foundation for planning treatment and discussing expectations. Ascertaining that clinicians are able to make a reliable diagnosis is also essential for research purposes.11 We aimed to investigate the inter-rater reliability of using standardised history and physical examination to diagnose MTSS.
Methods
Design
Cross-sectional study.
Setting
Four locations in The Netherlands (Inholland University of Applied Sciences, Haarlem; Academy for Physical Education, The Hague; the Sports Medical Advice Centre Haarlem and a handball club (HV Hellas) in The Hague), from March 2015 to August 2016.
Participants
Athletes (≥16 years) (ie, students at Inholland University of Applied Sciences, Haarlem and Academy for Physical Education, The Hague; athletes at the Sports Medical Advice Centre Haarlem and HV Hellas, The Hague) who presented with a gradual onset of any lower leg pain (ie, pain between the tip of the medial malleolus and the tibial plateau) for at least 1 week were potentially eligible for inclusion. No further restrictions with regards to the location of the pain were imposed. Exclusion criteria were a traumatic cause for the pain or a history of tibial fracture.
Potential candidates were informed about the study by a clinician or trainer/coach. Those athletes that were potentially willing to participate were provided with written and verbal information about the study by one of the medical professionals or trainers/coaches. All athletes were asked to sign informed consent after which they were included in the study. The Medical Research Ethics Committees United, Nieuwegein, The Netherlands (W15.029) provided approval.
Demographic information
After inclusion, athletes filled out a demographic information form: gender (male/female), height (cm), weight (kg) and injury duration (months). The MTSS score was filled out by athletes with MTSS to assess injury severity. The MTSS score is a new valid, reliable and responsive patient reported outcome measure. Scores range from 0 to 10, 0 means having no pain/limitation, whereas 10 is maximal severity.12 13
Procedure
Eight clinicians (five sports physiotherapists and three sports physicians; mean (SD) years of experience 8 (9); median (range) 5.5 years (1–23)) were available to assess the included athletes. For most cases, there were more than two clinicians available to make the diagnosis MTSS. In those cases, two clinicians were randomly selected by a computer from the pool of available clinicians. The assessment order was also randomly determined by a computer in all cases. The person that performed the randomisation procedure was not blinded to the clinician to be selected, nor to the athlete that was to be assessed. The two assessments took place on the same day to prevent the athlete’s condition changing.
Diagnosis MTSS, based on history taking and physical examination
We used a standardised history and physical examination to diagnose MTSS clinically (figure 1). We used six steps for the confirmation of the diagnosis MTSS, based on the previous work by Yates and White and Edwards et al. (2005).1 14 We explained this to the clinicians before the study commencement. The clinicians were not specifically trained for the study purpose.
History
The standardised history comprised questions on the onset and location of the pain. If there was exercise-induced pain along the medial tibial border, the athlete was asked what aggravated and relieved their pain. Athletes were also asked about pain in adjacent areas, or remote areas in the lower leg. Then, athletes were also specifically asked for the presence of any signs of chronic exertional compartment syndrome (CECS), which could be a concurrent injury or the sole explanation for their pain. Athletes were asked about cramping, burning and pressure-like calf pain; pain that was primarily present during exercise, which quickly decreased after exercise. Athletes were also asked whether they experienced any pins and needles in the foot or a cold foot during exercise, especially when pain in the calf area was reported.
Physical examination
If MTSS was suspected after the history, the posteromedial tibial border was palpated and the athletes were asked for the presence of recognisable pain (ie, from painful activities). If no pain on palpation was present, or the pain could be palpated over less than 5 cm, other lower leg injuries (eg, a stress fracture) were considered to be present and the athlete was labelled as not having MTSS. When recognisable pain was present on palpation over 5 cm or more and no atypical symptoms were present, the diagnosis MTSS was confirmed. When the length of perceived pain along the posteromedial border was equivocal, a tape measure was used to determine the exact length. During physical examination, athletes were specifically asked for pain in adjacent structures. If so, those structures were palpated and athletes were asked if recognisable pain was present.
