INTRODUCTION

Weight gain during gestation is an important predictor of newborn outcomes ¹. Insufficient gain increases the risk of preterm and low birth weight (LBW) births ^{2 , 3}, while excessive gain increases the risk of neonatal macrosomia ⁴. In turn, both LBW and macrosomia increase neonatal mortality and morbidity ¹, adverse maternal outcomes ², and the risk of chronic diseases during adulthood ^{2 , 5}, which means a significant economic burden on health services and society ⁶.

During the 1970s and 1980s, a single gestational weight gain of 12.5 kg was recommended in most countries, regardless of pre-pregnancy weight and height ⁷. In 1983, the Latin American Center for Perinatology (CLAP) of the Pan American Health Organization developed weight gain patterns when pre-pregnancy weight was unknown ⁸. While in 1990, the Institute of Medicine (IOM) of the United States published recommendations for weight gain during pregnancy, adopted by most countries worldwide, and updated in 2009 ¹.

In Peru, the recommendations from CLAP and the IOM are commonly used because of the absence of a national pattern of weight gain during pregnancy. However, both international patterns present limitations that affect their diagnostic accuracy, such as the origin and age of the basic epidemiological data, the representativeness of the reference population, and methodological aspects, such as the inclusion and source of information on pre-pregnancy body mass index (BMI) data. For this reason, a systematic review was carried out to evaluate the ability of both recommendations to predict LBW and neonatal macrosomia.

MATERIALS AND METHODS

This systematic review seeks to evaluate the best diagnostic parameter for pregnancy weight gain. The study was conducted following the recommendations of the PRISMA statement ⁹, except for the protocol’s registration in a database, which is available upon request to the corresponding author.

RESULTS

A total of 2,601 references were identified. The systematic review included 5 studies that met the previously established selection criteria ^{11 - 15} (Figure 1).

Figure 1 Study selection flowchart 

Quality evaluation

All studies presented a high risk of bias regarding patient selection and the reference test. Most studies showed a high risk of bias regarding the index, flow and time tests. As for the applicability of the studies, most of them presented high risk in the patient selection and low risk in the reference test; the results of the index test were variable (Table 4).

Table 4 Quality evaluation of the included studies according to QUADAS 2 

DISCUSSION

Nowadays, there is a greater understanding of the interaction between maternal diet, pre-pregnancy BMI, weight gain during pregnancy, maternal comorbidities, and immediate and long-term health outcomes ¹⁶. However, optimal gestational weight gain has not yet been clearly defined and remains a controversial and critical issue ¹⁷. There are weight gain recommendations from the IOM and CLAP, which are commonly used in prenatal care. In this context, the study aimed to evaluate the recommendations for gestational weight gain proposed by the IOM and CLAP based on their ability to predict birth weight and neonatal macrosomia knowing that they have been associated with a significant increase in the risk of adverse maternal and neonatal outcomes ¹⁸.

The findings suggest that the country where the study takes place influences both outcomes. With respect to low birth weight, the findings were consistent, showing similar values for sensitivity and specificity among the Brazilian cohort of adult pregnant women ¹⁴ and adolescent pregnant women ¹³; different from those obtained in a Chinese cohort, which presents lower sensitivity and higher specificity ¹⁵. These differences were also observed in the outcome of neonatal macrosomia, where the Brazilian cohort obtained less sensitivity and specificity ¹³, compared with the Italian ¹¹, American ¹², and Chinese ¹⁵ cohorts.

Some studies describe differences in pre-pregnancy nutritional status, weight gain during pregnancy, and the magnitude of risk of adverse gestational and neonatal outcomes around populations of different continents and ethnicities ^{19 , 20}. In that sense, the IOM proposed the development of additional studies to investigate whether the recommendations for weight gain during pregnancy differ according to the racial/ethnic group, suggesting that they could reflect not only the biological differences between the groups, but also the difference in their socio-economic conditions ¹.

On the other hand, the studies included in our review reported a higher proportion of negative predictive values, mostly above 90%, compared to the proportion of positive predictive values, mostly below 25% for both outcomes. Positive and negative predictive values indicate the probability that a diagnostic test will obtain the correct diagnosis if it turns out to be positive or negative, while sensitivity and specificity assess the probability of obtaining a positive or negative result, depending on the patient’s true condition. In a scenario where this condition is previously unknown, predictive values could be relevant to support clinical decisions ^{21 , 22}.

However, the prevalence of the disease or condition to be evaluated influences the predictive values. When it is low, the negative predictive value is higher and, consequently, the positive predictive value is lower ^{21 , 23}. In this context, this review reports higher negative predictive values, which could mean an important characteristic and would allow the presence of low birth weight and neonatal macrosomia to be ruled out with greater security and confidence in settings where both prevalences are relatively low.

The average age of the adult pregnancy cohorts included in our review ranged from 26 to 31 years. This epidemiological characteristic differs from the fecundity of Peruvian women. In Peru, the average age at the first birth is 22.3 years, and 12.6% of women between 15 and 19 years have been pregnant at some time ²⁴, which increases the risk of age-related adverse obstetric and neonatal outcomes ²⁵. The exact origin of this relationship has not been determined. One hypothesis suggests that the older the mother, the greater the body weight before pregnancy, due to the cumulative weight increase in adulthood ²⁶, while in older pregnant women the influence of the natural aging process and the existence of comorbidities cannot be ruled out ²⁵. Similarly, it has been demonstrated that adolescent pregnancy increases the obstetric and neonatal risk related to biological and social factors ²⁷.

