To explore their theory the researchers used two Swedish cohorts. The first included 61,433 women, and the second 45,339 men (both after all exclusions), and with them the authors explored the cohorts' relationships between milk intake (ascertained by way of just two food frequency questionnaires with the women and only one with the men) and mortality and hip fractures over a 20.1 year mean follow-up period. As far as possible confounders go, they considered hormone replacement therapy, menopausal status, and whether or not the women had children, smoking status, body mass index, vitamin D supplementation, alcohol consumption, a healthy dietary pattern, cortisone use, leisure time physical activity, education, and marital status.
Results wise the researchers found that the mean intake of milk was roughly a glass a day for both men and women but that some folks were drinking more than 3 glasses daily, and some less than 1. During the 1,231,818 person years of follow up 15,541 women died as did 10,112 men. Fracture wise there were 4,259 hip fractures among women and 1,166 among men. As far as milk goes, the researchers found those women consuming >3 glasses a day vs. women consuming <1 had a hazard ratio of total mortality of 1.93 (and 1.60 for hip fractures). Comparing male heavy vs. light milk drinkers, the risk found was less pronounced (though still significant according to the researchers) at 1.10 for mortality, but nothing for hip fractures. But bear in mind, those hazard ratio span an average of 20 years of follow-up.
So should you stop drinking milk because it's going to kill you after it leads you to have a hip fracture? These sorts of studies are brutal. No doubt the researchers do their utmost to try to ensure they control for confounders, but given the food data utilized are often single contact food frequency questionnaires (as was here) which are then extrapolated to represent decades worth of consumption, and that lives in general are complicated, it really is difficult to get too worked up about the results. For instance it's difficult to put a great deal of stock in this particular study's FFQ validities as they report the frequent milk drinkers as consuming nearly 40% more calories than the infrequent, and yet the weights of those two groups are within about 3lbs of one another.
So I remain unconvinced that milk is in and of itself markedly toxic, but that said, drinking 3 or more glasses of the stuff a day isn't something I'd be advising anyone to do simply on the basis of liquid calories not being particularly filling. Given that there are many studies that dance around milk providing ever so slight health benefits as well as many studies that dance around milk providing ever so slight health risk, I'm still comfortable with my belief that milk is neither a magic fairy brew nor a devil's broth and consequently should be consumed in the name of loving it (in the smallest quantities you need to like your life), but not in the name of health, and that eating your dairy, from a health perspective, is likely preferable to drinking it.