We did not specifically define other injury conditions, that is, we did not define CECS, a tibial stress fracture or soleus strain. The clinicians were free to use their own preferred terms to describe other diagnoses (eg, calf pain/CECS/suspicion of tibial stress fracture). This study solely focused on the reliability of making the diagnosis MTSS (present yes/no) and the presence of co-existed lower leg conditions (present yes/no).
Blinding
The raters made their clinical diagnosis independently, and were blinded to the other clinician’s assessment. Blinding of the raters was performed as follows: only one clinician was in the assessment room when the athlete’s injury was examined. The first clinician examined the athlete but did not relate their findings to them, and only the second clinician communicated the diagnosis to the athlete. Each athlete was also instructed beforehand, not to share the findings of the first clinician with the second.
Statistical analysis
Statistical analyses were performed using SPSS V.22.0 (IBM SPSS, Chicago, USA). Demographic data are presented with their estimates and appropriate measure of dispersion. For the reliability analysis, we used the two diagnoses of each set of clinicians and aggregated these for the analysis to one set of two clinicians. Specifically, we calculated the reliability statistics over data collected by all sets of clinicians, in contrast to the calculation for each set of clinicians. We calculated the prevalence, percentage of agreement, observed Ppos, Pneg and our primary outcome measure: the chance-corrected ratio for agreement, Kappa-statistic. Kappa was interpreted as follows: poor (k<0.00), slight (k=0.00–0.20), fair (k=0.21–0.40), moderate (k=0.41–0.60), substantial (k=0.61–0.80) or almost perfect (k=0.81–1.00).15 Bias between clinicians can inflate Kappa whereas a low or high prevalence can deflate Kappa.16 17 We calculated the Bias Index (BI) and Prevalence Index (PI) to evaluate how Kappa may have been affected by bias between clinicians, and by prevalence. BI ranges from −1 to +1. In case both clinicians label an equal proportion of the population as having MTSS (ie, ’prevalence') the BI is 0, and consequently, the Kappa-statistic is not affected by bias between clinicians. The closer to −1 or +1 the more the Kappa-statistic is inflated. PI ranges also between −1 and +1. Opposite to the BI, a value closer to −1 or +1 results in a deflated Kappa. A value of 0 (the average prevalence across the two clinicians is 50%) indicates that prevalence does not affect Kappa (see online supplementary appendix 1 for all calculations).18 The sample size calculation showed that 51 athletes with lower leg pain were required for an expected Kappa of 0.6 and the prevalence to be 50%, constructing a two-sided 95% CI, with a distance from the estimated Kappa to the limit of the 95% CI of 0.2.19 Missing demographic, continuous data were handled by imputing sample means. Missing sports activity data were labelled as ’unknown'. If athletes failed to attend their second assessment they were excluded from the reliability analysis.
Supplementary appendix
Results
A total of 52 athletes agreed to participate, of which 49 met our inclusion criteria. Three athletes were excluded; two due to lower leg pain after acute ankle trauma, one who had insertional Achilles tendinopathy. Three athletes failed to attend their second assessment. Figure 2 shows the study flow.
We included 14 (29%) males and 35 (71%) females. Injury severity was moderate in the athletes with MTSS; the mean (SD) MTSS score was 3.82 (1.42). Table 1 provides further demographic information.
Missing data
There were missing demographic data for eight athletes (3.2% of all data), as they did not fill out the demographic information form. No data regarding the diagnosis making process was missing.
Inter-rater reliability MTSS diagnosis
There were 34/46 (74%) athletes with MTSS, and 12/46 (26%) with other lower leg injuries. These other lower leg injuries were categorised as: anterior tibialis muscle pain (n=5), calf pain (n=3), tibial bony stress reaction (n=2) and peroneal muscle pain (n=2).
The percentage of agreement, Ppos and Pneg were 96%, 97% and 92%, respectively. The Kappa was almost perfect: k=0.89, 95% CI 0.74 to 1.00, p<0.000001 (see table 2A and table 3). Clinicians did not make the same diagnosis in 2/46 cases (4%). One was labelled as having only MTSS by one clinician, and as having pain in the flexor hallucis longus by the other. The second athlete was labelled as having MTSS and a tibial stress reaction by one clinician, and as only having a tibial stress reaction by the other. The reliability may have been deflated by the high prevalence of MTSS in our sample, that is, an underestimation of Kappa, PI=0.48. Kappa was not affected by bias between clinicians, BI=−0.04.
