Neovascularisation is one feature of tendinopathy among others at various anatomic sites, such as the Achilles tendon, the patella tendon, tendinopathy of the wrist as well as in tennis elbow. Contemporary ultrasound studies using colour and/or power Doppler ultrasound identified neovascularisation among patients suffering Achilles tendinopathy 5 6 7 as well as in histological specimens from Achilles tendon ruptures 8 (table 1).

Neovascularisation was also reported in ultrasound of patellar tendinopathy with the vessels typically arising from the Hoffa fat pad 9 10 . The same phenomenon has been described for lateral elbow tendinopathy 11 , flexor carpi ulnaris tendinopathy of the wrist 12 , posterior tibial tendon insufficiency 13 , and in supraspinatus tendon overuse 14 determined by colour and/or power Doppler ultrasound techniques. Currently, there is reasonable published evidence that the neovessels are at least part of the pathophysiological process in tendinopathy of the Achilles tendon in its mid-portion area, at the patella tendon and in tendinopathies of the upper extremity such as in tennis elbow or in tendinopathies at the wrist level.

The diagnosis of tendinopathy of the main body of the Achilles tendon is made if patients have Achilles tendon pain at rest or at exercise in the main body of the Achilles tendon, 2–6 cm proximal to the insertion, associated with tenderness and swelling. In contrast, insertional tendinopathy of the Achilles tendon might involve various distinct clinical entities besides mere insertional tendon problems associated with neovascularisation. This distinct entity such as Haglund's exostosis or bursitis subachillae does not necessarily involve neovascularisation. Therefore, all insertional Achilles tendon problems reported in this review are tendon problems with neovascularisation arising from tiny vessels from the ventral aspect of the Achilles tendon in the Karger triangle with increased capillary blood flow.

The importance of structures close to the Achilles tendon and the "communication" in between and the role of the skin barrier, subcutis, as well as the paratenon is importance in this regard 15 . However, currently one has to be aware that the cells and biology which controls these extra and intra tendinous processes are only poorly understood. We do not even know what type of cells we find in the diseased tendons or how they work, and several up and down regulating factors, extrinsic and extrinsic factors may be involved.

I use the term 'neovascularisation' as a descriptive term for the appearance of abnormal vessels 16 and 'angiogenesis' for the process by which this occurs. Angiogenesis is known to be controlled by several stimulatory and inhibitory proteins 17 18 19 (table 2). Inhibition of angiogenesis is necessary for the development and maintenance of hypo- or avascular tissues. This might be caused either by production of an inhibitory factor or by a reduction of the angiogenesis factor.

Mechanoreceptors and nerve-related components such as glutamate NMDA receptors are present in association with blood vessels in tendinopathic tendons 27 28 . In tennis elbow, substance P and its receptor (neurokinin-1-receptor) could be detected using immunostaining as well as interleukin-1 alpha and TGF beta 1 positive cells in small vessels 29 . Recently, the Umea research group described the distribution of general (PGP 9.5) and sensory (substance P/CGRP) innervations in the human patellar tendon 30 . They proposed that there was nerve-mediated regulation of the blood vessels supplying the tendon, at the level where they course in the loose paratendinous connective tissue. The same authors also demonstrated an up-regulation of the cholinergic system as well as the presence of autocrin/paracrine effects in patellar tendinopathy 31 .

Recent publications suggest that the vascular in-growth in tendinopathy, in other words the neovascularisation, is accompanied with a nerval in-growth facilitating pain transmission in Achilles tendionpathy 32 and patella tendinopathy 33 . In other words we encounter a neuro-vascular inflammatory reaction in tendinopathy. Currently, based on the published reports, we cannot determine whether the vascularisation or the neurogenic component or both are the predominant factor in tendinopathy. One could speculate that with a resolution of the neovascularisation by a given treatment option such as eccentric training or sclerosing therapy, which I refer to later, the closely associated nerve endings will be disturbed or even destroyed due to a lack of perfusion by their nutrient neovessels. However, currently this is mere speculation. Alfredson speculated that eccentric training might traction the area of neovessels and be responsible for the good clinical results 34 but this hypothesis remains untested based on the current published reports.

