Plant and Animal-Based Foods
A case-control study in India found a lower intake of fruits and a higher intake of green leafy vegetables in patients with knee OA (Sanghi et al., 2014).
Some fruits have also been clinically tested for their efficiency in patients with OA. 40g of freeze-dried blueberries have been tested in a randomized, double-blind, placebo-controlled manner in regard to their effectiveness in improving inflammatory markers beside pain and functional assessments in the WOMAC questionnaire. The blueberries reduced pain and difficulty to perform daily activities, improved gait performance, but didn´t influence the inflammatory markers after 17 weeks (Du et al., 2019).
50g of freeze-dried strawberries were also tested with a comparable research design and included several inflammatory markers and the ICOAP Pain Questionnaire. It found a significant reduction in pain and biomarkers IL-6, IL1β and MMP-3, but not in C-reactive protein and MMP-8 after 26 weeks (Schell et al., 2017). Strawberries might also decease serum TNF- α and lipid peroxidation products in adults with OA (Basu et al., 2018).
200ml of pomegranate juice was tested against a control group and showed a significant intra- group reduction in stiffness and total score but no reduction in pain in the WOMAC index (Ghoochani et al., 2016). Beside that, an increase in physical function score was observed. MMP- 13 levels increased in the control group and remained the same in the pomegranate group leaving a significant between-group difference. MMP-1 stayed the same in the control group while being reduced in the intervention group also leaving a significant between-group difference. MMP-13 and MMP-1 play a role in cartilage matrix breakdown via digestion of type 2 collagen and release of matrix proteoglycans. Another research group tested the effects of 150mg encapsulated passion fruit peel extract for 8 weeks against placebo (Faird et al., 2010). The intervention group showed a significant improvement in WOMAC total, physical function, and pain score while stiffness was not improved.
A 2008 meta-analysis with the 3 available studies at that time showed a reduction in WOMAC pain, stiffness, and physical function scores by administration of 5g of rosehip powder compared to placebo in patients with knee OA, but with conflicting results for the use of pain medications (Christensen et al., 2008).
In a 2015 meta-analysis of randomized controlled trials with a placebo group oral ginger was examined for its effects in patients with knee or hip OA (Bartels et al., 2015). 500mg to 1000mg of ginger extract showed a reduction in pain and disability. A newer 2017 study compared ginger extract with unknown dosage and 1% diclofenac gel after 12 weeks. Both groups improved WOMAC and Patient Global Assessment score with no differences between groups (Amorndoljai et al., 2017).
1000mg of garlic has also been tested in a randomized, double-blind, placebo-controlled trial in patients with knee OA, showing no effect of garlic for improving pain, stiffness, or physical function scores (Salimzadeh et al., 2018).
40g of sesame has also been tested against a control group for 2 months (Sadat et al., 2013). The sesame and control groups Knee Injury and Osteoarthritis Outcome Score (KOOS) and timed-up and go test improved similarly but no statistical test for a comparison of the two groups was done. VAS scale improved more in the sesame group again with no between-group comparison.
Black cumin is another spice tested in a randomized, double-blind trial against placebo. 2g of black cumin powder were given to patients with knee OA showing no between-group differences in KOOS (Salimzadeh et al., 2017).
In a case-control study, regular soy-milk consumption of over or under once a day showed a lower prevalence of osteophytes while not influencing joint space narrowing (Li et al., 2016). In a direct comparison between either 40g of soy or milk-based protein for 3 months both proteins improved knee range of motion, hindrance of activities of daily living and limitations to exercise while also improving work performance and productivity (Arjmandi et al., 2004). In Men soy protein supplementation also improved pain medication use and biomarkers of cartilage metabolism.
In a cohort study, no association was found with eating processed meat under 1.5 times to over 4, chicken under 1.5 times to over 3.5 times, or fish under 1 time to over 2.5 times per week and the risk of getting a hip or knee replacement done (Wang et al., 2011). Eating fresh red meat over 6.5 times per week compared to 3 – 4.4 times per week decreased the relative risk of needing a hip replacement by 6%, while not being associated with needing a knee replacement.
A case-control study found lower odds of radiologically diagnosed knee OA in daily milk consumers, while no association with cheese or yoghurt was found (Kaçar et al., 2004). Another case-control study also found a higher intake of milk, milk products and meat in patients without compared to patients with OA of the knee (Sanghi et al., 2014). In a cross-sectional study it was also found that the intake of over 3.3 servings per day compared to 1.9 servings a day of full- fat dairy reduced the relative risk for knee OA by 32%. The consumption of over 1.7 servings per day compared to under 0.7 servings per day reduced the risk by 27% were most of the effect comes from the consumption of full-fat dairy (Denissen et al., 2019). No association was found for total dairy, semi-skimmed milk, skimmed dairy, non-fermented dairy, fermented dairy, unflavoured milk, yoghurt, and total cheese.