Inter-rater reliability presence of concurrent lower leg injury
Of the 34 athletes with MTSS, 11 (32%) had a concurrent lower leg injury. These were anterior tibial muscle pain (n=5), calf pain (n=5) and a tibial stress reaction (n=1). The percentage of agreement, Ppos and Pneg for the identification of a concurrent lower leg injury (yes/no) were 88%, 82% and 91%, respectively. The Kappa for the identification of concurrent lower leg injuries (yes/no) was substantial, k=0.73, 95% CI 0.48 to 0.98, p<0.0001 (see table 2B and table 3).
In four athletes with MTSS, the clinicians did not agree whether there was a concurrent lower leg injury present. The first clinician identified three cases with MTSS plus a concurrent injury: one anterior tibial stress reaction, one calf pain and one anterior tibialis muscle pain. These concurrent injuries were not noted by the second clinician. The second clinician identified one MTSS athlete as having MTSS plus concurrent calf pain. This additional calf pain was not noted by the first clinician. Reliability may have been deflated by the low prevalence of concurrent lower leg injuries, PI=−0.35, but was not affected by bias between clinicians, BI=0.06.
Discussion
This is the first study to assess the inter-rater reliability of diagnosing MTSS using standardised history and physical examination. Our results show that MTSS can be diagnosed with almost perfect reliability in clinical practice. Concurrent lower leg injuries were often present (32%) in athletes with MTSS and the presence of concurrent injuries could also be identified reliably. Our findings support the use of standardised history and clinical examination for diagnosing MTSS in clinical practice and research settings.
Clinical diagnosis of MTSS: the logical approach?
Although MTSS is mainly considered a bony overload injury,4 20 21 some studies suggest it being related to traction periostitis,22–24 meaning evidence for its pathogenesis is equivocal.9 Previous studies investigated the accuracy of MRI and CT for diagnosing MTSS.5 6 In these studies, the clinical diagnosis of MTSS was used as the gold standard to determine it being present. In this approach, the diagnostic accuracy of imaging will always be lower than that of clinical examination.
The more common text book approach in diagnostic research is when clinical tests/diagnoses are compared with imaging, surgery or histological findings. This is useful when the pathogenesis of an injury is known. This is, however, not the case in most overuse sports injuries. An alternative approach in this paradigm is the use of imaging in the diagnosis of sports injuries to examine its ability to accurately discriminate symptomatic from asymptomatic subjects. In the majority of cases for overuse sports injuries, imaging has been found to have a poor discriminatory ability.7 8 Imaging leads to uncertainty in sports medicine practice, trying to identify which imaging ’abnormalities' are related to the clinical condition, rather than clarifying a patient’s condition. This has also been highlighted by others recently.10
The role of imaging could focus on whether it provides prognostic information or predicts treatment response rather than diagnostic accuracy. However, clinical findings should also be accounted for when assessing the prognostic value of imaging, as shown by a recent study of acute hamstring injuries. MRI did not add to the predictive value when clinical parameters were used to estimate the prognosis of time to recovery.25 For diagnostic purposes, imaging may be used to rule out other entities with a known pathogenesis (eg, stress fractures, or suspicion of another rare condition like osteosarcoma,26 ie, if there is doubt in the source of lower leg pain).
There seems a need for a paradigm shift in the diagnosis of clinical conditions, like MTSS. They can be diagnosed clinically, without wasting resources using additional investigations. This paradigm shift seems to be increasingly adopted in sports medicine, where the clinical diagnosis is now considered the cornerstone in the diagnosis making of many sports injuries.27 28
We consider MTSS a clinical diagnosis with mixed evidence for its pathogenesis. Therefore, making the diagnosis MTSS clinically seems the most logical approach. Our findings suggest that diagnosing MTSS clinically can be achieved reliably.
