Smoking has long been associated with an early onset of menopause (1–2 years earlier than nonsmokers on average) and has also been connected to POI [24–29]. Studies have also demonstrated a possible connection between stress and chronic disease and accelerated ovarian aging [23]. For example, the Study of Women Across the Nation (SWAN) analyzed cross-sectional data from 16,000 women and found statistically significant associations between POI and arthritis, diabetes, poorer self‐reported health, higher BMI, osteoporosis, severe disability, single marital status, lower education level, difficulty paying for basics, and smoking [23]. Additional studies have found that women with socioeconomic stress, HIV infection, and drug abuse tend to undergo menopause at earlier ages [30, 31]. A Chinese study using data on 31,955 women from the Shanghai Women’s Health Study found that women who were less than 46 years old at menopause were more likely to be current smokers and alcohol consumers, unemployed, single, and of lower socioeconomic status in addition to having fewer live births than women who underwent menopause at an older age [32].

Most women who develop POI have a history of normal puberty and menses, but family history of POI in a first-degree relative is fairly common and has been reported in 10–15 % of cases [33]. Data is limited on racial and ethnic differences in the development of POI. The SWAN study demonstrated a 1.0 % prevalence of POI in Caucasian women, compared to 1.4 % in African American women and Hispanics. POI was less common in Japanese (0.1 %) and Chinese women (0.5 %) compared to Caucasians, African Americans, and Hispanics; however, this was only statistically significant for Japanese women [29].

Approximately 5 % of cases of POI can be attributed to autoimmune disease, 60–80 % of which are thought to be of adrenal autoimmune origin [6]. One of the first associations made between autoimmune disease and POI was the discovery that POI could precede development of adrenal autoimmune disease (Addison’s) by 8–14 years [34]. Addison’s commonly occurs in concurrence with other endocrine diseases, collectively named the polyglandular autoimmune syndrome (PGA) [35]. Two distinct forms exist. PGA-I includes mucocutaneous candidiasis, hypoparathyroidism, and Addison’s, and has co-occurring ovarian failure in approximately 60 % of cases in women. PGA-II is the most common immunoendocrinopathy and includes Addison’s and hypothyroidism and/or insulin-dependent diabetes mellitus (IDDM). Approximately 10 % of women with PGA-II have POI [36, 37].

Two autoantibodies typically associated with Addison’s are adrenal cytoplasmic autoantibodies and steroid-cell antibodies [38]. The steroid-cell antibody found in Addison’s is thought to cross-react with the hilar, theca, granulosa, and corpus luteal cells in the ovary [39]. This can be observed histologically in the ovaries as lymphocytic and plasma cell infiltrate concentrated around steroid producing cells [6]. It is estimated that steroid-cell antibodies can be found in 100 % of Addison’s patients with primary amenorrhea and 60 % of Addison’s patients with secondary amenorrhea [36, 39, 40]. Furthermore, 40 % of Addison’s females with positive steroid-cell antibodies but normal menses developed ovarian failure within 10–15 years [36]. Positive steroid-cell antibodies are found in 60–80 % of patients with APGS I and 25–40 % with APGS II, somewhat consistent with the percentage of POI cases reported with these syndromes [36, 37].

Strong evidence exists for autoimmune POI in patients with co-existing adrenal disease. It is also thought that autoimmunity may play a larger and not yet fully understood role in POI, as 10–30 % of women with idiopathic POI have coexisting autoimmune disease [41]. For example, women with POI have been found to have a higher prevalence of thyroid disease than the general population (14–27 %), pernicious anemia (4 %), Type I diabetes (2–2.5 %) and myasthenia gravis (2 %) [42, 43]. Antinuclear antibodies, rheumatoid factor, and systemic lupus erythematosus (SLE) have also been found to occur with more frequency in POI patients compared to the general population. One study found the presence of antiovarian antibodies in 16 of 19 (84 %) patients with SLE [44]. Thus, evidence supports a link between POI and autoimmune disease, but the relationship has yet to be fully defined.

Over 80 % of POI cases are idiopathic; however, as yet unidentified genetic causes of POI are thought to make up a large portion of spontaneous POI cases [12]. Genetic causes could be chromosomal (involving the X-chromosome or autosomes), multigenic, or the result of a single gene mutation [45]. Of the known genetic causes that make up approximately 10 % of total cases of POI, it is estimated that X-chromosome abnormalities such as Turner Syndrome represent 13 % of cases [46, 47].

