Fig. 2

Famine exposure during the first 1000 days of life, alterations of gut microbial diversity, and type 2 diabetes. A Distribution of observed OTUs across different groups. Observed OTUs within each cohort were independently z-scored. Participants were classified into different groups: the no-exposed group (NE1, born after 1964), no-exposed control group (NE2, born between 1962 and 1964), three in utero exposed groups (E1–E3, born in 1959, 1960, and 1961, respectively), infancy and toddler exposed group (E4, born between 1956 and 1958), preschooler exposed group (E5, born between 1953 and 1955), school-aged child exposed group (E6, born between 1947 and 1952), adolescent exposed group (E7, born between 1942 and 1946), and adult exposed group (E8, born before 1942). Here, participants in the no-exposed control group were used as the reference group. Participants who were exposed to the famine during the first 1000 days of life (E1, born in 1959) were highlighted in red. B Left: association of early-life famine exposure with observed OTUs. Linear regression was used to estimate the difference in observed OTUs between other groups and reference group, with adjustment of age, sex, BMI, and the use of hypoglycemic and hypolipidemic medications (yes/no for each). Here, we highlighted the comparation between the E1 and control groups in the three cohorts. Other results were present in the supplemental materials. Right: association of observed OTUs with type 2 diabetes. Logistic regression was used to examine the association of observed OTUs (per SD unit) with type 2 diabetes, adjusted for age, sex, and BMI. We combined the effect estimates from the three cohorts using random-effects meta-analysis. C Bray–Curtis-based principal coordinate analysis for the genus-level profiles in the GNHS cohort. The circles and error bars indicate the mean and standard errors of the mean within each group, respectively. D As in C, but for the GGMP cohort. E As in C, but for the CHNS cohort. F Relationship between the first two principal components of the microbial variation and type 2 diabetes. Logistic regression was used to examine the association of principal components (per SD unit) with type 2 diabetes, adjusted for age, sex, and BMI