On this page
Key takeaways
● The right dose depends on what you are taking collagen for. There is no single "correct" dose.
● Bone (postmenopausal): 5 g/day — König 2018 trial-validated (2). Higher doses show no clear advantage for bone.
● Skin: 2.5–10 g/day — meta-analyses cover this range (1). Effect size does not track dose strongly.
● Knee osteoarthritis pain: 10 g/day — Yu 2023 meta-analysis dose (3).
● Tendon and recovery: 10–15 g/day, timed pre-training with vitamin C — Shaw 2017 (4).
● Sarcopenia (with resistance training): 15 g/day — Zdzieblik 2015 (5) (with magnitude caveats).
● Above 15 g/day: no evidence advantage and theoretical oxalate load rises (7). Do not exceed.
Quick answer
The trial-anchored dose depends on your goal. For postmenopausal bone, 5 g/day is enough — higher doses show no clear advantage. For skin, 2.5–10 g/day covers the trial range. For knee osteoarthritis pain, 10 g/day. For tendon recovery, 10–15 g/day timed pre-exercise with vitamin C. For sarcopenia with resistance training, 15 g/day. Above 15 g/day there is no evidence advantage and the theoretical oxalate load rises. If you have combined goals, 10 g/day is a reasonable compromise. Consistency of daily intake matters more than within-day timing.
The trial-anchored dose table
The core reference for this article is the table below. Each row lists a specific indication, the dose that trial evidence supports, the duration expected before assessment, and the anchor trial or meta-analysis. Use this to match your dose to your goal rather than defaulting to the standard 10 g scoop that dominates commercial products.
|
Indication |
Dose |
Duration |
Anchor trial |
|
Skin (hydration, elasticity) |
2.5–10 g/day |
8–12 weeks |
de Miranda 2021 meta-analysis (1); Proksch 2014 at 2.5 g (8) |
|
Bone (postmenopausal, osteopenia) |
5 g/day |
12 months |
König 2018 (2) |
|
Knee osteoarthritis pain |
10 g/day |
3–6 months |
Yu 2023 meta-analysis (3) |
|
Tendon / ligament recovery |
10–15 g/day |
8–12 weeks |
Shaw 2017 (4) — take pre-training with vit C |
|
Sarcopenia (with training) |
15 g/day |
12 weeks |
Zdzieblik 2015 (5) — see caveats |
|
General wellness / combined goals |
10 g/day |
3+ months |
Reasonable compromise dose across indications |
Two immediate implications from this table. First, if bone is your only goal, you are overpaying at 10 g/day — a 5 g dose delivers the same trial-validated effect for half the powder. Second, if muscle recovery around resistance training is your goal, 15 g may be worth reaching, taken pre-exercise rather than any-time-of-day.
Why 10 g is not universally better than 5 g
This is worth stating plainly because product marketing frequently implies otherwise. Doubling the dose from 5 g to 10 g does not double the effect — for any measured outcome in the trial literature.
The reason is pharmacokinetic. After a 5–10 g oral dose, plasma bioactive peptides (chiefly Pro-Hyp and Hyp-Gly, the dipeptides that survive intestinal digestion intact) reach measurable concentrations one to two hours later (6). But the area under the plasma concentration curve scales roughly linearly up to about 10 g and plateaus thereafter. The receptor-level effects on fibroblasts, chondrocytes, and osteoblasts are saturable — beyond a threshold peptide concentration, more input does not deliver more signal. The bulk of a larger dose becomes generic amino acids joining your general amino-acid pool, which is nutritionally useful but does not scale the specific collagen-signalling effects.
So the honest question when choosing between 5 g and 10 g/day is not "which is better?" but "which is enough for my specific goal?" For postmenopausal bone: 5 g. For knee OA pain: 10 g. For skin: either is defensible.
Why 15 g is the ceiling — and above is not worth it
The upper end of the trial-anchored range is 15 g/day, used primarily in tendon/ligament recovery contexts and in the Zdzieblik sarcopenia trial. Above 15 g/day:
● There is no evidence advantage for any indication studied to date. Trials at higher doses are rare and have not shown additional benefit.
● The theoretical oxalate load rises proportionally. Hydroxyproline (present in high concentration in collagen) is metabolised via glyoxylate to oxalate. Knight 2006 showed that a 30 g gelatin dose increased urinary oxalate excretion by approximately 43% (7). For most people this is inconsequential; for anyone with a personal or family history of calcium-oxalate kidney stones, it is a legitimate reason to keep the dose modest. See our kidney stones article for the fuller treatment.
● The protein displacement effect becomes non-trivial. At 20+ g/day, collagen is contributing 15–20% of a typical adult's total daily protein intake as an incomplete protein. This can meaningfully displace better-quality dietary protein in people already at protein-intake ceilings.