Conventional ultrasound for tendon assessment in tendinopathy reveals hypoechogenic texture within an enlarged tendon especially in the anterior-posterior diameter. Power Doppler technology is capable in identifying neovascularisation in tendinopathic tendons because it allows visualisation of low-flow vessels by far more accurate than conventional colour Doppler ultrasound. In the Achilles tendon, these neovessels typically arise from the ventral and paratendinous portion leading into the Achilles tendon body. In patella tendinopathy these neovessels often arise inferior to the patella from the Hoffa fat pad entering the patella tendon in a 60°–90° angle. Magnetic resonance tomography determines tendon signal changes as well as paratendinous fluid with the signal intensity being the important factor in current tendinopathy MRI. Intratendinous pattern changes may be also depicted using MRI. Furthermore, volume calculations can be done using MRI, as demonstrated for the Achilles tendon by Shalabi 35 .

Among 33 patients with chronic Achilles tendinosis (mean age 52 yrs) they found that a computerized 3-D seed growing technique demonstrates an overall excellent reliability to monitor and evaluate the volume of the Achilles tendon and the mean intratendinous signal. Furthermore, the same authors reported that both, eccentric and concentric loading of the Achilles tendon resulted in an immediately increased tendon volume and intratendinous signal in 22 patients with chronic Achilles tendinopathy 36 . The eccentric training regimen was performed with 3 sets of 15 repetitions of heavy-loaded eccentric training with an immediate MRI following this exercise within 30 minutes showing the above mentioned changes. However, one has to mention that acute tendon effects even within 30 minutes might not illustrate acute changes immediately after the exercise. Long term eccentric training decreases the Achilles tendon volume by 14% and the signal intensity in T1-weighted MRI scans from 6.6 ± 3.1 cm³ to 5.8 ± 2.3 cm³ (p < 0.05).

However, neither conventional ultrasound nor MRI is currently used for microcirculatory monitoring. Power Doppler is of qualitative use with visualisation of the course of the neovascularisation, but no quantitative data are derived by Power Doppler only in its current routine application.

Real-time microcirculation assessment is possible using a combined non-invasive Laser-Doppler and spectrophotometry system, the Oxygen-to-see System (LEA Medizintechnik, Giessen, Germany, figures 1, 2). Three distinct parameters of microcirculation can be determined using the Oxygen-to-see system 37 38 (table 3):

• Capillary flow

• Tissue oxygen saturation

• Postcapillary venous filling pressure

However, one has to consider that these three microcirculatory do not necessarily display the complete microcirculatory environment, since vascular factors such as clotting, adhesion, thrombus formation and several others are not addressed by the aforementioned mere non-invasive technique.

Laser-Doppler flowmetry has been introduced for determination of capillary flow in various disease states. Stern has applied the Doppler effect to study in the microcirculation as early as 1975 in his Nature paper 39 . Validation work has been done extensively in the following, stating that the Laser Doppler method is a promising tool for rapid monitoring of dynamic changes in tissue perfusion" 40 .

Piloting the laser Doppler application to the tendon scientific area, Astrom and Svensson from the Malmo General hospital in Sweden studied the Laser Doppler flowmetry to the Achilles tendon surface of ten mature albino rats 41 . Clamping the femoral artery resulted in a 60% reduction of tendon blood flow and consecutive hyperaemia following clamp release in reperfusion. In circulatory arrest, no tendon flow was determined in this pilot study of Laser Doppler application in the tendon area.

Three years later, in 1994, Astrom and Westlin 42 reported about their initial experience at rest, during vascular occlusion, and during passive stretch and isometric contraction of the triceps surae among 40 healthy volunteers. They used an invasive needle probe, which was placed 5 mm above the distal insertion of the Achilles tendon, at the midportion and the musculotendineus junction of both legs, one of the first studies ever to differentiate between insertional and mid-portion locations. Astrom reported a significantly lower tendon blood flow at the insertion, but otherwise even vascular distribution. Vascular occlusion reduced all Achilles tendon blood flow values. Interestingly, passive stretch and isometric contraction induced a progressive decline in capillary tendon blood flow determined by laser Doppler flowmetry in this initial study. Hyperaemia was often noted by Astrom following contraction with higher tendon blood flow among women and a decreasing blood flow with increasing age. However, one has to bear in mind that only healthy volunteers participated in this study.

The reduced capillary tendon blood flow with increasing age is of special interest, since a decreased capillary tendon blood flow with increasing age might imply a consecutive malperfusion with age, thus leading to tendinopathy and finally to tendon rupture. On the other hand, as will be demonstrated in the following, we found that in symptomatic tendinopathy neovascularisation is associated with a significantly increased capillary blood flow in the Achilles tendon at the point of pain 43 (figure 3). Tendon repair, such as minimal-invasive percutaneous in complete Achilles tendon rupture, changes Achilles microcirculation in a time-dependent manner (figure 4).

