Published online Jun 01, 2026.
https://doi.org/10.5646/ch.2026.32.e25
Famine and blood pressure trajectories across 2 generations: a prospective cohort study
Abstract
Background
While the transgenerational impact of famine on adult health is well-documented, its impact on blood pressure (BP) trajectories from childhood to adolescence remains poorly understood. We aimed to capture multi-generational BP trajectories and identify the primary drivers of adolescent BP elevation.
Methods
Utilizing data from the China Health and Nutrition Survey, we analyzed BP trajectories in the F1 (n = 5,357) and F2 (n = 2,349) generations. The F1 generation was categorized into 2 groups: famine exposed and non-exposed groups, while the F2 generation was divided into 4 groups: no parental exposure, father-only-exposed, mother-only-exposed, and both-parents-exposed. Linear mixed-effects models and causal mediation analyses were conducted to ascertain BP trajectories and the mediation of parental body mass index (BMI).
Results
In the F1 generation, BP of famine-exposed individuals surpassed that of controls after age 40. In the F2 generation, BP trajectory divergence occurred at approximately age 14. By age 18, all exposed groups had significantly higher BP than controls, independent of parental hypertension (P < 0.01). Notably, the mother-only-exposed group exhibited the highest velocity of BP elevation; however, after excluding offspring with hypertensive parents, the father-only-exposed group exhibited the fastest velocity of BP increase. Mediation analysis revealed that over 85% of this effect was direct, with parental BMI accounting for only 14.51% and 10.08% of the effect for systolic and diastolic BP, respectively.
Conclusions
Individuals with direct famine exposure possess higher rate of BP elevation. Furthermore, parental famine exposure programs an accelerated BP velocity in offspring that manifests during early adolescence. These findings suggest that the famine exposure creates a subclinical vascular risk that requires early monitoring, independent of adiposity and hypertension.
BACKGROUND
Blood pressure (BP) is a well-established risk factor for cardiovascular diseases (CVDs), which remain the leading cause of global morbidity and mortality. Between 2000 and 2023, deaths attributable to CVDs increased by 31.4% [1]. Alarmingly, the age distribution of CVDs is shifting toward younger generations [2], with data showing that younger generations account for a substantial portion of cardiovascular mortality [3]. Back in 2017, the American College of Cardiology and American Heart Association lowered the hypertension diagnostic threshold to 130/80 mmHg to better address early-stage risks [4].
Between 1959 and 1962, China experienced one of the most devastating famines in human history. Extensive research has since linked this famine exposure during early life to CVDs later in life, with data suggesting heightened risk of adult-onset hypertension by at least 24% [5, 6, 7]. To make things worse, emerging evidence further suggests these adverse effects are transgenerational, as offspring of exposed parents also face elevated risks of obesity and hypertension in adulthood [8]. Possible mechanisms for this transmission include fetal growth reprogramming due to malnutrition and persistent epigenetic modifications, such as altered DNA methylation of metabolic regulatory genes [9, 10, 11, 12].
Despite these insights, a significant gap remains in our understanding of the temporal manifestation of these risks. Existing transgenerational studies have largely focused on static BP measurements in adult offspring, effectively overlooking the critical period of childhood and adolescence. This oversight is particularly concerning given the fact that BP patterns established in childhood are known to track into adulthood [13, 14], and the global prevalence of pediatric hypertension has doubled in the last 2 decades [15].
This study addressed these gaps by exploring the longitudinal BP trajectories of both the directly affected generation (F1) and their offspring (F2) using data from the China Health and Nutrition Survey (CHNS). Unlike existing studies that examine BP as a simple diagnosis of hypertension, the BP trajectories over decades across 2 generations were explored to identify when the famine-exposed lineage began to diverge from the non-exposed one. By determining when these physiological differences emerged before the clinical manifestation of adult hypertension, we aimed to provide a missing link in the adverse transgenerational impact of nutritional stress. To the best of our knowledge, this is the first study to investigate transgenerational effects on the BP trajectories of offspring during childhood and adolescence.