Heart disease is the number one killer of women, and it has long been thought that premature menopause increases the risk of cardiovascular disease [58]. Studies in the 1970s and 1980s linked bilateral oophorectomy at an early age to heart disease but found an inconsistent link between natural menopause and heart disease [59–61]. However, a 2001 analysis of the SWAN cross-sectional survey of 14,620 women found that self-reported age of natural menopause occurred 1.4 years earlier in women with heart disease than in women with no reported heart disease [24].

Several additional studies have linked early menopause to increased heart disease but often these findings were confounded by patient’s smoking status. For example, in the Nurses’ Health Study (NHS) , there was a significant association between POI and myocardial infarction (age-adjusted RR: 1.95; 95 % CI, 1.21–3.13); however, the association was attenuated with adjustment for smoking and statistical significance was lost altogether when only nonsmokers were included in the analysis (age-adjusted RR: 1.07; 95 % CI, 0.26–4.34) [62]. A separate study by Løkkegaard et al. of a Danish cohort reported that women younger than 40 with natural, nonsurgical menopause had a twofold increased risk of CHD (hazard ratio [HR]: 2.2, 95 % CI, 1.0–4.9) [63]. Although they did adjust for smoking in multivariable analyses, they did not report analyses stratified by smoking status; thus, it is uncertain whether analysis of nonsmokers alone would have affected this association.

Most recently, a multicenter cohort study, Multi-Ethnic Study Atherosclerosis (MESA) , examined 2509 women with no history of cardiovascular disease at baseline. Of these women, 693 had early menopause (28 %) and were found to have a statistically significant increased risk of coronary heart disease, which persisted after adjusting for age, race/ethnicity, MESA site, and traditional cardiovascular risk factors (HR: 2.08, 95 % CI, 1.17–3.70). Although 247 (36 %) of the women had early surgical menopause, adjustment of analysis for type of menopause did not significantly alter results [64].

Data on the relationship between early menopause and cardiovascular disease is mixed. Recent evidence suggests there is a link; however, further research is needed.

Multiple studies have examined the relationship between age at menopause and cardiovascular mortality. A study of a Norwegian cohort identified a weak inverse relationship between age at menopause and cardiovascular mortality, while another group from the Netherlands reported a 2 % decrease in cardiovascular mortality risk for each year that menopause was delayed [65, 66]. In a study of an American Seventh Day Adventist cohort, investigators found increased mortality due to ischemic heart disease in women with natural menopause at age less than 40 [67].

Four large studies since 2005 have examined the relationship between early menopause and cardiovascular mortality. Two of these studies found a significant association between early menopause and cardiovascular mortality, while two found no relationship. The first was a study by Mondul et al., which examined 68,154 women from the Cancer Prevention Study-II who had experienced natural menopause, were nonsmokers and had no history of hormone replacement therapy [68]. Their group found a small but statistically significant relationship between early menopause (age ≤44) and mortality from coronary heart disease after multivariate adjustment (RR = 1.09; 95 % CI, 1.00–1.18). The second study was by Hong et al. of 2865 Korean women, who found a significant association between menopause and cardiovascular mortality after age and multivariate adjustments. This significance held for all cardiovascular disease (HR: 1.53, 95 % CI, 1.00–2.39) and coronary heart disease specifically (HR: 8.77, 95 % CI, 2.07–37.16) [69].

A 2006 Japanese study of 37,965 (age ≤44) and 2014 Chinese study of 31,955 women (age ≤46.64) found no statistically significant association between early menopause and cardiovascular disease or mortality [32, 70]. Thus, no consensus currently exists on the relationship between early menopause and cardiovascular disease and mortality, as large studies continue to produce mixed results.

