In 2023, a trial of more than 17,000 people delivered the headline the makers of semaglutide had been waiting for: the drug cut the rate of heart attacks and strokes in people with obesity and heart disease. It was, by any measure, a landmark.

Buried in the same mountain of data was a smaller number that nobody printed on a billboard. Among the oldest participants — those past 75 — hip and pelvic fractures were later reported at close to five times the rate seen in the placebo group.

One number is a triumph. The other is a signal. And the second one is the subject almost nobody in the weight-loss conversation wants to slow down for.

To be clear up front: a fracture signal in one age subgroup of one trial is not a verdict. It is a flag — the kind that should prompt scrutiny, not panic. But it points at something real, something that has been true since long before these drugs existed, something the entire modern weight-loss story is built to look past.

When fat leaves the body, bone tends to leave with it.

To understand why — and to understand why that is suddenly a mass-scale question rather than a footnote — you have to start with the hormone that was supposed to end obesity and didn’t.

The Hormone That Was Supposed to End Obesity

In 1994, a research team studying a strain of grotesquely obese mice found the broken gene responsible. The gene coded for a hormone no one had known existed. They named it leptin, from the Greek leptos, “thin.”

The discovery was electric because the logic seemed so clean. Leptin is made by fat cells. The more fat you carry, the more leptin you release. That leptin travels to the brain and says, in effect: the stores are full, you can stop eating now. Here, at last, was the body’s fuel gauge — and a mouse with a broken gauge ate itself enormous. Give it leptin, and it slimmed down. The press wrote the ending before the science did: inject the “thin hormone,” melt the fat.

It failed. Spectacularly. When leptin was given to people with ordinary obesity, almost nothing happened.

The reason is the first thing you need to hold onto, because it governs everything that follows. People with obesity are not short on leptin. They are swimming in it. All that fat is broadcasting the “you’re full” signal at full volume — and the brain has stopped listening. This is leptin resistance, and it is the central cruelty of the condition. The gauge is pinned at full, and the body still feels empty.

This is why “just eat less” is such useless advice, and why so much moral weight gets dumped on people who can’t follow it. High leptin that the brain ignores doesn’t feel like willpower failing. It feels like ordinary hunger that never resolves — because the off switch is jammed.

So much for high leptin. The story gets stranger, and more important, when leptin runs low.

Why the Body Fights Back

Start losing weight — by any method — and leptin falls. Not gently, and not in proportion. As fat shrinks, leptin drops faster than the fat itself, and the brain reads that plunge the way it was built to over four million years of scarcity: famine.

The response is not subtle. Hunger climbs. The metabolism banks down — the same daily activities now burn measurably fewer calories, an effect researchers call adaptive thermogenesis. Thyroid signaling softens. The body, convinced it is starving, defends the weight it used to have with everything it has. This is the machinery that makes the last ten pounds impossible and makes the lost pounds come roaring back. It is not a character flaw. It is a thermostat.

The most elegant proof came from studies at Columbia, where people who had lost weight were given back the leptin their bodies were now missing. The famine alarm quieted. Much of the metabolic slowdown reversed. Replace the signal, and the body stopped fighting.

That finding tells you exactly what a weight-loss drug would have to do to actually work: it would have to win an argument against a system engineered, at the hormonal level, to make you lose it.

For most of human history nothing could. Then something did — which is where the skeleton re-enters the story, because the same hormone that governs this whole defense also happens to talk to bone. And here the science gets honestly unsettled.

What Leptin Does to Bone — and What Nobody Can Say Cleanly

It would be tidy to write the next sentence as: low leptin weakens bone. It would also be wrong, or at least far past what the evidence can carry.

Leptin’s relationship with the skeleton is one of the strangest open problems in the field. Working directly on bone-building cells, leptin appears to help build bone. But acting through the brain — via the hypothalamus and the sympathetic nervous system, the body’s fight-or-flight wiring — it does the opposite, putting a brake on bone formation. The landmark mouse work in the early 2000s showed these two effects can be pulled apart, and that the brain-routed, bone-suppressing signal is powerful in its own right.

The upshot is a hormone pulling in two directions at once, and the net result depends on which pathway wins. The mouse data are openly contradictory: leptin-deficient mice have shorter, thinner long bones but denser spinal bone. And in humans, a consistent skeletal signature of leptin deficiency has never been nailed down.

So the honest statement is this: leptin ties fat tissue and bone tissue together into one system, but its net effect on the human skeleton is unresolved. Anyone who tells you the bone loss of weight loss is “a leptin story” is overselling a single hormone.

Because bone loss during weight loss is not one mechanism. It is several, arriving together:

  • Mechanical unloading. Bone is built to load. Carry less weight, and the skeleton — reading the drop in force the way a muscle reads disuse — quietly downsizes to match. Less load, less bone.
  • Less estrogen. Fat tissue manufactures estrogen, a hormone that protects bone in both women and men. Less fat means less estrogen at exactly the moment bone needs it.
  • Fewer building signals. Weight loss lowers IGF-1 and shifts the bone-remodeling balance toward breakdown faster than rebuild.
  • Nutrient shortfalls. Eat much less, and protein, calcium, and vitamin D often fall short of what bone maintenance demands.
  • The leptin question, layered on top of all of it — real, but muddled.

Every one of those fires during any substantial weight loss. Which brings us to the drugs that have made substantial weight loss ordinary.

