Across 41 studies (27 in quantitative synthesis) spanning children, adolescents and adults, no beverage category was significantly associated with bone mineral density; carbonated cola/non-cola drinks came closest (β = -0.653, 95% CI -1.310 to 0.004). Fracture risk told a different story: carbonated cola/non-cola beverages showed a log odds ratio of 0.442 (95% CI 0.184 to 0.699), roughly 1.56-fold higher odds, with near-identical estimates for caffeinated and non-caffeinated versions. Osteoporosis/osteopenia estimates were also elevated for cola. GRADE certainty was very low for all three primary outcomes.
The near-identical caffeinated and decaffeinated estimates argue against caffeine as the culprit, and the mismatch between a null BMD result and a clear fracture signal suggests something beyond density, or simply confounding. People who drink cola daily tend to have lower milk intake, less physical activity, higher smoking rates and poorer overall diets, and the pooled data are observational, so displacement of calcium-rich drinks is a more plausible explanation than phosphoric acid leaching bone. Heterogeneous exposure definitions, such as merging 'cola/non-cola' categories or grouping packaged juices with soda, further blur interpretation. The very-low GRADE rating means these estimates could shift substantially with better data. Practically, nothing here justifies alarm, but it adds modest weight to limiting habitual soda intake, especially where it replaces milk or fortified alternatives during skeletal growth. Editorially, this is confirmatory and hypothesis-generating, not decision-changing; the network meta-analysis framework is useful, but it cannot rescue weak primary studies.