Physical activity increases synovial fluid in ankle tendon sheaths: an adjustment of MR Criteria is needed

To compare the amount of fluid in synovial sheaths of the ankle before and after running. Our hypothesis was that this amount would increase and that the threshold for what is normally acceptable should be adjusted after physical activity. Twenty-one healthy volunteers (n = 42 ankles) ran for 40 min on a treadmill. They underwent 3 T MRI before and immediately after running using a dedicated ankle coil. The images were stored and subsequently measured in a standardized way and independently read by two readers for fluid in the tendon sheaths in the retro and inframalleolar area. Statistics were performed for each tendon (Wilcoxon signed rank test), and also for the pooled data. Intraclass correlation coefficients were calculated. For reader 1, for all tendons the values after running increased without reaching statistical significance. For reader 2 this was not the case for all tendons but for most. When all the data were pooled (n = 800 measurements), the statistical difference before and after running was significant (p < 0.001). Data pre and post-running show a trend of increasing synovial fluid, however, not significant for each individual tendon. The pooled data for all tendons, (n = 800) show a statistically significant increase after running (p < 0.001). The clinical implication is that the threshold for normally acceptable fluid should be adjusted if the patient undergoes an MR study after recent physical activity.

One question that arises is what are the limits of normal tendon sheath fluid in the ankle. Too much fluid may indicate mechanical overuse, infection, or a rheumatological condition. A few studies have addressed this topic of normal synovial fluid especially in the ankle [3,7,[10][11][12][13]15].
In this article, we focus on the ankle tendons. Previous reports have given indications of the amount of fluid that can be expected in a normal person without any known underlying condition. Fluid was quite common and rather important around the medial tendons. It was minimal and usually absent around the peroneal tendons. In the normal situation, fluid was never observed in the anterior tendon 1 3 group. Reference values are available in several publications [2,7,[12][13][14][15].
Our research question was whether increased athletic would increase tendon sheath fluid. Our hypothesis was that it indeed would increase tendon sheath fluid, and hence that the normal values could change [1,5,6,8,9]. We would try to demonstrate this with MR imaging.
To study this type of research question is rather complicated and an experimental context was designed. We decided to perform MR of both ankles on 21 volunteers before and after running for 40 min on a treadmill.

Materials and methods
Our study was approved by the IRB of our hospital. Written informed consent was obtained from the volunteers. They also filled out a questionnaire to make sure they had no underlying rheumatologic condition, ankle pain or trauma, or previous severe injury or surgery to the ankle. Conditions that would make the experiment unsafe were also assessed. The investigators were available to answer any questions and monitor the experiment.
The volunteers were also weighed, and their length was obtained, allowing subsequent calculation of BMIs. MRI in the three orthogonal planes was obtained with a PD FS sequence on a Philips 3 T Achieva system (Philips, Best, The Netherlands). Imaging parameters were as follows: TR 2969 ms; TE 30 ms; NA 2; slice thickness 2.5 mm, fat saturation. A dedicated ankle coil was used allowing maximum reproducibility of patient positioning before and after running.
After pre-exercise imaging the patient started running on a treadmill at intensity setting nr 3 (Domyos T520, Decathlon, Villeneuve-d'Ascq, France). The investigators continuously monitored the volunteers for injury, pain, or feeling unwell. After running, the same imaging sequences where repeated immediately with exactly the same imaging parameters. All images were stored on PACS (Agfa, Mortsel, Belgium). None of the investigators made any immediate assessments or interpretations of the obtained images. Positioning consistency and image quality were ascertained by a certified RT (three different RTs performed the studies; 15, 20, 25 years of work experience).
For the interpretation session, the transverse images were used, however, the other planes were used to ascertain that the measurements were performed in the most similar position possible for the pre and post-images. The measurements that were obtained included TP, FD, FH, PL, and PB. For each tendon, a retro-as well as inframalleolar measurement was obtained. The exact measurement positions are further detailed in Table 1. The largest dimension of fluid was used, in exactly the same position pre and post running. The images were interpreted by two MSK radiologists (8 and 20 years of experience) independently (Figs. 1, 2, 3, 4, 5).
Separate measurements were obtained for each tendon, but also all data were pooled.
The pre and post-image set was analyzed simultaneously to make sure each measurement was performed in exactly the same location for both image sets. The readers were blinded to which image was pre and which was posted. The images were thus presented to the viewers in a randomized fashion.
The data were compared by a Wilcoxon signed rank test. Intraclass correlation coefficients were calculated between both readers. BMI was calculated for each volunteer.