Three studies reported the characteristics of maternal body mass index before pregnancy ^(13-15). In these 3 studies, 2.5% to 11.2% of the patients presented with thinness, and 13.6% to 25.3% presented with overweight or obesity before gestation. Both conditions differ from the characteristics of women of childbearing age in Peru, where thinness measured by BMI is found in 1.6%, and excess weight (overweight or obesity) in 62.2% ²⁴. We believe that these differences could be explained by the age of the cohorts, between 1999 and 2009. These discrepancies could influence in the accuracy of the results of our review, given that there are studies that show that pre-pregnancy thinness is associated with an increased risk of low birth weight ^{28 , 29}, while pre-pregnancy overweight and obesity is associated with an increased risk of neonatal macrosomia ²⁹. For this reason, it is important that recommendations on weight gain during pregnancy should consider the pre-pregnancy body mass index.

Three studies included pregnant women with gestational diabetes, with an average prevalence of 1.3% to 2.6%, and 1 study excluded pregnant women with this condition. Two studies did not report controlling this variable. In all cases, this situation limits the representativeness of the results, given that the estimated prevalence of gestational diabetes affects up to 16% of pregnant women in Peru ³⁰. Gestational diabetes triples the risk of macrosomic births, while treatment of this disease reduces the incidence of macrosomia in neonates. The proposed physiological mechanisms to explain the relationship between gestational diabetes and macrosomia are explained by fetal hyperinsulinemia, increased glucose use, and increased fetal adipose tissue ³¹.

There is no gold standard that properly assesses the fetal development and, at the same time, provides an error-free classification of all subjects assessed, verifies all index test results, and that can be performed within a short time interval that avoids changes in the target condition ³². Measuring fetal growth by using ultrasound or MRI does not comply with the conditions described above ^{33 , 34}; this is why gestational weight gain is used for an approximate measure of fetal growth, it is simple to use and low cost. However, misclassification due to an imperfect standard inevitably leads to positive or negative biased accuracy.

Worldwide, it is estimated that 47% of pregnant women had a gestational weight gain higher than IOM’s recommendation, and 23% had a lower gain than those recommendations ¹⁸. In this context, it has been identified that social determinants configure a complex problem that goes beyond solely biological issues. Within this set of social determinants, parity could influence the pre-pregnancy body mass index ³⁵; periods between pregnancies under 18 months and over 59 months could increase the risk of low birth weight ³⁶; low educational level increases the risk of insufficient or excessive weight gain during pregnancy ³⁷; psychological factors, such as emotional distress or dissatisfaction with body image, influence in weight gain ³⁸, or low levels of physical activity during pregnancy increases the risk of excessive weight gain ³⁹. Each of these social determinants is accompanied by underlying factors that interrelate to form a complex causal model.

One systematic review reported that quantifying weight as an isolated intervention in antenatal care is not effective in reducing weight gain during pregnancy ⁴⁰. Thus, it should be part of the interventions made by health professionals in collaboration with mothers to achieve weight gain goals, such as to monitor and provide feedback on their progress ⁴⁰, to have standardized and evidence-based approaches ⁴¹, to improve knowledge and beliefs about weight gain in women of childbearing age ⁴², to incorporate a holistic approach integrating not only the pregnant woman, but also the family or her social environment ⁴³, to provide nutritional consultation on healthy food, oriented to calories, needed macronutrients and adequate levels of physical activity ⁴⁴, to accompany the dietary activities with strategies of behavior change, mainly in the provision of information, motivational interviews, self-monitoring and contingent rewards ⁴⁵, to provide prenatal care with a multidisciplinary approach ^{46 , 47}. The exact duration, frequency, intensity or an optimal environment for providing these interventions has not been determined ⁴⁸.

One of the most important limitations of this systematic review is that the majority of the included studies were retrospective cohorts that gathered information from medical records. Some other limitations, due to the study type: difficulty in controlling confounding factors (gestational age, mother’s age, socioeconomic level, pre-existing diseases, among others), often there are no variables that could narrow the analysis or that could be measured precisely; additionally, not ensuring a high percentage of participation and follow-up increases the risk of probable selection bias. For example, the included studies had between 20% and 40% of data loss, mostly related to the variables of interest. Only 1 study ¹³ analyzed the related characteristics between the studied population and the losses, and found significant differences between both groups. The heterogeneous methodology for measuring the variables of interest may not have been adequately controlled, which increased the risk of bias.

Our review focused on the ability to predict low birth weight and neonatal macrosomia as recommended by the IOM on gestational weight gain. However, inadequate gestational weight gain is also related to maternal outcomes: the lower probability of starting or continuing breastfeeding ⁴⁹, increased postpartum weight retention ⁵⁰, the risk of obesity, diabetes, heart disease, endocrine disease, and psychiatric disease in later years ⁵¹. In the newborn, it is associated with an increased risk of obesity, metabolic and cardiovascular disease in later life ⁵².

We did not find studies that evaluated the ability of CLAP weight gain recommendations to predict LBW and neonatal macrosomia. As for the IOM recommendations, sensitivity and specificity was low, especially in Latin American cohorts. However, the low methodological quality of the studies, the limited representativeness of the studied cohorts and the probable unadjusted confounding factors indicate that there is a need to develop studies with an adequate methodological quality to allow the establishment of recommendations that would fit the epidemiological characteristics of the Peruvian population. Likewise, it is absolutely necessary to prevent inadequate weight gain during pregnancy, providing quality antenatal care, based on a multidisciplinary approach, with interventions supported by scientific evidence and considering the different social determinants involved in its development.