The GLP-1 Muscle Protection Protocol

There is a version of GLP-1 therapy that goes extremely well. The patient loses significant weight, their A1c improves, their blood pressure drops, their cardiometabolic risk profile is transformed. And there is a version that goes badly — not in the dramatic sense of serious adverse events, but in a quieter, more metabolically insidious way. The patient loses weight, yes, but a disproportionate fraction of what they lose is muscle. They arrive at their goal weight lighter but functionally weaker, metabolically compromised, with a body composition that in some respects is more fragile than when they started. They step on the scale and feel success. Their bloodwork may even look better. But the body underneath has been quietly hollowed out in ways that will take years to reverse — if they ever do.

This guide is for people who want the first version of that story, not the second.

GLP-1 muscle loss is not a fringe concern or a scare story from people who don’t understand the medication. It is a documented, physiologically predictable consequence of aggressive caloric restriction, and GLP-1 agonists are among the most powerful appetite-suppressing tools ever deployed in clinical practice. The more effectively the drug works, the more aggressively it suppresses intake, the higher the risk that the deficit falls on lean tissue as well as fat. Understanding why this happens — and what can interrupt it — requires understanding some molecular physiology, a few landmark clinical trials, and a brutally honest assessment of what the supplement industry does and does not have right. All of that is here.

---

The physiology of muscle loss on GLP-1

Let’s start with the data you may not have heard if your prescriber spent ten minutes discussing Ozempic with you. The STEP 1 trial, the landmark phase 3 study that established semaglutide 2.4mg as an effective obesity treatment, showed that participants lost approximately 15.3% of body weight over 68 weeks — a result that rightly generated enormous clinical excitement. (Wilding et al., 2021, NEJM) What received considerably less emphasis was the body composition data: roughly 25% of that total weight lost was lean tissue, not fat. One in four pounds lost was muscle, bone density, organ mass, and fluid — not adipose.

Twenty-five percent sounds manageable. It isn’t, for two reasons. First, that figure comes from a trial population that was monitored, supported, and presumably counseled on diet. Real-world data from body composition studies in patients not following a structured protein and exercise protocol show lean tissue loss rates closer to 35–40% of total weight lost. Second, the figure compounds. If you lose 40 pounds and 10–16 of them are muscle, you have lost a physiologically meaningful quantity of the tissue that most directly governs your metabolic health, insulin sensitivity, and functional capacity.

The mechanism is not mysterious, and it is not specific to GLP-1 drugs. At the cellular level, a GLP-1-induced caloric deficit is functionally indistinguishable from any other caloric deficit. The body does not know you achieved your energy gap by taking a weekly injectable. It knows that energy intake has dropped below expenditure, and it responds to that signal in ways shaped by millions of years of evolution: by catabolizing whatever macronutrient substrates are most available, and by downregulating the molecular machinery responsible for building new protein.

The specific mechanisms are worth understanding, because they are also the molecular targets of the interventions described later. When caloric intake drops significantly — particularly when protein intake drops as a consequence of reduced meal volumes — the activity of mTORC1, the master kinase complex that drives muscle protein synthesis, falls. mTORC1 is activated by the amino acid leucine through the Sestrin2/GATOR2 complex; leucine essentially acts as a molecular sensor for amino acid sufficiency. (Churchward-Venne et al., 2014, J Physiol) When meal sizes shrink under GLP-1-induced satiety, the per-meal leucine content frequently falls below the threshold required to fully activate this pathway. The synthesis signal weakens. Simultaneously, the catabolic signal strengthens: myostatin, a negative regulator of muscle mass, is upregulated during caloric restriction; the ubiquitin-proteasome pathway, which degrades muscle protein, becomes more active; and if activity levels decline along with appetite — a common pattern in the early weeks on GLP-1 as fatigue and GI side effects peak — the loss of mechanical loading removes the most potent stimulus for muscle protein synthesis that exists.

This combination — reduced synthesis signal, increased degradation signal, reduced mechanical stimulus — is why muscle loss on GLP-1 therapy is not just possible but likely in the absence of deliberate countermeasures. Preserve muscle on semaglutide requires addressing all three simultaneously. That is what this protocol does.

---

Why muscle loss matters more than the scale shows

Here is a scenario worth sitting with. A patient starts semaglutide at 240 pounds, a BMI of 37, significant visceral adiposity, pre-diabetes. Over a year, they lose 30 pounds. Their doctor is pleased. Their A1c has improved. Their blood pressure is better. They look meaningfully different. And of those 30 pounds, 10 were muscle.