METHODS
Study design and participants
This study utilized longitudinal data from the CHNS, an ongoing prospective cohort across 15 provinces and autonomous regions (Guangxi, Guizhou, Henan, Hubei, Hunan, Jiangsu, Shandong, Heilongjiang, Liaoning, Yunnan, Zhejiang, Shaanxi, Chongqing, Shanghai and Beijing) in China. The CHNS employs a multistage random clustering sampling design to ensure a representative sample of the Chinese population [16]. Data from ten survey waves (1989–2015) were analyzed. Participants were eligible for the current study if they had at least one BP measurement between ages 24 and 61 for the F1 generation or ages 7 and 18 for the F2 generation. To ensure data integrity, BP and body mass index (BMI) values exceeding 5 standard deviations (SD) from the mean were excluded as outliers.
All participants in the CHNS provided written informed consent. The CHNS protocol was approved by both the University of North Carolina and the Chinese Center for Disease Control and Prevention (2015017).
Famine exposure definition
The Chinese Famine did not have a well-defined starting and ending time. Based on the mortality rate, which elevated in 1959, peaked in 1960 and returned to normal in 1962 [8], the F1 generation was categorized into a famine-exposed group (born 1959–1962) and a combined control group (born pre-famine 1954–1957 or post-famine 1964–1967). According to the meta-analysis in 2017 [17], existing studies often chose those born post famine as the control group. The observed positive association between famine exposure and adverse health outcomes could be explained by the age differences between the famine-exposed and non-famine-exposed groups. Combining pre- and post-famine cohorts was intended to minimize age-difference bias often found in famine studies. Furthermore, since the exact starting and ending time points of the Chinese Famine was still not well established, those born in 1958 and 1963 were not incorporated into the current study to minimize the possible misclassification bias.
Regarding the F2 generation, offspring born to the famine group and pre-famine group were divided into the following three groups: father-only-exposed, mother-only-exposed and both-parents-exposed groups. The remaining were considered the control group.
Outcome measures and covariates
After at least 10 minutes of rest, the participant BP was measured using a mercury sphygmomanometer with a suitable cuff on the right arm in a seated position. Each participant was measured three times, and the average BP was used in the final analysis. In this study, BP readings that were 5 SDs away from the average were considered outliers and excluded from further data analysis.
While the Chinese Famine impacted entire mainland China, there was significant geographic heterogeneity in its severity. In the present study, provinces within the CHNS were categorized based on the excess mortality rate (EMR) [18], Regions with an EMR exceeding 100% were classified as severely affected, while the remaining provinces were designated as less severely affected.
For the F1 generation, the covariates include in the analysis were as follows: hypertension (yes/no) defined as a definite physician diagnosis or taking anti-hypertension medicine, smoking (yes/no), drinking (yes/no), gender (male/female), ethnicity (Han/others), residence area (urban/rural), education (illiterate or primary/high school/technical or vocational/college or above), family income (low income/middle income/high income), famine severity (severely affected/less severely affected) and BMI (kg/m2, calculated as weight divided by height). For the subsequent F2 generation, the covariates included were as follows: gender (male/female), ethnicity (Han/others), residence area (urban/rural), family history of hypertension (yes/no), household income (low income/middle income/high income), parental education (illiterate or primary/high school/technical or vocational/college or above), famine severity (severely affected/less severely affected) and BMI (kg/m2).
Statistical analysis
Firstly, participant characteristics were presented as mean ± SD for continuous variables and counts (percentage) for categorical variables, by famine exposure status, for both F1 and F2 generations. Differences between groups were examined by analysis of variance and χ2 analysis, for continuous and categorical variables, respectively.
To determine the best-fitting median trajectories, fractional polynomials were performed based on the Akaike Information Criterion [19]. For the F1 generation, the best-fitting polynomials were
BP change was defined as the difference between measurements at each age and the established baseline, which was age 24 for the F1 generation and age 7 for the F2 generation. The BP for baseline age were estimated from aforementioned linear mixed-effects models. Mixed-effects models were then utilized to explore the velocity of BP changes relative to the baseline.
Since offspring of hypertensive parents tended to have higher BP, after excluding those F2 individuals with hypertensive parents, aforementioned mixed-effects models with fractional polynomials were applied to analyze the effect of famine exposure on offspring BP trajectories.
