Maternal Nutrition and Birth Outcomes: Effect of Balanced Protein-Energy Supplementation

Authors

  • Aamer Imdad,

    1. Division of Women and Child Health, Aga Khan University, Karachi, Pakistan
    Search for more papers by this author
  • Zulfiqar A. Bhutta

    Corresponding author
    1. Division of Women and Child Health, Aga Khan University, Karachi, Pakistan
      Zulfiqar A. Bhutta, Husein Lalji Dewraj Professor and Head, Division of Women and Child Health, The Aga Khan University, Karachi 74800, Pakistan. E-mail: zulfiqar.bhutta@aku.edu
    Search for more papers by this author

Zulfiqar A. Bhutta, Husein Lalji Dewraj Professor and Head, Division of Women and Child Health, The Aga Khan University, Karachi 74800, Pakistan. E-mail: zulfiqar.bhutta@aku.edu

Abstract

The nutritional status of a woman before and during pregnancy is important for a healthy pregnancy outcome. Maternal malnutrition is a key contributor to poor fetal growth, low birthweight (LBW) and short- and long-term infant morbidity and mortality. This review summarised the evidence on association of maternal nutrition with birth outcomes along with review of effects of balanced protein-energy supplementation during pregnancy. A literature search was conducted on PubMed, WHOLIS, PAHO and Cochrane library. Only intervention studies were considered for inclusion and data were combined by meta-analyses if available from more than one study. Sixteen intervention studies were included in the review. Pooled analysis showed a positive impact of balanced protein-energy supplementation on birthweight compared with control [mean difference 73 (g) [95% confidence interval (CI) 30, 117]]. This effect was more pronounced in undernourished women compared with adequately nourished women. Combined data from five studies showed a reduction of 32% in the risk of LBW in the intervention group compared with control [relative risk (RR) 0.68 [95% CI 0.51, 0.92]]. There was a reduction of 34% in the risk of small-for-gestational-age babies in the intervention compared with the control group [RR 0.66 [95% CI 0.49, 0.89]]. The risk of stillbirth was also reduced by 38% in the intervention group compared with control [RR 0.62 [95% CI 0.40, 0.98]]. In conclusion, balanced protein-energy supplementation is an effective intervention to reduce the prevalence of LBW and small-for-gestational-age births, especially in undernourished women.

The nutritional status of a woman before and during pregnancy is important for a healthy pregnancy outcome.1 Maternal malnutrition is a key contributor to poor fetal growth, low birthweight (LBW) and infant morbidity and mortality and can cause long-term, irreversible and detrimental cognitive, motor and health impairments.2–4 Undernutrition in females may occur during childhood, adolescence and pregnancy, and has a cumulative adverse impact on the birthweight of future babies.2

This review will discuss the following aspects of maternal nutrition:

  • 1an historical background of famine studies to describe the association between maternal malnutrition and birth outcomes (mostly observational studies);
  • 2to assess the effect of balanced protein-energy supplementation during pregnancy on pregnancy outcomes (experimental studies).

Famine studies

During the World War II, several studies were carried out in the Netherlands Leningrad and Germany, and knowledge about maternal malnutrition and its relation with birth outcomes mostly derived from these studies. The common aspect in all these studies was that there was an acute shortage of food as a result of war. Table S1 provides a brief description of famine studies that reported data on maternal nutrition and birth outcomes.

The Dutch famine lasted 6 months during the winter of 1944–45.5–7 The food intake of mothers was markedly decreased as the official rations fell as low as 590 calories a day. This food rationing resulted in maternal weight loss of as much as 2.5 kg from pre-famine levels. The birthweight was also affected and the mean birthweight fell by about 300 g at the height of the famine.5 The greatest effects on birthweight were among infants conceived before the onset of, but delivered during, the famine.5,6 Women in their third trimester of pregnancy were affected the most and there was little difference in impact between those exposed to the famine only during the third trimester and those exposed during both the second and the third trimesters. No adverse effect of famine was observed among infants conceived during, and exposed to, famine through the second trimester but whose mothers received inadequate nutrition during the third trimester. During the height of famine, perinatal mortality rates were raised sixfold, from about 4 per 1000 to as high as 24 per 1000 births. A recent study by Stein et al. showed that infants exposed to intrauterine famine may be predisposed to the development of hypertension, obesity and diabetes mellitus in middle age.8