The distribution of tendon capillary blood flow was performed in an Achilles tendon mapping technique, evaluating four tendinous and eight corresponding paratendinous location throughout the Achilles tendon. In contrast to histological data from staining suggesting that within the mid-portion part of the Achilles tendon the perfusion is limited 44 45 , which favours Achilles tendon ruptures in the mid-portion area due to its relative malperfusion. This was stated by the anatomical studies by Lang and supported by a plastination study from Heidelberg, Germany 46 . They studied the vascular anatomy of eight human specimens suing a plastination perfusion method through the femoral artery. They identified the well-vascularized paratendon with a large number of intra- and extratendinous anastomosis. Microcirculatory effects on the capillary flow have been described in cardiac surgery, where retrosternal capillary flow is reduced by 50% following harvesting of the internal thoracic artery for coronary revascularisation 47 . Recently we could demonstrate the successful combined sclerosing therapy with polidocanol followed by a 12-week eccentric training in a tennis player suffering tremendous pain due to flexor carpi ulnaris tendinopathy 12 .

Considering tendon oxygen saturation and postcapillary venous filling pressures besides capillary blood flow one has to acknowledge that each microcirculatory parameter is independent from each other. However, pathophysiological relations are evident such as in the case of ischemia, where a decrease of capillary blood flow due to arterial vessel obstruction is followed by a decrease of tissue oxygenation 48 . A venous stasis such as a venous thrombosis increases postcapillary venous filling pressures with consecutive decrease of oxygen saturation due to venous congestion and subsequent decreased capillary inflow in the further course.

Tissue oxygen saturation is attributed as the local oxygen content of the focussed tissue. Tissue oxygen saturation has been determined using different probes and techniques. In 1987, Stone and coworkers from the Brigham and Women's Hospital in Boston, MA determined the tendon and ligament oxygenation using invasive polarographic oxygen sensors 49 . First, they started with tendon oxygen saturation determination in the Achilles tendon of the sheep, where they found a 100% (13 of 13 events) response rate to changes of blood flow with consecutive oxygen saturation reduction. Second, they moved to the human anterior cruciate ligament, where they placed the invasive probe in five human knees during routing total knee replacement. In 83% of the cases (5 of 6 events) the oxygen saturation response was appropriate to the tourniquet placement.

Non-invasive tissue oxygen saturation is determined by spectrophotometry in the Oxygen-to-see System. Ischemia decreases oxygen saturation dramatically (figure 5). Repetitive ischemia and reperfusion, which is called preconditioning is capable in increasing tissue oxygenation (figure 6). Tendon oxygenation therefore is a marker for local oxygen content. A decrease of tendon oxygenation is potential harmful, since this is a sign for local acidosis. High intratendinous lactate levels have been reported in painful chronic Achilles tendinopathy by Alfredson¹⁷. Normal prostaglandin E2 levels have been identified by in vivo microdialysis as well] questioning the tons of non-steroidal anti-inflammatory drugs (NSAIDs) prescribed for this condition 50 . As aforementioned one has to acknowledge that tendon oxygen saturation may change independent of capillary blood flow as vice versa. Recently, an increase of Achilles tendon saturation has been reported after repetitive contractions among twelve men 51 .

Venous congestion causes venous stasis, which is part of inflammation. Capillary venous stasis deteriorates local capillary clearance of local metabolic end products. On the other hand, decreased postcapillary venous filling pressures are beneficial, since local clearance is facilitated. In disease states, increased postcapillary venous filling pressures have been encountered in the retrosternal region following removal of the internal thoracic artery and vein for coronary revascularisation 48 . Decreased postcapillary venous filling pressures of the mid-portion Achilles tendon are encountered using simultaneous cryotherapy and compression over 10 minutes 52 . Achilles tendon postcapillary venous filling pressures were significantly reduced following a 12 week eccentric training at the Achilles tendon insertion (51 ± 16 vs.41 ± 19, p = 0.001) and the distal mid-portion (36 ± 13vs.32 ± 12, p = 0.037) at 2 mm and at the insertion of the Achilles tendon at 8 mm (63 ± 19vs.51 ± 13, p = 0.0001) 53 .