The Longevity Desk
13 min readInsulin resistance

L-Carnitine Injectable: The Route With the Muscle Evidence Is the One You Swallow

Five hours of infusion, at roughly ten times the blood carnitine the label calls normal, moved muscle carnitine by nothing unless insulin was also pinned high by a second drip. The only protocol I found that raised it outside that laboratory setup was 24 weeks of capsules taken with 160 grams of sugar a day.


What it is

Carnitine is a small molecule the body builds from lysine and methionine and also takes from red meat. Levocarnitine is the L form, PubChem CID 10917. Not a peptide. It sits on peptide shelves because that is where injectables sell. The approved injection is CARNITOR, 1 gram in a 5 mL vial, approved 16 December 1992 under NDA 020182 for intravenous use only.

The mechanism claim is that carnitine shuttles long-chain fats across the inner wall of the mitochondrion, the compartment where fuel gets burned. It has to reach the inside of muscle cells to do that, and it does not drift in: a transporter called OCTN2 pumps it uphill. The label puts absorption of a swallowed dose at about 15 percent, and Harper and colleagues measured 16 percent at 2 grams and 5 percent at 6 grams by mouth in six healthy volunteers. (PMID 3234464) The gate is not the gut.

What the trials found

Stephens and colleagues, FASEB Journal 2006. Eight healthy men, two randomised visits, five hours of intravenous L-carnitine each time, holding blood carnitine near 500 micromoles per litre, roughly ten times the 40 to 50 the label calls normal. Insulin sat at fasting level on one visit, about 7.4 milliunits per litre, and was pinned at 149.2 on the other by a second infusion, a laboratory procedure called a clamp. Muscle carnitine, by biopsy, rose from 22.0 to 24.7 millimoles per kilogram of dry muscle on the high insulin visit, and not at all on the other. (PMID 16368715) The abstract's second sentence says studies to date had failed to raise it in healthy humans by either route, and the group's review a year later said the same. (PMC2075186)

Bruls and colleagues ran a six hour clamp of the same kind in eight healthy young lean sedentary men, infusing 28 mg per kilogram of carnitine, always on top of a fat emulsion. Plasma carnitine rose, muscle carnitine did not move. (PMID 32976523)

The one protocol I found that worked outside a clamp was swallowed: Wall and colleagues gave 14 healthy men 2 grams of L-carnitine L-tartrate with 80 grams of carbohydrate twice a day for 24 weeks. Muscle carnitine rose 21 percent from baseline and work output in the performance trial rose 11 percent, both against no change in controls. (PMID 21224234)

Every weight-loss meta-analysis I found is of capsules: 37 randomised trials, 2,292 people, a 1.21 kilogram difference, and no separation from placebo in waist circumference or body fat percent. (PMID 32359762)

Intravenously I found two, from one group in China, both inside a fast. The randomised one, single-blind, meaning the patients did not know which they got, put 30 patients with metabolic syndrome, 15 per arm, through a seven day protocol of five days of modified fasting after two days of calorie restriction, on 4 grams a day against saline, and weight fell 4.6 kilograms against 3.2. (PMID 25424121) The other, a retrospective chart review, 40 per group after matching, found 4.05 against 3.25 kilograms and gives neither the dose nor the fast's length. (PMID 31126550) Under interventions for levocarnitine, ClinicalTrials.gov lists 171 completed studies, and screening the titles I found no intravenous fat-loss or performance trial in healthy adults.

The rest of the intravenous record is hospital medicine. In the RACE trial, 250 patients in septic shock got 6, 12 or 18 grams over twelve hours against saline, and the 18 gram arm did not reduce organ failure at 48 hours. (PMID 30646314) The other half is cardiac, the one place the vein has a positive human signal. Thirteen controlled trials and 3,629 patients after a heart attack pooled to 27 percent lower all-cause mortality, on a margin that stops just short of no effect one way of measuring it and lands on it exactly the other. (PMID 23597877) The largest single trial there, 2,330 patients, missed its primary endpoint, death or heart failure at six months, 9.2 against 10.5 percent. (PMID 16685128) The vein there was a five day loading dose of 9 grams a day. (PMID 25044037)

The two routes carry different evidence, and the split runs both ways

The labels say it first: the tablet is indicated for primary systemic carnitine deficiency and the injection is not, nor is a generic from another sponsor. Chewcharat and colleagues then pooled eight randomised trials and 224 dialysis patients on low blood pressure during dialysis. Split by route, the five intravenous trials with 177 participants did not reach significance while the three oral trials with 47 did. (PMID 35834513) That oral subgroup is not a result to lean on: the statistic measuring how much trials disagree came out at 88.3 percent, about as much as three trials can produce. The intravenous side, with nearly four times the participants, did not clear the line.