That 10 pounds of muscle loss is not cosmetic. It is metabolically consequential in ways that are difficult to convey when the scale number is still moving in the right direction. Muscle tissue is metabolically active — each pound burns approximately 6 kcal per day at rest, which sounds trivial until you recognize that this represents the ongoing cost of maintaining that tissue’s structural and enzymatic machinery. Losing 10 pounds of muscle depresses resting metabolic rate by roughly 60 kcal per day, making weight maintenance incrementally harder for the rest of that person’s life. More critically, skeletal muscle is the primary site of insulin-mediated glucose disposal in the human body. When circulating insulin rises after a meal, the dominant clearance mechanism is uptake into skeletal muscle. (DeFronzo et al., 1981, J Clin Invest) A person who loses significant lean mass while treating insulin resistance has undermined the very mechanism they need to work better. The metabolic contradiction is stark: GLP-1 drugs improve insulin sensitivity partly through weight loss, but lean mass semaglutide depletion works directly against that improvement.

Then there is the rebound problem, and it is serious. The STEP 4 trial examined what happened to participants who discontinued semaglutide after 20 weeks of treatment. Over the subsequent 48 weeks, they regained weight — two-thirds of what they had lost, on average. (Rubino et al., 2021, JAMA) But the body composition of the regained weight was not the inverse of the loss. People regain fat preferentially, not muscle. The lean tissue lost during treatment is not restored proportionally when weight is regained. The metabolic outcome of the cycle — drug-induced weight loss followed by off-drug regain — is a net deterioration in body composition: less muscle, more fat, worse visceral distribution than baseline. This is not an argument against GLP-1 therapy. It is an argument for doing it correctly, so that the foundation you preserve during treatment is the one you keep.

The framing of “I just want to lose weight” fundamentally misunderstands what GLP-1 therapy is accomplishing at its best. The goal is not a lower number on a scale. The goal is a metabolically healthier body — less visceral fat, preserved or improved insulin sensitivity, functional strength, sustainable cardiovascular fitness. Muscle is load-bearing infrastructure for that goal. Losing it while claiming metabolic improvement is a category error.

---

The non-negotiable: resistance training

Nothing in the supplement section of this protocol substitutes for mechanical load on skeletal muscle. Not creatine, not HMB, not leucine, not protein shakes. This is not a caveat or a disclaimer — it is a physiological fact. The primary signal that drives muscle protein synthesis is mechanical tension. When muscle fibers are loaded under sufficient tension and that tension is progressively increased over time, the mechanosensing pathways (including FAK/Src signaling, mTOR activation via mechanical stimulation independent of leucine, and satellite cell activation) respond by building more contractile tissue. When mechanical tension is absent or insufficient, no amount of nutritional support can fully compensate. (Burd et al., 2010, J Physiol)

Progressive overload is the principle: the load placed on a muscle must increase over time for adaptation to continue. This does not mean adding weight every session — that is not sustainable indefinitely and is a common point of misunderstanding for beginners. It means that over weeks and months, the total mechanical stress imposed should increase: through more weight, more repetitions, more volume, shorter rest periods, or improved technique that more effectively loads the target tissue. Any of these mechanisms will work. The specific programming matters less than the principle.

The minimum effective dose for resistance training in a caloric restriction context is two sessions per week, using compound movements. Compound movements are exercises that load multiple large muscle groups simultaneously, producing the most anabolic stimulus per unit of training time. What this means in practice: a squat pattern (goblet squat, leg press, back squat, front squat), which loads quads, glutes, and hamstrings simultaneously; a hip hinge pattern (conventional deadlift, Romanian deadlift, hip thrust), which emphasizes posterior chain — glutes, hamstrings, erectors; a horizontal push (bench press, incline dumbbell press, push-up with added load), for chest, anterior deltoid, triceps; a horizontal pull (barbell row, dumbbell row, chest-supported row), for mid-back, lats, rear deltoids, biceps; a vertical pull (lat pulldown, pull-up, assisted pull-up), emphasizing lats and upper back. These five movement patterns, performed twice weekly with progressive overload, cover the major muscle groups and produce sufficient mechanical stimulus to preserve lean mass during a caloric deficit.

The landmark evidence for the combined diet-plus-exercise approach in exactly this population is Villareal et al.’s 2011 paper in the New England Journal of Medicine. (Villareal et al., 2011, NEJM) The trial randomized older obese adults to diet alone, exercise alone, diet plus exercise, or control. The diet-plus-exercise group had significantly better outcomes on lean mass preservation, physical function, and metabolic markers than either intervention alone. The practical lesson is not subtle: diet without resistance training produces inferior lean mass outcomes. This is true for older adults, and it is true for anyone undergoing significant caloric restriction.

One important adjustment for GLP-1 users specifically: training volume should be kept moderate, particularly during dose escalation. Training volume — sets per muscle group per week — drives muscle protein synthesis upward, but it also increases recovery demands. Recovery capacity is diminished during caloric restriction, and GLP-1 dose escalation often coincides with the most significant reduction in caloric intake. Eight to twelve working sets per major muscle group per week is a reasonable starting volume for most people. Three to four sets per movement, two movements per muscle group pattern, two to three days per week. This is not an advanced athlete’s program. It is not meant to be. It is meant to provide sufficient mechanical stimulus to prevent catabolism while being recoverable on reduced caloric intake.

