Methods
Search strategy
A systematic literature search was undertaken in August to October 2014 by 2 independent student dietitians (J.M. and H.J.). A starting date of 1990 was specified so that the outcomes reflected the current generation of women whose pregnancy advice may have been influenced by the IOM guidelines. We searched Ovid Medline, Cochrane Library, Excerpta Medica DataBASE (EMBASE), Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Science Direct for studies that reported energy intake in early and late pregnancy and GWG in singleton pregnancies in women of any age. Randomized controlled trials (RCTs) and observational, cohort, and longitudinal studies were eligible for inclusion. The following search terms were used: “pregnant” OR “pregnancy” OR “pregnant woman” OR “gestation” OR “maternal” AND “energy intake” OR “macronutrient” OR “dietary fat” OR “dietary proteins” OR “dietary carbohydrate” OR “dietary intake” OR “calorie intake” OR “kilojoule intake” AND “weight gain” OR “body weight” OR “weight change” OR “body mass index” OR “BMI.” Hand-searching was conducted to identify additional studies. Studies reported as withdrawn in the database, and retrospective studies that preceded 1990 were excluded.
Study selection
Full-term pregnancy was defined as 37–42 weeks’ gestation. Women were classed as underweight, normal, overweight, and obese category according to the IOM criteria. Countries were classified as “developed” or “developing” on the basis of the criteria from the United Nations. In relation to energy intake, early and later pregnancy were defined by time points (t 1 and t 2 ) at least 12 or more weeks apart, where t 1 <18 weeks and t 2 >30 weeks’ gestation (studies reporting data at intervals <12 weeks were excluded). GWG was recorded as the mean ± SD, where data was collected at <18 weeks (t 1 ) and >34 weeks’ gestation (t 2 ), except in 2 studies, where the value was calculated as the difference in weight at the 2 time points and the SD was calculated. Studies published in a language other than English were excluded if a translation was not available. In the RCTs, the control and intervention groups were analyzed as separate groups. Efforts were made to contact authors for additional data regarding their respective studies.
Data extraction
Data were extracted independently by the use of standardized forms in an Excel spreadsheet (Microsoft, Redmond, WA) that collected information on author, title, study type, year published, quality rating, population characteristics (country, age, number of participants, BMI, parity), dietary collection method, weeks’ gestation at time of data collection, energy intake at 2 time points (t 1 and t 2 ), macronutrient intake (g or % energy), weight (t 1 and t 2 ), and GWG. Data were cross-checked for accuracy and discrepancies resolved through discussion or involvement of a third party (J.C.B.M. or J.C.Y.L.).
Statistical analysis
The primary outcome measures were standardized mean difference (SMD) in energy intake and GWG from early to late pregnancy. Data were meta-analyzed collectively and stratified by developed and developing countries, BMI (underweight, normal, overweight, and obese), study design (observational and RCT), and dietary assessment methodology. A random-effect model assumed heterogeneity among studies. The Mood median test was used to test the equality of medians of SMD for energy intake and weight gain between developed and developing countries. Because of small sample sizes within each BMI group, the median GWG and interquartile range (IQR) were used to assess mean weight gain compared with the IOM recommendations.
To calculate SMDs of mean weight gain between 2 time points (t 1 and t 2 ), a Spearman correlation coefficient of 0.85 was applied. Similarly for the 26 subgroups with reported energy intake at t 1 and t 2 , a Spearman correlation coefficient of 0.74 was assumed. For the studies that provided a range for weight rather than SD, a value was imputed where r=0.85. Analyses were repeated with r=0.8 or 0.9 and r=0.7 or 0.8 for weight and energy, respectively; however, this did not alter findings. Data were analyzed with the Comprehensive Meta-Analysis (CMA) package, version 2.2 (Biostat, Englewood, NJ), and presented in the form of forest plots. P values of <.01 were considered statistically significant because 7 comparisons were made in this study, including BMI, country’s economic status, dietary collection method, study type, energy intake, macronutrient distribution, and GWG. This was achieved using Bonferroni correction, which divides the original P = .05 by the number of estimates made, producing a new P = .007, which was rounded to .01.
