August 7, 2019

Blueberries for Knee Osteoarthritis

Another clinical indication?
Study finds that blueberries improve pain, stiffness, and gait but do not reduce biomarkers of inflammation in patients with knee osteoarthritis.


Du C, Smith A, Avalos M, South S, et al. Blueberries improve pain, gait performance, and inflammation in individuals with symptomatic knee osteoarthritis. Nutrients. 2019;11(2):e290.

Study Design

Double-blind, randomized, placebo-controlled trial

Study Objectives

To examine the effect of blueberries on symptomatic knee osteoarthritis (OA)


Sixty-three men and women, aged 45-79 years, with self-reported, symptomatic knee OA were included in the study. All participants agreed not to take COX-2 inhibitors, chondroitin, glucosamine, glucosamine hydrochloride, or blueberries during the study period.

Exclusion criteria included the following: smoking more than 1 pack per day; insulin regimen that prohibited an additional carbohydrate in a routine diet; uncontrolled diabetes; congestive heart failure; history of knee replacement; allergy to blueberries.


The treatment group received 40 g of freeze-dried whole blueberry powder (Highbush Blueberry Council, Folsom, CA) that was reconstituted by adding it to 10-12 ounces (296-355 mL) of water. Each packet was 20 g; they were asked to consume 2 packets per day. Placebo powder was also in 20 g packets, to be consumed 2 packets per day. Placebo consisted of maltodextrin to mimic the composition of blueberries but without blueberries. Energy content and color of the placebo was similar to the blueberry powder. Trial duration was 4 months from the time of baseline measurements.

Study Parameters Assessed

A screening questionnaire was completed by phone. During visits at baseline, 2 months, and 4 months, height, weight, leg length, blood pressure, fasting blood tests, gait test, and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) questionnaire were assessed (WOMAC yields a standardized assessment of pain, stiffness, and performance of daily activities). Treatment compliance was tracked via a calendar provided at each visit. Gait was analyzed using GAITRite System (CIR Systems, Inc, Franklin, NJ), a portable 10 m electronic walkway. Each participant walked 3 times at normal speed and 3 times at fast speed, but not running, with 20 seconds between each walk.

Gait analysis measured speed, cadence, velocity, right and left step and stride length, and left and right leg support percentage. Overnight fasting venous blood samples were analyzed for interleukin (IL)-1-beta, IL-6, tumor necrosis factor (TNF)-alpha, IL-10, IL-13, matrix metalloproteinase (MMP)-3, MMP-13, and monocyte chemoattractive protein (MCP)-1.


During the study, 14 (22.2%) participants withdrew due to issues of taste, palatability, lack of interest, compliance, or conflicts with visit scheduling. The study population was 29% male and 71% female, with an average age of 56.5±1.2 years and an average BMI of 31.9±0.8. At baseline, systolic blood pressure was significantly greater in the blueberry vs the placebo group; diastolic blood pressures were not different.

The placebo group had a significant increase (+2.31 kg) in weight at midpoint and endpoint compared to the blueberry group. Systolic blood pressure in the blueberry group had decreased significantly by midpoint and remained so at endpoint, while diastolic blood pressure had decreased significantly at endpoint.

Total WOMAC scores were unchanged in the placebo group, but decreased significantly in the blueberry group at midpoint and endpoint. Pain decreased significantly at midpoint and continued to decrease to endpoint, stiffness decreased significantly at midpoint and continued to endpoint, and there was significant change in performance of daily activities at midpoint and endpoint. In the placebo group pain was reduced significantly at endpoint and stiffness decreased significantly at endpoint, but there was no change in performance of daily activities.

Gait velocity (cm/sec) in the blueberry group increased significantly at midpoint and endpoint, with significant changes in left and right leg support percentage at midpoint and endpoint. In the placebo group, velocity also increased significantly at midpoint and that change held to endpoint. In the blueberry group, left and right step length (cm) and stride length (cm) increased significantly at midpoint, and further improved at endpoint. In the placebo group, step length increased significantly at midpoint and held at endpoint, while stride length improved at midpoint and that change held at endpoint, improving again only on the right side. In the blueberry group, right and left leg support percentage while walking improved significantly at midpoint, and again at endpoint, only for normal pace. In the placebo group right and left leg support improved significantly, only for fast pace. Female-only data had similar results.

