The impact of nutrition on tendon health and tendinopathy: a systematic review

ABSTRACT Background Tendinopathy is a painful condition that is prevalent in athletes as well as the general human population, and whose management is challenging. Objective This systematic review aimed to evaluate the impact of nutrition on the prevention and treatment of tendinopathy. Methods Searches were conducted in PubMed, EMBASE, Web of Science, and SPORTDiscus without restriction to year of publication. Studies examining the impact of exposure to nutrient intake in an adult human population on 1) prevalence/incidence of tendinopathy, 2) clinical outcomes of tendinopathy, 3) structural changes in the tendon by imaging modalities. Experimental and observational study designs written in English, Dutch, or German were eligible. Results Nineteen studies met the inclusion criteria. The effects of the habitual diet were investigated in one study. Four studies examined the effects of exposure to alcohol. Alcohol consumption can be a potential risk factor associated with Achilles tendinopathy and rotator cuff tears, although findings were inconsistent. The use of dietary supplements was examined in fourteen studies. Among these, collagen-derived peptides were most often part of the supplements evaluated. Combining training and dietary supplements seems to induce better clinical and functional outcomes in tendinopathy. Conclusion This review demonstrates the paucity of high-quality studies and a wide variety among studies regarding nutrients, tendon location, study population, and reported outcome measures. Individual studies showed promising clinical implications for the use of dietary supplements, particularly those containing collagen-derived peptides. However, giving any definitive dietary recommendations on the prevention and treatment of tendinopathy remains elusive.


Introduction
Tendinopathy, which involves persistent tendon pain and loss of function related to mechanical loading [1], is common in athletes as well as in the general population [2,3]. Mechanical overuse is seen as the key initial trigger in the multifactorial etiology of tendinopathy [3], hence it is a common cause of injury in sports that involve exposure to high forces and repetitive movements, such as running, volleyball, and tennis [3][4][5]. Tendinopathy is also prevalent in workers exposed to monotonous repetitive work tasks [2], and is associated with a number of medical conditions such as obesity and diabetes mellitus [6][7][8]. Other intrinsic risk factors are age, sex, and genetics [9]. The potential suffering from pain and loss of function may impact health, sports performance, and work ability [10]. Also, the impact of tendinopathy on quality of life is substantial, specifically on the domains mobility, pain/discomfort, and usual activities [11].
There is only limited evidence for the efficacy of preventive interventions for tendinopathy [12]. Numerous treatment options for tendinopathy have been described [9], but convincing evidence of success of many common therapies is lacking. Hence the management of tendinopathy remains a challenging and often time-consuming process [9]. This stresses the relevance of additional strategies for both prevention and treatment of tendinopathy.
The pathological tendon is characterized by an altered tissue homeostasis [13]. Given that diet plays a major role in the homeostasis of all tissues and poor nutrition is one of the extrinsic factors that contributes to the development of tendinopathy [9], nutritional interventions, e.g. intake of collagen, are a plausible, potential strategy to improve the prevention and healing of tendinopathy.
Adequate intake of nutrients -macronutrients as well as micronutrients -is of great importance, especially for populations with specific demands such as athletes. This primarily entails consumption of a healthy habitual diet, but also sport-specific nutritional strategies as well as dietary supplements may be used to optimize intake in specific situations. A dietary supplement is defined as a food, food component, nutrient, or nonfood compound that is purposefully ingested in addition to the habitually consumed diet with the aim of achieving a specific health and/or performance benefit [14]. The use of supplements is widespread among athletes as well as the general population [15]. However, specific recommendations of dietary strategies for preventing or treating tendon injuries, either as part of the habitual diet or with additional dietary supplements, are lacking [16].
The potential benefits of nutrition on tendon health have been described in only a few reviews [17][18][19]. A recent short review identified various nutrients, including amino acids, vitamins, and trace minerals, as being potentially useful in improving tendon growth and healing [17]. It has additionally been suggested that nutritional interventions involving multiple nutrients, e.g. collagen combined with vitamin C, may be more effective than single-nutrient strategies, as many nutrients are involved in tendon and collagen metabolism [17]. So far, only evidence for effects of vitamin and amino acid supplements on tendon tissue healing has systematically been evaluated, but hardly any clinical studies are included [20,21]. No previous study has systematically synthesized the evidence of nutritional exposure for clinical outcomes on physical, psychosocial and overall life impact, or the risk of tendinopathies in relation to nutrition. The aim of this systematic review was therefore to evaluate the impact of nutrition on the prevention and treatment of tendinopathy in a general human population. This will guide future studies on directions of research toward evidence-based nutritional recommendations to prevent and treat tendinopathy, which ultimately leads to a lower prevalence and better management of this bothersome condition.