Assessment of risk of bias
Studies were assessed individually at a study level for bias and quality based on the Quality Criteria Checklist obtained from the Evidence Analysis Manual: steps in the academy evidence analysis process. Only the studies that obtained a positive or neutral rating were included. Publication bias was assessed by developing a funnel plot using standard difference in means and standard error as x and y-intercepts, respectively.
Results
The electronic search revealed 2440 articles with a further 47 identified by hand-searching. Of these, 2301 did not meet inclusion criteria because they did not report energy intake at 2 time points at least 12 weeks apart. Three potential studies were excluded because of missing data. The screening and selection process resulted in 18 studies of 2644 women published between 1992 and 2013 ( Figure 1 ). Fourteen studies were observational studies, and most conducted in a representative population with a mean BMI in the normal range. One study was in an overweight population and 1 study reported data by BMI category. Of the 4 RCTs, 1 was an intervention in an overweight population and 2 in an obese population, all aimed at limiting GWG. One study had a population group with a mean BMI in the normal range with interventions comparing pregnancy outcomes on a low glycemic index diet vs healthy eating advice.
Of the 18 studies that met inclusion criteria, 2 obtained a positive quality rating and 16 were neutral ( Table 1 ). Within studies, the number of study participants ranged from 10 to 620, with a mean (SD) age 29.6 (1.7) years, and BMI 25.3 (4.9) kg/m 2 . Seven studies had a retention rate of >82%. 6 ranged from 63 to 78%, 1 had 55%, and 3 did not report. Sixteen subgroups reported total weeks’ gestation with an average of 39.6 (0.43) weeks. Only 6 studies reported parity mean = 1.8 (2.32). Dietary data were collected on average at 12 (2.6) weeks’ and 35 (2.1) weeks’ gestation. The most frequent dietary collection method was a weighed or estimated food record (n = 18), but 3 studies used a food diary, 2 used repeat 24-hour recall, and 2 used a diet history. Characteristics of the included studies are summarized in Table 2 .
| Was the research question clearly stated? | Was the selection of study subjects free from bias? | Were study groups comparable? | Was method of handling withdrawals described? | Was blinding used to prevent introduction of bias? | Were intervention procedure and comparison(s) described in detail? Where intervening factors described? | Were outcomes clearly defined and the measurements valid and reliable? | Was the statistical analysis appropriate for the study design and type of outcome indicators? | Are conclusions supported by results with biases and limitations taken into consideration? | Is bias due to study’s funding or sponsorship unlikely? | Overall Quality Rating | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Alberti-Fidanza | √ | (–) | (–) | √ | (x) | √ | √ | √ | √ | √ | Neutral |
| Carbone et al. | √ | √ | √ | √ | (–) | (–) | √ | (–) | (–) | √ | Neutral |
| Conway et al. | √ | √ | √ | √ | (–) | (–) | √ | √ | √ | (–) | Neutral |
| De Vriese | √ | √ | √ | (–) | (x) | (–) | √ | √ | √ | √ | Neutral |
| Fung et al. | √ | √ | √ | √ | (–) | √ | √ | √ | √ | √ | Positive |
| Goodarzi Khoigani et al. | √ | √ | √ | (–) | (–) | (–) | √ | (–) | √ | √ | Neutral |
| Guelinckx et al. | √ | √ | √ | √ | (–) | (–) | √ | √ | √ | √ | Neutral |
| Hronek et al. | √ | √ | (–) | (x) | (x) | (–) | √ | √ | √ | √ | Neutral |
| Jansson et al. | √ | √ | √ | √ | (x) | (–) | √ | √ | √ | √ | Neutral |
| Kopp-Hoolihan et al. | √ | (–) | √ | √ | (x) | (–) | √ | √ | (–) | √ | Neutral |
| Korpi- Hyövälti et al. | √ | √ | √ | √ | (x) | (–) | √ | √ | √ | √ | Neutral |
| Kubota et al. | √ | √ | √ | (x) | (–) | (–) | (–) | (–) | √ | √ | Neutral |
| Lof et al. | √ | (–) | √ | √ | (x) | (–) | √ | (–) | √ | √ | Neutral |
| Martinez et al. | √ | (–) | (x) | (–) | (x) | (–) | (–) | (x) | (–) | √ | Neutral |
| Moses et al. | √ | √ | (–) | √ | (x) | (–) | √ | √ | √ | √ | Neutral |
| Piers et al. | √ | (–) | √ | (–) | (x) | (–) | √ | √ | (–) | √ | Neutral |
| Tabrizi and Saraswathi | √ | √ | √ | (–) | (x) | (–) | (–) | (–) | √ | √ | Neutral |
| Wolff et al. | √ | √ | √ | √ | √ | √ | √ | √ | √ | √ | Positive |
| Key: | √ = yes, (x) = no, (–) = unclear | ||||||||||
| Observational studies | |||||||
