New evidence about the subcutaneous Achilles bursa - should we care?

Peter Malliaras30th of July 2016home / blog / tendinopathy-updates / new-evidence-about-the-subcutaneous-achilles-bursa-should-we-care

Dear all

Welcome to tendinopathy blog 30.

The blog has evolved to a fortnightly format. The main reason for this is that my balance of academic work has changed, aka I have a real job now! Leading the musculoskeletal and sports research at Monash Uni in Melbourne. So the blog lives on, not only to assist translation of tendinopathy research into practice (the initial aim), but also as a way to meet potential research collaborators and students – if that is you, please make yourself known now!

The blog (subscribe here) this week we focus on a study that sheds light on the potential role of the subcutaneous bursa in insertional Achilles tendinopathy, but the question has to be asked, what do these association studies actually add and do they change our management? The second study I picked because although some of the science was good I struggled with their arguments about the importance of pathology in tendon management. Always a topic that divides people and brings up lot’s of opinions and debate.

I hope you enjoy.


Shedding light on the subcutaneous Achilles bursa, but should we care?

What they did: Andersson et al. were interested in the subcutaneous Achilles bursa which is superficial to the Achilles tendon insertion – see figure below. They argue that the retrocalcaneal bursa and tendon is often considered as a pain source in insertional Achilles issues, but the subcutaneous bursa is not. They collected tissue samples of the subcutaneous and retrocalcaneal bursae, the posterosuperior prominence of the calcaneal tuberosity and the anterior part of the distal Achilles tendon among 10 people undergoing surgery for insertional Achilles tendinopathy.

 Screen Shot 2016-09-13 at 12.45.57 pm.png

Immunohistochemical staining was performed on tissue samples. Samples were given a semi-quantitative grade depending on the amount of PGP 9.5-positive structures, ranging from 0 to 3 (0 = very few or no PGP 9.5-positive structures, 1 = a few PGP 9.5-positive structures were seen in the section, 2 = PGP 9.5-positive structures were frequently occurring throughout the section, 3 = PGP 9.5-positive structures were seen in all fields of view throughout the section). PGP 9.5 is a general nerve marker and can visualize nerve fibres and nerve fascicles, but does not differentiate between different nerve types. The microscopists were blinded to the tissue being studied.

What they found: There were few nerve fibres detected in the tendon, and more in the peritendinous tissue, as you would expect. The grading of the biopsies resulted in a median value of 1 (IQR=0–1.5) for the retrocalcaneal bursa; 2 (IQR=1–2.3) for the subcutaneous bursa; 0 (IQR=0–1) for the tendon, and 1 (IQR=0–1) for the calcaneus – the differences were not statistically significant.

Clinical interpretation: Although not significant, the authors speculate that the innervation pattern of the subcutaneous bursa is clinically relevant. They go on to suggest it is a ‘neglected structure’. The obvious question is, are these nerves relevant at all to the pain presentation of an individual with insertional Achilles tendinopathy? That is the BIG issue with the many cross sectional association studies we have in the area of tendon pain – there is no way of knowing whether these factors are actually relevant. These studies have 'linked' biochemical with tendon pain, eg substance P, acetylcholine, etc, etc, the list is endless. Aside from only showing associations, often the science is poor, ie they compare apples and oranges, i.e. end stage pathological and painful people vs completely asymptomatic and normal tendon controls. Don’t get me wrong, I think exploring pain mechanisms is really really critical, but a focus on the tissue and biochemicals has, so far, in my opinion, not changed our clinical practice one bit, and we are talking about hundreds of studies in this area. Pain biochemicals research may lead to a huge break through and is important, but it is not yet clinically relevant. What is clinically relevant is understanding the multidimensionality and broad spectrum of tendon pain.

So will this study change our management? Will we be more likely to refer for surgery to have the subcutaneous bursa removed or recommend steroid injections for this bursa? Probably not because these interventions have huge negative implications and in my opinion can be avoided in a very vast majority of patients.

Then there are more specific issues with the design of these studies, eg the semi-quantitative assessment of the presence of positive staining, but the primary and more important issues relate to what we are actually getting out of them.


