Many recent publications regarding tendinopathy have demonstrated an increase in matrix metalloproteinases (MMPs) in tendinopathic tissue [1, 17, 21, 26, 29], particularly the collagenases (MMP-1, MMP-8, and MMP-13) and gelatinases (MMP-2 and MMP-9). We reasoned if an excess of collagenases (enzymes which break down collagen) represents an important part of the pathologic process of tendinopathy it may be a reason for delayed recovery in certain patients [31]. With this presumption, the previously reported [9, 10] local injection of a collagenase inhibitor seemed a sensible option for treating chronic tendinopathy. Aprotinin (pronounced a-PRO-tin-in) is a broad spectrum serine protease inhibitor, with particular inhibition of plasmin (along with trypsin and kallikrein) [19]. It is a strongly basic polypeptide with a half life of approximately 7 hours [19]. In vitro it is a strong inhibitor of MMPs, including the collagenases, with a likely mechanism of inhibition of the plasmin-activation pathway of the MMPs [5, 12, 14, 16, 28].

Aprotinin was first manufactured in the 1960s by Bayer (Leverkusen, Germany) and has primarily been used in medicine for preventing blood loss during major surgery [6, 13] and promoting soft tissue healing after surgery (as a component of “fibrin glue”) [27]. The doses used in anesthetics (for limiting blood loss) are high, up to 980 mg (7 million kallikrein inhibitor units [KIU]), with well-described actual and potential side effects [13, 32]. The major side effect is anaphylaxis, which is particularly seen after repeated use of the drug [4, 18, 36].

Aprotinin has also been used for over 35 years as an off-label injection for treatment of chronic tendinopathy. This was first described by Genety and Pernin [22] and Geudj in the early 1970s in France (where it is still used [3, 38]) and more recently has been popularized by Maffulli and colleagues [9, 10, 30, 31, 39]. The doses used to inject tendons are far smaller than those used in anesthesia, typically 4.2 to 8.5 mg (30,000–62,500 KIU) (which compares to the “test dose” in anesthesia) [36]. Anaphylaxis is definitely a potential side effect when aprotinin is used to treat tendinopathy [36, 39]. However, it is not known whether any of the other reported side effects are likely or possible in the far smaller doses used for tendinopathy management, with most side effects reportedly dose-related [32]. Although the side effect profile when using aprotinin in small doses to treat tendinopathy is almost certainly less pronounced, whether aprotinin actually acts as a collagenase-inhibitor in these low doses in vivo has not been studied.

Because aprotinin is still a relatively novel treatment for tendinopathy, our study objective was to describe results from a large clinical case series. We hypothesized aprotinin injection treatments for the common forms of tendinopathy would lead to good clinical improvement.

This retrospective cohort study is an extension of a previous investigation [35, 36] followed up with identical methods using a standard questionnaire. The dose used in the published studies and the current one was 3 mL of Trasylol (Bayer, Leverkusen, Germany) containing 30,000 KIU (4.2 mg) aprotinin, combined with 2 mL of the local anesthetic lignocaine 2%. This was the same volume of fluid injected (5 mL), but a lower dose of active agent, by Capasso et al. [10] (62,500 KIU), because of the more diluted concentration of the available aprotinin brand. The injections were made without ultrasound guidance in the clinic and we attempted to use a peri- rather than intratendinous technique.

The first published study followed up 155 cases of tendinopathy for an average of 9 months (minimum of 3 months) with an improvement rate of 69%, the majority of whom were given multiple aprotinin injections over a few weeks [35]. This study also reported a 6% case rate of systemic allergic reaction [35]. As a result, we recommended to subsequent patients that if multiple injections were to be used, there should be a delay between them to minimize the risk of allergy. A reduced rate of allergy was observed in 119 cases which were followed up for an average of 10 months (minimum of 3 months) [36].

Our results are primarily descriptive. Where comparisons between groups were required, Pearson chi-squared tests were used to determine differences between groups.