Results
All participants performed the experiment successfully without pain, injury, or becoming unwell.
Planning of imaging was organized as such that each participant immediately underwent MR when his running session ended.
All the measurements obtained for all tendons pre and post are shown in Table 2 and 3. This Table also shows the 95% confidence interval, the delta (and confidence interval), and the p-value for this specific tendon. Statistical significance was calculated for each tendon separately. Since the number of cases per tendon remained small, we also calculated statistical significance for all measurements pooled. This calculation thus related to 800 cases. The first table relates to reader 1, and the second table to reader 2 (Figs. 1, 2, 3, 4, 5).
All the measurements pre and post were largest for both readers for the TP and FH, followed by the FD, and peroneal tendons but in their inframalleolar location.
For all tendons pooled the delta was small and ranged from 0.00 to 0.14 mm. None of the observed differences for both readers per tendon was statistically significant. For reader 1 all measurements post running showed a trend to be higher than pre-running. For reader 2 most values after running showed a trend to be higher, but not all. There was a trend for the post measurement to be larger, although not reaching statistical significance per tendon. When the calculation was performed on the pooled data (n = 800 cases) the result was significant, however (p < 0.001), with the measurement post running being higher.
The mean BMI of the volunteers was 23, 2 (N). The BMI range was 19-32. Three individuals were in the obese range, and 3 in the overweight range, all the others having a BMI in the normal range.

Discussion
Several synovial sheaths, in the normal situation contain some synovial fluid. Generally, a sheath is present where the tendon(s) make a turn or are subject to increased mechanical forces. The presence of tendon sheath fluid therefore in itself is not necessarily abnormal. The normal amount of fluid in tendon sheaths has been the subject of several investigations, some using MR, some US. In our opinion, the ones using US may have been flawed as it may be quite difficult to differentiate normal hypoechoic tendon retinacula from hypoechoic tendon fluid. This is less of a problem with MRI, where fluid is hyperintense and retinacula and sheaths exhibit hypointense signal on all pulse sequences [6].
Our measurements pre-running is in accordance with those of previous investigations [3,7,10,12,13,15], pointing to the fact that fluid is particularly common in the medial tendons in a normal situation, and to a lesser degree in peroneal tendons. In the anterior tendon, group fluid is never present in a normal situation.
Our present investigation has a physiological implication, but more so a clinically relevant implication. If physical   Although it may be argued that for each tendon individually our pre and post-imaging values were not statistically significant, significance was obtained when the data were pooled (to 800 cases). Our number of cases per tendon is not sufficient to show a significant difference per tendon. However, devising an experiment such as this is quite time consuming, and it would not be realistic to image e.g., 150 volunteers in this way. Probably an even better experiment 1 3 would be to image this number of marathon runners before and after their contest, but this also is not realistic.
Our study has several limitations. Our number of volunteers was limited. Nevertheless, there was a consistent trend for the measurements to increase after running and when the data were pooled, they became statistically significant.
Second, our setup was experimental using a treadmill. Running outside on uneven terrain or for longer distances would likely be more significant. It could be argued that the experiment simulates a patient hurrying from an away parking lot to the MR suite.
We did not analyze the anterior tendons, as based on previous studies, it was shown they never show fluid in the normal situation. Hence fluid at this level would always be abnormal [15].
We did not analyze changes in joint fluid before and after running, as these can potentially change too. This has, however, been the subject of previously published studies [4].
MR positioning before and after running may have changed slightly, although every effort was made to avoid this by meticulous positioning by experienced RN's.
The distribution of fluid in the sheath is a three-dimensional phenomenon and ideally the total volume in the entire sheath could have been measured. This would imply heavy computing and segmentation of images which was beyond our capabilities. It can be criticized that we obtained only one measurement in the transverse plane. However, we defined the measurement site precisely and also evaluated the paired images to measure exactly at the same location.
Our study population was also rather heterogeneous, some individuals having a BMI out of the normal range. We believe; however, this in itself would not affect our measurements and is rather representative of a general population. We also omitted the anterior tendon group since it has been repeatedly shown that fluid never occurs in a normal situation. Finally, the observed differences were often small, and it could be argued that measurement bias played a role. Nevertheless, as mentioned before there was a quite consistent trend for an increase post running.
In conclusion, there was a statistically significant difference in synovial sheath fluid before and after running, when all obtained data were pooled (n = 800).
Therefore, we suggest that if a patient is imaged after physical activity (such as running, jogging, fitness), our threshold for considering synovial fluid as pathological is not fully supported by our observations. Second column, measurement region, third column reference planes for determining measurement region.
Author contributions MD: study design, manuscript writing, SD: data collection, VDG: data collection, NB: statistical analysis, JDM: study design, manuscript review, MS: manuscript writing and review, SP: study design, data collection, IW: study design, experiment, data analysis.

Funding No funding was received.
Availability of data and materials All the available measurement data are provided in the manuscript.

Declarations
Competing interests The authors declare no competing interests.

Conflict of interests No conflicts of interest or competing interests to declare by the authors.
Ethical approval Ethical committees Vrije Universiteit Brussel. Consent to participate and consent to publish were obtained from the subjects.