Data from the famine in Leningrad, the former Soviet Union was reported by Antonov.9 The siege in Leningrad was extremely severe and of longer duration. It lasted from August 1941 to January 1943, while the worst period was from September 1941 to February 1942. Food intake was markedly decreased as bread rations fell to 250 g per day among manual and 125 g per day among non-manual workers. A review of data from one of the clinics showed that mean birthweight of liveborn infants during the famine was 2789 g, a decrease of almost 550 g. Perinatal mortality was substantially high as over a quarter of babies died during January–June 1942.10

There was a shortage of food in many parts of Germany, after the end of World War II. Data from a town named Wuppertal, Germany were reported by Dean.11 The siege in Germany was not as severe as those in Holland and Leningrad. Official rations fell as low as 1052 kcal a day. Data from one clinic showed that during the worst deprivation, mean birthweight was lower, on average, by 170 g among private patients and by 227 g among public patients. This decrease was later reversed; when official rations were as high as 1550 kcal per day, birthweight drop was, on average, 81 g among private patients and 117 g among public patients.11

Very recently, Huang et al. reported data from the Chinese Famine (1959–61).12 The Chinese Famine was much longer and more severe than those described above. It is considered to be one of the longest in human history, causing up to 30 million deaths.13 Although all parts of China were affected by famine, its severity and duration varied across different areas. In rural areas affected by famine, birthweights were greater by 72 g in the famine group than the offspring of women born in 1963 and unexposed to the famine. There was no association of famine with offspring birth size in urban areas. The authors proposed that markedly increased mortality in rural areas may have resulted in the selection of hardier mothers with greater growth potential, which becomes expressed in their offspring.12

In conclusion, acute severe maternal malnutrition may adversely affect the birthweight of the fetus especially when the exposure is during the third trimester of the pregnancy. There appears to be an increased risk of perinatal mortality when there is exposure to intrauterine famine and the risk may be as high as sixfold in comparison with those not exposed to famine.

Effect of balanced protein-energy supplementation during pregnancy on birth outcomes

This section will focus on a review of studies on macronutrient food supplementation for pregnant women. During pregnancy extra energy is required for the growth of the fetus, placenta and various maternal tissues, such as in the uterus, breast and the fat stores.14 The ideal situation for a woman is to enter pregnancy with a normal weight and good nutritional status. Pre-pregnancy weight is a strong predictor for LBW.1 Major determinants for LBW in low- and middle-income countries are poor maternal nutritional status [low body mass index (BMI)] at conception, inadequate gestational weight gain due to poor dietary intake and short maternal stature due to mother's own childhood undernutrition.15

Differences in the body size, life style and nutritional status of the mother3,16 underscore the need for population-specific estimates of energy requirements and recommendations for energy intake of a pregnant woman. Several nutrition interventions have been proposed and evaluated in accordance with the maternal needs during pregnancy.17 Some of these include dietary advice to pregnant women (as discussed in another paper in this supplement), balanced protein/energy supplementation (protein provides <25% of total energy content), high protein (the protein provided 25% of the total energy content), isocaloric protein supplementation (the protein replaces an equal quantity of non-protein energy) and prescribing low-energy diet to pregnant women who either are overweight or exhibit high weight gain earlier in gestation.17–19 Among these interventions, balanced protein-energy supplementation is considered as one of the most promising macronutrient interventions in prevention of adverse perinatal outcomes including intrauterine growth restriction.17,20

Studies from the UK21,22 and Chile23 found that there was no positive effect on pregnancy outcomes when maternal energy intakes were isocalorically replaced with 10–11% of protein. Even higher levels of protein supplementation (>25% of energy) in relatively well-nourished populations did not show any benefits on pregnancy outcomes.24,25 Findings from a recent meta-analysis by Kramer and Kakuma20 also showed no effect of unbalanced protein supplementation on pregnancy outcomes like birthweight and small-for-gestational-age babies. Balanced protein-energy supplementation is designed to provide less than 25% of total energy content and is believed to be the most suitable supplement for malnourished pregnant women. In this section we review data from intervention studies examining balanced protein-energy supplementation during pregnancy and its effects on adverse birth outcomes.