---

Protein: the foundational intervention

Before creatine, before HMB, before leucine powder, before any supplement discussed in this guide — protein. Adequate protein intake is the single most evidence-supported nutritional intervention for lean mass preservation during caloric restriction, and it is the one most frequently underachieved by GLP-1 users.

The target: 1.6 to 2.2 grams of protein per kilogram of bodyweight daily. This is not a range to aim for the middle of. The lower bound — 1.6g/kg — is the floor below which research consistently shows meaningful lean mass loss regardless of other interventions. (Morton et al., 2018, BJSM meta-analysis of 49 RCTs) The upper end — 2.2g/kg — is where research in caloric restriction contexts suggests the best lean mass outcomes. For practical purposes, shoot for 1.8–2.0g/kg minimum. For a 180-pound (82kg) person, that is 148–164 grams of protein per day. On an unrestricted diet, that is achievable without deliberate planning. On a GLP-1-induced caloric restriction of 1,200–1,500 kcal per day, it requires deliberate, active management of every meal.

The molecular reason protein matters at the per-meal level, not just the daily total, is the leucine threshold. mTORC1 activation — the signal that initiates muscle protein synthesis — requires approximately 2.5 to 3 grams of leucine per meal to reach maximal stimulation. (Norton & Layman, 2006, J Nutr) Leucine does not accumulate across meals; it must be present in sufficient quantity at each feeding to trigger the synthesis signal. A meal providing 40 grams of high-quality protein (say, chicken breast) contains roughly 3.2 grams of leucine — above threshold. A meal providing 20 grams of protein contains roughly 1.6 grams of leucine — below threshold. The synthesis response is substantially blunted. Daily protein targets matter, but distribution across meals matters nearly as much.

This creates the central GLP-1 nutrition challenge. Appetite suppression shrinks meal sizes. Nausea during dose escalation makes some foods intolerable. The per-meal protein content that previously was achieved without thought now requires strategic planning. Practical strategies that work: protein-first eating, where protein is consumed before vegetables or carbohydrates at every meal, ensuring that even if satiety cuts the meal short, the leucine-rich component has been consumed; liquid protein sources at meals when solid food volume is limited — Greek yogurt (17–20g per cup), cottage cheese (25g per cup), and protein shakes (20–35g per serving) are volume-efficient; scheduling the largest meals for the time of day when appetite is highest (for most GLP-1 users, this is morning or early afternoon, as appetite tends to be most suppressed in the evening, when social convention suggests the largest meal).

Protein sources ranked by leucine content: whey protein isolate (approximately 10–11% leucine by weight, the highest commercially available), eggs (8.5%), beef and chicken (7–8%), salmon and tuna (7–8%), Greek yogurt (8%), cottage cheese (7%). These are the sources to build meals around. Collagen protein — which is marketed aggressively and finds its way into many GLP-1 “support” products — contains approximately 1.5% leucine and lacks tryptophan entirely. It cannot meaningfully stimulate muscle protein synthesis. Its role in preserving lean mass during GLP-1 therapy is essentially zero. This point will be revisited.

---

Creatine monohydrate: the highest-evidence supplement

[Strong evidence]

Creatine monohydrate is the most studied ergogenic supplement in the history of exercise science. It has been examined in over 500 peer-reviewed studies spanning more than four decades. Its safety profile is exceptionally well-established, its mechanism is well-understood, and its effects in the context of resistance training — including during caloric restriction — are consistent across trial populations. For GLP-1 users, it belongs in the protocol from week one.

The primary mechanism: creatine is stored in muscle as phosphocreatine, where it serves as an immediate energy buffer during high-intensity effort. When ATP is depleted rapidly — as occurs during the initial seconds of a heavy resistance training set — phosphocreatine donates a phosphate group to ADP to rapidly regenerate ATP, extending the duration of maximal-intensity effort. (Rawson & Volek, 2003, J Strength Cond Res) The practical effect is that creatine supplementation allows more total training volume at a given intensity — more reps at a given weight, more total sets before fatigue becomes limiting — and more total training volume drives more muscle protein synthesis. This is the indirect mechanism. The direct mechanism is cellular volumization: creatine increases intramuscular water content, and cell swelling is itself an anabolic signal that upregulates protein synthesis and downregulates protein degradation. (Haussinger et al., 1993, Biochem J)

Dosing is straightforward: 5 grams daily, taken at any time, with or without food. The timing myths — creatine must be taken immediately post-workout, creatine must be taken with fast-digesting carbohydrates, creatine is more effective in the morning — are not supported by the evidence. (Antonio & Ciccone, 2013, J Int Soc Sports Nutr) Consistency matters; timing does not. Loading phases — 20 grams per day divided into four doses for five days — do accelerate muscle creatine saturation and will produce faster initial results, but are entirely optional. Standard dosing at 5 grams per day reaches saturation in approximately 28 days.