The other direction runs against the capsule. Koeth and colleagues showed in Nature Medicine that gut bacteria convert swallowed carnitine into trimethylamine, which the liver turns into trimethylamine N-oxide, TMAO for short, a compound associated in observational work with cardiovascular risk. In a human challenge with a 250 mg capsule, omnivores made more TMAO than vegans or vegetarians, and in mice dietary carnitine increased atherosclerosis unless the gut bacteria were suppressed. (PMID 23563705)

The label already knew, warning that chronic high oral doses in badly damaged kidneys may accumulate trimethylamine and TMAO. Bain and colleagues gave six dialysis patients 1 gram by mouth daily for twelve days and TMAO roughly doubled, still climbing at the end. (PMID 17073580)

So the infusion should skip a pathway the capsule feeds, which is inference from two literatures and one commentary rather than a measurement. (PMID 25878774)

How it compares

Approved?Human evidence behind itWhat that evidence covers
The injection, for its labelYes, since 1992, for two usesDialysis trials and the septic shock trial. I did not read the inborn-error literatureCarnitine deficiency, not fat in healthy people
The clinic dripNoTwo studies, 110 patients, both inside a supervised fastRoughly 1 kg of extra weight loss over five days of fasting
The compounded shot into muscleNo, compounded under section 503ANone I found for the carnitine componentNothing I could find
CapsulesApproved as an oral drug, also sold as a supplement37 randomised trials, 2,292 people, pooledAbout 1.2 kg of weight difference

Where the dose came from

The label doses are the only ones with a derivation: 50 mg per kilogram for the inherited metabolic diseases, and for dialysis 10 to 20 mg per kilogram of dry body weight after each session, tied to a pre-dialysis plasma carnitine below the normal 40 to 50 micromoles per litre. That is 700 to 1,400 mg per session in a 70 kilogram adult, three times a week. Medicare will not pay to start it without that blood test, three months on dialysis and one of two named clinical signs, and stops paying if nothing has improved six months after starting.

The clinic number I could not verify at all, and I am reporting rather than endorsing it. Neither clinic page I read publishes a milligram figure. The vendor pages that do give numbers say 250 to 1,000 mg into muscle, daily or three times a week, a different route and a far higher frequency than the weekly or fortnightly drip the clinics describe. One low-quality vendor-facing aggregator states that no peer-reviewed randomised trials evaluate that range, attributing it to protocols passed around on forums.

My reading, and it is inference rather than derivation, is that these are packaging. The compounded product sold to clinics is 500 mg per millilitre in a 30 mL vial, making 250, 500 and 1,000 mg half a millilitre, one and two; the approved product is 1 gram in a 5 mL vial, where round gram figures are whole and half vials. Empower's Lipo-C carries carnitine at 50 mg per millilitre, 1 mL into muscle once or twice weekly, so 50 to 100 mg a week, a seventh to a fourteenth of one dialysis dose in a 70 kilogram adult even at the top, on a page saying the FDA has not evaluated it.

What could go wrong

The label's only serious warning is allergic reaction: anaphylaxis, swelling of the voice box, and airway spasm, mostly in dialysis patients, some within minutes of infusion. On overdose it reports no cases of toxicity in people and warns of diarrhoea at large doses, and its only numbers are rodent: 5.4 grams per kilogram intravenously kills half of rats, 19.2 grams per kilogram by mouth half of mice.

My take

The regulatory gap is ordinary. The interesting part is that the pitch has the routes the wrong way round. On the endpoint the rationale rests on, muscle carnitine in a healthy person, the capsule is the route with the result and the drip the route with the null outside a laboratory clamp. Fixing absorption when the bottleneck is uptake is like widening the driveway when the garage is already full.