Methods

To assess the effect of balanced protein-energy supplementation, a literature search was performed on different electronic databases. The search strategies used in different databases are given in Appendix S1. We synthesised the collective evidence from different interventional studies with the help of meta-analysis.

Inclusion and exclusion criteria

  • • Balanced protein-energy supplementation was defined as nutritional supplementation during pregnancy in which protein provided less than 25% of the total energy content.
  • • Those studies were excluded where the main intervention was dietary advice to pregnant women for increase in protein/energy intake, high protein supplementation (i.e. supplementation in which protein provides at least 25% of total energy content), isocaloric protein supplementation (where protein replaces an equal quantity of non-protein-energy content) or low energy diet to pregnant women who either are overweight or exhibit high weight gain earlier in gestation.
  • • Only intervention studies (randomised, quasi-randomised trials and before after design) were included in the meta-analysis.
  • • The comparison group include either routine diet or no intervention.

Quantitative data synthesis

The primary outcomes of interest included LBW, small-for-gestational age and mean birthweight (g). We abstracted data on other neonatal and maternal outcomes including neonatal mortality, stillbirth, birth length, head circumference, pre-eclampsia, gestational weight gain/week and Bayley mental score at 1 year. The summary estimates from the meta-analyses were presented as relative risk (RR) for dichotomous data and mean difference (MD) for continuous data along with their corresponding 95% confidence interval (CI). Assessment of statistical heterogeneity in the pooled data was done by visual inspection of forest plots, by the chi-square (P-value) and by calculating the I2 statistic [calculated as I2 = (Q − d.f.)/Q, where Q is Cochrane's heterogeneity statistic and d.f. is the degrees of freedom]. Heterogeneity was assumed to be substantial when the P-value of chi-square test was <0.10, and/or I2 exceeded 50%. Reasons for heterogeneity were explored by doing a sensitivity analyses that included removing studies with large attrition and/or small sample size.

Estimates of pooled effect measures were generated from either the fixed-effects models or the random-effects models, the latter used when there was substantial heterogeneity across the pooled studies (I2 ≥ 50%). Data from cluster randomised trials were pooled with individually randomised trials. In this case, cluster adjusted values were used as given in the original study; however, if results were not adjusted for cluster randomisation, sample sizes were adjusted by using an estimate of the intracluster correlation coefficient derived from the trial, or inferred from similar studies.26

All analyses were conducted using software Review Manager version (version 5).27 The quality of overall evidence was assessed by GRADE criteria.28 According to this grading system, the quality of overall evidence was graded as ‘high’, ‘moderate’, ‘low’ or ‘very low’. A score of ‘high’ means that further research is very unlikely to change the results of the intervention. A score of ‘moderate’ means further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate, and a score of ‘low’ means that further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. A score of ‘very low quality’ means that we are very uncertain about the estimate.28

Results

Figure 1 shows the results of literature search. We identified 5606 titles from searches conducted on all databases. After screening the titles and abstracts, 24 studies were identified that addressed protein-energy supplementation during pregnancy. A total of 16 interventional studies were chosen for data extraction to conduct meta-analyses.21–23,25,29–40Table 1 presents the characteristics of included studies. Nine studies were excluded and the reasons for exclusion are provided in Table 2. Eight of the included studies were from developing countries30–32,34,35,37,38,40 and eight were from developed countries.21–23,25,29,33,36,39 In 10 of the included studies, women were undernourished (as defined by authors) and/or were at risk of delivering a LBW baby.23,25,29,31,34–36,38,39,41

Figure 1.