Clinical measures that predict patellar tendon pathology

What they did: Mendonca et al. assessed the diagnostic accuracy of the VISA and single leg decline squat (SLDS) tests, as well as patellar tendon pain history, age and years of sports participation in predicting the presence of patellar tendon pathology (PTP). They argued that knowing whether the person has PTP is important in their management, something I don’t agree with but we will pick up on this later.

They recruited 52 people from sports with a high prevalence of patellar tendinopathy. Data on ultrasound imaging, VISA-P questionnaire, pain during SLDS (rated yes or no) and information about patellar tendon pain history (i.e. history of pain localized to the patellar tendon that interfered with sport, rated as yes or no), years of sports participation and age were collected. PTA was defined as a hypoechoic region within the patellar tendon – see figure below which shows a normal patellar tendon on the top and hypoechoic region shown with arrows on the bottom image. The ultrasonographer was blinded to clinical data and vice versa. They tested reliability of SLDS and VISA (ICC = 0.99 for VISA and Kappa of 1 for SLDS). However, Kappa of ultrasound detection of a hypoechoic region was 0.75, highlighting even intra-rater repeatability of a seemingly simple measure on on US is far from perfect.

Screen Shot 2016-09-13 at 12.46.11 pm.png 

A receiver operating characteristic (ROC) curve was used to determine clinically relevant cut-off points for the VISA-P questionnaire, age and years of sports participation. They then used a regression model to predict (with the binary version of all variables) people with and without PTP. A contingency table and positive and negative likelihood ratios were calculated from the model.

What they found: Forty-three athletes were included (volleyball (n=26), basketball (n=14), soccer and running (n=3)). 27 tendons had PTA (31% of the 86 tendons in the cohort). 11 athletes had bilateral PTP, meaning that 16 athletes (37% of the cohort) had PTP on either 1 (n=5, 5 tendons affected) or both (n=11, 22 tendons affected) sides.

Cut offs for age were 25.5 years, sports participation 9.5 years, and VISA 88 points. The ROC derived cut off is based on maximizing the sensitivity and specificity, but it has to be said these were pretty terrible for age and years of sport participation, not as bad for VISA. The regression model only retained variables with a strong relationship to PTP, and these were VISA (cut off of 88), tendon pain history (yes/no) and SLDS pain (yes/no). Whether you use history of patellar tendon pain, pain on the SLDS or the combined factors in the regression model, there was a similar increase in likelihood (25%) and decrease in likelihood (15-25%, slightly better in this instance for the model).

Clinical interpretation: The authors suggest the regression model was more accurate than history of patellar tendon pain or pain on SLDS alone in reducing likelihood of PTP, and therefore recommend using VISA, SLDS and history of patellar tendon pain in combination with their suggested cut offs to identify PTP.

There are two main issues with this. First, the improvement in prediction of PTP was very small using the model than using any of the measures in isolation (aside from VISA). I’m not convinced it is worth the extra effort.

Second, and the key issue for me, is to question whether we actually need to be identifying PTP in the first place! The authors argue that confirming PTP guides treatment, i.e. eccentric exercise to ‘recover the tendon’s structural normality’.  They go on to suggest that if there is no PTP treatments directed ‘towards symptom reduction’ may be more appropriate. There are flaws in this thinking as the mechanisms underpinning success of exercise are unknow, but likely to be broad, and it could be argued that the least likely mechanism seems to be structural tendon changes. The presence of PTP should NOT guide our rehab.

This paper also brings up questions about how we diagnose patellar tendinopathy. Remember ‘tendinopathy’ is a very general term. If we are talking about pathology then it should be imaging. Imaging does have a place, eg differential diagnosis. If we are talking about pain, then diagnosis should be clinical. There are NO imaging tests that have a close correlation with tendon pain. What is debated is what clinical factors and definitions we should use to diagnose tendinopathy, mine are changing all the time as I gain greater appreciation for the spectrum of pain presentations that is tendinopathy.

See you next time


Peter Malliaras
Tendinopathy Rehabilitation