To investigate the mechanisms underlying the transgenerational transmission of BP risk, a causal mediation analysis was performed using the F1 generation BMI as a mediator. The covariates in the F1 and F2 BP trajectories analysis mentioned above were adjusted in the mediation analysis. This allowed us to distinguish between the direct biological programming effect of famine and the indirect metabolic influence of the parental environment. The choice of parental BMI as the primary mediator was based on its established role as a proxy for the metabolic environment that is frequently programmed by early-life nutritional stress and subsequently transmitted to offspring. This analysis was conducted under the following causal assumptions: 1) no unmeasured confounding of the exposure-outcome, mediator-outcome, or exposure-mediator relationships, adjusted for the covariates previously mentioned; 2) temporal precedence, where famine exposure occurred prior to the measurement of parental BMI, which preceded the assessment of F2 BP; and 3) the absence of mediator-outcome confounders affected by the exposure.
Several sensitivity analyses were conducted to evaluate the robustness of our findings. First, for the F1 generation, the BP trajectory analyses were repeated using the pre-famine and post-famine groups separately as the control groups. Second, recognizing that the temporal boundaries of the Chinese famine were not universally defined, an alternative exposure window of 1959–1961, the most widely recognized period in mainland China, was used to re-analyze BP trajectories for both the F1 and F2 generations, while maintaining the original pre- and post-famine year bands. Finally, parental waist circumference was incorporated as an alternative mediator in the mediation analysis.
All statistical analyses were performed with the Statistical Analysis System version 9.4 (SAS Institute Inc., Cary, NC, USA) and R (version 4.5.2; R Foundation for Statistical Computing, Vienna, Austria). A 2-tailed P-value < 0.05 was considered statistically significant.
RESULTS
For the F1 generation, among the initial 5,609 individuals, 63 were excluded due to missing BP measurements, 183 were excluded for having outlier measurements (defined as exceeding 5 SDs from the mean at each survey age), and 6 were removed due to abnormal BMI. This resulted in a final analytic sample of 5,357 individuals. From the offspring of this F1 cohort, 2,568 individuals were initially identified for the F2 generation. After excluding 111 individuals with missing BP data, 12 with outlier measurements, 15 with abnormal BMI, and 81 with unknown exposure status, the final F2 analytic sample consisted of 2,349 individuals.
Within the F1 generation, 1,522 individuals were exposed to the famine, while 3,835 were unexposed. Those in the exposed group tended to exhibit a higher BMI and a greater prevalence of alcohol consumption and smoking. Specifically, those in the post-famine group tended to have the highest BMI, followed by those in the famine exposed and pre-famine group, respectively. Regarding the F2 generation, 1,419 individuals had at least one parent with famine exposure with 367 in the father-only exposed group, 556 in the mother-only exposed group and 496 in the both-parents-exposed group. Compared to their counterparts without such a background, individuals with parental famine exposure demonstrated a higher prevalence of family history of hypertension (Tables 1 and 2, Supplementary Table 1).
Table 1
Characteristics of subjects for analysis of famine exposure and blood pressure trajectories
Table 2
Characteristics of subjects for analysis of famine exposure and blood pressure trajectories
F1 generation: long-term BP trajectories and velocity
Linear mixed-effects models revealed that BP trajectories for both F1 generation groups followed J-shaped curves. A significant crossover effect was observed: before age 40, the famine-exposed group exhibited lower SBP and DBP than the control group (P < 0.01). However, beyond age 40, BP in the famine-exposed group became significantly higher and remained elevated through age 61 (P < 0.01). At age 40, SBP and DBP for the exposed group were 114.96 mmHg (95% confidence interval [CI], 114.83 to 115.09) and 76.48 mmHg (95% CI, 76.40 to 76.56) compared with 114.97 mmHg (95% CI, 114.88 to 115.06) and 76.60 mmHg (95% CI, 76.55 to 76.65) in the control group, respectively. At age 60, SBP and DBP for the exposed group rose to 133.78 mmHg (95% CI, 133.59 to 133.98) and 84.31 mmHg (95% CI, 84.19 to 84.44), while the control group reached 131.33 mmHg (95% CI, 131.21 to 131.46) and 82.75 mmHg (95% CI, 82.68 to 82.83), respectively (Fig. 1).
Fig. 1
Trajectories of BP and BP change for F1 generation with famine exposure and no famine exposure. (A) SBP, (B) DBP, (C) SBP change, (D) DBP change. All models were adjusted for gender, ethnicity, residence area, smoking, drinking, educational level, hypertension, famine severity and body mass index.
BP, blood pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure.