Two patients experienced a rupture of the Achilles tendon. One partial rupture occurred many months after the use of aprotinin in an elite hurdler, who had subsequently sought treatment elsewhere with a cortisone injection for a recurrence of the condition. The second patient was a high-level rugby league player, who underwent a course of three aprotinin injections over 3 weeks for long-standing Achilles tendinopathy, and suffered a complete rupture of the Achilles tendon in a game 5 days after the third injection. He also had a past history of panhypopituitism that was being treated at the time with anabolic steroid supplementation. He is included in this cohort twice as two distinct cases occurring on different sides as he subsequently sought aprotinin treatment for the contra-lateral side.

In one trial aprotinin injections appeared superior to both corticosteroid and saline injections in patellar tendinopathy, but the results reported for similar treatment in Achilles tendinopathy have been mixed. We therefore retrospectively reviewed a large clinical case series and hypothesized aprotinin injection treatments for the common forms of tendinopathy would lead to good clinical improvement.

The major limitation of this study is a lack of control group, resulting in a low level of evidence (Level 4). The followup is relatively short with a minimum of 3 months and an average of 12 months. Given patients with chronic tendinopathy can have recurrence after many months, we would not able to ascertain the rate of recurrence. Our questionnaire for assessing patient impressions has not been validated against some standard and accepted instrument measuring patient status or function, but we presume the responses are representative.

The best quality trials to assess efficacy are intervention trials, particularly randomized control double-blind trials, as this form of study design minimizes bias and confounding. Two such studies have been performed for aprotinin in tendinopathy. The study of Capasso et al. [10] involved athletes being injected every other week with two to four injections for patellar tendinopathy, and reported superior results for aprotinin at 12 month followup (72% good or excellent) compared to both cortisone (59%) and saline injections (28%). On the other hand, a study by Brown et al. [7] using aprotinin injections for Achilles tendinopathy reported no improvement over saline and local anesthetic injection. However the power of the study was low, with the aprotinin group achieving generally greater improvement on raw values [7]. Possible explanations of the findings in the Brown et al. [7] study are that: (1) aprotinin injections have no beneficial effect in the treatment of Achilles tendinopathy; (2) saline injections and aprotinin injections both have a beneficial prolotherapeutic effect in Achilles tendinopathy; or (3) aprotinin injections are slightly more efficacious than saline for Achilles tendinopathy but the study was underpowered to determine differences.

Our study included over 300 patients with a minimum followup of 3 months demonstrating a rate of patient improvement of 75%, with 64% of patients considering aprotinin injections helpful and none believing aprotinin injections were harmful for their tendinopathy. However, it is uncertain to what extent these results can be attributed to the drug itself. In such an uncontrolled clinical study, both the placebo effect and natural improvement of the condition almost certainly contribute to the group of beneficial results. In addition, there is a school of thought that any injection with an irritant agent can be beneficial in treating the condition [37].

The rate of allergic reaction when using repeat injections of aprotinin (bovine-derived) is higher than for most medications and this represents a major factor to consider when choosing this drug [36]. If aprotinin works simply as a form of prolotherapy, it would be a better choice to use dextrose or autologous blood for treatment of tendinopathy. However, if aprotinin works specifically as a collagenase inhibitor, then it may have advantages over more inert substances. This is an important question with mixed results demonstrated to date in the RCTs [7, 10].