Methods
This systematic review complied with the PRISMA guidelines [22]. The study protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) on 29 June 2020 (CRD42020189773).

Eligibility criteria
Studies were eligible if they investigated the impact of exposure to nutrient intake (either as part of the habitual diet or in the form of specific dietary supplements) in an adult (>18 years) human population, using at least one of the following outcome measures: prevalence/incidence of tendinopathy, a clinical outcome that captures one of the tendinopathy-related core domains as established by the ICON group [23] (see Additional file 1), structural changes in the tendon identified by imaging modalities such as magnetic resonance imaging (MRI) or ultrasound imaging (USI) [24]. Randomized and nonrandomized intervention studies and cohort, case-control, cross-sectional, and case studies written in English, Dutch, or German were included. Reviews, letters, and editorials were excluded. There was no restriction with regard to year of publication.

Search strategy
We searched the electronic databases PubMed, EMBASE, Web of Science and SPORTDiscus in June 2020 for eligible studies. The specific search strategies were created by a health science librarian with expertise in systematic review searching and had three components: Nutrition, Tendinopathy, and Human. The PubMed search strategy was adapted to the syntax of other databases and is presented in Additional file 2. In addition to the database search, reference lists of included studies as well as relevant reviews were manually checked to identify additional studies for inclusion.

Study selection
The records were imported into Endnote X9, where duplicates were removed. Two reviewers (AH, JZ) independently screened the titles and abstracts from the identified articles for eligibility, followed by full-text evaluation for final study inclusion. Any disagreements about inclusion/exclusion were discussed between the reviewers and a final decision was made by all authors.

Data extraction and analysis
Data was extracted from the included papers using a spreadsheet prepared in Microsoft Excel. We extracted publication details, study design, study aim, population characteristics, type of tendinopathy/site of interest, exposure and comparator details, relevant outcome measures and results, and conclusion with respect to nutritional exposure. The study characteristics and results are presented in tables and summarized semi-narratively.

Risk of bias assessment
Two reviewers (AH, CR) independently assessed risk of bias of the included studies. The revised Cochrane risk-of-bias tool (RoB 2) was used to assess the quality of randomized trials [25]. We assessed risk of bias on a per-protocol basis for all five domains: (1) randomization process, (2) deviations from intended interventions, (3) missing outcome data, (4) outcome measurement, and (5) selection of the reported result. Other intervention, cohort, case-control, and cross-sectional studies were assessed using the ROBINS-I tool [26]. Bias was assessed for the following domains: (1) confounding, (2) selection of participants into the study, (3) classification of interventions, (4) deviations from intended interventions, (5) missing data, (6) measurement of outcomes, and (7) selection of the reported result.
An overall level of certainty in the evidence for clinical outcomes and for the occurrence/prevalence of tendinopathy was rated using the GRADE approach for systematic reviews in which only a narrative summary of the effect across studies is available [27]. Figure 1 shows the study selection process. A total of 8618 records were retrieved from the database and manual searches. After duplicates were removed, 6538 records were screened for eligibility, of which 89 were assessed in full-text. Of these studies, nineteen met the inclusion criteria. Seventy articles were excluded because they did not meet the language (n = 3), study design (n = 37), study population (n = 5), exposure (n = 11) or outcome (n = 9) criteria, or were not available (n = 5). Details of the included studies are provided in Tables 1 and 2. Five studies investigated the effect of the habitual diet ( Table 1). The use of dietary supplements was examined by fourteen studies ( Table 2), one of which evaluated an intervention of supplement use combined with habitual dietary changes [28].