|---|---|---|---|---|---|---|---|
| Source | Aim | Study population | Dietary assessment method | Inclusion/exclusion criteria | Conclusion | Limitations | Quality rating & retention |
| Alberti-Fidanza et al. | To examine longitudinally the ToAC of women from early pregnancy to delivery and of their newborns and relate the results to the dietary intake of the same women during pregnancy. | Italy, developed n=12 BMI 22.6 (5.5) kg/m 2 Age 31.1 (4.2) y Parity 1.8 (0.9) nil smokers | Collected by qualified and experienced dietitians using the diet history method during the First, second, and third trimesters. | Not stated | It is important to monitor ToAC values during the entire period of pregnancy, and we suggest large intakes of fruit and vegetables and, if necessary, antioxidant vitamin and pro-vitamin supplements. Data regarding the antioxidant status of mothers and newborns, particularly if preterm, may offer valuable information for increasing the chance that pregnancies proceed successfully to term and achieve physiological deliveries. | Small sample size; possible selection bias because inclusion criteria not reported; factors such as pollution, ionizing radiation, and smoking habits may interfere with antioxidant status. | Neutral Not reported |
| Carbone et al. | Present patterns of leukocyte energy metabolism during normal pregnancy and puerperium and its possible relationship to fetal growth and maternal nutrient intake. | Spain, developed n=33 BMI 22.7 (2.2) kg/m 2 Age 28 (3.6) y Parity of 1, n=18; 2,3,4, n=15 | 3-day weight food record including 1 holiday completed at 11, 19, and 35 weeks’ gestation. | Gestational age less than 12 weeks, normal menstrual cycles before conception, no personal or family history of metabolic, vascular and/or genetic disorders, no apparent disease present at the time of visit and pregnancy classification as low risk. | There was a correlation between protein/DNA ratio and head circumference at 36 weeks’ gestation. Findings may suggest a relationship between the metabolism of maternal leukocytes and fetal development in utero. | Small sample size; selection bias attributable to 66% retention rate. | Neutral 66% |
| Conway et al. | To explore the relationship between dietary restraint and appropriateness of weight gain during pregnancy using AIM criteria and to assess dietary intake during pregnancy in relation to dietary restraint. | England, developed Restrained: n=32 BMI 22 (2.4) kg/m 2 Age 31.2 (4.6) y Parity of 1, n=28; 2, n=4 Unrestrained: n=30 BMI 20.7 (1.96) kg/m 2 Age 30.6 (3.6) y Parity of 1, n=26; 2, n=4 | 7-day weighed food record at 12 and 30 weeks’ gestation | White women, expecting their first or second singleton baby, >18 y of age, and free from any medical condition that might affect nutrition or fetal outcomes. | Providing pregnant women with more guidance about appropriate weight gains may be beneficial. Because cessation of smoking during pregnancy was associated with large weight gains, it would be prudent to accompany any advice about stopping smoking during pregnancy with advice about eating and weight gain. | Small sample size; not a representative sample, nonresponders had a greater BMI than responders. | Neutral 81% |
| De Vriese et al. | To determine the relative validity and usefulness of a Dutch FFQ adapted to the Belgian diet by comparing dietary fat intake data collected by this FFQ with the 7-d food record in pregnant Belgian women during the First and third trimesters. | Belgium, developed n=26 BMI 22 (17.6–29.3) kg/m 2 Age 30 (25–37) y All pregnant women were nulliparous. | 7-d EFR during the first (median 15 wk) and third (median 35 wk) trimesters. A FFQ containing 180 of the most common fat-containing foods was conducted at the same time points. Data recorded was based on the EFR. | First pregnancy, diastolic blood pressure <90 mm Hg, not diabetic, no proteinuria, and not suffering from renal or cardiovascular disease. | The FFQ in conjunction with the individual fatty acid composition database of Belgian foods is an adequate method to reasonably rank subjects according to their dietary fat intake. | Small sample size; food composition database used to analyze the 7-d EFR lacks a lot of data concerning the linoleic acid content of different foods. | Neutral 78% |