Inflammatory biomarkers did not significantly change in the blueberry group during the trial. In the placebo group, TNF-alpha increased significantly at midpoint and returned to baseline at endpoint, and IL-1-beta increased significantly at midpoint and then returned to baseline at endpoint. The other biomarkers did not change significantly in the placebo group. In the blueberry group, MMP-3 and MMP-13 decreased at midpoint and endpoint, but not significantly, while MMP3 decreased significantly at midpoint in the placebo group, but had increased nearly to baseline by endpoint. MCP-1 decreased in the blueberry group at endpoint and increased in the placebo group at endpoint, but neither change was significant. Female-only data revealed no significant changes in any biomarkers.

Key Findings

Freeze-dried whole blueberry powder consumption twice a day was achievable and it did reduce pain, stiffness, and performance of daily activities compared to placebo. This extract also improved gait velocity, step length, stride length, and left and right leg support percentage better than placebo. The blueberry extract had no significant effect on inflammatory biomarkers, and there was a positive but not significant trend for decreasing levels of MMP-3, MMP-13, and MCP-1.

Practice Implications

Long-term dietary consumption of polyphenols can have positive effects on human health and inflammation, presumably because of their antioxidant effects.1-3 Polyphenols are found in green tea, turmeric, red wine, citrus fruits, pomegranates, cherry juice, and other foods with primarily red, blue, or dark colors. One serious weakness of this study was the lack of control for these foods and supplements (eg, epigallocatechin gallate [EGCG], resveratrol, quercetin, sulforaphanes, blackberries, strawberries, dark cherries) in participants’ diets during the 4-month trial. The study did not control for the use of acetylsalicylic acid (ASA) or Tylenol, drugs commonly used for joint inflammation. In a recent 6-month cardiovascular study in patients with high-risk metabolic syndrome, consumption of non-blueberry sources of polyphenols was tightly controlled; investigators observed a significant reduction (11%) in risk of cardiac events.3 The level of polyphenols in the blueberry extract used in the present study was not reported.

In the present study, the gait analysis demonstrated that as pain and stiffness decreased, a more natural, longer, and stable stride could be maintained. In the blueberry group, improvement was demonstrated at the 60-day midpoint and continued to the endpoint of 120 days. This parallels or confirms the WOMAC findings and again confirms the use of WOMAC as a clinical tool to assess patient function when osteoarthritis is present.

The biomarkers of TNF-alpha, IL-1-beta, and IL-6 did not change with blueberry treatment, but are known to be elevated in osteoarthritis.4 Elevation of MCP-1 is felt to contribute to both initiation and worsening of both rheumatoid arthritis and OA.4,5 In the blueberry group, MCP-1 decreased, but not significantly. Would a higher dose of blueberries or a longer treatment period have a significant effect? We do not know. In a trauma-induced OA study in mice, EGCG reduced MMP-1, MMP-3, MMP-8, MMP-13, IL1-beta, and TNF-alpha, but we lack human confirmatory data.6 MMP-13, which degrades type II collagen and cartilage, has been deemed essential for the progression of OA; when present the prognosis is poor.7,8 MMP-1 degrades collagen I, II, and III and is high in arthritic cartilage; MMP-3 degrades collage II, III, IV, IX, X, fibronectin, laminin elastin, and other proteoglycans and is absent in normal joint tissues, while MMP-8 is present in dentin and, like MMP-3, is present in wounded joints, especially acute wounds.8 All of these metalloproteinases are zinc-dependent, and hyperglycemia leads to higher circulating levels in the bloodstream.8 MMP-13 has been linked to cognitive impairment, renal malformation and impairment, obesity, hypertension, and type II diabetes.8 In rheumatoid arthritis (RA), MMP-9 is higher than in OA, suggesting that it is the initiating enzyme for higher rates of tissue turnover in RA versus OA.8

Currently, knee OA is diagnosed by history and physical examination, with joint aspiration, x-ray, and MRI reserved for questionable cases or cases where knee replacement is a clinical consideration.9 In the present study, presumably the over-the-phone screening questionnaire and WOMAC scores at baseline confirmed the diagnosis, but this was not addressed in any detail in the paper. None of the authors was a rheumatologist and no rheumatology consultation was cited. The cohort’s mean BMI (31.9 ± 0.8) indicates that participants in this study were overweight or obese, a confounding factor for the role of inflammation in knee OA. Participants were allowed to smoke up to 1 pack of cigarettes per day, a further source of inflammation. Another confounding factor is not knowing how many joints were affected. Presumably, participants with 2-knee OA and multiple-joint OA would have a greater elevation of inflammatory factors than those with single-joint involvement.

