Menopause before 40 and depression: hazard of death 5.04 in US study

Picture a woman in her early fifties at a routine visit. Her periods ended at 38, a fact that went into her chart years ago and mostly stayed there. She has also been low for a long while, the kind of low a nine-question screening form would score as clinically significant. Nobody has ever put the two facts side by side and asked what they might add up to over the next decade or two.

A paper published online on 29 September 2026 in Menopause, by Cheng-Yu Hou, Yung-Chieh Tsai, Tian-Ni Kuo and Ing-Luen Shyu, does exactly that arithmetic on a national scale. Its authors note that premature ovarian failure and depression are each linked to poorer health outcomes, but that their combined association with long-term mortality was unclear. The study is an attempt to fill that gap, and its headline number is large enough that the way it was built matters as much as the number.

The data come from the National Health and Nutrition Examination Survey, known as NHANES, for the cycles from 2005 to 2018, linked to national mortality records that run through 2019. NHANES is a recurring US survey designed so that its participants stand in for the whole population, and the authors report that the 4,754 women analysed represent 27.2 million US women.

Eligible women were postmenopausal, aged 40 or older, and had usable information on three things: the age at their final menstrual period, their depression status, and follow-up for mortality. Premature ovarian failure was defined as a last menstrual period before age 40. Depression was assessed with the Patient Health Questionnaire-9, a nine-item symptom questionnaire, with a score of 10 or higher counted as clinically significant depression.

Women were sorted into four groups: neither condition, premature ovarian failure alone, depression alone, and both. The researchers used Cox proportional hazards models, a standard method for comparing how quickly deaths accumulate across groups over follow-up, to estimate adjusted hazard ratios for death from any cause. Death from cardiovascular disease was examined as an exploratory secondary outcome. The abstract does not list which factors the models adjusted for, how many women were in each group, or how long follow-up ran on average.

Compared with women who had neither condition, women with premature ovarian failure alone had an adjusted hazard ratio for all-cause mortality of 2.50 (95% confidence interval 1.72 to 3.64). Women with depression alone had a ratio of 1.64 (1.14 to 2.38). Women with both had a ratio of 5.04 (2.18 to 11.62).

A hazard ratio compares the rate at which deaths occur in one group with the rate in the reference group at any given moment. A ratio of 2.50 means that, in these data and after the authors' adjustment, deaths accumulated at about two and a half times the pace seen in the reference group; it does not mean a 2.5-fold lifetime chance of dying, and it is not a probability for any single woman. The abstract gives no absolute figures, such as how many women in each group died or the share who died during follow-up, so the size of the gap in everyday terms cannot be read from it.

Each interval excludes 1.0, the value that would mean no difference, and the authors describe the two conditions as independently associated with higher mortality. For cardiovascular death the abstract says only that exploratory analyses showed a similar direction of association, and that the findings require confirmation in larger studies with sufficient numbers of cardiovascular deaths. No cardiovascular hazard ratios appear in the abstract.

The 5.04 figure is the one likely to be repeated, and it is the least precise number in the paper. Its confidence interval runs from 2.18 to 11.62, more than a five-fold spread between the low and high ends. A wide interval of this kind usually reflects a small group in a statistical model: when few women have both an early final period and depression, and fewer of those die during follow-up, the estimate swings widely. The phrase the authors use, highest observed risk, is accurate and carefully chosen. It describes what was seen in this sample, not a fixed property of the combination.

The design adds a second limit. NHANES observes women once and then follows them through death records; nobody was assigned to have an early final period or to be depressed. Women who stop menstruating before 40 may differ from other women in ways the model cannot fully capture, from earlier health problems to the reason the periods stopped, and depression measured by a questionnaire may reflect either a long-standing condition or a passing one. The abstract does not say how age at the final period was obtained or whether it distinguishes natural from surgical causes, so those questions stay open here.

There is also a direction-of-effect question. Depression can follow serious illness as well as precede it, and a shorter remaining life can come from conditions that are themselves tied to both exposures. An association of this kind is a reason to look more closely, not a demonstration of mechanism.

The authors say the combined association of early loss of ovarian function and depression with long-term mortality had been unclear. A national sample in which the two can be examined together, as this one does, is a way to begin answering that, and the group with both conditions is one that larger studies can look at separately rather than averaging it into either condition alone.

The paper's own conclusion is measured: the two conditions were independently associated with higher all-cause mortality, women with both had the highest observed risk, and the cardiovascular results are exploratory. Read at that strength, it adds a hypothesis that larger studies, with more women in the combined group and more cardiovascular deaths, are in a position to test.

This is a paper whose main strength and main weakness sit in the same sentence. Using a nationally representative survey linked to death records gives the estimates a reach that a clinic-based series could not, and the authors behave carefully with the results: they call the combined group's risk the highest observed, they label the cardiovascular findings exploratory, and they say plainly that confirmation needs larger studies. The two single-exposure results, 2.50 and 1.64, have intervals that sit clearly above 1.0 and are the steadier part of the picture. The 5.04 for women with both conditions is the number that will travel, and its interval, 2.18 to 11.62, is the reason to hold it loosely. What the paper cannot supply, from what the abstract shows, is the machinery to turn any of this into a statement about a particular woman: no absolute risks, no group sizes, no list of adjustments. The sensible reading is that an early final period and depression each tracked with higher mortality in this US sample, that having both tracked with more, and that the size of that extra is still unknown.