Now the caveat that genuinely matters and is consistently mishandled by both patients and practitioners: creatine supplementation raises serum creatinine by approximately 0.1 to 0.2 mg/dL. This elevation is not a sign of kidney damage. It does not represent reduced glomerular filtration. It is a predictable, mechanistically understood, reversible artifact of increased phosphocreatine metabolism in muscle — creatinine is the breakdown product of creatine and phosphocreatine, so more muscle creatine means more creatinine in circulation. (Poortmans & Francaux, 1999, Int J Sports Med) The literature on creatine and renal function in healthy individuals is unambiguous: no adverse effect on kidney function has been demonstrated at standard doses, including in long-term studies of up to five years.

The problem is not the biology. The problem is what happens when a patient who has been supplementing creatine for two months shows up for a metabolic panel and the creatinine reads 1.3 when their baseline was 1.1. The prescriber, not knowing about the supplement, flags it. The patient becomes frightened. The creatine is discontinued. Occasionally, nephrology gets involved. This entire cascade — which consumes clinical resources and produces genuine patient anxiety — is completely avoidable by doing one thing: telling your prescriber you are taking creatine before your next blood draw, not afterward. This is the most important administrative action associated with creatine supplementation. Do it proactively, in writing if possible so it is in the chart, before the blood draw. Not after the creatinine comes back elevated.

---

HMB (β-Hydroxy β-methylbutyrate): the anti-catabolic agent

[Moderate evidence]

HMB is less well-known than creatine and is surrounded by a thicker layer of marketing noise, which makes it harder to evaluate fairly. Let’s try. HMB is a metabolite of leucine — specifically, approximately 5% of ingested leucine is converted to HMB in the liver via the KIC (alpha-ketoisocaproate) pathway. Its mechanism of action is meaningfully different from both creatine and dietary protein, and that difference is precisely what makes it relevant to the GLP-1 context.

While creatine and leucine primarily act to stimulate anabolic pathways (ATP regeneration and mTORC1 activation respectively), HMB’s primary documented mechanism is inhibition of catabolism — specifically, inhibition of the ubiquitin-proteasome pathway, the major cellular system by which muscle protein is degraded during catabolic states such as caloric restriction, illness, bed rest, and disuse. (Eley et al., 2008, Clin Nutr) This anti-catabolic mechanism is most clinically meaningful precisely when catabolism is most active — which is exactly the scenario GLP-1-induced caloric restriction creates. An agent that slows the breakdown side of the protein turnover equation is a different intervention than one that increases the synthesis side, and in a situation where synthesis signals are already suppressed by energy deficit, the catabolic side may be the more productive target.

The Wilson 2014 trial (Wilson et al., 2014, J Int Soc Sports Nutr) is the most cited RCT for HMB in a training context: a 24-week study using the free acid form of HMB in trained adults showing significant lean mass preservation versus placebo during caloric restriction. The results were impressive enough to drive significant commercial interest. However, intellectual honesty requires noting that one of the lead authors had disclosed commercial interests in HMB products at the time of publication. This does not invalidate the findings — the disclosure was made, the methodology was peer-reviewed — but it means the study should be read with the same critical skepticism applied to any industry-adjacent research, rather than cited with the automatic authority it often receives in supplement circles.

The evidence picture for HMB is this: meaningful mechanistic plausibility, positive RCT data that is somewhat heterogeneous in quality, and a specific relevance to caloric restriction contexts that distinguishes it from supplements primarily tested in hypercaloric or maintenance-calorie conditions. That is a “moderate evidence” designation, not “strong,” and not “preliminary.” (Nissen & Sharp, 2003, J Appl Physiol)

Dosing: 3 grams per day, ideally split across two to three doses taken with protein-containing meals. The calcium salt form (standard in most supplements) has somewhat weaker evidence than the free acid form used in the Wilson trial, but it is substantially cheaper and more widely available. The free acid form has better bioavailability but costs considerably more. Starting with calcium HMB and assessing response is reasonable. Most importantly, timing relative to the GLP-1 protocol matters: HMB is most valuable when catabolism is most active. Begin during or shortly before dose escalation, when the caloric deficit is deepening.

---

Leucine enrichment: the mTORC1 trigger

[Moderate evidence]

Leucine occupies a somewhat unusual position in sports nutrition: it is both an essential amino acid (present in dietary protein) and a molecular signaling agent with effects that can, in some contexts, be separated from its contribution to total protein intake. The distinction matters for GLP-1 users.

mTORC1 is the central anabolic kinase complex that initiates muscle protein synthesis. Among the essential amino acids, leucine is the most potent activator of mTORC1, acting through the Sestrin2/GATOR2 complex to signal amino acid sufficiency to the mTORC1 pathway. (Anthony et al., 2000, J Nutr) The critical feature of this signaling is its threshold nature: there is a minimum leucine concentration per meal required to maximally stimulate mTORC1 activity, estimated at approximately 2.5 to 3 grams. Below this threshold, even modest leucine intake blunts the synthetic response. Above it, additional leucine provides diminishing returns. This is not a continuous dose-response relationship — it is more like a switch that is either on or off, with the threshold determining which state applies. (Norton & Layman, 2006, J Nutr)