The velocity of BP elevation relative to the age 24 baseline was significantly faster in the famine-exposed group (P < 0.01). By age 60, the mean SBP and DBP change reached 27.47 mmHg (95% CI, 27.33 to 27.60) and 15.72 mmHg (95% CI, 15.65 to 15.78) in the exposed group versus 23.01 mmHg (95% CI, 22.93 to 23.10) and 12.72 mmHg (95% CI, 12.68 to 12.76) in the control group, respectively (Fig. 1).
F2 generation: adolescent BP divergence
Similar to their parents, F2 BP trajectories followed J-shaped patterns. A critical divergence was observed at approximately age 14. Before age 14, offspring of exposed parents had lower BP, but after this point, their trajectories accelerated, resulting in higher SBP and DBP by age 18 compared to the control group (P < 0.01). The velocity of BP change from age 7 to 18 was significantly heightened in the F2 famine-exposed lineage (P < 0.01).
At age 14, SBP and DBP for the non-exposed, both-parents-exposed, father-only-exposed and mother-only-exposed groups were 103.12 mmHg (95% CI, 102.94 to 103.30) and 67.65 mmHg (95% CI, 67.53 to 67.77); 103.07 mmHg (95% CI, 102.84 to 103.31) and 67.43 mmHg (95% CI, 67.27 to 67.59); 103.98 mmHg (95% CI, 103.68 to 104.27) and 67.80 mmHg (95% CI, 67.61 to 67.99); and 103.33 mmHg (95% CI, 103.10 to 103.56) and 67.81 mmHg (95% CI, 67.65 to 67.96), respectively. By age 18, SBP and DBP for these groups increased to 107.58 mmHg (95% CI, 107.33 to 107.82) and 70.16 mmHg (95% CI, 70.00 to 70.33); 109.09 mmHg (95% CI, 108.76 to 109.42) and 71.35 mmHg (95% CI, 71.12 to 71.58); 110.80 mmHg (95% CI, 110.39 to 111.21) and 71.99 mmHg (95% CI, 71.72 to 72.27); and 110.23 mmHg (95% CI, 109.92 to 110.55) and 72.28 mmHg (95% CI, 72.06 to 72.49), respectively (Fig. 2).
Fig. 2
Trajectories of BP and BP change for F2 generation with different famine exposure status. (A) SBP, (B) DBP, (C) SBP change, (D) DBP change. All models were adjusted for gender, ethnicity, residence area, family history of hypertension, household income, parental education level, famine severity and body mass index.
BP, blood pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure.
The velocity of BP elevation relative to the age 7 baseline was fastest in the mother-only-exposed group, followed by father-only-exposed group, both-parents-exposed group and non-exposed group (P < 0.01). By age 18, the mean SBP and DBP change reached 23.07 mmHg (95% CI, 23.00 to 23.15) and 15.40 mmHg (95% CI, 15.34 to 15.46); 22.84 mmHg (95% CI, 22.76 to 22.92) and 14.63 mmHg (95% CI, 14.57 to 14.70); 20.63 mmHg (95% CI, 20.55 to 20.71) and 13.89 mmHg (95% CI, 13.82 to 13.96); and 16.38 mmHg (95% CI, 16.32 to 16.44) and 10.04 mmHg (95% CI, 10.00 to 10.09), respectively (Fig. 2).
After excluding F2 offspring with hypertensive parents, the remaining sample sizes for the non-exposed, father-only-exposed, mother-only-exposed, and both-parents-exposed groups were 782, 277, 434, and 337, respectively. The previously described mixed-effects models were performed to re-evaluate BP trajectories. These exclusions did not significantly alter the observed BP patterns in the F2 generation; the crossover effect persisted at approximately age 14. Notably, the father-only-exposed group exhibited the most rapid rate of BP elevation, followed by the mother-only-exposed, both-parents-exposed, and non-exposed groups, respectively (Fig. 3).
Fig. 3
Trajectories of BP and BP change for F2 generation with different famine exposure status after excluding F2 offsprings with hypertensive parents. (A) SBP, (B) DBP, (C) SBP change, (D) DBP change. All models were adjusted for gender, ethnicity, residence area, family history of hypertension, household income, parental education level, famine severity and body mass index.
BP, blood pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure.
