Vitamin C

Vitamin C occupies an odd position among supplements: its mechanism is not in dispute at all. Ascorbate is the reducing agent that keeps the iron centre of collagen prolyl 4-hydroxylase and lysyl hydroxylase in the ferrous state. Deprive a person of it and procollagen goes under-hydroxylated, the triple helix loses stability, and connective tissue comes apart — which is what scurvy is. The question was never whether vitamin C matters to collagen. It is whether more of it does more.

How ascorbate reaches a chondrocyte

Cartilage is avascular, so it is fair to wonder whether anything swallowed gets there at all. For this vitamin the route is mapped in unusual detail.

Human chondrocytes concentrate ascorbate 960-fold over the fluid around them, by sodium-dependent active transport. Silencing the transporter SVCT2 abolishes the active component entirely, which identifies it as the door. A second door exists for the oxidised form, dehydroascorbate, through the glucose transporters — and that one opens wider at low oxygen tension, which is precisely the condition inside cartilage. Modelling from measured synovial-fluid ratios, the oxidised route supplies about a quarter of intracellular ascorbate in osteoarthritic chondrocytes and almost all of it in rheumatoid joints.

The gradient shows up in a live animal. In guinea pigs, which like us cannot make their own ascorbate, concentrations run cartilage above synovial fluid above plasma, with normalized cartilage levels sitting between liver and adrenal gland. Once inside, ascorbate raises transcription of type II collagen, prolyl 4-hydroxylase and aggrecan in cartilage explants, and does so only for the L-forms that SVCT2 carries.

One finding complicates the picture. SVCT2 messenger RNA is significantly lower in grade 3 osteoarthritic cartilage than in grade 1, with much less transporter visible on the chondrocyte membrane. If ascorbate needs that door, the cells that need it most are the least able to open it.

The human joint evidence

One randomized trial has tested vitamin C against placebo with an osteoarthritis endpoint. Ten Danish general practices randomized 133 patients with radiographically verified hip or knee osteoarthritis to 1 g of calcium ascorbate or identical placebo for two weeks, then crossed them over after a washout. On intention to treat, ascorbate beat placebo on pain by a mean 4.6 mm on a 100 mm scale (95% CI 1.2–8.0, p = 0.0078), with the Lequesne function index and patient preference following. The author's own reading is printed in the paper: the effect is less than half what is commonly reported for anti-inflammatories. It is also below the 15–20 mm usually treated as the smallest difference a patient would notice, and each arm lasted fourteen days.

Two cohorts point the same way. In the Framingham osteoarthritis cohort, among people who already had knee osteoarthritis, the middle and highest tertiles of vitamin C intake each carried a threefold lower risk of radiographic progression — adjusted odds ratios of 0.3, driven predominantly by less cartilage loss, with less new knee pain as well. Incident osteoarthritis showed no association with anything, and the non-antioxidant control vitamins showed nothing, which is the internal control that makes the result worth taking seriously.

The Osteoarthritis Initiative followed 1,785 participants for four years with 3T MRI and found higher vitamin C intake associated with lower cartilage T2 — the imaging measure of matrix quality — and lower medial tibial damage scores. The standardized coefficients are −0.07 to −0.05: real, and very small. Vitamin D, measured in the same subjects with the same questionnaire, produced coefficients three to five times larger and was the only vitamin whose consistent supplementation predicted less worsening over the four years.

On tendon, a single randomized trial gave 500 mg a day for 45 days after arthroscopic rotator cuff repair. Non-healing on ultrasound was 11% with vitamin C against 23% without, which did not reach significance (p = 0.2), and clinical scores were identical. The control arm received nothing rather than a placebo.

The guinea-pig result

The mechanism runs one way; the best long-duration animal experiment runs the other, and it deserves to be stated at full strength rather than buried.

