Ten years of hip scans place coffee's bone-density signal above five cups a day

A steady dose-response curve would suggest caffeine acts on bone roughly in proportion to how much of it you drink. A flat stretch followed by a drop suggests something else, or possibly nothing beyond the noise of a heavy-intake subgroup, and the arithmetic frames the question plainly: the intake that tracked with lower density sits at about twice the intake this study cleared ^[1].

Neither mechanism the release offers was measured in these women. Co-author Ryan Liu points to catechins, abundant in tea, as compounds that may promote bone formation and slow bone breakdown ^[8]. For coffee he cites laboratory work in which caffeine interferes with calcium absorption and bone metabolism, with effects he describes as small and offset by adding milk ^[9]. That caveat is doing quiet work. A cohort that counts cups is not counting what went into them, and the release does not say whether milk was accounted for ^[15].

The design's strength sits in the paper's own title, a 10-year repeated-measures analysis ^[16]. Drinking habits were reassessed across the decade rather than frozen at enrolment, and density was read at the total hip and femoral neck, sites the authors chose for their link to fracture risk ^[3]. What repeated measures cannot do is randomise. Women who drink more than five cups a day differ from women who drink two in ways a beverage questionnaire captures imperfectly, so the analysis can rank associations without showing that cutting back would move any individual scan.

The two conditional findings are the most fragile in the set. The negative coffee association was stronger among women with higher lifetime alcohol consumption, and the tea association was more pronounced among women with obesity ^[7]. Interactions are where thin cell counts and multiple comparisons do their damage, and the release does not report whether those subgroups were specified in advance ^[15].

The population argument the authors make is fair on its own terms. Osteoporosis affects roughly one in three women over 50 and is responsible for millions of fractures a year ^[13], and with billions of people drinking coffee or tea every day, small average differences scale ^[14]. Enwu Liu frames it as small improvements in density translating into fewer fractures across large groups ^[10], while the team also says the differences were too modest to justify major changes in an individual's habits ^[11], and that nobody needs to give up coffee or start drinking tea by the gallon ^[12]. My read: the tea result is the one worth following, since it has a candidate mechanism and costs nothing to test, and the coffee result is more useful as a dose marker than as advice. What was reported is density; whether any of these women broke fewer bones went unmeasured ^[15].