The GLP-1 complication is straightforward once you understand the threshold mechanism. A person who previously ate a chicken breast providing 40 grams of protein (approximately 3.2 grams of leucine — above threshold) is now, under semaglutide, eating half that amount: 20 grams of protein from half a chicken breast, providing 1.6 grams of leucine — below threshold. The total protein may be adequate if distributed correctly across multiple meals. But each individual meal is providing a subthreshold leucine stimulus, and the synthetic response is blunted at each feeding. The person can be hitting their daily protein target while still experiencing inadequate per-meal muscle protein synthesis stimulation.

There are two practical solutions. The first, and preferable, is to choose leucine-rich protein sources and maintain adequate per-meal protein even as total volume shrinks: whey protein shakes, Greek yogurt, eggs, and meat provide enough leucine per gram of total protein that moderate servings still approach the threshold. The second is to supplement leucine powder — typically 2 to 3 grams per dose — taken alongside a protein-containing meal when serving sizes have become small enough to fall below threshold. Leucine powder has a strong and somewhat unpleasant taste; it mixes adequately into smoothies or yogurt. For people who are eating very small volumes and struggling to maintain muscle, this direct supplementation approach is evidence-supported and inexpensive.

What will not solve this problem is collagen protein, and the fact that it is sold aggressively in the GLP-1 supplement market needs to be addressed directly. Collagen is approximately 1.5% leucine by weight — one-sixth the leucine density of whey. It lacks tryptophan, making it not a complete protein at all. Consuming collagen as a primary protein source while attempting to preserve lean mass is not a neutral choice — it actively uses up appetite capacity and digestive resources that should be occupied by complete protein sources. (Shaw et al., 2017, Nutrients)

---

Vitamin D3 and K2: bone density and muscle function

[Moderate evidence]

The STEP trials did not measure bone density. This is a genuine and underappreciated gap in the literature on GLP-1 outcomes — one that will take years to properly fill as longer-term follow-up data accumulates. (Jensen et al., 2022, Obesity) What we do know from the broader caloric restriction literature is that significant weight loss impairs bone mineral density, particularly in older adults, and that the risk is amplified when the weight lost includes lean mass (muscle provides mechanical loading stimulus for bone maintenance). Adding GLP-1-induced caloric restriction to the established obesity-BMD relationship — where BMD tends to be elevated due to load — creates a situation where bone health may deteriorate even as other metabolic markers improve.

Vitamin D3 is relevant to this concern through two distinct mechanisms. The first is its established role in calcium absorption: vitamin D3 deficiency reliably impairs intestinal calcium uptake, and calcium is load-bearing infrastructure for bone mineral density. The second is its direct role in muscle function: vitamin D receptors (VDRs) are expressed in skeletal muscle tissue, and vitamin D signaling influences muscle protein synthesis and contractile function. (Bischoff-Ferrari et al., 2004, JAMA) Clinical studies consistently show that vitamin D deficiency is independently associated with reduced muscle strength and increased fall risk in older adults — not just through bone fragility, but through the muscle pathway directly.

The obesity-vitamin D connection is particularly relevant for GLP-1 patients: vitamin D is a fat-soluble vitamin, and excess adipose tissue acts as a sequestration reservoir, sequestering vitamin D in fat cells and reducing circulating 25-OH-D concentrations. This means that many patients beginning GLP-1 therapy are already vitamin D insufficient before the drug reduces their dietary vitamin D intake further through caloric restriction. The target serum 25-OH-D is above 40 ng/mL — most endocrinologists now agree that the old threshold of 20 ng/mL was too conservative. Achieving this typically requires 2,000–5,000 IU of D3 daily depending on baseline serum levels.

Vitamin K2 in the MK-7 form is included here not because it independently builds muscle, but because it complements D3 supplementation for bone health through the carboxylation of osteocalcin and matrix GLA protein, processes that direct calcium to bone rather than vascular tissue. (Vermeer et al., 2004, Thromb Haemost) When supplementing D3 at doses sufficient to correct deficiency, adding K2 at 100–200 micrograms per day is a reasonable precaution against calcium being deposited in soft tissue rather than bone — a theoretical but mechanistically supported concern with long-term high-dose D3 use.

---

Magnesium glycinate: the overlooked deficiency

[Moderate evidence]

The conversation about GLP-1 therapy and micronutrient deficiency tends to focus on the dramatic — protein targets, creatinine concerns, the leucine threshold. Magnesium gets less attention, which is unfortunate because subclinical magnesium insufficiency is almost guaranteed in patients eating significantly reduced food volumes, and its symptoms overlap substantially with side effects commonly attributed to the drug itself.

