The association between dietary patterns before and in early pregnancy and the risk of gestational diabetes mellitus (GDM): Data from the Malaysian SECOST cohort.

Generally, dietary patterns (DP)s have been linked to the risk of diabetes mellitus, however, only few studies examined the associations between DPs in early pregnancy and the risk of gestational diabetes mellitus (GDM). This study aims to determine the association between DPs before and during pregnancy and risk of GDM in Malaysian pregnant women. DPs were derived using principal component analysis of consumed 126 food and beverage items assessed using a validated semi-quantitative food frequency questionnaire collecting data retrospectively for pre-pregnancy, but prospectively for the first and second trimester. Three different DPs were identified at each time point and labelled as DP 1-3 (pre-pregnancy), DP 4-6 (first trimester), and DP 7-9 (second trimester). About 10.6% (n = 48) of pregnant women were diagnosed with GDM in our cohort. Women with high adherence (HA) to DP 2 (adjusted OR: 0.45, 95% CI: 0.20-0.91) and DP 5 (adjusted OR: 0.28, 95% CI: 0.11-0.68) showed a significantly reduced risk of GDM compared to women with low adherence (LA). Other DPs were not significantly associated with GDM risk. Compared to women with GDM, non-GDM women showed HA scores for all DPs throughout pregnancy. Overall, a relative low percentage of women with GDM was found in this cohort. The risk was lower in women with HA to a relatively unhealthy dietary pattern, i.e. DP 2 and DP 5. The lower body mass index (BMI) status and energy intake of women showing a HA to DP 2 in the first trimester may underlie the observed association with a lower GDM risk. Additionally, genetic variance might explain the less susceptibility to GDM despite HA to unhealthy DPs among non-GDM women.

Introduction

Gestational Diabetes Mellitus (GDM), an increasingly common type of hyperglycemia during pregnancy, follows the increasing trends of obesity and Type 2 Diabetes Mellitus (T2DM). In 2017, the International Diabetes Federation (IDF) estimated that 20.9 million (16.2%) of live births were affected by hyperglycemia in pregnancy, and about 85.1% of these live births were due to GDM [1]. The National Health and Morbidity Survey (2016) reported that the prevalence of GDM among Malaysian mothers aged 15–49 years old was 13.5% [2]. The GDM rate in Malaysian population (8.7–29.7%) [3-5] was significantly higher than the reported rates in many Western (2.0–9.2%) [6,7] and Asian countries (2.8–25.0%) [8-11].

Dietary intake during pregnancy is commonly assessed through the intakes of energy, macronutrients, micronutrients or food groups [12,13], which may then be examined in relation to an imbalanced maternal diet and poor maternal nutritional status. Dietary pattern is a relatively new approach that describes a combination of commonly consumed foods [14] that allows for the diet to be described as a whole [15]. Nutritional health outcomes are often the result of multiple synergies among nutrients and foods rather than just the sum of the individual food [16]. Although there are several approaches to identifying dietary patterns of pregnant women, the posterior-approach derived from principal component analysis (PCA) is the most commonly used for deriving dietary patterns during pregnancy [17-20]. PCA is a technique to reduce a large of correlated variables into a smaller number of components [21,22], revealing the underlying structure within diets of the population. Numerous studies examining the role of PCA derived DPs in adverse pregnancy outcomes, such as GDM, pre-eclampsia, and pre-term birth [23-25] have been published.

Associations between DP before and during pregnancy with risk of GDM has been examined [26-30]. Most studies showed that an adoption of the prudent pattern, diet that was rich in vegetables, fruits, whole grains and legumes showed significant lower risk of GDM [26-28], whereas adhering to the Western pattern, which is characterized by high intakes of red meat, processed meats and refined foods, was associated with increased risk of GDM [23,26,31]. There are several suggested mechanisms by which Western pattern may increase the risk of GDM through inflammation, and placental dysfunction [26,29,30].

Despite the numerous studies on the association between DPs and risk of GDM, findings are mostly inconsistent. These inconsistencies could be due to different study designs and diagnostic criteria of GDM, methods to measure dietary intake (e.g. diet recall, diet record, food frequency questionnaire), and socio-demographic factors (e.g. age, ethnicity, and BMI). It could also be attributed to the different food items reported for the respective DPs. For example, the “prudent” DP, which is a diet delivers health benefits, observed in a prospective cohort study of women in Northern China was characterized by higher intakes of dark-colored vegetables and deep-sea fish [26], while the “prudent” DP of women in the Born in Guangzhou Cohort Study (BIGCS) was predominantly dairy products, nuts, eggs, fish, soups, and fruits [27].