Causal mediation analysis
The total effect of parental famine exposure on offspring SBP and DBP was significant at −0.98 mmHg (95% CI, −1.77 to −0.19; P = 0.03) and −0.94 mmHg (95% CI, −1.56 to −0.32; P < 0.01), respectively. This negative total effect reflected the lower BP baseline characteristic of the famine-exposed lineage during the childhood and early adolescent years of the observation period. Decomposition of this effect revealed that parental BMI served as a significant but partial mediator (indirect effect: −0.15 and −0.09 mmHg, respectively, P < 0.01). However, most of the transgenerational impact was accounted for by the direct effect, which remained robust at −0.84 mmHg (95% CI, −1.63 to −0.04; P = 0.03) and −0.84 mmHg (95% CI, −1.47 to −0.23; P < 0.01), respectively. Overall, parental BMI accounted for only 14.51% and 10.08% (P < 0.01) of the total effect on offspring SBP and DBP, respectively (Table 3).
Table 3
Mediation effects of body mass index on the transgenerational associations of famine exposure with F2 offspring blood pressure
Sensitivity analysis
When comparing the pre- and post-famine groups separately against the famine-exposed group, a crossover effect remained evident between the ages of 40 and 50. Notably, the famine-exposed group exhibited the most rapid elevation in BP, followed by the post-famine and pre-famine groups (Supplementary Fig. 1). Even when the exposure window was strictly defined as 1959–1961, the crossover effect in BP trajectories persisted for both F1 and F2 generations at approximately ages 40 and 14, respectively (Supplementary Figs. 2 and 3). Mediation analysis indicated that parental waist circumference accounted for 15.96% and 9.62% of the total effect on offspring SBP and DBP, respectively (Supplementary Table 2).
DISCUSSION
Drawing on data from the CHNS, this prospective cohort study found that individuals with direct famine exposure maintained higher BP after age 40 and demonstrated a greater velocity of change from age 24 to 61. Similarly, the subsequent generation—those born to exposed parents—had higher BP between ages 14 and 18, while exhibiting a higher velocity of change from age 7 to 18, with the mother-only-exposed group exhibited the fastest velocity, followed by father-only-exposed, both-parents-exposed and non-exposed group. This increased velocity of BP change in the F2 generation was independent of parental hypertension. Also, this effect was partially mediated through parental BMI by less than 15%.
Existing research has tended to focus on a narrow definition of hypertension when examining the negative impact on BP for both directly affected individuals and subsequent generations [5, 6, 7, 8]. This approach often overlooks those with “high-normal” BP. Notably, the American College of Cardiology and the American Heart Association recently lowered the diagnostic threshold for hypertension from 140/90 to 130/80 mmHg, a change primarily driven by the increased risk of CVDs [4]. To the best of our knowledge, only one other study has explored the effects of famine on BP trajectories, yielding results consistent with our current findings [8]. Furthermore, the current study is the first to investigate the transgenerational effects on the BP trajectories of the adolescent offspring of directly exposed parents. Our findings provided new evidence that, even before reaching adulthood, these offspring experienced distinct BP trajectories compared to those without parental famine exposure. Given that BP patterns established in childhood often persist through adolescence and into adulthood [13], our results suggest that physiological conditions differed significantly long before the clinical manifestation of hypertension. This further supports the argument that focusing solely on a narrow definition of hypertension is insufficient for capturing the full scope of risk.
Age is a well-established risk factor for CVDs like hypertension [20]. Below a certain age, the protective effect of youth may outweigh the adverse effects of famine exposure. A meta-analysis of famine exposure surprisingly found it to be “protective” against high BP when the control group consisted solely of those born before the famine [17]. That study concluded that an age difference of just three years between the 2 groups was the primary driver of this result. Consequently, in the current study, pre-famine and post-famine cohorts were combined into a single control group to minimize this age-difference effect. By minimizing this three-year age gap, the effect size identified in this study likely represented a conservative estimate of the actual impact.
For the F1 generation, individuals with direct famine exposure exhibited elevated BP trajectories only after approximately age 40. This crossover phenomenon aligns with existing literature, including another CHNS-based study indicating a similar cut-off age [8]. Comparison with international data shows varying timelines: the Dutch famine study found no effect on SBP and DBP until age 50 [21], while the Leningrad Siege study showed that adverse effects did not manifest until age 70 [22]. Several mechanisms could support this adverse BP effect. Famine exposure programs the expression of the renin-angiotensin system, triggering an irreversible increase in angiotensin II receptor expression. This leads to chronic vasoconstriction and subsequent hypertension [23, 24]. Second, malnutrition—the most direct consequence of famine—can cause a congenital reduction in the number of nephrons. This nephron deficit impairs sodium excretion, resulting in lifelong salt sensitivity and hypertension [25]. The crossover phenomenon may imply a strong survivor bias where only the most physiologically resilient individuals could survive the famine. Under a certain age, the physiologically more resilient trait may have outweighed the adverse famine effect, while beyond that age, the biological scar of famine exposure overrode this genetic advantage as the main driver of accelerated BP change.
