How much of the fatigue is actually drug-attributable, and what did the placebo arms report?

Accepted answer

The trial data indicate that fatigue is modestly drug-attributable and substantially deficit-attributable, because the placebo arms reported a great deal of it. The drug-attributable increment is roughly 3-6 percentage points of incidence, against a placebo background of 5-11%. That is a real signal and a small one, and it means most of what you are describing is probably not pharmacological.

The reported figures

Read the placebo column carefully. Somewhere between one in twenty and one in twelve people randomised to a saline injection and a lifestyle intervention spontaneously reported fatigue over a year. Those people were also on an energy deficit, also being weighed fortnightly, also filling in diaries, and also often losing a few kilograms. The placebo arms in these trials are not untreated controls; they are lifestyle-intervention arms, which is exactly why their fatigue rate is not zero.

So the arithmetic on your situation: of the people reporting fatigue in the STEP 1 active arm, roughly half would have reported it on placebo. The remainder is the drug-attributable share, and even within that share most is plausibly mediated by the drug's effect on intake rather than by a direct effect on anything.

Is there a direct pharmacological fatigue mechanism?

No well-established one, and the honest answer is that the evidence for a direct central fatigue effect is thin. What exists:

• Nothing in the receptor distribution obviously predicts fatigue. There is no analogue of the area postrema story for nausea, where a specific structure with a specific function explains the symptom.
• Plausible indirect routes are numerous and adequate. Reduced energy availability, reduced carbohydrate intake and hence lower muscle glycogen, dehydration, electrolyte drift, sleep disruption from reflux or nocturnal nausea, reduced iron and B12 intake, loss of lean mass, and reduced habitual activity. Any of these will produce exactly what you describe.
• The dose relationship is weak for the established agents, which argues against a direct effect. It appears stronger for the more aggressive multi-agonists, where a larger deficit is achieved faster, which is consistent with the indirect explanation rather than the direct one.

The one caveat worth naming: fatigue is a non-specific symptom that is poorly captured by spontaneous adverse-event reporting. Nobody in these trials was administered a validated fatigue instrument at fixed intervals. So the figures above are soft, and it is possible the true incidence is higher in both arms. What is more robust is the ratio between arms, because both arms were ascertained the same way.

Your specific numbers, which are where the answer probably lies

Three things in your post are more likely explanations than the drug:

1. The deficit size. 16 kg in five months is roughly 0.74 kg/week. At an approximate 7,700 kcal per kg of tissue, that implies a mean daily deficit of about 0.74 × 7,700 / 7 ≈ 814 kcal. Your intake of 1,400 kcal therefore implies an expenditure around 2,200 kcal. A sustained 800 kcal deficit is a large physiological signal: it reduces resting metabolic rate, lowers T3, lowers spontaneous activity, and reduces training capacity. Fatigue on that is not a side effect, it is the expected consequence.

2. Carbohydrate availability. Protein at 110 g is 440 kcal. If fat is a typical 30-35% of your intake, say 450 kcal, that leaves roughly 500 kcal, or about 125 g, of carbohydrate per day. That is low for anyone training, and low muscle glycogen presents specifically as everything costing more, with afternoon and late-session collapse. Your description of afternoons being worst fits this better than it fits a drug effect.

3. Fluid. 1.8 L of drinks on a 1,400 kcal intake means total water availability is materially lower than it was before, because food water fell with intake. Under-replacement of a few hundred millilitres a day produces fatigue and reduced exercise tolerance well before it produces thirst.

What I would do, in order

1. Get a panel. Five months, 16 kg, no bloods. Full blood count, ferritin, B12, folate, sodium, potassium, magnesium, calcium, urea, creatinine, thyroid function, HbA1c and liver enzymes. Not because a drug effect is likely but because iron deficiency, B12 deficiency and thyroid change are all common, all present exactly like this, and all treatable. Do it before optimising anything else.
2. Raise fluid to a measured target for two weeks and see what changes. Cheapest intervention with the fastest read-out.
3. Add carbohydrate around training without changing total intake much, if training performance is a large part of what bothers you.
4. Consider slowing the rate of loss. 0.74 kg/week at your intake is aggressive, and the fatigue is the price signal. A smaller deficit is a legitimate response and does not mean abandoning the outcome.
5. Only then consider whether the drug itself is contributing, which is the hypothesis that is both hardest to test and least likely to be the main term.

edited 25 Oct 2024 by nkem_obiora — clarified the distinction between purity and content