The evidence linking age at menopause and increased risk of stroke is currently inconclusive. Studies such as the NHS found no relationship between natural menopause and hemorrhagic or ischemic stroke [62]. A Japanese study by Baba et al. found an association between POI and stroke (HR: 2.18, 95 % CI, 1.20–5.49), but this finding may have had to do with their definition of POI, as a large proportion of these women had surgical menopause. When women with early natural menopause were analyzed independently, there was no statistically significant association with stroke (HR: 0.94, 95 % CI, 0.12–7.07) [71]. A 2009 study of the Framingham cohort found that women with natural menopause before age 42 had twice the stroke risk compared with women without early menopause after adjustment for multiple confounders including smoking (HR: 2.03, 95 % CI, 1.16–3.56). The association persisted in separate analysis of women who never smoked [72]. The MESA study also found that women with early menopause and no history of cardiovascular disease had more than twice the risk for stroke compared to women with normal menopause (and HR: 2.19, 95 % CI, 1.11–4.32) [64].

Elevated total cholesterol, low-density lipoprotein (LDL), and triglycerides, as well as low high-density lipoprotein (HDL) have all been implicated as risk factors for cardiovascular disease [73–75]. Normal menopause has been strongly linked to changes in lipid levels independent of those that occur with natural aging, placing women at increased risk for cardiovascular disease. This is particularly true for changes in LDL and total cholesterol [76–80].

Lipid abnormalities are also thought to occur in early menopause. For example, one study noted that cholesterol increased by a higher magnitude for women in early menopause compared to women in normal menopause. In the postmenopausal time period, women with early menopause had cholesterol levels similar to women who had gone through menopause at later ages [81]. Thus, it is possible that early menopause results in a more abrupt rise in lipids as well as an increased duration of lipid derangement.

There is little known about lipid profiles in women with POI. A 2008 cross-sectional study of 90 women with POI found elevated triglyceride levels and borderline low HDL cholesterol in POI women when compared to controls [82]. However, results from two recent cross-sectional studies (with 47 and 43 women with POI, respectively) were more consistent with the abnormalities noted in women with early menopause. When compared to controls, women with POI had statistically significant elevated LDL and total cholesterol, whereas no difference was found in triglycerides. One of the studies found HDL-C to actually be significantly higher in women with POI, while the other found no relationship [83, 84]. Further research is needed to better elucidate the relationship between POI and changes in lipid levels.

Worsening of insulin resistance is frequently reported with menopause, but it is difficult to tease out metabolic changes associated with aging from changes directly attributable to menopause [85]. The risk of developing metabolic syndrome (central obesity, dyslipidemia, and insulin resistance) after menopause is estimated to increase by 60 %, independent of age, BMI, income, and physical activity [86]. Worsened insulin resistance is thought to be associated with an increase in visceral adiposity during menopause; however, there is no documented association between menopause and increased risk of type II diabetes [85, 87].

Some studies haves shown that women with type I and II diabetes undergo menopause at a younger age, although others have refuted this [88–90]. Given the autoimmune pathogenesis of type I diabetes and its association with PGA-II, women with type I diabetes may be at higher risk for autoimmune premature ovarian failure. Women with POI may also be at increased risk for insulin resistance; however, evidence is limited. For example, a 1997 study of 119 women with spontaneous POI (46,XX) revealed an increased prevalence of glucose abnormalities (2.5 %) compared to the estimated population prevalence of diabetes for women within the same age range [43, 91]. A more recent cross-sectional study of 43 women with POI compared to 33 women with normal menses found that women with POI had statistically significantly elevated insulin and glucose [84]. This suggests that women with POI may be slightly more burdened by glucose abnormalities than the general population but only mildly so.

Ovarian secretion of estrogen plays an important role in maintaining bone health throughout a woman’s life and women entering menopause lose a significant amount of bone mineral density (BMD) [92]. In a cohort study of 862 women followed over 10 years, the rate of BMD and cumulative loss was greatest during “transmenopause”: 1 year prior to the final menstrual period (FMP) through 2 years after the FMP. During “postmenopause” (2–5 years after FMP) bone loss persisted but was less. Over the 10-year timeframe these women were followed, cumulative BMD loss of the lumbar spine was 10.6 %, with 7.38 % lost during transmenopause. Loss at the femoral neck was 9.1 % with 5.8 % lost during transmenopause [92]. Although this study was exclusively in women undergoing normal menopause, these findings are particularly concerning for women with POI, as accelerated bone loss will begin at a much younger age, and despite BMD loss slowing after transmenopause, POI women will spend a longer duration of their lifespan with a lower BMD.