Study designs and populations
All studies examining habitual dietary intake had an observational design; there were two prospective cohort studies [29,30], two cross-sectional studies [31,32] and one casecontrol study [33]. The number of people included in these studies ranged from 50 to 80,106 (Table 1).
Although there was a wide age range within the study populations, they mainly comprised middle-aged (40-60 years) and older (>60 years) non-athletic adults. The participants in three studies examining athletes or active persons were younger [37,42,45]. Also, one study among military personnel had a large proportion of young adults [30]. Most studies included both men and women, but some included relatively more [29,37] or only men [31,42,45,46].
Three studies calculated odds ratios as a measure of association between exposure to alcohol and occurrence/prevalence of tendinopathy [30,32,33]. Structural changes of the tendon by MRI or USI were examined in seven studies [31,34,36,39,42,44,45].

Risk of bias
The risk of bias assessment of all nineteen eligible studies, including the most important sources of bias, is presented in Tables 3 and 4. For the nine RCTs this was evaluated with the RoB 2 tool (Table 3). Two studies were judged at low risk of bias [38,41]; both were double-blinded, placebo-controlled, and included several outcome measures for which the results were reported adequately. However, one of these studies [38] raised some small concerns because baseline characteristics were not presented. Six studies expressed 'some concerns' [28,[34][35][36][37]39]. Main   [44] Serious No control group High drop-out rate Limited participant information Baar [45] Serious One participant Farup et al. [42] Low Small study population Hjerrild et al. [31] Serious Cross-sectional design Incomprehensive exposure assessment Not all confounders were taken into account Jain, et al. [29] Moderate/Serious Small study population Participant-reported outcome Incomplete follow-up Owens et al. [30] Moderate Not all confounders were taken into account Only severe cases were identified Inappropriate statistical adjustments Passaretti et al. [33] Serious Potential of selection bias Potential for underreporting of alcohol consumption Rechardt et al. [32] Moderate Cross-sectional design Data of association not shown Schneider et al. [46] Critical Small study population Selection bias No quantification and qualification of exposure Single participant-reported outcome measure Vitali et al [43].
Moderate No placebo Mainly participant-reported outcomes *Possible judgments are: Low, Moderate, Serious, Critical aspects that raised concerns were no inclusion of placebo treatment [34,36], small study populations [36,37,39], unclear or imbalance of baseline comparison and/or other confounding factors [36,39], and mainly presenting participant-reported outcomes [35,37]. Additional concerns, involving judgment tending toward high risk of bias, were raised by the absence of intention-to-treat analyses, assessment of limited outcomes and short follow-up [37], and a cross-over design without a wash-out period [39]. One study had a high risk of bias due to poor clarity of the randomization process, participant characteristics, study protocol and analyses, and selective reporting of results [40]. The ROBINS-I tool was used to assess the risk of bias of the ten remaining studies ( Table 4). One study was judged at low risk of bias, although the study population was small [42]. Other studies presented limitations as a result of their observational study design [29][30][31][32][33], small study populations [29,45], incomplete follow-up [29,44] and inappropriate exposure or outcome assessment [29][30][31]33,43], and were therefore judged at moderate [29,30,32,43] or serious [31,33,44,45] risk of bias. The case series [46] was judged as critical because of selection bias in its small study population, insufficient information about exposure, and assessment of one single outcome measure.
The overall level of certainty in the evidence for clinical outcomes was based on fourteen studies that examined the effect of a dietary supplement (Additional file 5). For the occurrence/prevalence of tendinopathy, the overall assessment of certainty included three studies that examined the effect of alcohol intake (Additional file 6). The summary of findings regarding the judgments of the certainty in evidence is presented in Additional file 7. For the clinical outcomes, the certainty in the evidence was judged at low. There was a very low level of certainty for occurrence/prevalence of tendinopathy.

