Prevalence and temporal trends in prepregnancy nutritional status and gestational weight gain of adult women followed in the Brazilian Food and Nutrition Surveillance System from 2008 to 2018

Abstract Prepregnancy body mass index (BMI) and gestational weight gain (GWG) are the most investigated indicators of maternal nutritional status, which is a modifiable factor that plays a vital role in maternal and infant health. This study describes prepregnancy BMI and GWG of 840,243 women with 2,087,765 weight observations in the Brazilian Food and Nutrition Surveillance System from 2008 to 2018. Prepregnancy BMI was classified according to the World Health Organization cut‐offs. Total GWG was calculated from weight measurements taken after 36 weeks of pregnancy and classified according to the Institute of Medicine guidelines. Temporal trends in prepregnancy BMI status were examined, and maps were used to evaluate changes in excessive GWG in each Brazilian federation unit. On overall, prepregnancy overweight and obesity increased from 22.6% to 28.8% and from 9.8% to 19.8%, respectively, between 2008 and 2018. The prevalence of excessive GWG rose from 34.2% to 38.7% during the same period and in 11 of the 27 Brazilian federation units between 2008 and 2016. Women with underweight showed the highest values for mean total GWG for all the compared years (overall variation from 12.3 to 13.1 kg), followed by those with normal weight (11.9 to 12.5 kg), overweight (10.1 to 10.9 kg) and obesity (from 8.2 to 8.9 kg). Within each BMI group, values remained fairly stable throughout the studied period for first‐ and second‐trimester GWG and total GWG. These results help to fill a significant gap in understanding the distribution of prepregnancy BMI and GWG in Brazilian women.


| INTRODUCTION
Prepregnancy weight status is an important consideration during prenatal care because of its association with maternal and child adverse health outcomes (Aune et al., 2014;Yu et al., 2013). The prevalence of overweight or obesity according to body mass index (BMI) has increased among women of reproductive age in the last 30 years, especially in high-income countries (HIC) (Popkin & Slining, 2013;Poston et al., 2016). This is also true for low-and middle-income countries (LMIC), such as Brazil, where the prevalence of excessive weight (overweight and obesity) for non-pregnant adult women is estimated to be over 50% in several states (Ministério da Saúde, 2020). However, national-level data on temporal changes in prepregnancy BMI are scarce.
One of the main manifestations of pregnancy is an increase in maternal weight. Gestational weight gain (GWG) is normal and reflects both adequate growth and development of the fetus and the changes in women's bodies to support pregnancy and subsequent lactation (Institute of Medicine [IOM] [US] and National

Research Council [US] Committee to Reexamine IOM Pregnancy
Weight Guidelines, 2009). Health professionals often monitor GWG during prenatal care because excessive and inadequate weight gain is linked with adverse outcomes (Goldstein et al., 2017(Goldstein et al., , 2018Kominiarek & Peaceman, 2017). show the prevalence of GWG below IOM guidelines for LMIC are scarce. The few available studies show that GWG below these guidelines is common, which indicates that low GWG is a serious problem for many LMIC (Asefa & Nemomsa, 2016;Esimai & Ojofeitimi, 2014;Gondwe et al., 2018;Young et al., 2017). In contrast, the prevalence of GWG above these guidelines is higher in HIC (Goldstein et al., 2018;Kominiarek & Peaceman, 2017).
Since 2008, the Brazilian population's nutritional status has been monitored using the National Food and Nutrition Surveillance System (SISVAN) (Ministério da Saúde, 2013). The system was designed to record population-level information on nutritional status and food consumption, including pregnant women who used public prenatal care services (Ministério da Saúde, 2011). The data collected are used as a basis for the development and evaluation of public health nutrition policies for the country.
To the best of our knowledge, there are no studies with population-based data that describe prepregnancy BMI and GWG per trimester and their changes over time in LMIC. Monitoring the trends in those indicators is essential to formulate and redefine policies to reduce and prevent adverse outcomes associated with the deviation in the maternal nutritional status before and during pregnancy. Thus, we aimed to describe the prevalence and temporal trends in prepregnancy BMI and GWG in adult pregnant women followed in the SISVAN from 2008 to 2018.

| Data and main variables
We used data from the SISVAN, an administrative system run by the Prepregnancy weight was self-reported and, together with the first measurement of height (in metres) taken during pregnancy, was used to calculate prepregnancy BMI (kg/m 2 ). The World Health Organization (WHO) cut-offs were used for BMI classification as underweight (<18.5 kg/m 2 ), normal weight (18.5-24.9 kg/m 2 ),