Multiple studies have demonstrated an association between early menopause and lower BMD, and several have focused specifically on POI [93–95]. A review article by Gallagher found that, overall, the younger that menopause occurs, the higher the fracture rate and the lower the bone density will be in later life [94]. In a study of 357 women with POI, 58 % had evidence of osteopenia or osteoporosis [96]. Another study by Uygur et al. noted that women diagnosed with POI for a median of 2 years already demonstrated significantly lower BMD when compared to age-matched controls. Interestingly, these findings were present despite the fact that 91 % of participants had taken estrogen and/or progestin replacement at some point prior to the study [97]. Similarly, in a study by Leite-Silva et al. of 50 women with POI who had taken an average of 28 months of hormone replacement therapy, BMD of the femoral neck and lumbar spine was significantly lower in women with POI [98].

A larger study by Popat et al. examined 442 women with POI using 70 concurrent controls and 353 from the National Health and Nutrition Examination Survey III (NHANES III) and found that women with POI had 2–3 % lower bone density at the spine, hip, and neck [99]. Modifiable risk factors for accelerated bone loss included delay in diagnosis of POI greater than 1-year, nonadherence to estrogen replacement, vitamin D and calcium deficiency, and lack of exercise. Race was also a risk factor for low BMD in women with POI. African-American and Asian women were 3.18 and 4.34 times more likely to have Z-scores on DEXA bone scan of -2 [97]. However, the study by Greendale et al. on BMD in women undergoing normal menopause found that African-American and women with larger BMI had slower rates of bone loss compared to Caucasians, while Japanese and Chinese women experienced accelerated bone loss. Therefore, although African-American women may start at a lower BMD, the rate of bone loss during menopause may be slower than for other women. In contrast, Japanese and Chinese heritage appears to have unfavorable BMD in addition to accelerated loss during menopause [92].

Thus, there is strong evidence that women with POI are at significant risk for accelerated bone loss compared to women undergoing age-appropriate menopause. Differences among ethnic groups in BMD and rate of bone loss may exist, but further research is needed.

Evidence suggests that women with early menopause may be at reduced risk for breast and ovarian cancers. The study by Mondul et al. of 68,154 nonsmoking women with natural menopause and no history of hormone replacement therapy found that women with menopause occurring younger than age 44 had significantly reduced risk of breast and ovarian cancers after multivariate adjustments (RR = 0.68, 95 % CI, 0.56–0.82) [68]. Later age of menopause is thought to increase the risk of breast cancer due to the increased duration of exposure to ovulatory hormones such as estrogen and progesterone [100]. In women with early menopause, lifetime exposure to these hormones is less and therefore breast cancer risk is less.

Although risk of breast cancer may be decreased for women with POI and early menopause, several studies have demonstrated a positive correlation between POI and other types of cancer. For example, the Korean study of 2865 women by Hong et al. found that women with POI had double the risk of mortality from cancer, overall, compared to women with age-appropriate menopause (HR: 2.01, 95 % CI, 1.06–3.82). After multivariate adjustment, women less than 40 had 3.53 times higher risk for stomach cancer and 14.26 times higher risk for colon, rectum, and anal cancers (95 % CI; 1.30–9.56, p = 0.01) and (95 % CI; 2.42–83.96, p < 0.01) [69]. The Chinese study by Wu et al. of 31,955 women also found an association between early age of menopause and cancer risk [32]. Risk of gynecologic cancer appeared to be lower in women younger than age 46.64 at menopause but was not statistically significant (HR: 0.67, 95 % CI, 0.411.11). Risk of digestive system cancer and respiratory system cancer was significantly increased (HR: 1.43, 95 % CI, 1.11–1.84 and HR: 1.49, 95 % CI, 1.01–2.18, respectively). This relationship persisted after multivariate adjustment, which included smoking status.

Researchers have investigated the association between early age at menopause and longevity in multiple large-scale epidemiological studies, primarily in white European cohorts. In the late 1980s, Snowden et al. reported a modest increase in all-cause mortality in Seventh-Day Adventist women with natural menopause at ages less than 40, as compared to women with natural menopause at ages 50–54 (age-adjusted odds ratio: 1.95; 95 % CI: 1.24–3.07) [101]. In a separate publication using the same data, researchers estimated that each 1-year decrease in age at natural menopause before age 47 was associated with a 0.47-year earlier age at death (p = 0.04) [102].