Discussion
This systematic review aimed to evaluate the potential impact of nutrition on the prevention and treatment of tendinopathy. The majority of the included studies investigated the use of dietary supplements. Only a limited number of studies examining the effect of the habitual diet on tendon outcomes were identified. Overall, there was insufficient high-quality data available to enable meta-analyses as a result of the considerable variation in study design, nutritional exposure, concurrent exposure, outcome measures, and risk of bias.

Habitual diet
Only one study was identified that investigated the habitual intake of several foods and type of diet [31]. This study did not find an association between any of the dietary parameters and Achilles tendon thickness, but no other clinical outcomes were assessed. Considering the serious risk of bias, firm conclusions about the impact of the habitual diet from this single study are not possible.

Alcohol
Findings of the four studies examining the impact of alcohol intake varied for effect on tendinopathy. No associations were found between alcohol consumption and chronic rotator cuff tendinitis [32]. Moderate weekly alcohol consumption (men: 7-13 drinks, women: 4-6 drinks) was associated with a modest increased risk for Achilles tendinopathy, but not with patellar tendinopathy [30]. Excessive alcohol intake (men: >13 drinks, women: >6 drinks) was a significant risk factor for the occurrence and severity of rotator cuff tears [33]. By contrast, Jain et al. [29] found a positive association between alcohol consumption and less shoulder pain and better function after rotator cuff repair. However, it was suggested that alcohol use may be a proxy for another variable that was not captured in the study. In addition, differences in study design, population, and classification into categories for amount of alcohol consumption (light, moderate, heavy) limit comparability across these studies. Despite inconsistent findings on the risk and severity of tendinopathy in humans, there is evidence that alcohol may inhibit collagen synthesis through toxic effects [47].

Collagen
As tendons are mainly composed of collagen, whose ongoing synthesis is required to maintain a healthy extracellular matrix, it is not surprising that collagen-derived peptides, including gelatin and hydrolyzed collagen, were most often a component of the dietary supplements evaluated. The majority of these supplements were found to improve clinical and/or structural outcomes in the treatment of tendinopathy [34,36,[38][39][40][43][44][45]. Merolla et al. [35] showed only short-term effectiveness for pain reduction after supraspinatus tendon repair, while long-term pain was unchanged. This could be addressed by increased dosage and treatment duration. Thus, in the study of Gumina et al. [34], an extended treatment of a similar preparation in rotator cuff patients showed improvement in pain after six months. However, both studies did not find a better physical function capacity after surgical repair as a result of the supplement intake. In all other studies examining the effect of a supplement containing collagen, participants were not treated surgically and benefits for various clinical outcomes were found [36,[38][39][40][43][44][45]. Although most studies showed improvement after 2-3 months of supplement use, the daily dose as well as the type of collagen varied among interventions. Tendon is mainly composed of type I, while cartilage contains type II. Most tendon studies supplemented type I, although some combined with type II [35,43], or type II only [40], which can explain some of the variation in outcome. All in all, collagen seems to be beneficial in the treatment of tendinopathy, but conclusions about optimal dosage, timing, duration, and type of collagen supplementation cannot be drawn yet. Also, because many interventions implemented collagen supplementation with concurrent treatment, the specific benefits of collagen alone remain unknown.