Male Hartley guinea pigs develop knee osteoarthritis spontaneously. Given low, medium or high dietary ascorbate for eight months — the medium dose producing plasma levels equivalent to a person eating five servings of fruit and vegetables — cartilage collagen content rose with intake, exactly as the cofactor biology predicts. So did the severity of the osteoarthritis. Osteophyte size and number increased as intake increased, active TGF-β was expressed in those osteophytes, and synovial-fluid cartilage oligomeric matrix protein corroborated the histology. The authors' stated conclusion is that ascorbic acid should not be supplemented above the recommended dietary allowance.

Two things keep this from being the last word. It is a guinea pig, in a spontaneous, osteophyte-heavy model that may not mirror human post-traumatic cartilage loss. And the same laboratory had earlier found ascorbate slowing progression in surgically induced osteoarthritis in the same species — so the direction of effect may depend on the kind of joint damage. Nobody has resolved that in the twenty years since.

What a gram a day does

Vitamin C has one of the best-characterised dose-concentration curves in nutrition, and it explains why 1 g is not five times 200 mg.

Plasma concentration is sigmoid against dose. The steep part runs from 30 to 100 mg a day. The curve flattens by 200 mg. Plasma is completely saturated at 1,000 mg. Circulating white cells saturate lower still — at 100 mg in men and 400 mg in women. Bioavailability is complete for a single 200 mg dose and starts declining from 500 mg upward, with whatever is absorbed above that excreted. A 1.25 g oral dose peaks at about 135 µmol/L in plasma; the same dose given intravenously peaks near 885, and even 3 g every four hours is modelled to reach only 220. Beyond about a gram, absorption and the kidney decide the exposure, not the label.

So 1 g/day sits at the top of the curve. It is a defensible place to sit — it guarantees saturation with room to spare, and it is the dose the one randomized osteoarthritis trial actually used. It is not, on the pharmacokinetics, a higher tissue concentration than 400 mg would give. Worth noting separately: the trials that pair vitamin C with collagen or gelatin before loading exercise used 50 mg, not a gram.

Oxalate and kidney stones

Ascorbate is metabolised in part to oxalate, and this is the one claim about gram-level vitamin C that has real data behind it rather than folklore.

The depletion-repletion study that mapped the dose curve found urinary oxalate and urate elevated specifically at 1,000 mg a day, not below it. A 2026 meta-analysis of nine interventional studies puts numbers on it: vitamin C supplementation raised 24-hour urinary oxalate by 9.72 mg against baseline and 6.45 mg against placebo, in stone formers and non-stone formers alike, while urinary citrate and calcium — the other two levers on stone chemistry — did not move at all.

Whether that translates into stones depends on who you are. Across 197,271 people followed for more than a decade, men whose total vitamin C intake reached 1,000 mg a day or more had a hazard ratio of 1.43 for a first kidney stone (95% CI 1.15–1.79), with a graded trend across intake categories; counting supplements alone the figure was 1.19. Among the 156,735 women in the same analysis, neither total nor supplemental vitamin C was significantly associated with stones. Dietary vitamin C was not associated with stones in either sex — the signal attaches to supplements, not to fruit.

There is a second entrance to the same pathway that matters for anyone taking collagen. Hydroxyproline, the amino acid that makes collagen collagen, is itself an oxalate precursor. In a controlled feeding study, 30 g of gelatin a day against a matched 30 g of whey raised urinary glycolate 5.3-fold and urinary oxalate 43%, with the threshold for a detectable change falling between 2 and 5 g. A smaller study at the same dose found the glycolate rising without the oxalate following, so the size of the effect is unsettled. Neither study measured a stone.

The precise statement is this: gram-level vitamin C reliably raises urinary oxalate, gram-scale collagen probably does too, and someone taking both is loading one pathway from two directions. That is a reason for anyone with a history of calcium-oxalate stones, hyperoxaluria or reduced kidney function to keep vitamin C nearer 200 mg. It is not a reason for anyone else to stop.

What would change this entry

The cofactor biology is settled, the transport into cartilage is demonstrated, and one randomized trial found a small real analgesic effect at 1 g a day. What is missing is a trial long enough to matter with a structural endpoint — both of the imaging signals here are observational, and the animal work points the other way above the RDA. A randomized cartilage-imaging trial would resolve more of this entry in one study than the last thirty years have.