Longitudinal data from the OA Initiative revealed a dose-response relationship for a reduction in joint space width and milk consumption with the best outcome at over 7 glasses a week. This association was not found for men (Lu et al., 2014). A randomized, double-blind, active controlled trial gave participants a milk-based beverage with milk protein concentrate fortified with some vitamins and minerals (Colker et al., 2002). WOMAC total score and daily activities improved compared to the control group, while sport function, knee symptoms and knee pain were unchanged.
The Microbiome and Osteoarthritis
The microbiome is defined as a characteristic microbial community occupying a reasonable well-defined habitat which has distinct physio-chemical properties and communicates with the different systems and organs of the human body, so one might also assume a connection with the joints. As shown in part 1 of this blog series, it has been shown that certain strains of probiotics might differentially influence OA.
In the study by Coulsen on green-lipped mussel extract it was shown that OA symptoms improved, and Clostridium and Staphylococcus species tended to decrease while Lactobacillus, Streptococcus and Eubacterium species tended to increase. In the Rotterdam cohort study, it was shown that the abundance of Streptococcus species is associated with an increase in knee pain (Boer et al., 2019). It is important to know that “Streptococcus species” refers to the general genus of this bacteria that include different species in that genus which might have different effects. It was also shown that the abundance of lipopolysaccharides as a pro-inflammatory secondary metabolite of microbiome gram-negative bacteria found in the joints of patients with knee OA are associated with an increased abundance of activated macrophages in the knee, osteophyte and joint space narrowing severity, WOMAC total score and self-reported knee pain score (Huang et al., 2016).
In terms of nutrition, dietary fibre and phytonutrients exert the most beneficial effects for the microbiome because those can be metabolized be certain parts of the microbiome which thrive on a fibre and phytonutrient rich diet (30+ grams of fibre with lots of variety in whole plant-based foods per day) and produce secondary metabolites which have physiological effects outside the gut. A diet-modulating effect of fibre has been shown earlier, but it has not yet been researched if the mediating effect comes from an effect of fibre on the microbiome.
Dietary Patterns and Osteoarthritis
As mentioned above, inflammation plays a role in the aetiology of OA. The dietary inflammatory index (DII®) is a dietary tool developed to assess the overall inflammatory potential of an individual’s diet with higher values to be associated with some elevated inflammatory markers (Shivappa et al., 2014). It was shown that a more pro-inflammatory diet increased the risk for symptomatic knee OA (Veronese et al., 2017). In another cross-sectional study, it has been shown that a good overall diet quality was not associated with knee structure and OA symptoms but associated with greater limb muscle strength, lower depressive symptoms and better quality of life which was mainly driven by vegetable intake (Ruan et al., 2021). For the progression of existing knee OA, it has been shown that a standard western diet is associated with more severe radiographic and symptomatic progression of knee OA. On the other hand, an overall healthier dietary pattern decreased the risk of radiographic and symptomatic OA progression (Xu et al., 2020).
A dietary pattern that has generally been advised as health promoting is the Mediterranean Diet. Two cross-sectional studies assessed how following a Mediterranean Diet might influence OA. In one study a higher adherence to a Mediterranean Diet decreased the prevalence for knee OA, especially driven by the consumption of healthy cereals. In another study quality of life, physical function and mental health increased with increased adherence to the Mediterranean Diet while pain, stiffness and disability and depressive symptoms decreased (Veronese et al., 2016; Veronese et al., 2016). A 2019 longitudinal study found that a higher adherence to a Mediterranean Diet reduced the risk of pain worsening in people with knee OA and reduced the risk of symptomatic incident radiographic knee OA but not incident asymptomatic radiographic knee OA after 4 years (Veronese et al., 2019). In an interventional trial, patients with OA were randomized to either following a Mediterranean Diet or their normal diet for 16 weeks (Dyer et al.., 2017). The Arthritis Impact Measurement Scale questionnaire showed no differences between groups and only the inflammatory biomarker IL-1α decreased in the intervention group while also a significant improvement in knee flexion and hip rotation in the Mediterranean Diet group was seen. A small decrease of 1,5 Kg of body weight was also seen. It is also important to notice that only 30 of the 50 participants in the intervention group achieved a predefined high compliance with the Mediterranean Diet.
See below a summary of common diets / dietary patterns.