Other nutrients
Many other nutrients may contribute to collagen synthesis or may have anti-inflammatory effects, so most dietary supplements used in the different studies contained multiple substances. In four studies a similar preparation was used, as they all comprised methylsulfonyl-methane and arginine in addition to hydrolyzed collagen [34,35,38,43]. Curcumin and Boswellia serrata were only used by Merolla et al. [35] and Vitali et al. [43], while the preparations used by Gumina et al. [34] and Notarnicola et al. [38] contained bromelain. All these compounds were found to be associated with improvement in pain, but a specific evaluation of each nutrient is difficult because of the multiple nutrients within a single supplement.
Also, the amino acid leucine may have exerted a stimulatory effect on collagen synthesis. A single study in healthy young men found that tendon hypertrophy was augmented with high-leucine whey protein supplementation (providing 19.5 g amino acids of which 2.77 g leucine on training days) in addition to resistance exercise [42]. Although findings from this small single study are not conclusive, they may have important clinical implications. Leucine-induced tendon hypertrophy may lead to relatively less mechanical stress on the tendon during exercise, which may assist in tendon rehabilitation.
Two studies investigated the potential role of essential fatty acids in the treatment of tendinopathy because of their anti-inflammatory properties [48]. In the high-quality trial of Sandford et al. [41], eight weeks of Omega 3 PUFA supplementation (daily providing 1530 mg of EPA and 1035 mg of DHA) was found to have a modest improvement on disability and pain outcomes in patients with rotator cuff-related shoulder pain after 3 months, but not after one year. Mavrogenis et al. [37] used supplements for only 32 days but at a higher dose, and reported a significant improvement in pain. However, the short follow-up and other methodological concerns limit the potential impact of these findings. Current evidence does not support the use of essential fatty acids, but further research is warranted to assess the potential impact of higher-dose and longer-duration interventions.
The intervention evaluated by Szczurko et al. [28] consisted of multiple components, including supplementation of hydrolytic enzymes (bromelain, trypsin, rutin), individual dietary counseling, and acupuncture. This naturopathic treatment showed clinically significant improvement in shoulder pain and quality of life compared with standardized physical exercise. Although they raised only small concerns with respect to risk of bias, their study is of limited evidence for the effectiveness of dietary supplements and/or dietary changes on tendinopathy, because the effects of the individual components cannot be established.

Study quality and limitations
Despite the broad inclusion criteria, the total number of eligible studies was relatively low. In addition, the overall quality of studies was poor. Evidence for the effects of nutrition results from limited high-quality studies. An overall rating of the certainty in the evidence was only provided for clinical outcomes and for occurrence/ prevalence of tendinopathy.
Another limitation resulting from the inclusion criteria is the heterogeneity among studies. This is why the rating of quality of evidence needs to be interpreted with caution. Especially the evidence from studies on the effects of nutrition in healthy human tendons was scarce. We also included studies that did not primarily aim to investigate nutritional exposure or one of the eligible outcomes, and therefore did not assess or report this in much detail. Nevertheless, this review provides an overview of the evidence for effects of any nutritional exposure and is the first study to systematically synthesize the findings for clinical outcomes on physical, psychosocial and overall life impact, and the risk of tendinopathies in relation to nutrition.
A limitation of the included studies is that supplement use was often combined with several types of interventions (e.g. exercises, physiotherapy, shockwaves, surgery). Although these interventions were similar in the control group, it cannot be ruled out that the concurrent treatment supported the effect of the nutritional treatment. The results show that combining supplement use with other treatments provides further benefits than the treatment alone. It would be interesting to see what effects are induced by taking the dietary supplement alone. Nevertheless, nutritional strategies do not interfere and can easily be implemented in combination with other interventions.
Another barrier in the synthesis of findings from the different studies is that many disparate clinical outcomes were reported. Many studies rely on a limited number of outcomes and were mainly participant-reported. Pain measured by the VAS was an outcome reported in most studies, but referred to a different activity or timeframe, or rated pain without further specification. As recommended by the ICON group, clinical trials should include a measure for each of the nine core domains at a minimum [23]. However, from the studies identified in this review that of Sandford et al. [41] captured six domains. Other studies reported even less. This stresses the need to determine a core outcome set that should be adopted widely in tendinopathy research.