Magnesium is required as a cofactor for over 300 enzymatic reactions in human metabolism, including virtually every step of ATP synthesis and all of the enzymes of the glycolytic and oxidative phosphorylation pathways. (Vormann, 2003, Mol Aspects Med) It is also required for muscle contraction and relaxation — the SERCA pump that returns calcium to the sarcoplasmic reticulum after a muscle contraction is magnesium-dependent. Clinical manifestations of magnesium insufficiency include fatigue, impaired sleep quality, muscle cramps (particularly nocturnal leg cramps), anxiety, constipation, and headache. Look at that list and compare it to the commonly reported side effects of GLP-1 therapy: fatigue, sleep disruption, GI disturbance, muscle cramps. The overlap is substantial. Many GLP-1 users who are suffering from these symptoms and attributing them to the medication may actually be experiencing, in part, a correctable micronutrient deficit.

The form of magnesium supplementation matters significantly, and this is where most people go wrong. Magnesium oxide is the most common form in cheap supplements and fortified foods — it has approximately 4% bioavailability, meaning 96% of the elemental magnesium in the supplement is excreted without being absorbed. Magnesium citrate has intermediate bioavailability and is commonly used as a laxative, which is sometimes useful but not ideal given existing GI sensitivity on GLP-1. Magnesium glycinate is a chelated form — magnesium bound to glycine — with substantially higher bioavailability, significantly lower GI irritation risk, and the additional benefit that glycine itself has mild anxiolytic and sleep-supportive properties. (Abbasi et al., 2012, J Res Med Sci) For GLP-1 users, magnesium glycinate taken at 200–400mg of elemental magnesium before bed is the appropriate choice.

This is not a muscle-building supplement. Correcting magnesium insufficiency will not produce the hypertrophic or anti-catabolic effects of creatine or HMB. What it will do is restore normal ATP synthesis, improve sleep quality (which has significant downstream effects on hormonal milieu and recovery capacity), reduce muscle cramps that interfere with training, and potentially ameliorate fatigue symptoms that otherwise would limit training adherence. It is a deficiency-correction intervention, and in GLP-1 users eating reduced volumes, it is almost universally warranted.

---

What doesn’t work: the supplement industry’s GLP-1 products

This section is going to name problems directly, because the supplement industry has identified GLP-1 users as a lucrative market and is deploying a predictable set of tactics: rebranding existing products with new GLP-1 marketing, creating “support stacks” that combine reasonable and unreasonable ingredients in sub-clinical doses, and exploiting the genuine anxiety patients feel about muscle loss and metabolic disruption to sell products with weak or absent evidence for the specific concern being sold.

Let’s go through the main offenders.

Collagen protein. This has been mentioned already but deserves its own treatment because its marketing is particularly aggressive in the GLP-1 space. Collagen protein is sold as a “joint-supporting,” “gut-healing,” “skin-firming” protein that is “gentle” on the digestive system — all of which appeals to GLP-1 users dealing with GI side effects. None of these claims bear on the specific problem of lean mass semaglutide depletion, and collagen’s fundamental limitation as a protein source makes it actively counterproductive when used as a primary or significant protein source. It lacks tryptophan — making it an incomplete protein — and its leucine content of approximately 1.5% by weight is one-sixth that of whey. (Shaw et al., 2017, Nutrients) A 20-gram serving of collagen protein provides approximately 0.3 grams of leucine — about 10% of the per-meal threshold for mTORC1 activation. Collagen is fine as a minor skin and joint supplement. It has no meaningful role in muscle preservation.

BCAAs (branched-chain amino acids). BCAA supplements — leucine, isoleucine, and valine — were enormously popular in sports nutrition for decades. The theoretical basis was reasonable: leucine activates mTORC1, so supplementing BCAAs around training should enhance muscle protein synthesis. The problem, which took years to properly establish in the literature, is that BCAAs alone cannot fully stimulate muscle protein synthesis because complete protein synthesis requires all essential amino acids, not just three. (Wolfe, 2017, J Int Soc Sports Nutr) When protein intake is adequate — which is the goal of this protocol — adding BCAAs provides no additional lean mass benefit. The leucine threshold issue is better addressed by choosing leucine-rich complete protein sources or supplementing leucine powder, not by consuming incomplete BCAA blends that lack the remaining essential amino acids needed for protein construction.

Commercial GLP-1 “support stacks.” Several companies have launched products specifically marketed to GLP-1 users, typically featuring 10–15 ingredients in a proprietary blend. The individual ingredients are often not unreasonable in isolation — one might find creatine, HMB, leucine, vitamin D, magnesium, berberine, electrolytes, and digestive enzymes in the same capsule. The problem is not necessarily the ingredient list. The problem is dosing, sequencing, and diagnostic utility. Proprietary blends frequently underdose individual ingredients to keep costs manageable while hitting a palatable price point. More importantly, starting everything simultaneously eliminates any ability to identify which ingredient is providing benefit, which is causing GI irritation (a genuine concern given the existing GI burden of GLP-1 therapy), or which is interacting with other components. The phased introduction approach outlined in this guide — starting creatine first, adding magnesium, then HMB, then assessing — is more rational and produces actionable information.