Key messages
• In data from the national surveillance system in Brazil, prepregnancy BMI increased substantially from 2008 to 2018. Overweight increased from 22.6% to 28.8%, whereas obesity doubled from 9.8% to 19.8%.
• The prevalence of GWG within the IOM guidelines was fairly stable throughout the studied period at approximately 33%. Still, the mean total GWG presented a slight increase (~1 kg) between 2008 and 2018 for all BMI categories.
GWG was calculated as the difference between weight measured in each visit and self-reported prepregnancy weight. Whenever there was more than one weight measurement in a trimester for a given woman, the last measurement taken within the trimester was used.
Total GWG was classified as 'below', 'within' or 'above' the guidelines, according to the values proposed by the IOM (US) and National Research Council (US) Committee to Reexamine IOM Pregnancy Weight Guidelines (2009). Only measurements taken after 36 weeks of pregnancy were used for calculating this variable. The classification of total GWG considered the expected GWG at the exact gestational age of the measurement and not at 40 weeks (which is the value used in the guidelines). As an example, the upper limit of the recommendations for a normal-weight woman who had her last weight measured at the 38th week should be calculated as 2 kg for the first trimester (13 weeks) + 0.5 kg/week (the IOM rate for normal-weight women) Â 25 weeks (38-13 weeks, the gestational age difference between the weight measurement and the end of the first trimester) = 14.5 kg. This adjustment in the IOM ranges aims to consider the gestational age at the measurement and is essential because the misclassification of GWG resulting from varying lengths of gestation is possible (Hutcheon et al., 2012).
For variables that remained constant during pregnancy (e.g., maternal age, education, prepregnancy BMI and participation in the PBF), only information from the first visit in each pregnancy was used.

| Data cleaning process
Because the system is based on data collected in routine prenatal care, several steps to clean the data were implemented. To improve data quality, we removed duplicates and identified implausible mater- We implemented two methods to identify outliers for weight at each clinical appointment. These steps considered the plausibility of the measurement relative to previous and subsequent weight measurements. Specifically, we employed the conditional approach method proposed by Yang and Hutcheon (2016)

| Statistical analyses
Continuous variables were summarized using mean and SDs and categorical variables with absolute and relative frequencies. Maps to evaluate changes in the prevalence of excessive GWG in each Brazilian state in triennials (2008-2010, 2011-2013 and 2014-2016) were created. In these maps, only states with n > 100 women in each triennial were included. We did not perform any statistical tests to compare the variation in the key variables throughout the years because the large sample size would mean that even small differences would be statistically significant while not being clinically important.
The analyses were performed in JupyterHub from 'Plataforma de Ciência de Dados aplicada à Saúde' (ICICT/Fiocruz), using R Version 3.6. For the creation of the maps, R package 'brazilmaps' was used.

| Ethical considerations
The Research Ethics Committee of the Rio de Janeiro Federal University Maternity Teaching Hospital approved the project from which this study is part (protocol number: 85914318.2.0000.5275).
All analyses were conducted with deidentified data, and security procedures to protect the access to the data were taken.

| RESULTS
The initial dataset contained 5,178,974 different participants and 22,291,787 weight records. We removed the 10,613,874 duplicate records that resulted from combining data from three sources (i.e., the SISVAN data, data from e-SUS and the PBF database) and 603 records from 2007 that appeared in the registers, which left 11,677,310 records. After identifying and removing outliers and selecting women with prepregnancy weight information, there were 840,243 women with 2,087,765 weight records for analysis ( Figure S1). The high proportion of women removed in the final step (approximately 84% of the initial dataset) primarily resulted from missing data in the prepregnancy weight field, which is not mandatory in the SISVAN form and was often left blank.  Figure S2). However, only 48% of the women presented ≤12 weeks in the first registered visit, and half of the women had only one visit registered in the system.
Prepregnancy BMI increased substantially throughout the studied period. Overweight increased from 22.6% to 28.8% and obesity from 9.8% to 19.8%, respectively, from 2008 to 2018. These results show a 21% and 202% increase in the prevalence of overweight and obesity, respectively, in the period. In contrast, the prevalence of women classified as normal weight decreased from 60.7% to 46.5%, whereas the prevalence of underweight decreased from 6.8% to 4.8% (Figure 1).
According to the IOM, approximately one third of the participants had GWG classified as 'within guidelines', and this proportion was fairly stable throughout the studied period. There was a slight increase in the prevalence of GWG above the guidelines from 34.2% to 38.7%, which coincided with a reduction of women gaining weight below the guidelines from 29.7% to 25.9%, from 2008 to 2018, respectively ( Figure 2).
Mean total GWG varied according to prepregnancy BMI. Women with underweight women showed the highest values (variation from 12.3 to 13.1 kg between 2008 and 2018), followed by normal weight (variation from 11.9 to 12.5 kg), overweight (variation from 10.1 to 10.9 kg) and obesity (variation from 8.2 to 8.9 kg). Within each BMI category, the variation was below 1 kg throughout the studied period.
The pattern for the first and second trimesters was similar to total GWG ( Figure 3).
Excessive GWG became more common in the vast majority of the Brazilian states throughout the study period. Brazil's South Region was the only region with a prevalence of excessive GWG above 45% since the first triennial (2008-2010), a situation that has remained the same throughout the studied period. The Southeast Region had the second highest prevalence of excessive GWG, even though a reduction in the most recent triennial was observed in three of the four states of the region. Excessive GWG increased for almost all states in the Northeast and Center-West regions (Figure 4).