Strengths and limitations
A strength of this study is that our methods allowed for an unbiased estimate of effect. We blinded our clinicians to each others' diagnoses and randomised the assessment order to control for a possible ’clinical experience' effect, which could have been present due to the great variation of experience in our sample of clinicians. We did not specifically train the clinicians to make the diagnosis of MTSS. This allows for a true estimation of the clinical diagnosis' reliability in daily practice. The Kappa’s found in our study are likely an underestimation of the true Kappa-value, for two reasons: (1) we used eight clinicians to form a pair of clinicians, this may have added variation in perception among clinicians of what MTSS really is; (2) the Kappa-statistic is usually an optimal presentation of agreement when the prevalence is around 50%. For the MTSS diagnosis, the PI showed that Kappa was deflated due to a high prevalence, whereas for the presence of concurrent lower leg injuries, Kappa was also deflated but in this case due to a low prevalence. A further strength of this study is the generalisability of our findings to multiple professions and years of clinical experience. MTSS is a clinical diagnosis, and as such, sports physicians and sports physiotherapists seem able to reliably diagnose the condition, irrespective of their years of clinical experience.
This study also has some limitations. Firstly, some of the participants also participated in two other studies.29 This may have led to an increased risk of a type 1 error, due to multiple testing. However, considering the very high Kappa and subsequent p value (p<0.000001) found, we are confident that making the diagnosis clinically is truly reliable.
We did not reach the a priori calculated sample size (n=51). However, we found a Kappa-value much higher than we estimated when planning the study. Therefore, we are confident that this sample size enabled for a robust estimation of inter-rater reliability, which is confirmed by the 95% CI, k=0.74–1.00. We used an arbitrary cut-off value (5 cm) to differentiate between focal pain (suspected of having a tibial stress fracture) and diffuse pain (MTSS) along the posteromedial tibial border, for the purpose of this study. Although this criterion is based on previous literature,1 14 there is no evidence for this specific cut-off value. One might consider imaging to rule out a tibial stress fracture when an athlete presents with <5 cm of pain in clinical practice. It is of note that no athlete was clinically suspected of having a tibial stress fracture, one of the more important differential diagnoses when assessing overuse injuries along the medial aspects of the tibia. However, tibial stress fractures are extremely rare in The Netherlands, even in the Dutch Royal Army.30 We acknowledge that in other geographical areas (eg, Australia,31 Great Britain,32 Israel33 and the USA34) the prevalence of tibial stress fractures seems much higher, and, possibly this may affect the ease to distinguish between MTSS and tibial stress fracture. Future studies should investigate the reliability of the clinical diagnosis MTSS in other geographical areas and in military populations.
Conclusion
The clinical diagnosis MTSS can be made reliably using history and physical examination. Concurrent lower leg injuries were often present (32%) in athletes with MTSS and the presence of concurrent injuries could also be identified reliably. Our study supports the use of standardised history and clinical examination for diagnosing MTSS in clinical practice and research settings.
What are the findings?
Medial tibial stress syndrome (MTSS) can be reliably diagnosed clinically, based on history and physical examination.
Co-existing injuries are common in athletes with MTSS.
Clinicians can identify co-existing injuries in athletes with MTSS reliably.
How might it impact on clinical practice in the future?
MTSS should be diagnosed clinically.
Clinicians should be aware that about 1/3 of the athletes with MTSS have co-existing lower leg injuries.
Acknowledgments
We would like to thank the following clinicians for making diagnoses in this study: Sarah Kager, PT; Melissa Mes, PT; Floor Groot, MD; Feikje Rietstra, MD and Jan-Willem Dijkstra, MD. We thank Pim Noordam, PT and Rick de Regt, PT, for their contribution to the study.
References
Footnotes
Competing interests MW received a small grant from The Dutch National Olympic Committee during the course of this study for the performance of a prospective cohort study in athletes at risk for medial tibial stress syndrome (MTSS), investigating the relation between local tibial bone changes and MTSS.
Provenance and peer review Not commissioned; externally peer reviewed.