| Fung et al. | To determine whether fractional absorption of a stable isotope of zinc from a standardized meal is altered in well-nourished women followed from before conception through lactation and, determine whether the change in FZA is related to indicators of maternal zinc status. | USA, developed n=13 BMI 22.3 (2.9) kg/m 2 Age 30 (2.9) y Parity of 1, n=10; 2, n=3 | 3-day weighed food record, nonconsecutive days, 2 weekdays, and 1 weekend at 8−10 weeks, 24−26 weeks and 34−36 weeks’ gestation. | Aged 22−40 y, BMI of 19–26 kg/m 2 , nonsmoking, nondiabetic, nonvegetarian, no drug and alcohol use, and no previous obstetric or gynecologic complications. | Well-nourished women met the additional need for zinc during pregnancy by increasing zinc intake and by a 30% increase in zinc absorption that was not significant. The increase in dietary zinc was due largely to an increase in intake of dairy foods. FZA increased 75% early in the lactation period, presumable as an adaptation to the lactation process. These data indicate that mechanisms regulating zinc homeostasis differ between pregnancy and lactation. | Small sample size | Positive 87% |
| Hronek et al. | To evaluate the DIEN of Czech pregnant women and compare it with RDA. | Czech Republic, developed Participants were randomly recruited from both rural and city regions. n=152 BMI 21.1 (3.6) kg/m 2 Age 28.9 (3.6) y | 7-d food record on consecutive days using scales and household measures at 0−20 weeks, 21−29 weeks, 30−36 weeks, and 37−39 weeks’ gestation. | Nonusers of chronic medication, nonsmokers and nonabusers of alcohol or drugs and had parity ≤2. Subjects were euthyroid, normoglycemic, and not anemic. | Lower intake of energy and intakes of some nutrients relative to the corresponding RDA during pregnancy. Evaluated DIEN corresponded with body size variables. Modification of food intake or alternatively supplementation is recommended, for folic acid, iron, vitamin D, zinc, iodine, and fiber. | Possible selection bias as retention rate not reported; to prevent distortion of the data, intake of supplements was not included in the evaluation, because of irregular intake; deficiencies were not evaluated; only health women with prepregnancy BMI in the range of 17.5−24.7 were recruited. | Neutral Retention not reported |
| Jansson et al. | To identify hormonal factors that can explain the link between early pregnancy BMI, maternal dietary intake, and birth weight. | Sweden, developing n=49 BMI 25.5 (6.9) kg/m 2 Age 30 (4.5) y | Diet history covering dietary intake over a 24-h period, collected by a registered dietitian at 8−12 weeks and 32−35 weeks’ gestation. | Inclusion criteria: Scandinavian heritage, healthy and ≥20 y. Exclusion criteria: smoking, vegetarianism, assisted reproduction, concurrent disease such as eating disorder or diabetes, development of pregnancy complications such as gestational diabetes, preeclampsia, or intrauterine growth restriction. | High first trimester maternal serum resistin and low third trimester IGFBP-1 were correlated with increased birth weight. We propose that low serum concentrations of IGFBP-1 represent a link between high BMI and increased fetal growth by increasing bioavailability of IGF-1, regulating placental nutrient transport. | Small sample size decreased the ability to detect biologically relevant differences. | Neutral 88% |
| Khoigani et al. | To assess the association between preeclampsia risk and the intake of 40 macro- and micronutrients during the first, second, and third trimesters based on demographic and reproductive characteristics and PA of pregnant women. | Iran, developing n=620 (t 1, n=584; t 2, n=510) BMI 23.5 (3.9) kg/m 2 Age 25.6 (4.4) y Parity 1.58 (0.76) | 48-h dietary recall completed at 11−15 weeks, 26 weeks, and 34−37 weeks’ gestation. Interviewers were trained. | Pregnant women who did not have conditions such as factors causing preeclampsia, preterm delivery, low birth weight, and factors that may affect pregnancy outcomes such as smoking, drug addiction, digestive and metabolic disease, hemoglobinopathies, eating disorders, allergies, mental diseases, and malignancy. | Mean value of saturated fat in the first trimester in subjects who experienced preeclampsia later in pregnancy was greater than in other pregnant women. Intakes of manganese, vitamin C, vitamin E, fiber, and carbohydrate during the third trimester were significantly less among pregnant women who developed preeclampsia. No significant associations between other micro and macro nutrients and preeclampsia risk. | Reasons for withdrawal not reported; data were not collected for all subjects at each time point, and reasons for this not reported. | Neutral 89% |