Do not exceed 15 g/day. If you are considering higher doses for any reason, revisit the goal and the underlying evidence rather than escalating the dose.
Split dose or single dose?
Both work. A single 10 g serving in the morning and a split 5 g morning + 5 g evening produce broadly similar plasma peptide concentration profiles over 24 hours. Consistency of daily intake matters more than within-day distribution.
Three practical considerations. GI tolerance: some people report mild fullness or bloating with a single 10 g dose; splitting into two 5 g doses reduces this if it occurs. Convenience: a single serving is easier to remember and stick with; split dosing requires two touchpoints per day. Timing-specific goals: if tendon synthesis is your specific goal, a single 10–15 g dose one hour before training is the trial-anchored protocol (4), not a split.
For most readers, one serving per day at a consistent time is the pragmatic choice. See the timing article for the detailed timing discussion.
With food or on an empty stomach?
Both work. The Pro-Hyp and Hyp-Gly dipeptides that survive intestinal digestion do so whether the dose is taken fasting or with a meal. Most trial protocols do not specify fasting. Take it whenever fits your routine and helps consistency.
Two exceptions worth knowing. First, if you take levothyroxine, oral bisphosphonates, or tetracycline antibiotics, do not take collagen in the same window as these medications. Protein meals reduce their absorption; a 30–60 minute separation from levothyroxine and tetracyclines is sufficient, and bisphosphonates require their full 30–60 minute empty-stomach window before any food or drink other than water. Second, for tendon synthesis specifically, the Shaw 2017 protocol has the collagen taken about one hour before exercise, which happens to coincide with an empty-stomach state for many people (4).
Body weight and dose scaling
The trial evidence does not obviously scale dose to body weight. König 2018 used 5 g/day across postmenopausal women regardless of body mass (2); Yu 2023 meta-analysis trials used ~10 g/day without body-weight scaling (3). The peptide receptor effects on target cells (fibroblasts, chondrocytes, osteoblasts) appear to saturate at plasma concentrations that a 5–10 g oral dose achieves in most adults.
A practical implication: a 60 kg woman and a 90 kg man taking the same 10 g/day for skin outcomes are both operating within the trial-validated range. No adjustment is required for typical adult body sizes. For very small or very large individuals (below 45 kg or above 120 kg), no trial evidence guides adjustment either way; use the standard trial dose and assume it applies.
Timeline expectations by dose and indication
The dose determines what you can expect in theory; the duration determines when you should reasonably assess whether it is working.
● Skin outcomes: 8–12 weeks minimum. Objective changes measured by corneometer or cutometer take about this long to develop; subjective changes may be reported earlier but should be treated as suggestive until the trial timeframe is met.
● Knee osteoarthritis pain: 3–6 months. Trials assess at 12–24 weeks. Do not conclude the intervention is ineffective before three months.
● Bone density: 12 months. König 2018 assessed at 12 months (2). Bone remodelling operates on a slow timescale. Turnover markers may shift earlier (weeks to months), but DXA-measurable BMD requires the full year.
● Tendon synthesis around exercise: acute effects within hours (Shaw 2017 detected marker changes within a single session (4)). Cumulative clinical effect over weeks to months.
● Sarcopenia lean-mass changes: 12 weeks, per Zdzieblik 2015 (5). Combined with resistance training.
Special populations — dose considerations
Kidney stone formers
Anyone with a personal or family history of calcium-oxalate kidney stones should keep the dose at the lower end of what their indication needs (5 g/day if bone, 10 g/day if joints — do not exceed) and consider the kidney stones article for the fuller mitigation protocol (calcium co-administration, hydration).
Pregnancy and lactation
No specific trial evidence on dose during pregnancy. Default to caution and discuss with your obstetrician before starting any non-essential supplement. See the pregnancy article.
Older adults
Same trial-anchored doses apply. Consider the higher end of ranges (10–15 g/day) if muscle preservation is a goal alongside skin or bone.
Athletes at high training loads
The 10–15 g/day range with pre-training timing (with vitamin C) is the trial-anchored protocol for tendon-heavy training contexts. Higher doses do not have evidence support.
What we still don't know
● Whether specific molecular-weight fractions (e.g. 2–3 kDa vs 5–6 kDa) require different doses to achieve equivalent effect. Manufacturers vary; explicit dose-adjustment guidance is not available.
● Whether the plateau above 10 g/day is truly firm or whether specific indications (very high-load athletes, elderly with severe sarcopenia) might benefit from higher doses. Trials at higher doses are rare.
● Whether long-term (multi-year) daily supplementation requires periodic dose changes or breaks. No evidence guides this.
● Whether combined multi-indication users benefit from splitting the dose across timing windows aligned with each goal. Not tested.