Previous studies on DPs and GDM have focused on assessment of such relationship at a single time point, either pre-pregnancy [29,30,32-34] or during pregnancy, before diagnosis of GDM [28,35-37]. Most studies on DPs were limited to Western countries [28,35-37], with only two studies conducted in Asian populations [27,38]. As there are considerable differences in diet and lifestyle behaviors, the DPs of Malaysian pregnant women could be different from those reported in Western as well as other Asian populations. Most people eat a combination of healthy and less healthy foods, but the impact of mixed dietary patterns on the risk of GDM remains unknown. Thus, the objectives of this prospective study are to describe the DPs before and during pregnancy in a sample of Malaysian pregnant women and to determine the association between DPs and the risk of GDM.

Materials and methods

Study design and location

The SECOST (Seremban Cohort Study) project is an on-going prospective study in which pregnant women are followed-up up to 1 year postpartum, and their infants are followed-up every 6 months until 2 years old. Women in the first trimester (8 – 10th weeks of gestation) of pregnancy were recruited from three maternal and child health (MCH) clinics in Seremban District, Negeri Sembilan, Malaysia. Detailed descriptions of the study methodology have been previously published [39], and only a brief overview is provided here. All pregnant women were eligible to participate unless they had one or more exclusion criteria as published previously.

A total of 737 pregnant women were enrolled in the SECOST study. About 22.7% (n = 167) and 16.0% (n = 118) women were excluded at first and second visit, respectively (Fig 1). The present study reported on data of 452 pregnant women only. The adequacy of the sample size of pregnant women was checked to ensure a 5% statistical significance level and 80% of power.

Fig 1

Flowchart of study respondents.

The study protocol was approved by the Medical Research Ethics Committee (MREC), Universiti Putra Malaysia (UPM/FPSK/100-9/2-MJKEtika), and the Medical Research Ethics Committee (MREC), Ministry of Health Malaysia (KKM/NIHSEC/08/0804/P12-613). Permission to conduct this study was also obtained from the Seremban District Health Office. All women provided written informed consent and a study manual.

Dietary patterns (DPs)

A validated 126-food item semi-quantitative Food Frequency Questionnaire (SFFQ) was utilized to assess food consumption patterns [40]. The dietary assessments were conducted at three time points: at the first prenatal visit (9.82 ± 2.51 weeks of gestation to collect data on DPs before pregnancy), the first trimester (12.26 ± 1.58 weeks of gestation) and the second trimester (26.73 ± 1.64 weeks of gestation), respectively. The dietary questionnaire used for this study has shown a good relative reliability and validity of food intakes of Malaysian pregnant women [41]. The food items were categorized into 17 food groups based on similar nutrient characteristics and considering similar in culinary usage [42]. They include rice, noodles & pasta; bread, cereal & cereal products; poultry & meat; processed meat; fish & seafood; eggs; nuts, seeds & legumes; milk & dairy products; green leafy vegetables; other vegetables; fruits; tea and coffee; high energy beverages; sweet foods, sugar, spread & creamer, condiments & spices, and oils & fats.

Principal component analysis (PCA) was performed based on 17 food and beverage items to derive dietary patterns. PCA was performed separately for each time point and the factors were rotated by an orthogonal transformation (varimax rotation) to maintain uncorrelated factors and greater interpretability. The number of factors was determined by eigen values greater than 1.5 [43]. The data suitability was assessed before performing PCA. Inspection of the correlation matrix on the food groups revealed the presence of several correlation coefficients of 0.2 and above. Food groups with absolute factor loading with values < 0.30 were not listed in the final model. The Kaiser-Meyer-Olkin measure of sampling adequacy was used to ensure that the sample size was adequate for these analyses. The Kaiser-Meyer-Oklin values were 0.68 (pre-pregnancy), 0.71 (first trimester), and 0.65 (second trimester). Three characteristic DPs were identified at each time point and labelled as DP 1–3 (pre-pregnancy), DP 4–6 (first trimester), and DP 7–9 (second trimester). The factor scores for each DP and for each individual were calculated by summing the food item intakes weighted by their factor loadings [44]. The factor score was then categorized into tertiles (1st tertile = low adherence (LA), 2nd tertile = moderate adherence (MA), and 3rd tertile = high adherence (HA)) for ease of interpretation in the subsequent analysis. More detailed description of DPs before and during pregnancy was published elsewhere (DOI: 10.4162/nrp.2019.13.3.230) [42].

