Interpretation of main findings
An analysis of almost all full-term births in Hungary in 1998-2018, clearly showed an increasing birthweight trend of 4.1 g/year until 2008, followed by a less steep decline of 2.5 g/year in 2008-2018. During the same period, important changes in maternal and delivery related characteristics were observed: gestational age at delivery decreased, maternal age increased, the proportion of first parities decreased, and the frequency of both caesarean sections and induced deliveries increased.
According to our multivariate models, most of the increase in birthweight in the first period was explained by the increasing maternal age, while a substantial part of the decrease in the second period was explained by decreasing duration of pregnancies (i.e., decreasing gestational age at delivery).
When we investigated interactions between pregnancy related factors and calendar time (i.e., subgroups with the least and most changes over time), we found that the most pronounced difference between the first and second period was in mothers over 30 years of age, who had faster increase in the first period and faster decrease in the second period compared to younger mothers and the mean yearly change. Furthermore, the increase of birthweights in the first period was faster in newborns delivered by caesarean sections compared to vaginal deliveries, however no such interaction in the second period was found. Similarly, the increase in birthweights in the first period was more pronounced in multiparas compared to primiparas, while no interaction by parity in the second period was found.
Validity of results
Birthweight Trends
The increasing birthweight trend observed in the first period (1999-2008) parallels with similar observations from other high income countries(6,14).
During the second period we found declining birthweight trends. This is in line with observations from the U.S., where the average birthweight of term pregnancies declined from 3,315g in 1990 to 3,247g in 2013, a decrease of 67 g (15). The validity of this observation was confirmed by other reports from the U.S. (9,16,17) and Germany(10). Overall, a similar decrease to the one observed in Hungary was also found in most developed countries, however the decrease started approximately a decade earlier than in Hungary.
Decreasing Gestational Age at Delivery
Gestational age at delivery declined by two days between 1999 and 2018. This trend is similar to other surveys, however the magnitude of the decline varies between less than 1 to almost 3 days between 1990 and 2013 in the different studies (9,10,15,16).
Increasing Maternal Age over Time
We found that median maternal age at delivery increased from 26.2 years in 1999 to 30.5 years in 2018, corresponding to an increase in the proportion of older mothers (≥30 years) from 24% to 53%. An increasing trend in maternal age is reported from most countries worldwide (18). For example, the mean age of primiparas increased from 24.9 years to 26.3 years in the U.S. between 2000 and 2014 (19).
Decreasing Parity over Time
During the 20-year observation period, the proportion of primiparas increased from 46.4% to 49.6%. Our results are somewhat different from those in other developed countries. For example, the proportion of primiparity remained constant (43.3%) in France between 1998-2003, (7) while it decreased (37.3% to 33.7%) in the U.S. between 2000 and 2008 (17).
Increasing Rates of Caesarean Sections and Induced Deliveries
The rate of caesarean sections and labor inductions more than doubled (from 17.6% to 39.7% and from 12.7% to 26.2%, respectively) in Hungary between 1999 and 2018. This is in line with observations from almost all countries. The rate of scheduled or induced deliveries almost tripled reaching over 30% in the U.S. between 1990 and 2013 25,27. Similar, but smaller increase (25.9% to 33.6%) was observed in Scotland in 1988-2012 (21). The rate of caesarean sections increased similarly in India (from 28.2% to 42.0% in 2010-2017) (22) and Brazil (from 34.1% to 57% in 1997-2014) (23) (24), while the increase was minimal in Norway (13.6% to 16.3% in 1999-2016) (25).
Potential Explanation for the Increasing Birthweight Trends in the First Period
According to our hierarchical logistic regression models, maternal age explained a large proportion (5.4g/year vs. 2.4g/year – 55.5%) of the increasing birthweight trend over time.
While maternal age may be directly related to birthweight, it could be a marker of other determinants, such as anthropometric or social factors. For example, maternal smoking might decrease with maternal age (6,26). Similarly, maternal weight increases with aging and maternal BMI is a known predictor of newborn weight (27). Indeed, there is an increasing trend in obesity among fertile aged women in Hungary in the last decades (28). Furthermore, older age is associated with better socioeconomic circumstances that is associated with larger birthweights (29). As advanced maternal age is also associated with higher risk of adverse obstetrical and perinatal outcomes (30), as well as elective deliveries (17) the changes observed during the first period could be associated with worsening pregnancy outcomes.
Potential Explanation for the Decreasing Birthweight Trends in the Second Period
We found that a large proportion of the decreasing birthweight trend was explained by gestational age at delivery (i.e. length of pregnancy). The decreasing length of gestation over time is strongly related to the fact that the proportion of induced deliveries and caesarean sections more than doubled over the examination period. Other authors that found similar decreasing birthweight trends explained this observation by the increasing rates of caesarean deliveries and induced labors. This is supported by the fact that births became much less likely to occur beyond gestational week 40 and much more likely to occur during weeks 37-39 (14).
It is plausible that the worsening short term pregnancy outcomes associated with advanced maternal age is compensated by early term pregnancies (16).
However, the question remains how the approach to early term deliveries will modify long-term consequences. It is known that caesarean sections are associated with an increased risk of severe acute maternal morbidity and mortality, and a higher risk of adverse outcomes in subsequent pregnancies (31). In terms of newborn outcomes, caesarean sections are associated with increased risks of fetal respiratory problems (32) and long-term consequences (i.e. asthma, overweight, obesity, allergy) (31).
Subgroups Driving Increasing and Decreasing Birthweight Trends
We found significantly faster increase in birthweight compared to the overall increase among older mothers (≥30 years of age), among those with multiparity, and among newborns delivered by a caesarean section in the first part of the observation period. These findings may suggest that the approach to deliveries was reactive by obstetricians: wait in the high-risk groups (older mothers, multiparas) for delivery induction or caesarean delivery until the fetus becomes large. This notion is supported by the Spanish observation that term newborns from caesarean deliveries were larger than from vaginal deliveries and newborns of multiparas were larges than those of primiparas (33).
We found significantly faster decline in birthweight compared to the overall decrease among older mothers (≥30 years of age) in the second part of the observation period. Furthermore, newborns of multiparas and those of caesarean deliveries were no longer associated with faster increases in birthweights. These findings are compatible with the hypothesis of a proactive management of delivery, where pregnancy is terminated in high-risk women before fetal weight reaches abnormal levels.
Strengths and limitations
Our analysis includes most Hungarian pregnancies with an ascertainment rate of 94.8%. The huge number of records allowed adjustment for several risk factors and to provide narrow confidence intervals. The data entry software comes with detailed instructions that assures high quality of the collected variables. (34)
Our analyses are limited in several ways. First, there is no way to measure changes in the obstetric decision-making process in official administrative data. As with other administrative databases, other limitations have to be acknowledged: no data is available regarding race, social status, bodyweight, and smoking habits – important determinants of birthweight. Although there is a possibility of misclassification, it should be noted that the Tauffer database is not used for reimbursement limiting the role of selective over- or under-reporting. The role of unmeasured confounding cannot be downplayed. It is possible that the increases and decreases in birthweights were responding to unobserved factors. Individual measures of maternal behaviors, characteristics, and other risk factors for obstetric interventions were also quite limited. Potentially key details about maternal health risk factors related to obstetric decisions (such as obesity) may also be missing.