Vitamin K Dosage: An AI, Not RDA — 90/120mcg, MK-7/MK-4 Dose Gap

By Erin Rose · Updated July 19, 2026 · Methodology

Informational summary of published research — not medical advice. ⚠️ If you take warfarin, do NOT start, stop, or change your vitamin K intake without your doctor managing it. This statement has not been evaluated by the FDA. This product is not intended to diagnose, treat, cure, or prevent any disease.

As an Amazon Associate we earn from qualifying purchases. Picks are ranked by form, dose, and cost, never commissions.

Adequate Intake, not RDA — why the distinction matters

Most vitamins on this site have an RDA (Recommended Dietary Allowance) — a number derived with enough evidence to state a precise requirement that covers 97-98% of healthy people. Vitamin K only has an Adequate Intake (AI): 90 mcg/day for women, 120 mcg/day for men (19+), set by the National Academy of Sciences Food and Nutrition Board (StatPearls, PMID 30725668). An AI is set when the evidence isn't strong enough to calculate a full RDA — it's a reasonable estimate based on what healthy populations appear to consume, not a rigorously derived minimum. Practically, this means the AI is a decent baseline target, but it's a softer number than an RDA, and it's specifically calibrated around K1's clotting role — not around any bone or artery benefit from K2. There is no established Tolerable Upper Intake Level (UL) for vitamin K in either form — unlike some other fat-soluble nutrients, no toxicity ceiling has been defined.

Food-first: K1 dominates the US diet

K1 is genuinely a food nutrient. Leafy greens — kale, spinach, collards, broccoli — are rich, reliable sources, and most people who eat any of them regularly land reasonably close to the 90-120 mcg AI without trying. This is a real point of contrast with, say, choline, where most of the US population runs meaningfully under target: vitamin K1 status is generally better. K2, on the other hand, is nearly absent from a typical Western diet — natto (fermented soybeans) is by far the richest source, with much smaller amounts in some aged cheeses and egg yolk. That's the practical reason K2, not K1, is the form people actually supplement.

The MK-7 vs MK-4 dose gap

This is the single biggest dosing trap in the category. MK-7 works at 90-180 mcg/day — the two landmark 3-year Knapen trials on bone density and arterial stiffness both used 180 mcg/day (Knapen 2013; Knapen 2015 — both in healthy postmenopausal women, and both industry-funded by VitaK/MenaQ7 vitamin-K interests, so read them as suggestive rather than settled), and a dose-response study found doses around the EU RDA improved carboxylation of both bone and artery Gla-proteins with no adverse effect on thrombin generation (Theuwissen 2012). MK-4 is a different story entirely: it clears within hours, and a bioavailability study found nutritional-dose MK-4 wasn't detectable in serum at any timepoint in any subject (Sato 2012). The Japanese osteoporosis-drug trials that showed a MK-4 bone benefit used 45 mg/day (45,000 mcg) — a pharmacologic dose (Miki 2003, n=20, 2 weeks, prescription-only in Japan). A typical retail MK-4 bottle (45-100 mcg) sits roughly 450-1,000x below that drug dose depending which MK-4 number you compare — never assume a nutritional MK-4 capsule replicates that evidence.

By goal (what the evidence actually used)

• General food-gap insurance: K1 at 45-120 mcg/day is a reasonable, low-stakes add if you're not confident you eat enough leafy greens. This is not a targeted intervention — it's closing a plausible gap near the AI.
• Bone density mechanism: MK-7 180 mcg/day for 3 years hit its primary endpoint at the lumbar spine and femoral neck but missed at the total hip (Knapen 2013). K1 at a much higher 1 mg/day improved femoral neck BMD but not lumbar spine (Braam 2003). A high-dose K1 trial (5 mg/day) was a clean null on bone density at every site (ECKO, Cheung 2008).
• Arterial stiffness mechanism: MK-7 180 mcg/day for 3 years met its primary endpoint for two stiffness measures in the full cohort; other stiffness measures improved only in a subgroup with above-median baseline stiffness (Knapen 2015).
• Honest limit: these are the doses used in trials with mixed primary-endpoint results — not a guarantee. See our hub page for the full evidence breakdown, and don't confuse a dose that "moved the biomarker" with proof of a clinical outcome.

Frequently asked questions

Related

• Vitamin K: what the evidence shows
• Best Vitamin K Supplement
• Vitamin K and Warfarin
• Vitamin K2 Dosage Guide — MK-7-specific depth
• Vitamin D Dosage Guide — the common partner

Sources