Thermogenics and fat burners. This may seem too obvious to merit mention, but fat-burning supplements are aggressively cross-marketed to GLP-1 users as “amplifiers” of the drug’s weight loss effects. Taking stimulant-based thermogenics on top of GLP-1-induced energy deficit is redundant at best and physiologically concerning at worst. GLP-1 agonists, particularly semaglutide and tirzepatide, modestly elevate resting heart rate — a class effect seen in clinical trials. Combining this with the sympathomimetic cardiovascular effects of stimulant fat burners (caffeine at high doses, synephrine, yohimbine) creates unnecessary cardiovascular stress without meaningful additive fat loss beyond what the GLP-1 is already producing. The caloric deficit on GLP-1 therapy is already substantial. There is nothing to amplify.

---

Tracking: how to know if the protocol is working

The scale is a profoundly inadequate metric for assessing the outcomes of this protocol. It provides one number — total body weight — that conflates fat mass, lean mass, water, and everything else in a single figure that tells you nothing about the composition of what you are losing or retaining. A person who loses three pounds of fat and gains one pound of muscle over a month is making excellent progress. The scale shows a net loss of two pounds, which looks mediocre. The same scale might show a loss of four pounds in someone who lost three pounds of muscle and one pound of fat — a terrible outcome that looks better.

Body composition measurement options exist on a spectrum of accuracy and practicality. DEXA (dual-energy X-ray absorptiometry) is the gold standard for outpatient assessment: it produces separate measurements of fat mass, lean mass, and bone mineral density across body regions, with a precision error of approximately 1–2% for lean mass in good conditions. It takes about 10 minutes, delivers minimal radiation exposure (less than a transatlantic flight), and costs $50–150 at imaging centers and some university athletic facilities. For anyone on GLP-1 therapy who is serious about tracking lean mass outcomes, a DEXA scan at baseline and at three-month intervals is the appropriate monitoring tool. (Glickman et al., 2004, Obesity Research)

Bioelectrical impedance devices — including high-quality consumer devices like the Withings Body Comp or InBody units at gyms — are directionally useful for tracking trends over time but should not be treated as precise measurements. Hydration status substantially affects readings, and the algorithms used by most consumer devices are population averages that may not reflect an individual’s actual composition accurately. Consistent measurement conditions (same time of day, same hydration state, same phase of menstrual cycle for women) improve their utility as trend monitors if DEXA is not accessible.

Functional strength markers are underused and highly valuable. The question is not just whether body composition is being preserved, but whether functional capacity is. Can you lift 10% more on the compound movements you started with three months ago? Can you do more repetitions at the same weight? Can you walk up stairs without fatigue that wasn’t there before? These are valid clinical endpoints. Losing 30 pounds while maintaining all your compound lift numbers is an excellent outcome. Losing 30 pounds while your squat dropped 40% is a warning sign regardless of what the scale says.

Protein tracking, at minimum for two to four weeks, is non-optional for the majority of GLP-1 users. Research consistently shows that people overestimate their protein intake by 30–40% when relying on recall alone. A person who believes they are hitting 140 grams per day is typically eating 90–100 grams. Apps like Cronometer, Protein Tracker, or MyFitnessPal make this easy, and a two-to-four week logging period — even if not maintained indefinitely — establishes a concrete baseline and reveals the protein gaps that need to be addressed.

For labs: document your creatine supplementation in the chart before any metabolic panel. Track creatinine at baseline and quarterly. If you are over 45 and on long-term GLP-1 therapy, a baseline DEXA with annual follow-up is appropriate for monitoring bone density, given the absence of bone density data from the major GLP-1 trials.

---

The protocol at a glance: phased implementation

One of the consistent failures of supplement protocols in the GLP-1 space is the “start everything immediately” approach — the commercial stack that arrives in a box and is meant to be initiated all at once. This creates multiple problems: GI side effects cannot be attributed to any specific agent; if something works, you don’t know what; if something doesn’t work, you don’t know what; adherence suffers because the overall burden of pills and powders is high. The phased approach below is rational, evidence-sequenced, and diagnostically useful.

Week 1 of protocol (or week 1 of GLP-1 therapy, ideally concurrent): Start creatine monohydrate at 5 grams per day. Start protein tracking with a food logger. Implement protein-first eating at every meal. Establish baseline body weight and, if possible, circumference measurements (waist, mid-thigh, hip) or DEXA if accessible. Notify your prescriber in writing that you are starting creatine supplementation and ask them to note it in your chart prior to your next metabolic panel.