|Standard western diet||High in energy, red and processed meat, sugar, desserts, sweets, and sugar sweetened beverages, refined grains, salt, and high fat dairy products|
|Low in fish, fibre, fruits and vegetables, whole grains, legumes, nuts, and seeds|
|Synthesized guidelines world-wide recommendations||High in fruits and vegetables, whole grains, legumes, nuts, and water|
|Moderate in in energy, fish, white meat, eggs, and dairy products|
|Low in red and processed meat, sugar, desserts, sweets, and sugar sweetened beverages, refined grains, salt, and alcohol|
|Mediterranean diet||High in fruits and vegetables, whole grains, legumes, nuts and seeds, native oils, fish|
|Moderate in energy, white meat, dairy products, eggs, refined grains, and wine|
|Low in red and processed meat, sugar, desserts, sweets, and sugar sweetened beverages, and salt|
Researchers also examined whether a low-carb of low-fat diet would improve functional pain, self-reported pain, quality of life and depressive symptoms in patients with knee OA after 3 months (Strath et al., 200). Patients were assigned to either a control, low-carb or low-fat diet. The low-carb dieters were instructed to restrict their total carbohydrate intake to 20g to 40g per day with a restriction in fruits and while vegetables were permitted in limited quantities of two cups of leafy greens and one cup per day of non-starchy vegetables. The low-fat dieters should eat a maximum of 50g-67g of total fat with 800-1200 kcal per day. Protein was matched to 100g day. Because of reduced appetite a low-carb diet can lead to weight loss which is why men in the low-fat group were meant to eat 500 kcal and females 250-300 Kcal less than maintenance calories to induce weight loss. Body weight in the low carb group was reduced by around 9 Kg in the low-carb and 7 Kg in the low-fat group. Pain as assessed with the Brief Pain Inventory (BPI) and KOOS did not change between groups as did depressive symptoms, several inflammatory markers like c-reactive protein, oxidative stress, and leptin as a hormone of fat tissue. Some unvalidated tests for physical function were also taken with some improvements within the low-carb group, only one significant difference between groups and many null findings within and between groups.
While interpreting these results, one must keep in mind that this meaningful weight loss alone can induce symptom improvement, especially when starting from a higher body weight as did the low-carb group. The authors also state that the low-fat group ate 20% of total daily calories from fat. 50g-67g of fat are 450-603 kcal. When eating 800-1200 kcal this would indicate at least 37,5% of daily calories from fat but 75% are also possible which is not a low-fat diet. There is also no overview over the specific foods consumed within the diets which is important since a low-carb and low-fat diet are defined by a low consumption of a macronutrient while the rest of the diet can vary substantially and even fat subtypes might play a role for OA as mentioned earlier. It is also important to know that an energy matched low-fat and low-carb diet results in similar weight loss over time, so choosing one diet over the other solely should be based on personal preferences (Hall et al., 2015; Hall et al., 2016; Hall et al., 2017).
Another example of how diet might be able to influence OA is not by eating, but by not eating! In a 2018 study without a control group, patients with OA of different joints underwent fasting for 8 days (Drinda et al., 2018). The patients did 3 preparatory days with a lacto-vegetarian diet and 800 kcal, followed by 8 fasting days with 300 calories and 4 refeeding days during the interventions period. They drank over 2.5L of liquids of which 250ml were fruit or vegetable juice and stayed away from caffeine, alcohol, and nicotine. They also received magnesium supplementation and probiotics. The participants were also advised to do aerobic physical exercise and activities promoting a mental and emotional balance. Pain was reduced by WOMAC and VAS scale and the physical subscale of the SF-36 quality of life questionnaire also improved. The participants also lost weight while lowering waist circumference which might indicate a loss of visceral fat. Another uncontrolled study with a comparable design had comparable results and did also reduce weight (Schmidt et al., 2010).
Conclusion and Application to Clinical Practice
While dietary advice is by no means the primary role of a physiotherapist, modern physiotherapy often includes aspects of various health-related fields like medicine, psychology, sports science, and nutrition. Patients might even ask if this new diet trend or supplement they have heard of can really help them or is a waste of money and effort.
Nutrition made it into the 2019 American College of Rheumatology/Arthritis Foundation Guideline for the Management of Osteoarthritis of the Hand, Hip, and Knee in the form of strongly recommended weight loss in patients with hip and knee OA; while glucosamine, chondroitin, fish oil and vitamin D are strongly not recommended (Kolasinski et al., 2020). Vitamin D and fish oil might work best in people with low overall intake and status which is not uncommon and can be tested by the measurement of 25-Hydroxy-Vitamin-D and the Omega-3 Index, with optimal ranges around 30-60ng/dl (75-150nmol/dl) and over 8%, respectively.
The overall lack of recommendations for nutritional interventions in osteoarthritis is based on a sparsity of evidence despite the presented studies. The presented studies are often of low to moderate quality with a moderate to high risk of bias, not enough participants, of short duration in people with different age and disease stage, not tested against standard care, and mostly in patients with knee OA, making it difficult to make definitive statements regarding effectiveness for the patient in front of the practitioner. What´s positive is the safety profile of the tested supplements as they show similar or less adverse effects in the short- and middle-term compared to standard medications, while mostly not being researched for safety in the long-term usage. One must also keep in mind that a recommendation for a supplement can lead to less engagement in other forms of effective therapy like exercise, which might lead to unfavourable results. If a person is already taking a supplement, I would check the dosage and if it has at least been tested in a clinical trial. I personally do not discourage the use of an already- taken supplement (that is safe), because it might take away a beneficial or placebo effect.
The safest, cheapest but also most complex way of treating symptomatic osteoarthritis with diet might be the adoption of a healthy dietary pattern in which the synthesized dietary guidelines world-wide or a Mediterranean diet are adopted, and might or might not be paralleled with weight loss. To support adaptations of exercise therapy, it might also be feasible to recommend a relatively high protein intake (1.6g/kg bodyweight). Importantly however, specific dietary advice should not be given by a physiotherapist and requires referral to a dietician or other nutrition professional.