Gestational diabetes mellitus (GDM)

Women with pre-existing diabetes mellitus (DM) (fasting plasma glucose (FPG) >7.0 mmol/l), or abnormal glycaemia (FPG <3.0 mmol/l or FPG >6.0 mmol/l) at study enrolment were excluded. A standard two-point diagnostic 75g oral glucose tolerance test (OGTT) was performed at 28 – 32nd weeks of gestation. A 2-ml fasting venous blood was drawn by a clinic staff nurse before ingestion of a standard glucose solution to obtain fasting plasma glucose (FPG). Another 2 ml of venous blood was drawn at 2-hours after ingestion of standard glucose solution. All blood samples were sent for analysis on the same day to determine FPG and 2-hr plasma glucose (2hPG) concentration. GDM was diagnosed if either or both FPG was ≥ 5.6 mmol/l or 2hPG is ≥ 7.8mmol/l according to Ministry of Health Malaysia guideline [45].

Other variables

Socio-demographic information included age, education level, ethnicity, occupation status, and monthly household income. Obstetrical information (e.g. gravidity, parity, medical history GDM, and family history of diabetes mellitus) was obtained from medical records. Height was measured at study enrolment and was further categorized into tertiles based on the overall distribution, while weight was measured at each study visit using a standard instrument (SECA digital weighing scale and SECA body meter) and standard procedures. Women were requested to recall pre-pregnancy body weight. Pre-pregnancy body mass index (BMI) (kg/m2) was calculated as pre-pregnancy weight divided by the square of height and classified according to the recommendation of World Health Organization [46].

Statistical analysis

All analyses were performed using SPPS version 24. Exploratory Data Analysis (EDA) was carried out to determine the normality and homogeneity of the data. All continuous variables were normally distributed. Therefore, no transformation was performed. Basic descriptive statistics were generated such as means and standard deviations for the continuous variables, while for categorical variables, frequency, and percentage distribution. Chi-square test of independence or Fisher’s exact test and Independent t-test were used to assess the association between women characteristics with risk of GDM, respectively for continuous and categorical variables.

Generalized Linear Mixed Model (GLMM) was performed to assess the association between tertile of adherence to DPs and the risk of GDM, with MCH clinics and gestational week at OGTT as random effects. Potential covariates included in the multivariable model were age (continuous), ethnicity (Malays vs Non-Malays), medical history of GDM (yes vs no) and family history of DM (yes vs no). The lowest tertile of adherence was set as the reference group. Adjusted odds ratio (OR) with 95% confidence interval (CI) of the association between DPs and GDM were presented. A further stratified analysis was performed to determine the associations between DP 2 and DP 5 with GDM by pre-pregnancy BMI. Repeated measures ANOVA was used to plot the DP trajectory for GDM and non-GDM groups. Three major patterns were identified at each time point: pattern 1 (DP 1, DP 4, and DP 7), pattern 2 (DP 2, DP 5, and DP 8), and pattern 3 (DP 3, DP 6 and DP 9). Sensitivity analyses were run between acceptable reporters and under-reporters of energy intake. The findings did not change the main findings (S1 Table). The cut off point for statistical significance was set at p<0.05.

Results

Table 1 compares the characteristics of 404 non-GDM and 48 GDM women. Age was similar in both groups, with the mean of 30.04 ± 4.50 years in non-GDM and 29.80 ± 4.39 years in GDM women. Majority of the women were Malay (88.9–89.6%), had secondary and lower education level (41.7–46.6%), currently employed (68.6–68.7%), and had low monthly household income (58.3–63.4%). Women with GDM had a significantly higher percentage of history of GDM (16.7%) and family history of diabetes mellitus (37.5%) compared to non-GDM women (history of GDM = 5.7%; family history of diabetes mellitus = 23.0%). Most of the women had a height below 1.55 m (37.2–45.8%) and normal pre-pregnancy BMI (47.9–56.2%). We observed no significant differences in height and pre-pregnancy BMI between non-GDM (mean height = 1.56 ± 0.06 m; mean pre-pregnancy BMI = 23.64 ± 4.81 kg/m2) and GDM women (mean height = 1.56 ± 0.06 m; mean pre-pregnancy BMI = 24.45 ± 4.75kg/m2).

Table 1

Characteristics of women (N = 452).