Recommendations for future research
More knowledge is required on the impact of habitual dietary exposures on tendon health, as a healthy habitual diet is the basis for adequate nutrient intake. Improving the habitual intake should be the main focus of athletes rather than the intake of dietary supplements. High-quality studies with extensive dietary intake assessment are needed to examine this association in tendinopathy patients as well as in healthy populations, to determine the role of nutrition in preventing tendinopathy. Research should be conducted specifically in athletes and active populations and by assessing core clinical outcome measures to enable future meta-analyses.

Conclusion
Due to the limited scientific quality and variety among studies on nutrient intake, tendon location, study population and reported outcome measures, it is impossible to draw definitive conclusions and formulate dietary recommendations on the prevention and treatment of tendinopathy. Findings on alcohol intake were inconsistent. Individual studies present important clinical implications for the use of dietary supplements on tendon health, of which especially those containing collagen-derived peptides seem to be beneficial in the treatment of tendinopathy. Also, methyl-sulfonyl-methane, arginine, bromelain, curcumin, and Boswellia were present in supplements that showed clinical improvements. Future clinical studies considering nutritional intake should use standardized dietary assessment methods, adopt the core domains for tendon research and report a core outcome set for each tendinopathy, in order to synthesize findings from different studies.

Availability of data and materials
The datasets generated during the current study are available from the corresponding author on reasonable request.  Table 2 Search strategy in PubMed. Table 3 Rating of the certainty of evidence for clinical outcomes.

GRADE domain Judgment
Concerns about certainty domains Methodological limitations of the studies Among the nine RCTs, the majority expressed 'some concerns' with respect to the risk of bias. Two studies were judged at low risk of bias and one study had a high risk of bias. The risk of bias of the remaining five intervention and observational studies was judged at low or moderate for two studies, and serious or critical for three studies. Main aspects that raise concerns were reporting participant-reported outcomes, incomplete or unclear reporting of methods or results, and small study populations. Therefore, we judged the studies to have serious methodological limitations.

Indirectness
Most studies were primarily aimed to investigate the effect of the dietary supplement on clinical outcomes, but often in combination with other treatments. We judged the evidence to have moderate indirectness.

Imprecision
The total number of participants included in all studies was 819. We judged the evidence to have moderate imprecision.

Inconsistency
The majority of the studies found a beneficial effect of the supplement intake on one or more of the clinical outcomes. There is inconsistency in the effects on different time points, but this could be addressed by variation in study protocol. We judged the evidence to have moderate inconsistency.

Moderate
Publication bias Some studies are commercial studies. We found no commercial studies without effect.

Potential
Additional file 4 Table 4 Rating of the certainty of evidence for occurrence/prevalence of tendinopathy.

GRADE Domain Judgment
Concerns About Certainty Domains Methodological limitations of the studies The risk of bias was judged at moderate for two out of three studies. One study was judged at serious risk of bias, but this study was smaller compared to the other two (605 vs. 80,106/6237). All studies had an observational design, which involves several limitations. In addition, sources of bias were inappropriate statistical adjustments, potential underreporting of intake and incomplete reporting of results. Therefore, we judged the studies to have serious methodological limitations.

Indirectness
Only one study primarily aimed to investigate the association between the intake of alcohol and the risk of tendinopathy. In the other two studies, alcohol consumption was only one of many factors that were investigated to find an association. One study did not even report data with regard to alcohol consumption. Therefore, we judged the evidence to have serious indirectness.

Imprecision
The total number of participants included in all studies was 86,948. This is a large number, but this is mainly due to one large cohort study with relatively low number of cases identified. We judged the evidence to have moderate imprecision.

Inconsistency
The studies reported either a positive association or no association between alcohol consumption and the risk of tendinopathy. One study found a marginal association for moderate weekly alcohol consumption and Achilles tendinopathy, but not for heavy weekly alcohol consumption or patellar tendinopathy. Another study found significant risks of rotator cuff tears for excessive drinkers. We judged the evidence to have moderate inconsistency.

Publication bias
We do not suspect publication bias, taking into account that we have few studies Not suspected Additional file 5