| Kopp-Hoolihan et al. | To assess how well-nourished women meet the energy demands of pregnancy and to identify factors that predict an individual’s metabolic response. | USA, developed n=10 BMI 23.1 (2.1) kg/m 2 Age 29.1 (5) y | 3-d weighed food record completed before pregnancy and at 8−10, 24−26, 34−36 weeks’ gestation and 4−6 weeks’ postpartum. | Healthy, nonsmoking women. | Well-nourished women use different ways to meet the energy demands of pregnancy, including reduction in diet induced thermogenesis, increased energy intake, deposition of less fat mass than predicted. | Small sample size; subjects were new to WFR food collection method at t 1 , it would have been prudent to ask the subjects to repeat measurements to verify intake. | Neutral 100% |
| Kubota et al. | To investigate the associations among changes in dietary intake, maternal bodyweight, and fetal growth during pregnancy. | Japan, developed Underweight n=32 BMI 17.5 (0.1) kg/m 2 Age 29.7 (5.2) y Normal n=94 BMI 21 (1.8) kg/m 2 Age 30.8 (5.1) y Overweight n=9 Age 28.6 (3.6) y BMI 33.4 (6.5) kg/m 2 | Digital images taken before and after meals on 3 consecutive days at 14−16, 25−27, and 32−34 weeks’ gestation; a dietitian reviewed photos and recorded intake; validation of method not reported. | Singleton pregnant Japanese women were included; those which had suffered from obstetrical complications such as premature delivery, gestational diabetes, and preeclampsia or did not submit digital images were excluded. | Dietary intake was similar throughout pregnancy and did not correlate with fetal growth, despite Japanese national recommendations advising extra energy intake. | Possible selection bias because of low retention rate; no differentiation is made between participants excluded because of complications and those who withdrew from the study; characteristics of withdrawals not described; dietary collection method not validated. | Neutral 55% |
| Lof et al. | Investigate whether intakes of total dietary fat, types of fat, and weight gain are associated with estradiol and progesterone levels in plasma during pregnancy. | Sweden, developed n=226 BMI 22.9 (3) kg/m 2 Age 32 (4) y | 3-d weight food record on consecutive days including 1 weekend at 12, 25, and 33 weeks’ gestation. Calculations conducted by a nutritionist. | Women with multifetal pregnancies, missing data at baseline questionnaire or for measurements of either body weight or dietary intake were excluded. | No association found between gestational weight gain, maternal dietary fat intake (total or subtypes), and plasma estradiol levels. Progesterone levels correlated with weight gain in pregnancy. | Findings are only relevant to the white population; dietary intake was self-reported thus misclassifications of dietary intake cannot be excluded. | Neutral 78% |
| Martinez et al. | To determine how gestational weight gain varies according to BMI in a developing nation, Mexico. | Mexico, developing n=36 BMI 23.5 (2.6) kg/m 2 Age 31 (5.4) y | Diet history collected by trained interviewers twice per month and reported as mean first and third trimester values. | Women with an 18 month old or schooler (a person/child that attends school) 7−8 years in the Solis Valley, who became pregnant at the time of initial recruitment provided that they were no more than 5 mo pregnant. | In this sample of women, the relationship between BMI at conception, weight gain, and pregnancy outcome was similar to that of women in United States. Lower maternal BMI predicted smaller birth weight and size through at least 6 mo of life. | Small sample size; large standard deviation in data may reflect the small sample size; withdrawals not reported. | Neutral 86% |