Flow diagram for identification of interventional studies evaluating balanced protein-energy supplementation during pregnancy.

Table 1.  Characteristics of studies included in the meta-analyses Thumbnail image of
Table 2.  Characteristics of excluded studies from meta-analyses
Study ID (reference)Reason for exclusion
Kardjati 198843Both the groups received the nutrition supplement
Lechtig 197546The control group received a low-energy drink
Anderson 199547Nutritional education only
Briley 200248Nutritional education only
Hankin 196249Nutritional education only
Hunt 197650Nutritional education only
Kafatos 198951Nutritional education only
Sweeney 198552Nutritional education only
Iyengar 196724Isocaloric protein-energy supplementation
Rasmussen 201053High protein-energy supplement

Sixteen studies reported data on birthweight and the pooled estimate showed that babies born to women who received balanced protein-energy supplementation had a higher birthweight compared with controls (MD 73 [95% CI 30, 117]) (Figure 2). There was a substantial heterogeneity in the pooled data (I2 = 80%) and thus the random-effects models were used. A subgroup analysis based on nutritional status of the mothers showed that balanced protein-energy supplementation was more effective in malnourished women (MD 100 [95% CI 53, 147]) than adequately nourished women (MD 37 [95% CI −34, 99]). Data on the incidence of LBW (birthweight <2500 g) were available from five studies and the pooled results showed that balanced protein-energy supplementation resulted in a 32% reduction in LBW prevalence in the intervention compared with the control group (RR 0.68 [95% CI 0.51, 0.92]) (Figure 3). Random-effects models were used as there was heterogeneity in the pooled data (I2 = 52%).

Figure 2.

Effect of balanced protein-energy supplementation on birthweight (g).

Figure 3.

Effect of balanced protein-energy supplementation on risk of low birthweight. M-H, Mantel-Haenszel.

Pooled results from nine studies reporting data on the prevalence of small-for-gestational age showed a reduction of 34% in the intervention compared with the control (RR 0.66 [95% CI 0.49, 0.89]) (Figure 4). As the pooled estimate had significant heterogeneity (I2 = 87%), random-effects models were used. The risk of stillbirth in the intervention group was 38% lower based on pooled data from three studies (RR 0.62 [95% CI 0.40, 0.98]) (Figure 5). Balanced protein-energy supplementation had virtually no effect on preterm birth rates (Figure 6). Table 3 shows summary estimates of some of the other maternal and neonatal outcomes. Table 4 shows ‘summary of findings’ table for selected outcomes according to the GRADE criteria.

Figure 4.

Effect of balanced protein-energy supplementation on risk of small-for-gestational-age baby. M-H, Mantel-Haenszel.

Figure 5.

Effect of balanced protein-energy supplementation on risk of stillbirth. M-H, Mantel-Haenszel.

Figure 6.

Effect of balanced protein-energy supplementation on risk of preterm birth. M-H, Mantel-Haenszel.

Table 3.  Summary estimates of other maternal and neonatal outcomes
OutcomeNo. of studiesNo. of participantsSummary estimateFixed/random model
  1. RR, relative risk; MD, mean difference.

Neonatal mortality43361RR = 0.68 [0.59, 0.82]Fixed
Pre-eclampsia3516RR = 1.20 [0.77, 1.89]Fixed
Gestational age (weeks)93087MD = −0.03 [−0.26, 0.21]Random
Birth length (cm)73698MD = 0.16 [0.02, 0.31]Fixed
Birth head circumference (cm)73680MD = 0.07 [−0.02, 0.16]Fixed
Gestational weight gain/week (g)102571MD = 20.74 [1.46, 40.02]Fixed
Bayley mental scores at 1 year1411MD = −0.74 [−1.95, 0.47]Fixed
Weight at 1 year (g)2623MD = 30.43 [−139.67, 200.53]Fixed
Table 4.  Results of pooled analysis and qualitative grading according to GRADE criteria
Quality assessmentSummary of findings
No. of studiesDesignLimitationsConsistencyDirectnessNo. of eventsEffect
Generalisability to population of interestGeneralisability to intervention of interestInterventionControlRelative risk [95% CI]
  1. RCT, randomised controlled trial.