Weeks 3–4: Add magnesium glycinate at 200–400mg elemental magnesium taken before bed. This timing coincides with the period when GI side effects from GLP-1 escalation are often peaking and appetite suppression is deepening — the magnesium addresses emerging subclinical insufficiency from reduced food volume. Assess your protein log results: are you hitting 1.6g/kg minimum? If not, identify the specific meals where leucine-rich protein is lowest and address those first.

Weeks 5–8: Add HMB at 3 grams daily split across two doses with meals. This is the period when dose escalation is typically advancing and the caloric deficit is deepest — the catabolic window where HMB’s anti-proteolytic mechanism is most relevant. Initiate or intensify resistance training if not already underway: two sessions per week, compound movements, progressive overload as described above.

Month 2: Add vitamin D3 (2,000–5,000 IU based on estimated baseline) with K2 at 100–200 micrograms MK-7 form. Get a DEXA baseline if available, or establish circumference measurements. Assess training progress: are you maintaining or increasing strength on compound movements?

Month 3: First comprehensive assessment. Blood draw — ensure creatine use is documented. Review protein log data for the month. Compare body composition metrics to baseline. Assess functional strength numbers. Optimize based on what the data shows: if protein intake is consistently sub-threshold, prioritize protein-first strategies more aggressively; if lean mass is dropping despite protein targets being met, increase training volume or frequency moderately; if fatigue is limiting training, check magnesium and D3 status.

What to skip: Collagen as a primary protein source. BCAAs (unless total daily protein is genuinely inadequate and BCAA supplementation is a stopgap — but fix the protein first). Commercial GLP-1 stacks. Fat burners or thermogenic supplements. Any product that cannot tell you clearly which mechanism it is targeting and at what dose.

The protocol at scale is not complicated. The complexity is in the execution: being consistent about protein at every meal when appetite is suppressed, training twice a week when fatigue is real, taking the creatine every day without missing, getting the DEXA done, logging the food. The supplements are simple. The behavioral execution is where the work is, and where most people do not do it.

---

A note on long-term perspective

Muscle preservation during GLP-1 therapy is not a twelve-week project. Many patients will be on these medications for years — possibly indefinitely, given the weight regain data following discontinuation. The interventions described here are not short-term acute measures. They are long-term practices: resistance training twice weekly, adequate protein at every meal, creatine every morning, annual or semi-annual DEXA monitoring. The framing of “doing the protocol” implies a finite effort with a defined endpoint. The more accurate framing is that these are maintenance practices for a body under ongoing metabolic intervention.

This matters because it changes the calculus on effort and sustainability. A resistance training program that is optimal but unsustainable is worse than one that is merely adequate but maintained for years. Two sessions per week at moderate volume, done consistently for three years, produces better lean mass outcomes than four sessions per week for six months followed by dropout. Find the protein sources you actually like eating. Find the time of day for training you will actually keep. Find the form of tracking — DEXA, circumference, strength logging — that you will actually use.

The worst outcome in GLP-1 therapy is not the one where the drug doesn’t work well enough. It is the one where it works too well, the scale moves impressively, and the patient arrives at goal weight having lost the metabolic infrastructure that would allow them to sustain that outcome. Preserved muscle is not a bonus. It is the foundation. Everything else — the cardiometabolic improvements, the insulin sensitivity, the functional mobility, the ability to maintain weight without pharmacological support — rests on it.

Do the protocol. Track the data. Adjust over time. This is not more complicated than that.

---

All citations reference peer-reviewed publications. PubMed IDs available on request. This guide does not constitute medical advice. Consult your prescribing physician before starting any supplement regimen, particularly creatine given its effects on serum creatinine levels.

References cited: (Wilding et al., 2021, NEJM) · (Churchward-Venne et al., 2014, J Physiol) · (DeFronzo et al., 1981, J Clin Invest) · (Rubino et al., 2021, JAMA) · (Burd et al., 2010, J Physiol) · (Villareal et al., 2011, NEJM) · (Morton et al., 2018, BJSM) · (Norton & Layman, 2006, J Nutr) · (Rawson & Volek, 2003, J Strength Cond Res) · (Haussinger et al., 1993, Biochem J) · (Antonio & Ciccone, 2013, J Int Soc Sports Nutr) · (Poortmans & Francaux, 1999, Int J Sports Med) · (Eley et al., 2008, Clin Nutr) · (Wilson et al., 2014, J Int Soc Sports Nutr) · (Nissen & Sharp, 2003, J Appl Physiol) · (Anthony et al., 2000, J Nutr) · (Jensen et al., 2022, Obesity) · (Bischoff-Ferrari et al., 2004, JAMA) · (Vermeer et al., 2004, Thromb Haemost) · (Vormann, 2003, Mol Aspects Med) · (Abbasi et al., 2012, J Res Med Sci) · (Shaw et al., 2017, Nutrients) · (Wolfe, 2017, J Int Soc Sports Nutr) · (Glickman et al., 2004, Obesity Research) · (Haussinger et al., 1993, Biochem J)