The Override

GLP-1 drugs — semaglutide, tirzepatide, the names now traded like gossip — are imitations of a natural gut hormone. They slow the stomach’s emptying and, more to the point, reach straight into the brain’s appetite centers and turn the volume down on hunger.

Read that again against everything above, and you see what they actually are. They are the first tool that reliably wins the argument the leptin system is built to make you lose. Where willpower is trench warfare against a hormonal defense, these drugs simply quiet the defense. People who spent decades white-knuckling against a jammed off-switch describe, for the first time, an off-switch that works. Semaglutide drives roughly 12 to 15 percent of body weight off. Tirzepatide pushes toward 20. Those are numbers that used to require surgery.

This is a genuine medical achievement, and nothing that follows is meant to take it back.

But the body’s defense existed for a reason, and overriding it fast has consequences the marketing does not dwell on. The weight comes off — and a meaningful share of what comes off is not fat.

The imaging studies are consistent enough to take seriously. In detailed body-composition data from a large tirzepatide trial, participants lost roughly a third of their fat mass over 72 weeks — and about eleven percent of their lean mass, the muscle that, among other jobs, pulls on bone and keeps you upright when you stumble. On the bone side, a 52-week randomized trial in people already at elevated fracture risk found that semaglutide lowered hip and spine bone density by a few percent versus placebo, with markers of bone breakdown rising while bone building failed to keep pace. A separate cohort of older adults — average age 68, three-quarters of them already osteopenic or osteoporotic before they began — lost bone at the hip, spine, and femoral neck over about three years, and the hip loss tracked the amount of weight each person shed.

That last detail is the crux, and it leads to the most important finding of all.

The Skeleton Doesn’t Ask How

If GLP-1 drugs were poisoning bone through some direct chemical effect, you would expect the bone loss to track the dose of the drug. It doesn’t. It tracks the weight.

The cleanest evidence comes from an accidental natural experiment. Researchers compared people on these drugs who had diabetes against people on them who didn’t. The two groups differ in a useful way: people without diabetes tend to lose more weight on the same drug.

The result: in the people without diabetes — the bigger losers — hip bone loss outpaced comparison groups. In the people with diabetes — the smaller losers — bone loss was no different from controls. Same drug. Different weight loss. Different skeletal outcome.

It is the weight loss driving the bone loss, not the molecule.

Which means the drug is not the villain here. The villain, if there is one, is rapid, unsupervised weight loss — and that has been extracting a skeletal toll for as long as people have crash-dieted or gone under the bariatric surgeon’s knife. GLP-1 drugs didn’t invent the problem. They industrialized it, by making dramatic weight loss available to millions of people who could never before achieve it — including a great many older adults who have the least bone to lose in the first place.

The fat was always going to cost the bone something. The drugs just sent the bill to far more people, far faster, than the medical system was watching for.

The Case for Not Panicking

It would be easy to end there, on the alarm. That would also be dishonest, so here is the other side, given its due.

These drugs are among the most consequential advances in metabolic medicine in a generation. The heart-attack and stroke reductions are real. So are the benefits for kidney disease, sleep apnea, and — increasingly — overall mortality. For a great many people, the calculus isn’t close.

Obesity is not kind to bone either. Excess weight degrades bone quality even as it pads bone density, and a heavier body that falls is its own fracture risk. Staying obese to protect your skeleton is a bad trade, and not one any serious clinician recommends.

The fracture data are mixed. For every alarming subgroup signal, there are analyses showing no clear increase, and a few hinting at protection. Bone density on a scan is a proxy; what actually breaks is a harder, slower thing to measure, and the long-term fracture numbers on these drugs are still maturing. Honest verdict: unsettled, leaning toward caution, not toward alarm.

And the emerging expert consensus is not don’t take the drug. It is: don’t take it naked. Prescribing a powerful weight-loss agent without a plan to protect muscle and bone isn’t wrong because the drug is dangerous. It’s incomplete because half the prescription is missing.

The Missing Half of the Prescription

Here is the part that turns a warning into something you can act on, and it is the most encouraging finding in the whole literature.

The bone and muscle loss of weight loss is not fixed. It bends — hard — in response to two boring, unglamorous inputs.

Load the skeleton. In studies of older adults deliberately losing weight, the ones who did resistance training — who made their muscles pull hard against bone — held onto bone and muscle that the diet-only group simply surrendered. Walking helps the heart; it does comparatively little for the hip. What preserves the hip is force: lifting, loaded carries, anything that makes bone work against meaningful resistance. Bone downsizes when unloaded and defends itself when loaded. You get a vote.

Feed the rebuild. Protein high enough to defend muscle — well above the outdated minimums, especially past 60 — plus adequate calcium and vitamin D. You cannot rebuild tissue from materials you never ate.

Neither of these is exotic. Neither requires anything the drug doesn’t already make room for. They are simply the half of the plan that rarely gets written on the script.

Picture the person in that cohort study. Sixty-eight years old. Three-quarters of them already had thinning bone before they started — they just didn’t know it, because thinning bone announces nothing until it breaks. Six months in, forty pounds lighter, feeling better than they have in a decade, they get a scan they were never told to ask for.

For some of them, that scan shows the quiet subtraction nobody warned them about.

For the ones whose clinician treated the prescription as half a plan — who sent them out with a barbell and a protein target alongside the pen — the scan looks very different. Same drug. Same weight gone. A skeleton that got a vote.

The fat was always going to leave. Whether the bone left with it was the part that was up for negotiation all along.