Outcome: Birthweight (g): Quality of evidence – moderate
16RCTs/cluster RCT/quasi-RCTs, before after designMethods of sequence generation and allocation were inadequate in some of the included studiesExcept three studies, all the studies showing a positive effect in favour of intervention. Significant statistical heterogeneity (I2 = 76%). Random models usedStudies conducted in both developed and developing countries. All the studies did not included undernourished womenProtein content of Supplement for intervention group ranged from 30 to 44 g per day. The protein content provided <25% of total energy content in all the studies35942880Mean difference 73.78 [30.42, 117.15]
Outcome: Small-for-gestational age: Quality of evidence – moderate
9RCTs/cluster RCT/quasi-RCTs, before after designTwo studies were quasi-experimental trial. Sequence generation and allocation concealment was not adequate in some of the included studiesAll the studies showing a positive effect in favour of intervention. Significant statistical heterogeneity in the pooled data (I2 = 85%). P = 0.005. The sole contributor to this study was study by Mardones-Santander et al. 198823Studies conducted in both developed and developing countries. All the studies did not included undernourished womenProtein content of supplement for intervention group ranged from 30 to 44 g per day. The protein content provided <25% of total energy content in all the studies28632387Relative risk 0.68 [0.51, 0.92]
Outcome: Low birthweight (<2500 g): Quality of evidence – moderate
5RCTs/cluster RCT/quasi-RCTs, before after designHigh risk of bias because of inadequate sequence generation and allocation concealment in one of the included studies. One of the studies had before after designAll the studies showing a positive effect in favour of the intervention. Significant statistical heterogeneity (I2 = 52%). Random models usedStudies conducted in both developed and developing countries. All the studies did not included undernourished womenProtein content of supplement for intervention group ranged from 30 to 44 g per day. The protein content provided <25% of total energy content in all the studies226296Relative risk 0.68 [0.51, 0.92]
Outcome: Stillbirth: Quality of evidence – low
3RCTs/cluster RCT/quasi-RCTs, before after designHigh risk of bias because of inadequate sequence generation and allocation concealment in one of the included studiesAll the studies showing a positive effect in favour of the intervention. No significant statistical heterogeneity (I2 = 20%)Studies conducted in both developed and developing countries. All the studies did not included undernourished womenProtein content of supplement for intervention group ranged from 30 to 44 g per day. The protein content provided <25% of total energy content in all the studies1831Relative risk 0.55 [0.31, 0.97]
Outcome: Neonatal mortality: Quality of evidence – low
3RCTs/cluster RCT/quasi-RCTsOne quasi-experimental design. Allocation concealment was not adequate for one of the included cluster RCT. Large loss to follow-up in included studiesNo heterogeneity (I2 = 0). P = 0.81One study from a developed country and two from developing countriesProtein content of supplement for intervention group ranged from 30 to 44 g per day. The protein content provided <25% of total energy content2333Relative risk 0.63 [0.37, 1.06]

Comments

The effects of balanced protein-energy supplementation during pregnancy on adverse birth outcomes have been evaluated before, including a Cochrane review20 and the other being a review for the Live Saved Tool (LiST) model.42 Our results are in accordance with these reviews with some additions and modifications for certain outcomes. We updated literature search and used slightly different inclusion/exclusion criteria by including quasi-experimental trials and before–after studies. This led to the inclusion of five more studies compared with the previous reviews.23,37–40 Two of these studies were randomised trials,23,40 two were quasi-randomised37,39 and one was a before–after study.38 Of the two new randomised trials included in this review, one is a new study40 and the second was a previous study.23 In the second study we had compared group of balanced protein-energy supplements with control (taken as those who were non-complier) instead of that with isocaloric protein-energy supplementation. In the review by Kramer and Kakuma,20 this study has been included as that with isocaloric protein-energy supplementation. We have conducted a new meta-analysis for the prevalence of LBW that was not attempted before and have updated other outcomes.