A novel finding of this study is the demonstration that descendants of the famine-exposed generation also exhibited a delayed negative effect, with elevated BP manifesting around age 14. Epigenetic mechanisms likely facilitate the persistence of these adverse effects in offspring. Famine exposure triggers a decrease in the methylation of the AT1a receptor gene, leading to heightened expression and an increased risk of hypertension in the subsequent generation [26]. The initial analysis of the F2 generation indicated that offspring with mother-only-exposed group exhibited the most rapid rate of BP elevation, followed by the father-only-exposed, both-parents-exposed, and non-exposed groups. This primary finding supports the hypothesis that maternal nutritional deprivation directly impacts oocyte formation in the F1 generation, potentially inducing mitochondrial dysfunction and oxidative stress that impairs endothelial function in the offspring [27, 28]. However, after excluding offspring with hypertensive parents, the father-only-exposed group emerged as having the fastest velocity of BP elevation. This suggests that while the maternal influence in the total population may be partially driven by shared genetic susceptibility or the transmission of the mother's clinical hypertensive phenotype, the paternal effect appears to be a more robust manifestation of epigenetic programming. This aligns with the Paternal Origins of Health and Disease framework [29], which posits that environmental stressors can induce stable modifications in the male germline such as altered sperm DNA methylation or small RNA profiles that program cardiovascular set-points in subsequent generations independently of the father's own clinical status. Thus, the accelerated BP trajectories in the F2 generation may reflect a sex-specific transgenerational response, where paternal exposure provides a more stable epigenetic signal once clinical confounders are removed. Finally, the fact that the both-parents-exposed group exhibited slower BP change than those with only one parent exposed likely reflects a survivor bias, wherein the offspring of 2 famine survivors possessed a distinct genetic advantage.
A recent meta-analysis indicates that global childhood hypertension has become an increasingly significant public health concern, with prevalence nearly doubling between 2000 and 2020. Specifically, pediatric hypertension rates rose from 3.4% to 6.5% in boys and from 3.02% to 5.82% in girls [15]. This issue is further compounded by substantial underdiagnosis and under-recognition, with recognition rates potentially as low as 26% [30, 31]. Given the well-established fact that high BP tracks from childhood into adulthood, these findings underscore the need for pediatric healthcare practices to place a greater emphasis on monitoring BP—a factor currently overlooked in routine care, offering significant individual and societal benefits [13]. Effective management of pediatric hypertension is critical to mitigating long-term arterial damage and the subsequent development of CVDs [32]. Broadly speaking, children with parental nutrition deprivation should receive more intensive monitoring of BP to mitigate the risk of pediatric hypertension and possible CVDs in adulthood.
A primary strength of this study is its reliance on the CHNS, which features a highly representative sample of the Chinese population due to its scale and multistage random sampling design. Given that the Chinese famine affected the entire mainland, a representative sample is essential for generalizability. Furthermore, the prospective cohort design and the large sample size of the CHNS ensured high precision in our findings.
However, several limitations warrant consideration. First, regarding the F1 generation, individuals born both before and after the famine were combined into a single control group to mitigate the “younger age effect” that would have occurred if only pre-famine participants had been used [17]. Because the pre-famine cohort was included in the control, the resulting findings likely represent an underestimation of the actual effect size. Furthermore, when comparing the pre- and post-famine groups separately against the famine-exposed group, the crossover effect remained evident between the ages of 40 and 50. Second, despite the prospective design and adjustments for various covariates, the possibility of residual confounding cannot be entirely ruled out. Third, categorizing participants as famine-exposed or unexposed based solely on birth year may introduce misclassification bias, particularly as the exact start and end dates of the Chinese famine remain a subject of debate. While our exclusion of the years 1958 and 1963 may have alleviated some of this bias, the current findings likely remain a conservative estimate of the true impact of famine exposure. In addition, an alternative famine exposure window of 1959–1961 did not significantly alter the main findings. Moreover, a meta-analysis revealed the non-significance of switching to alternative exposure windows [17]. Fourth, data on the puberty status for the F2 generation were not available, making it impossible to explore whether hormone shifts during puberty acted as a catalyst for the observed risk divergence. Lastly, data on variables such as offspring salt sensitivity were not available, which made it impossible to explore the mediation effect of these factors.