We were surprised to find superior clinical results for treating mid-substance Achilles tendinopathy compared to patellar tendinopathy, given the superior results for aprotinin in the RCT for patellar tendinopathy than that involving Achilles tendinopathy. In clinical practice, midsubstance Achilles tendinopathy generally has a more benign prognosis than patellar tendinopathy. However, compared to patellar tendinopathy, we found comparable results for insertional Achilles tendinopathy and proximal hamstring tendinopathy, both of which have been resistant to conventional treatments [8]. It is generally easier to inject the Achilles tendon as a tendon sheath is usually present (which it is generally not for the patella tendon). It is not known in vivo whether aprotinin remains at the site of injection long enough to act as a collagenase inhibitor, but this may be more likely if the drug is injected into a tendon sheath. It is also likely the current Level 1–2 evidence for aprotinin in tendinopathy is incomplete and therefore it should not be concluded, simply based on Level 1–2 evidence, that aprotinin injections are effective in patellar tendinopathy but not in Achilles tendinopathy.

A disadvantage of using polidocanol sclerotherapy, inert agents, or dry needling to treat tendinopathy is these therapies probably or certainly require tendon penetration in order to get a beneficial effect, which theoretically increases the risk of tendon damage. In the belief that aprotinin is a therapeutic agent itself, it can be injected around the tendon (in the same fashion as cortisone) so the risk of iatrogenic tendon damage is reduced. This large case series supports the notion that the risk of tendon damage with aprotinin is very low, despite one Achilles tendon rupture that occurred soon after the use of aprotinin.

Because aprotinin is derived from bovine lungs, a further proposed complication of aprotinin treatment is the potential to contract bovine spongiform encephalopathy (BSE) [11]. Since this potential complication was raised, Bayer (Germany), the major manufacturer of bovine aprotinin, has outlined the precautions taken to ensure aprotinin does not contain viral prions [24]. These steps include verifying the product is only sourced from countries with no BSE, the tissue used is in the lung (rather than neural tissue), and a purification process is undertaken. There have been no suspected cases of BSE transmission in over 40 years of aprotinin use [4].

A further disadvantage of aprotinin is its image as an off-label treatment for tendinopathy, with the manufacturer marketing the drug for high-dose use in surgery. The primary indication is itself under threat with recent vigorous debate about whether the known benefits of aprotinin in terms of preventing blood loss during surgery outweigh the possible risks [25, 33]. Because of recent controversies, some countries have restricted the availability of aprotinin as a treatment. In Australia, for example, aprotinin is only available in 50- or 100-mL vials, which are not recommended for multidose use due to potential risk of contamination [20]. Although smaller dose presentations are available in other countries, these are generally more diluted and hence also less suitable for use as local injections. Either aprotinin vials should be used, as recommended, for single-use (which is very expensive), or great care must be taken to avoid contamination (ie, only fresh needles should ever penetrate the seal). A similar dilemma applies to the use of botulinum toxin in clinical practice, and many practitioners (and patients) choose to multidose from a single vial in order to keep the treatment costs down.

There have been recent attempts to manufacture aprotinin-like polypeptides in a recombinant fashion that could potentially give similar clinical effects yet not lead to nearly the same degree of allergic reactions [2]. If recombinant aprotinin is successfully introduced to the market at a competitive price (and hence allergy becomes a far less likely complication), aprotinin may become a first-line treatment for tendinopathy. On the other hand, if surgical complications related to its primary use cause aprotinin to be withdrawn from the worldwide market, other collagenase inhibitors may need to be tested in injection form to ascertain whether they may be safe and effective treatments for tendinopathy.

We believe patients must be warned of the risk of allergy from aprotinin injections for tendinopathy and be prepared to remain under medical surveillance (where anaphylaxis can be managed) for 30 to 60 minutes after injection. Because of this risk, aprotinin should be used as second-line therapy only, for chronic conditions where more basic measures (eccentric exercise, topical glyceryl trinitrate) have failed. In vitro, aprotinin acts as a collagenase inhibitor (via inhibition of the plasmin-activation pathway of matrix metalloproteases) and therefore it may theoretically assist in managing chronic tendinopathy when collagenase excess has been consistently demonstrated. For major load-bearing tendons (eg, Achilles, patella, hamstring tendons) in active individuals, aprotinin is a more appropriate second-line injection option than cortisone preparations.