Menarche and the Risks of Incident Cancers and Mortality by Cause

If the present proposal is approved, we will apply to the National Institutes of Health for extramural funding through the PAR-12-039 mechanism. The R03 grant proposal is due November 18, 2014.

Menarche is the beginning of menstrual function and female reproductive life. Over nearly 200 years, the average age at menarche has declined in European countries and the United States, beginning as early as the 1830’s, while attained heights in women and men have increased (1). Declining trends in menarche have also been observed in many other countries, including recent reports from Israel, urban South Africa, rural China, and the Brazilian Amazon (2-5). The underlying causes of these secular trends remain unknown. While genetic factors have been found to accounting for much of the variation in menarche within populations (6), secular trends are unlikely to be explained by genetic factors alone; however, of many environmental factors studied, only psychosocial stress has been consistently shown to be associated with earlier menarche (7).

Some studies have suggested that reported age at menarche is a predictor of mortality in middle-aged and elderly women. In their study of 15,807 women from the EPIC-Norfolk cohort, aged 40–79 yr at baseline and followed between 1993-2008, Lakshman and colleagues demonstrated a 4.0% reduced risk of all-cause mortality for each year of delay in the onset of menarche, and increased risks of both cardiovascular and cancer-associated deaths in women with early menarche (8). Similarly, inverse associations have been observed between menarche and both all-cause mortality (9-12) and cardiovascular mortality (10,13-14). Early menarche has also been associated with increased risk of cardiovascular disease (8) and some of its component causes including obesity (15), hypertension (8), and diabetes (16). Early menarche is an established risk factor for breast and endometrial cancers (17, 18), while late menarche has been associated with
increased risks of ovarian cancer (19) and gliomas (20, 21). Age at menarche has also been associated with the risks of colorectal, bladder, kidney, and lung cancers, although findings for these cancers are somewhat inconsistent (22-32).

Mechanisms underlying the associations of early menarche and disease risks are not well understood and have typically been framed in terms of roles for reproductive hormones in chronic disease (13). We hypothesize instead that early menarche and increased height are markers of an early life marked by a less diverse gut microbiome, and that reduced gut microbial diversity is itself a predisposing risk factor for many chronic diseases. Recent studies have demonstrated that the composition of the gut microbiome can influence host energy balance (33), and that early life exposure to oral antibiotics can alter patterns of growth and development, with lasting impacts on adult phenotypes (34). Numerous studies have pointed to reduced diversity of gut microflora in association with obesity, metabolic syndrome, diabetes, and autoimmune and allergic diseases; links of gut microbes to cancer risks have yet to be established in humans but are plausible based upon animal models (35, 36). In a pilot study of 60 postmenopausal women (described in 37), we observed an inverse association (p=0.05) between fecal microbial diversity and menarche.

Having observed that in developed nations the gastric pathogen Helicobacter pylori is dramatically lower than in the developing world, Blaser has suggested that the gut microbiome has undergone a cumulative and progressive decline as a result of improvements in sanitation, adoption of hygienic practices, reductions in parity, and more recently, increasing dissemination of antibiotics (38). H. pylori may be considered a correlate of early life exposures to commensal microbiota and positivity for the pathogen has been associated with shorter stature in adults and children (39, 40) and with later menarche (41).

Based upon the view that there has been a cumulative and progressive loss of gut microbial diversity, we hypothesize that diseases whose pathogenesis is influenced by the gut microbiome may 1) be associated with early menarche and increased height, 2) show increased incidence in successive birth cohorts, and 3) show weakened associations with menarche in geographically-defined populations with more intensive antibiotic use, which can alter even the adult microbiome. Therefore we propose a large pooled study of menarche and height in association with risk of incident cancers, all-cause and cause-specific mortality in order to identify diseases with etiologic roles for the gut microbiome; these can then be investigated further in hopes of developing preventive interventions.

To investigate the associations between age at menarche and height with risks of incident cancers and with mortality endpoints, including all-cause mortality and cause specific mortality.

Based upon the view that there has been a cumulative loss of gut microbial diversity which affects growth and disease risks, diseases influenced by the gut microbiome may be identified as outcomes 1)associated with menarche and height, showing increased incidence in successive birth cohorts, and 3) showing weakened associations with menarche and height in populations with greater antibiotic use.

Therefore we propose the following specific aims:
1. To test for associations of age at menarche and height with risks of cancers and cancer-associated mortality.
2. To test for associations of age at menarche and height with all-cause mortality and mortality by cause.
3. To assess age, period, birth cohort, study, geographic location and adult BMI as potential modifiers of the risk associations.

This study will include women (>18 years of age) from multiple cohorts with data on age at menarche, height, and longitudinal data on incident cancers and mortality events. For each outcome of interest, time to event data will be modeled using Cox proportional hazards regression. Attained age will be the timescale of interest. We will estimate associations of menarche with risks of incident cancers and death by cause. We will adjust for major risk factors and assess modification by race, cohort, study, period, geographic location and BMI.

The cohort consortium approach would allow us to describe the association of menarche and height with risks of cancer and mortality in a large database that includes diverse populations observed over varying time periods; this is important because we would like to know whether the disease associations differ over time and by geographic location.

1000 deaths

Baseline year, baseline history of cancer, vital status (Alive/Dead), time under surveillance for mortality, time under surveillance for cancer, primary cause of death (ICD code with version number), secondary causes of death (ICD codes with version number), first incident cancer (date of diagnosis and primary cancer site).

Age at menarche, height.

Age at baseline; race/ethnicity; menopausal status at baseline; parity; weight at study baseline; cigarette smoking (ever smoked, currently smoke, total pack years); alcohol use(ever drink alcohol, currently drink alcohol, alcohol intake in grams/week); history of diagnosis of coronary heart disease (CHD) at baseline; history of diagnosis of diabetes at baseline; history of diagnosis of hypertension at baseline; history of diagnosis of stroke at baseline; geographic measures (country, state).

No