CONCLUSIONS
In summary, famine exposure appears to exert adverse effects not only on the directly exposed generation but also on their offspring. Individuals with direct exposure tend to experience elevated BP between the ages of 40 and 61. Similarly, their offspring exhibit higher BP starting as early as age 14. Notably, both generations demonstrate an increased velocity of BP change over time. Specifically, offspring of the mother-only-exposed group exhibit the fastest velocity of BP change, followed by those with father-only-exposed and both-parents-exposed groups. After excluding those with hypertensive parents, the father-only-exposed group shows the fastest rate of BP increase, followed by the mother-only-exposed and both-parents-exposed groups.
SUPPLEMENTARY MATERIALS
Characteristics of subjects for analysis of famine exposure and blood pressure trajectoriesSupplementary Table 1
Mediation effects of waist circumference on the transgenerational associations of famine exposure with F2 offspring blood pressureSupplementary Table 2
Trajectories of BP and BP change for F1 generation with famine exposure, pre- and post-famine exposure. (A) SBP, (B) DBP, (C) SBP change, (D) DBP change. All models were adjusted for gender, ethnicity, residence area, smoking, drinking, educational level, hypertension, famine severity and body mass index.Supplementary Fig. 1
Trajectories of BP and BP change for F1 generation with famine exposure (1959–1961) and no famine exposure. (A) SBP, (B) DBP, (C) SBP change, (D) DBP change. All models were adjusted for gender, ethnicity, residence area, smoking, drinking, educational level, hypertension, famine severity and body mass index.Supplementary Fig. 2
Trajectories of BP and BP change for F2 generation with different famine exposure status with parental exposure window of 1959–1961. (A) SBP, (B) DBP, (C) SBP change, (D) DBP change. All models were adjusted for gender, ethnicity, residence area, family history of hypertension, household income, parental education level, famine severity and body mass index.Supplementary Fig. 3
Funding:None.
Competing interest:The authors declare that they have no competing interests.
Availability of data and materials:Current research is based on the China Health and Nutrition Survey (CHNS). All data is publicly available and can be accessed at https://www.cpc.unc.edu/projects/china.
Ethics approval and consent to participate:The CHNS protocol was approved by both the university of North Carolina and the Chinese Center for Disease Control and Prevention (2015017). All participants in the CHNS provided written informed consent.
Consent for publication:Not applicable.
Abbreviations
| BMI | body mass index |
| BP | blood pressure |
| CHNS | China Health and Nutrition Survey |
| CI | confidence interval |
| CVD | cardiovascular disease |
| DBP | diastolic blood pressure |
| EMR | excess mortality rate |
| SBP | systolic blood pressure |
| SD | standard deviation |
Acknowledgements
The author gratefully acknowledges all the participants from the China Health and Nutrition Survey (CHNS). Also, the gratefulness goes to CHNS team to open source all their data.
References
-
GBD 2023 Causes of Death Collaborators. Global burden of 292 causes of death in 204 countries and territories and 660 subnational locations, 1990–2023: a systematic analysis for the Global Burden of Disease Study 2023. Lancet 2025;406:1811–1872.
-
-
GBD 2019 Adolescent Mortality Collaborators. Global, regional, and national mortality among young people aged 10–24 years, 1950–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet 2021;398:1593–1618.
-
-
Whelton PK, Carey RM, Aronow WS, Casey DE Jr, Collins KJ, Dennison Himmelfarb C, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on clinical practice guidelines. Circulation 2018;138:e426–e483.
-
-
Luo Z, Mu R, Zhang X. Famine and overweight in China. Rev Agric Econ 2006;28:296–304.
-
-
Royston P, Altman DG. Regression using fractional polynomials of continuous covariates: parsimonious parametric modelling. J R Stat Soc Ser C Appl Stat 1994;43:429–467.
-
Publication Types
MeSH Terms
Figures
1 / 3
Tables
1 / 3