What nags at me is that the one argument for the injection is a safety argument, not a fat loss one, so nobody selling it makes it. I could not reach the FDA's warning letter database to see whether it has acted here.

If you have seen an L-carnitine drip advertised, I want to know whether the page gave a milligram figure at all, and whether it named the route of the studies it pointed to.

Frequently asked

Is injectable L-carnitine FDA approved?

Yes, and for two things that are not fat loss. CARNITOR levocarnitine injection was approved in December 1992 for an inherited metabolic disease that causes carnitine deficiency, and a later supplement added carnitine deficiency in people with end stage kidney disease who are on dialysis. Nothing about weight, energy or performance appears in the label.

Does an L-carnitine IV drip raise carnitine in muscle?

Not on its own, in the one experiment that measured it directly. Eight healthy men were given five hours of intravenous L-carnitine on two separate visits, enough to hold blood carnitine at roughly ten times the 40 to 50 micromoles per litre the label calls normal, and muscle carnitine did not change unless insulin was also held artificially high by a second infusion. The only protocol I found that raised muscle carnitine outside that laboratory setup was swallowed: 14 healthy men taking 4 grams a day with 160 grams of carbohydrate a day for 24 weeks, reported as a 21 percent rise.

Is injecting L-carnitine better than taking capsules?

It depends which problem you mean. Injecting does put far more carnitine into the blood, since the FDA label puts absorption of a swallowed dose at about 15 percent, but the blood is not where the shortfall is, and the one infusion experiment that measured muscle directly moved it only while insulin was held artificially high by a second drip, which is a laboratory procedure and not something a clinic does. The argument nobody in the sales channel makes is that an injection should bypass the gut bacteria that turn swallowed carnitine into a compound linked in observational work to cardiovascular risk. That is inference from the mechanism and from one published commentary, not something I found measured head to head.

Where does the L-carnitine injection dose used in clinics come from?

I could not find a derivation for it, and this is reporting rather than a recommendation. The two clinic pages I read publish no milligram figure at all, and vendor protocol pages that do publish numbers give 250 to 1,000 milligrams into muscle, daily or three times a week, as a circulating convention with no dose finding trial behind it. One low-quality vendor-facing aggregator states outright that no peer-reviewed randomised trials evaluate that range and attributes it to community protocols on internet forums. The only doses with a derivation are the label's: 50 milligrams per kilogram for the inherited metabolic diseases, and 10 to 20 milligrams per kilogram of dry body weight after each dialysis session, adjusted against a blood test.

What are the reported risks of levocarnitine injection?

The label's one serious warning is allergic reaction, including anaphylaxis, swelling of the voice box and airway spasm, reported mostly in dialysis patients and sometimes within minutes of an intravenous dose. Seizures have been reported by either route, in people with and without a prior seizure history. On overdose the label reports no cases of toxicity in people, says the drug is easily removed by dialysis, and notes that large doses may cause diarrhoea; its only numeric toxicity values are lethal doses in rodents, 5.4 grams per kilogram intravenously in rats and 19.2 grams per kilogram by mouth in mice.

Is L-carnitine banned for athletes?

The molecule is not on the World Anti-Doping Agency's 2026 Prohibited List, and the word carnitine does not appear anywhere in that 26 page document. The delivery method can be prohibited on its own: any intravenous infusion or injection totalling more than 100 mL in a twelve hour period is banned at all times outside hospital treatment, surgery or a diagnostic investigation, and levocarnitine's own label describes mixing the drug into 500 mL bags.


This content is for educational and informational purposes only and is not medical advice. Compounds discussed may be research compounds, or approved drugs discussed outside their approved uses. Nothing here should be used to diagnose, treat, cure, or prevent any disease. Always consult a qualified healthcare professional before starting any peptide, supplement, or protocol. Individual responses vary. Do not self-administer compounds without proper medical supervision.

ShareEmailBlueskyXReddit

Get the next one in your inbox

Free. One email a week at most. Unsubscribe in one click.

Double opt-in · No spam · Unsubscribe anytime

More in Metabolic and GLP-1