Pooled data from 16 studies showed that balanced protein-energy supplementation has a positive impact on birthweight (MD 73 [95% CI 30, 117]). Similar results were found in the LiST review (MD 59 g [95% CI 33, 86]); however, the results of Cochrane review were statistically non-significant (MD 37 [95% CI −0.21, 75]). The differences in the magnitude and statistical significance of the summary estimates in the current review compared with LiST and Cochrane review are the addition of five more studies23,37–40 and exclusion of one study43 included in the Cochrane but excluded in this review and also that in the LiST review. The main reason for exclusion of this study was that both the study groups received the supplement (high vs. low energy) and it was difficult to ascertain the true effect of the intervention.

The results of the pooled estimate on the prevalence of LBW were consistent with the positive impact on mean birthweight and maternal weekly gestational weight gain. Combined results from five studies showed that balanced protein-energy supplementation reduced the prevalence of LBW by 32%. This analysis was not attempted before and contributes an important data of the effectiveness of balanced protein-energy supplementation in reducing LBW in developing countries. Similarly, the prevalence of small-for-gestational age was also reduced by 44%. The direction of effect was the same as that of LiST and Cochrane review; however, the magnitude of effect was different because of addition of three more studies in this analysis.23,38,40

The included studies were of variable quality with most of the studies conducted in the 1980s and 1990s; the most recent study was conducted in 2002 in Iran.37 Methods of randomisation and allocation concealment were inadequate in most of the studies. Even though the quality of methods varied across these studies, the direction and magnitude of effect was quite consistent. The qualitative assessment of the pooled estimates for birthweight and small-for-gestational-age babies was that of moderate level based on limitation of methods in some of the included studies.

Participants in 11 of the studies were categorised as malnourished as defined by the authors (Figure 2). In most of the studies this assessment was based on general nutritional status of the study population. Only five of studies used defined criteria to recruit women who were undernourished or were at risk of having a LBW baby.22,23,25,32,44 The criteria used across these studies included, for example, pre-pregnancy weight, low weight-for-height, history of a LBW baby, low maternal weight gain and triceps skinfold thickness. None of the studies used BMI as the recruitment criterion. In any case, the subgroup analysis based on nutritional status of the mothers showed that balanced protein-energy supplementation was more effective in malnourished women (MD 100 [95% CI 53, 147]) than adequately nourished (as defined by authors) women (MD 37 [95% CI −34, 99]). It can therefore be inferred from this analysis that a food supplement with balanced protein-energy content seems the most suitable intervention for malnourished women to increase birthweight and can subsequently reduce the risk of LBW and small-for-gestational-age babies in these women. This finding should; however, be interpreted carefully as most of the studies did not use a defined criterion to define undernourished women and no study used low BMI as the inclusion criterion. The implementation of intervention should also be considered according to the social, cultural and economic context of the target population.

The fact that most of the studies did not used standardised criteria to define undernourished women calls for future studies with more rigorous inclusion criteria and improved methods for conduct of the studies. Another important consideration for future research is the combination of balanced protein-energy with micronutrient supplementation. A recent review by our team for the LiST45 has shown that multiple micronutrient supplementations can reduce incidence of small-for-gestational-age babies by 9%. A recent study from Burkina Faso has shown that combined supplementation with balanced protein-energy and appropriate multiple micronutrients has more pronounced effect on birth length than multiple micronutrients alone.40 Future studies should be encouraged to replicate these findings in other parts of the world.

In conclusion, balanced protein-energy supplementation seems an effective intervention to reduce the risk of LBW and small-for-gestational-age births, especially in undernourished women in underdeveloped countries.

Conflicts of interest

The authors declare that they do not have any conflicts of interest.

Ancillary