I drink my coffee black and I drink a lot of it, so a headline announcing that too much coffee may weaken my bones is exactly the kind of thing that stops me mid-sip. Then I read the study, and the scare fell apart in two places at once. The alarming coffee finding came from a curve the model drew through the data that failed its own test for statistical significance. And the paper everyone is quoting about brittle bones never analyzed a single fracture. It measured density on a scan and stopped there.
The dataset is genuinely good, which is what makes the spin so frustrating. Ryan Yan Liu and Enwu Liu at Flinders University pulled the Study of Osteoporotic Fractures, a cohort of 9,704 American women aged 65 and up, and tracked their total hip bone density with repeat DEXA scans across four visits over about ten years. At each visit the women reported one thing: whether they drank coffee or tea, and how many cups a day. The team then lined those cup counts up against the women’s total hip bone density readings. No broken bones were in the analysis. Bone density was the sole measured outcome, start to finish.
Here is where the numbers get quietly deflating. The tea result is the one the authors stand behind most, and I want you to sit with how small it is. Women who drank tea had higher total hip bone density by about 0.003 g/cm², which cleared statistical significance at p = 0.026. It also, in the authors’ own words, “falls below the clinical meaningfulness threshold for individual patient management.” So it is real, it is measurable, and it is almost certainly too tiny for your skeleton to notice.
The coffee alarm is flimsier still. That warning about heavy drinkers rests on a spline, a squiggly line the model bends through the scatter, that visually dipped somewhere past five cups a day. When the authors tested whether that dip was a genuine pattern or just noise, it came back non-significant (p for nonlinearity greater than 0.05), and there was no significant overall association between coffee and hip bone density at all. The single most-quoted claim in the coverage is a shape someone eyeballed on a graph that the statistics refused to back.
So why would a morning habit touch your skeleton at all? This is the part I actually enjoyed chasing. Caffeine does nudge calcium the wrong way, prompting a little more of it to leave in your urine and making your gut slightly less efficient at pulling it in from food. On paper that is a small net leak. In a living body it is so faint that a splash of milk more than covers it. Tea pulls the other way: its catechins, in lab work, seem to coax the cells that build bone and slow the cells that tear it down. The biology is plausible in exactly the directions the study saw. It is just whispering, and someone turned the whisper into a headline.
Zoom out to the whole literature and the panic has nowhere to stand. A meta-analysis in Medicina pooled 20 studies and 508,312 people, and here fracture is the endpoint, not a stand-in for it. Coffee’s hip fracture hazard ratio came in at 1.008 (95% CI 0.760 to 1.337), which is a statistician’s way of saying no effect that anyone can distinguish from chance. Tea leaned a hair protective at 0.929 (95% CI 0.836 to 1.033), and that interval crosses 1.0 too, so the hint dissolves under scrutiny. On bone density the story repeats: coffee’s mean difference was 0.020 with a confidence interval running through zero. Half a million people, and your daily cup is doing essentially nothing to your bones in either direction.
Which brings me to the move I cannot let slide. The lead author’s takeaway was that “even small improvements in bone density can translate into fewer fractures across large groups.” Fractures are the thing that actually matters, the broken hip that puts an 80-year-old in a hospital bed she too often never leaves. But this analysis did not examine one. It watched a 0.003 g/cm² flicker on a scan and let the press reach the rest of the way to broken hips. That leap, from a surrogate reading to bodies in beds, is the whole engine that turns a modest observational paper into a fright. To their credit, the authors took no industry money and declared no conflicts, and I read the science itself as honest. The spin is a separate machine, and it runs on whatever earns the click.
Every limit in the study points the same way, against the scare. The cup counts came from women recalling their own habits, with no record of how large the mug or how strong the brew. The cohort was almost entirely White women over 65, so it says nothing about a 40-year-old’s espresso. And a density reading is a single snapshot, not a verdict on where a bone is headed. Stack those together and the honest headline is not “coffee weakens bones.” It is “we measured something too small to feel and cannot connect it to a single break.”
So this is my conclusion, cup in hand. I am not letting a graph squiggle that flunked its own significance test change how I start my morning. If I were 70 with thinning bones, I would put my energy into calcium, vitamin D, protein, and lifting something heavy, and spend none of it counting whether my coffee was two cups or four. I am keeping every one of mine.
Sources
- Nutrients – Liu & Liu, coffee and tea intake and total hip bone mineral density in the Study of Osteoporotic Fractures, 10-year cohort (2025)
- Medicina – meta-analysis of coffee and tea intake with BMD and hip fracture, 20 studies, 508,312 people (2023)
- ScienceDaily – “Too much coffee may weaken bones, study finds,” with author quote (2026)
- News-Medical – coverage of the Flinders University coffee and tea bone-health analysis (2025)