| Piers et al. | To determine changes in maternal energy metabolism during pregnancy and lactation by comparing a group of well-nourished pregnant and lactating Indian women at 12, 24, and 34 weeks’ gestation and at 12 and 24 weeks’ postpartum with a nonpregnant, nonlactating control group. | India, developing n=18 BMI 21.7 (2.4) kg/m 2 Age 29.6 (5.2) y | 5-d EFR using household measures of known volume at 12, 24, and 34 weeks’ gestation and 12 and 24 weeks’ postpartum; participants were trained to use the measure by a dietitian who also checked all records. | Pregnant subjects of good health, were nonsmokers, had no appetite affected by morning sickness before the initial metabolic measurement at 12 weeks’ gestation. | BMR is significantly greater during pregnancy compared with nonpregnant, nonlactating data, and remains high even when differences in body weight are accounted for. Well-nourished Indian women have weight and fat gains similar to those of well-nourished Western women. Birth weights of infants born appeared to be lower. Energy cost of pregnancy estimated to be 303 MJ, close to the 335 MJ estimated by FAOIWHO/UNU. | Small sample size; withdrawals not discussed; methods used for the estimation of energy intake were not as precise as the weighed-intake, but it was less cumbersome; the possibility of a systematic underestimation of energy intake in studies reporting low increments in energy intake during pregnancy cannot be ruled out. | Neutral 82% |
| Tabrizi et al. | To assess the relationship between energy, protein, and mineral intake of pregnant women and birth weight of their neonates. | Iran, developing n=450 BMI 23.9 (3.8) kg/m 2 Age 26.1 (5.8) y | 24-h recall collected at the end of the first, second, and third trimesters. Qualification/ training of interviewer not reported | Women 16−40 y who continuously visited health care centers during the 3 trimesters in Khoy city. | Maternal energy, protein, calcium, iron, and zinc intake along with greater maternal serum calcium iron and zinc influenced birth weight of neonates. | Withdrawals not reported; actual consumption of supplements was not monitored. | Neutral Not reported |
| Randomized controlled trials | |||||||
|---|---|---|---|---|---|---|---|
| Source | Aim | Study population and intervention | Dietary assessment method | Inclusion/exclusion criteria | Conclusion | Limitations | Quality rating & retention |
| Guelinckx et al. | Study of which degree of intervention can improve dietary habits according to the National Diet Recommendations, increase PA level in obese pregnant women, and control gestational weight gain. | Belgium, developed; randomly assigned using block randomization. Control (routine): n=43 BMI 33.5 (3.9) kg/m 2 Age 29.4 (4.4) y Parity of 1 n=17 Passive (nutrition and PA brochure): n=37 BMI 33.4 (3.1) kg/m 2 Age 28.7 (4) y. Parity of 1, n =15 Active (brochure + group dietary counselling): n = 42 BMI 34.1 (4.5) kg/m 2 Age 28 (3.6) y. Parity of 1, n =20 | 7-d food records including both weighed and household measures during each trimester of pregnancy. Records were checked by a nutritionist. | White women attending the prenatal clinic before 15 weeks’ gestation. Exclusion criteria: preexisting diabetes or developing gestational diabetes, multiple pregnancy, premature labor (<37 weeks’ gestation), primary needs for nutritional advice in case of metabolic disorder, kidney problems, Crohn disease, allergic conditions, and inadequate knowledge of the Dutch language, because this language was used for both the brochure and group discussions. | A lifestyle intervention based on a brochure alone or group sessions combined with individual advice can improve dietary habits throughout pregnancy in obese women. Even in the absence of medical or obstetric complications, maintaining or increasing PA during pregnancy is difficult. To obtain a significant decrease in GWG, an individually designed caloric intake restriction based on energy expenditure data should be included. | Patients in the control group were aware of being included in a study aiming at promoting a healthy lifestyle and decreasing GWG may have influenced dietary records resulting in underestimation of the intervention effect. Nutritional data was incomplete for 27 participants (14%) who were excluded from analysis. | Neutral Control: 66% Passive: 57% Active: 65% |
