Berries & anthocyanins

Most of the trials in this section test an extract, a capsule or a concentrate. The berry trials tested fruit — freeze-dried and reconstituted, but whole fruit with its sugar and fibre intact — and one of them did something almost nobody else in the foods literature bothers with: it matched the control beverage for calories and fibre, so the comparison is berry against not-berry rather than berry against nothing.

Strawberry

Seventeen obese adults with radiographic knee osteoarthritis (mean BMI 39.1) drank 50 g a day of reconstituted freeze-dried strawberry, or the matched control, for twelve weeks each way with a two-week washout between.

Against the control, IL-6, IL-1β and MMP-3 all fell (p < 0.05). So did pain — constant, intermittent and total on the ICOAP questionnaire — and disability on the HAQ-DI. CRP, nitrite, glucose and lipids did not move. A companion paper from the same trial ran a 24-marker inflammation panel: twelve were detectable, and of those, high-sensitivity TNF-α and soluble TNF receptor 2 fell, along with two lipid-peroxidation products. Body weight, obesity-related hormones and osteocalcin were unchanged, which matters — the effect is not weight loss in disguise.

MMP-3 is the interesting one. It is a matrix-degrading enzyme rather than a general inflammation marker, and it moved in the same people whose pain moved.

The size is the problem. Seventeen people, four of them men, two papers from one trial — which is a second report, not a replication. And twelve markers detectable with two moving, unadjusted for multiplicity, is a result that deserves the word preliminary.

Blueberry

Sixty-three adults aged 45 to 79 with symptomatic knee osteoarthritis took 40 g a day of freeze-dried whole blueberry powder or a matched placebo powder, double blind, for four months.

WOMAC total and its pain, stiffness and daily-activity sub-scores fell significantly in the blueberry group; the placebo group did not improve on WOMAC total or on daily activities. Gait was measured on an instrumented walkway rather than reported by the participants: single-support percentage rose and double-support percentage fell in both limbs (p ≤ 0.007 and p ≤ 0.003), which is the pattern of someone spending more time confidently on one leg.

Two caveats sit on top of that. The reported comparisons are within-group changes arm by arm, and in a double-blind trial the comparison that counts is between arms — a placebo group that improves less is not the same as a treatment group that improves more. And every one of the eight plasma inflammatory markers was flat: TNF-α, IL-1β, IL-6, IL-10, IL-13, MMP-3, MMP-13 and MCP-1, with only trends in two. The symptom result has no mechanistic support inside its own study.

What anthocyanins do, and the gap underneath

The mechanism belongs to cyanidin, the pigment berries share with tart cherries. In human osteoarthritic chondrocytes it suppresses ADAMTS5 and MMP13 and preserves aggrecan and type II collagen through Sirt6, and it slows progression in mice with destabilized menisci. Anthocyanins as a class protect cartilage explants from the glycosaminoglycan loss driven by advanced glycation end-products. The same literature also contains a warning: in differentiating stem cells, cyanidin suppressed the chondrogenic markers along with the hypertrophic ones.

And the concentrations in those experiments are not the concentrations a person reaches. Anthocyanin plasma levels after eating are low and short-lived, mostly present as metabolites rather than the parent compounds. That gap is why the symptom results and the mechanism results in this entry cannot yet be joined up, and why a flat inflammatory panel in the blueberry trial is unsurprising rather than damning.

Practical notes

The doses were 50 g a day of freeze-dried strawberry — roughly half a kilo of fresh fruit — and 40 g a day of blueberry powder. Both are meaningful sugar loads, which is exactly why the strawberry trial's calorie- and fibre-matched control is the design detail worth copying.

What would change this entry

A between-group analysis of a blueberry-sized trial, run long enough to look at structure. The strawberry work already shows that a matched control and a matrix-degrading enzyme endpoint are achievable in a small trial; scaling that design up, with imaging at the end of it, is the study this topic is one step away from.