The authors argued that those in the blueberry group benefitted from its effects on satiety, similar to a prior rat study.10,11 However, this study was not designed to assess satiety. One very concerning issue was the high dropout rate (22.2%); dropout can seriously affect statistical significance of a study. Apparently, participants found the treatment and placebo less palatable over time, mixed it in milk, or added it to breakfast cereal, potentially reducing bioavailability of the polyphenols. However, in a recent blueberry powder trial (0g, 13g, 26g per day), participants with high-risk metabolic syndrome were given suggested recipes that included yogurt and dairy smoothies, and outcomes were positive and diverse metabolites from their blueberry powder were measurable.3

In the present study, the daily trial dose of 40 g is approximately equivalent to 1.35 cups of blueberries per day. Single intervention trials, like the one reviewed here, seek the ideal dose of a single treatment. However, intervention with a combination formula that addresses several aspects of knee OA inflammatory pathways could be effective at a lower dose and taste as good or better. For example, EGCG, glucosamine, chondroitin, quercetin, bromelain, and blueberry polyphenols could be combined for potential clinical benefit.

The final issue with this paper was the occasional disagreement between the text and the figures, but I believe my overall interpretation is correct. A more parallel writing style in the text would have made this author’s task easier. Some references were incomplete and others contained spelling errors. A final thorough review seemed to be lacking. Such errors call into question the integrity of the data. Calculated P values were not reported, only that a result was <0.05 and thus significant. Nonetheless, given the relatively benign intervention, clinical awareness of possible benefits of blueberries to OA seems warranted.


Blueberries as a freeze-dried powder in water, 20 g twice a day, equivalent to about 1.35 cups of fresh berries per day, reduced knee joint pain and stiffness and improved performance of daily activities and gait at both 2 and 4 months in women and men with self-reported knee OA. The biomarkers associated with osteoarthritis did not change significantly. Osteoarthritis may become another clinical indication for blueberry consumption.

Categorized Under


  1. Pandey KB, Rizivi SI. Plant polyphenols as dietary antioxidants in human health and disease. Oxid Med Cell Longev. 2009;2(5):270-278.
  2. Angeloni C, Pirona L, Vauzour D, Maraldi T. Dietary polyphenols and their effects on cell biochemistry and pathophysiology. Oxid Med Cell Longev. 2012;2012:583901.
  3. Curtis PJ, van der Velpen V, Berends L, et al. Blueberries improve biomarkers of cardiometabolic function in participants with metabolic syndrome—results from a 6 month, double-blind, randomized controlled trial. Am J Clin Nutr. 2019;109:1535-1545.
  4. Yuankun X Sr, Yan K, Bin W, Jian-Hao L. Monocyte chemoattractant protein-1 induced chondrocytes degeneration and cartilage degradation in osteoarthritis. Osteoar Cartil. 2016;24(Supp1):S140-S141.
  5. Scanzello CR. Chemokines and inflammation in osteoarthritis: insights from patients and animal models. J Orthoped Res. 2017;35(4):735-739.
  6. Leong DJ, Choudhury M, Harstei R, et al. Green tea polyphenol treatment is chondroprotective, anti-inflammatory, and palliative in a mouse post-traumatic osteoarthritis model. Arthritis Res Ther. 2014;16:508.
  7. Wang M, Sampson ER, Jin H, et al. MMP-13 is a critical target gene during the progression of osteoarthritis. Arthritis Res Ther. 2013;15:R5.
  8. Rose BJ, Kooyman DL. A tale of two joints: the role of matrix metalloproteases in cartilage biology. Dis Markers. 2016;2016(12):1-7.
  9. Arthritis Foundation. Osteoarthritis diagnosis. Accessed June 13, 2019.
  10. Molan AL, Lila MA, Mawson J. Satiety in rats following blueberry extract consumption induced by appetite-suppressing mechanisms unrelated to in-vitro or in-vivo antioxidant capacity. Food Chem. 2008;107(3):1039-1044.
  11. Bannuru RR, Schmid CH, Kent DM, Vaysbrot EE, Wong JB, McAlindon TE. Comparative effectiveness of pharmacologic interventions for knee osteoarthritis: a systematic review and network meta-analysis. Ann Intern Med. 2015;162(1):46-54.