| DISCUSSION
Using routine prenatal care data available in Brazil's SISVAN public health database, we observed a 21% increase in the prevalence of prepregnancy overweight and a 202% increase in prepregnancy obesity between 2008 and 2018. GWG above the IOM guidelines increased 11.6% throughout the years in this sample of Brazilian women. We also found a slight increase in the mean total GWG in all BMI categories during the study period. The use of this national sur-  Although the prevalence of GWG outside of the guidelines was stable in the period, the proportion of women gaining weight above the recommended ranges was high (varying from 34% in 2008 to 38% in 2018). The prevalence of GWG above the guidelines was more pronounced in several states and was above 40% in some regions. GWG above the IOM guidelines has been linked to post-partum weight retention and maternal obesity (Nehring et al., 2011). This becomes more relevant because of the increase in the prevalence of overweight and obesity observed among women, which may be partially connected to the high prevalence of women who exceeded the weight gain guidelines during pregnancy. On the other hand, the prevalence of women who gained below the guidelines was also high, that is, a (1-9 years) or more (Cruz et al., 2018). The increase in the percentage of women who are part of the PBF could have helped in the variation observed in prepregnancy BMI and GWG. This may be because recipients of conditional cash transfer programmes seem to be more vulnerable to the nutritional transition and more affected by the epidemic of chronic diseases (Fernald et al., 2008;Martins et al., 2013).
The mean number of prenatal visits increased from 3 to 4. However, this value is still below 6, the number recommended by the Ministry of Health and the WHO (Ministério da Saúde, 2012). This finding reinforces the need to increase the coverage of the SISVAN because recent studies show that the majority of Brazilian women attend six or more prenatal visits during pregnancy (Tomasi et al., 2017;Viellas et al., 2014). This poses a significant limitation when generalizing the study findings because not all the prenatal visits were registered in the system, and it is not possible to identify the registration selection criteria; that is, it is not possible to know how the health care professional decides when or not to include the information of the visit in the system.

| Strengths and limitations
The large sample size, the availability of data for 10 years and  (Nascimento et al., 2017), but it is still far from ideal. Unfortunately, because the system is based on routinely collected data on prenatal care, the protocol that recommends standardization for anthropometric data collection is not always followed.
This results in a lack of high-quality measurements, which could be both imprecise and inaccurate. To address these challenges to the reliability of the data, we implemented several methods to identify extreme and implausible values before analysing the anthropometric data. These procedures ensured that we worked with more reliable data with plausible distributions.
The form used to collect the data in routine prenatal care reported in the SISVAN includes some non-mandatory fields, such as prepregnancy weight. This led to a high proportion of missing data in the variables based on those fields. In the future, it is important that this type of information should become mandatory and that the system to enter the information should not accept implausible options, such as '9999' or zero, to reduce the proportion of missing data in key variables.
Finally, we calculated both prepregnancy BMI and GWG based on self-reported prepregnancy weight, and bias could have been introduced due to this measurement. However, the results from a previous study conducted by our team in a subsample of the same SISVAN data revealed that self-reported prepregnancy weight has a good agreement with the weight measured in the first trimester, especially when the latter was performed up to 6 weeks of pregnancy (Carrilho, Rasmussen, et al., 2020).

| CONCLUSION
The trends observed in the prevalence of prepregnancy BMI and GWG outside of the IOM guidelines indicate that the nutritional status of these women has worsened in recent years. These findings reveal the need for continuous monitoring of these indicators and the urgent development of strategies to reverse the observed trends.
The changes in sociodemographic and prenatal care characteristics observed in the present study suggest an improvement in maternal education, participation in a conditional cash transfer programme and access to prenatal care in early pregnancy.
The use of administrative data for research purposes, although challenging, represents a unique opportunity to evaluate trends in the prevalence of important indicators among pregnant women, such as BMI and GWG. Thus, effective strategies to improve the quality of the data, such as full completion of the data forms, should be made by the Ministry of Health so that the SISVAN data can be used with confidence to generate evidence-based policies for the country. Finally, the quantity of data and the ability to examine trends over time help to fill an important gap in understanding the distribution of prepregnancy BMI and GWG in women who live in LMIC.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the Brazilian Ministry of Health. Restrictions apply to the availability of these data, which were used under licence for this study. Data are available upon request to the Ministry of Health, following specific Brazilian laws.