Bottom line
Match your dose to your goal. Five grams per day for postmenopausal bone. Ten grams per day for knee osteoarthritis pain, skin, or general combined goals. Ten to fifteen grams per day for tendon recovery (taken pre-training with vitamin C) or sarcopenia prevention with resistance training. Do not exceed 15 g/day — there is no evidence advantage and the theoretical oxalate load rises. Consistency of daily intake matters more than within-day timing. Give any protocol at least three months (twelve months for bone) before assessing whether it is working. If cost matters and bone is your primary goal, do not default to the 10 g scoop — 5 g/day is what the evidence supports. See our pillar guide for the full picture.
Frequently asked questions
How many grams of collagen should I take per day?
It depends on your goal. Bone: 5 g/day. Skin: 2.5–10 g/day. Knee OA pain: 10 g/day. Tendon recovery: 10–15 g/day. Sarcopenia: 15 g/day. General wellness across combined goals: 10 g/day is a reasonable compromise. Do not exceed 15 g/day.
Is 20 g of collagen too much?
Above 15 g/day is beyond the trial-anchored range. There is no evidence advantage for higher doses and the theoretical oxalate load rises proportionally. For most healthy adults this is not acutely harmful, but there is no reason to go there and it may matter for kidney stone formers.
Is 5 g of collagen enough per day?
For postmenopausal bone: yes, 5 g/day is the trial-validated dose (König 2018) (2). For skin: within the trial-supported range (Proksch 2014 was positive at 2.5 g/day) (8). For knee osteoarthritis pain: probably not enough; trials use 10 g/day.
Can I split my collagen dose across the day?
Yes. Split dosing (5 g morning + 5 g evening) and single dosing (10 g at once) produce broadly similar plasma peptide profiles. Consistency of daily intake matters more than within-day distribution. One exception: for tendon synthesis specifically, a single pre-training dose is the trial-anchored protocol (4).
Should I take collagen with food or on an empty stomach?
Either works. The bioactive dipeptides survive intestinal digestion in both conditions. Take it whenever helps you remember consistently. Exception: separate from levothyroxine, bisphosphonates, and tetracycline antibiotics.
Does dose scale with body weight?
Not according to trial evidence. König 2018 and Yu 2023 trials used the same doses across adult body sizes. Peptide-receptor effects saturate at plasma concentrations that a standard 5–10 g dose achieves in most adults.
How long before I see results?
Skin: 8–12 weeks. Joints: 3–6 months. Bone: 12 months. Sarcopenia: 12 weeks. Give the intervention its full trial timeframe before deciding it is not working.
References
1. de Miranda RB, Weimer P, Rossi RC. Effects of hydrolyzed collagen supplementation on skin aging: a systematic review and meta-analysis. Int J Dermatol 2021. https://pubmed.ncbi.nlm.nih.gov/34553487/
2. König D, Oesser S, Scharla S, Zdzieblik D, Gollhofer A. Specific collagen peptides improve bone mineral density and bone markers in postmenopausal women. Nutrients 2018. https://pubmed.ncbi.nlm.nih.gov/29337906/
3. Yu Y, Cheng K, Zhao W, et al.. Analgesic efficacy of collagen peptide in knee osteoarthritis: a meta-analysis of RCTs. J Orthop Surg Res 2023. https://pubmed.ncbi.nlm.nih.gov/37715244/
4. Shaw G, Lee-Barthel A, Ross ML, Wang B, Baar K. Vitamin C-enriched gelatin supplementation before intermittent activity augments collagen synthesis. Am J Clin Nutr 2017. https://pubmed.ncbi.nlm.nih.gov/27852613/
5. Zdzieblik D, Oesser S, Baumstark MW, Gollhofer A, König D. Collagen peptide supplementation in combination with resistance training improves body composition and increases muscle strength in elderly sarcopenic men. Br J Nutr 2015. https://pubmed.ncbi.nlm.nih.gov/26353786/
6. Shigemura Y, Kubomura D, Sato Y, Sato K. Dose-dependent changes in the levels of free and peptide forms of hydroxyproline in human plasma after collagen hydrolysate ingestion. Food Chem 2014. https://pubmed.ncbi.nlm.nih.gov/24054241/
7. Knight J, Jiang J, Assimos DG, Holmes RP. Hydroxyproline ingestion and urinary oxalate and glycolate excretion. Kidney Int 2006. https://pubmed.ncbi.nlm.nih.gov/16988700/
8. Proksch E, Segger D, Degwert J, et al.. Oral supplementation of specific collagen peptides has beneficial effects on human skin physiology. Skin Pharmacol Physiol 2014. https://pubmed.ncbi.nlm.nih.gov/23949208/