| Korpi-Hyovalti et al. | To evaluate the effect of intensive dietary therapy on quality of diet, weight gain, and birth weight in women at high risk of GDM. | Finland, developed; randomized into group 1 or 2 using computed randomization Close follow-up (general information on diet and PA) n=27 BMI 25.5 (3.4) kg/m 2 Lifestyle intervention (individualized nutrition advice) n = 27 BMI 27.3 (6) kg/m 2 | 4-dweighed food record on consecutive days including 1 weekend completed at 8−12, 26−28, and 36−40 weeks’ gestation. Records checked by a nutritionist A 3-factor eating questionnaire was also used at the Fir st and thi rd trimester. | Women with one or more risk factors for gestational diabetes, venous plasma glucose concentration after 12-h overnight fasting was 4.8–5.5 mmol/L and the 2-h oral glucose tolerance test plasma glucose <7.8 mmol/L were recruited. Women diagnosed with GDM at 8–12 weeks’ gestation were excluded. | There were no clear differences in saturated fat and fiber intake; however, polyunsaturated fat did increase in the lifestyle intervention group. Intensive weight gain education led to a somewhat lower weight gain during pregnancy, and higher birth weights of the infants in lifestyle intervention but no differences in macrosomia compared with the close follow-up group. | Small sample size; the Three Factor Eating Questionnaire, which was used to measure 3 dimensions of eating behavior (cognitive restraint of eating, disinhibition, and hunger); however, its benefit was limited in the present study because all women received informative education regardless of the results of the questionnaire. | Neutral Lifestyle intervention: 70% Close follow-up: 67% |
| Moses et al. | To determine whether offering low glycemic index (LGI) dietary advice at the first antenatal visit would result in a lower fetal birth weight, birth percentile, and PI than providing HE advice. | Australia, developed randomly assigned using computer-generated random numbers 1 or 2 Healthy Eating (counseled to follow a healthy diet based on the AGHE ) n = 280 BMI 24.7 (5) kg/m 2 Age 29.9 (5) y. LGI (individualized nutrition advice on a LGI diet) n = 296 BMI 24.3 (5.2) kg/m 2 Age 29.9 (5.2) y. | 3-d EFR using household measures at 16 and 36 weeks’ gestation; records reviewed by a dietitian. | Enrolment at <20 weeks’ gestation with a singleton pregnancy, at least 18 y, ability to read and understand English language, and ability to comply with visit schedules. Subjects excluded if they had: diabetes or previous gestational diabetes, special dietary needs, the presence of medical conditions that could compromise their metabolic status, or the use of medications that were likely to influence body weight. | A low-intensity dietary intervention with a LGI compared with HE diet in pregnancy did not result in significant differences in birth weight, fetal percentile of PI. | Study personnel not blinded to the dietary assignment; PA was not controlled or measured; the study lacked a control group that followed standard prenatal protocol. | Neutral Healthy eating: 83% LGI: 84% |
| Wolff et al. | To determine whether a 10-h dietary consultations restricts weight gain in obese women and whether this restriction impacts on pregnancy-induced changes in glucose metabolism. | Denmark, developed; Randomized into group 1 or 2 using computed randomization. Control (general information on diet and PA) n=27 BMI 25.5 (3.4) kg/m 2 Age 30 (5) years Intervention (10 × 1-h consultations) n =27 BMI 27.3 (6) kg/m 2 Age 28 (4) y | 7 day weighed food record at inclusion (15 ± 3 wks.), 27 and 36 weeks gestation. | Pregnant obese women (BMI >30) in their early pregnancy were recruited. Exclusion criteria included: smoking, <18 y or >45 y, multiple pregnancy, or medical conditions that impact fetal growth. | Restriction of gestational weight gain in obese women is achievable and reduces the deterioration in glucose metabolism. | Small sample size; Limited generalizability of results because of the scientific settings of the trial with time-consuming extra ultrasound scans and blood samples that may have increased the number of dropouts; unrestricted control group knew they were participating in a maternal weight restriction study, which could have influenced gestational weight gain. | Positive 76% |
Mean reported energy intake in the 18 studies was 8130 (1100) kJ/day and 8600 (1230) kJ/day in early and late pregnancy, respectively. The SMD between the 2 was 0.266 ( P = .016), a difference equivalent to ∼475 kJ/day, which did not reach the a priori level of significance ( P = .01) ( Figure 2 ). The mean GWG was 12.0 (2.8) kg, representing a large statistically significant increase (SMD = 1.306, P < .0005) ( Figure 3 ). There was no correlation, however, between mean incremental energy intake and GWG ( Figure 4 , A; r = 0.321, P = .11). Only 1 study (18 women, BMI 21.7 [3] kg/m 2 ) reported a mean increase in energy intake in line with pregnancy guidelines, that is, 1700 kJ/day (GWG 11.4 [3.7] kg).