	Maternal glycemiaa		p-valueb
	Non-GDM (n = 404)	GDM (n = 48)
Age (years)
• < 35	340 (84.2)	39 (81.3)	0.61
• ≥ 35	64 (15.8)	9 (18.7)
Mean ± SD	30.04 ± 4.50	29.80 ± 4.39	0.73
Ethnicity
• Malay	359 (88.9)	43 (89.6)	0.88
• Non-Malay	45 (11.1)	5 (10.4)
Education level (years)
• Secondary and lower	188 (46.6)	20 (41.7)	0.57
• STPM/ Matriculation/ Diploma/ Certificatec	133 (32.9)	15 (31.3)
• Tertiary and above	83 (20.5)	13 (27.0)
Mean ± SD	12.93 ± 3.53	13.21 ± 2.70	0.45
Occupation status
• Unemployed	127 (31.4)	15 (31.3)	0.98
• Employed	277 (68.6)	33 (68.7)
Monthly household income (RM)d
• Low (< 3860)	256 (63.4)	28 (58.3)	0.08
• Middle (3860–8319)	138 (34.1)	16 (33.3)
• High (≥ 8320)	10 (2.5)	4 (8.4)
Mean ± SD	3683.63 ± 2015.56	4089.58 ± 2319.50	0.20
Parity
• Nulliparous	145 (35.8)	15 (31.3)	0.73
• Primiparous	121 (30.0)	14 (29.2)
• Multiparous	138 (34.2)	19 (39.5)
Medical history
• GDM	23 (5.7)	8 (16.7)	0.01*
Family history
• Diabetes mellitus	93 (23.0)	18 (37.5)	0.03*
Height (m)
• < 1.55	150 (37.2)	22 (45.8)	0.33
• 1.55–1.58	123 (30.4)	10 (20.8)
• ≥ 1.59	131 (32.4)	16 (33.4)
Mean ± SD	1.56 ± 0.06	1.56 ± 0.06	0.34
Pre-pregnancy BMI (kg/m2)	23.64 ± 4.81	24.45 ± 4.75	0.27
• Underweight (< 18.5)	44 (10.9)	4 (8.3)	0.47
• Normal (18.5–24.9)	227 (56.2)	23 (47.9)
• Overweight (25.0–29.9)	88 (21.8)	15 (31.3)
• Obese (≥ 30.0)	45 (11.1)	6 (12.5)

Note.

aGDM was classified according to Ministry of Health Malaysia criteria as either or both FPG ≥ 5.6 mmol/l or 2-hour plasma glucose ≥ 7.8 mmol/l.

bp-value for differences between GDM and non-GDM groups were examined by independent t-test for continuous variables and chi-square for categorical variables.

cSTPM–Malaysian Higher School Certificate

dEconomic Planning Unit, Prime Minister’s Department, 2014.

*p<0.05

Table 2 presents the absolute factor loadings ≥ 0.30 for each DP. Three specific DPs were observed at each time period before and during pregnancy: pre-pregnancy (DP 1–3), the first trimester (DP 4–6), and the second trimester (DP 7–9). DP 1, DP 4 and DP 7 were considered as the ‘prudent, healthy pattern’, which was predominantly plant-based with high factor loadings for other vegetables, nuts, seeds & legumes, green leafy vegetables, fruits (DP 1 and DP 4), eggs (DP 1), and rice, noodles & pasta (DP 7). DP 2, DP 5 and DP 8 had high factor loadings for condiments & spices, and sugar, spread & creamer, with DP 5 having additional oils & fats. This pattern could be considered as a DP compliant to the local food culture. DP 3, DP 6 and DP 9 were characterized by high protein (poultry, meat, processed meat, dairy, egg and fish), sugars (mainly as high energy beverages, and sweet foods), and energy (bread, cereal & cereal products, rice, noodles & pasta), and only for DP 9 with additional fruits This pattern represented a combination of various food groups, but may be more equivalent to a Western unhealthy diet.

Table 2

Dietary pattern from pre-pregnancy to second trimester (n = 452).

Food Groups	Pre-pregnancy			First trimester			Second trimester
	DP 1	DP 2	DP 3	DP 4	DP 5	DP 6	DP 7	DP 8	DP 9
Other vegetables	0.76			0.81			0.75
Nuts, seeds & legumes	0.65			0.30			0.41
Green leafy vegetables	0.65			0.78			0.83
Fruits	0.48			0.64					0.56
Eggs	0.46					0.65			0.56
Milk & dairy products	0.35					0.48			0.52
Sugar, spread & creamer		0.98			0.97			0.97
Condiments & spices		0.98			0.97			0.97
Rice, noodles & pasta			0.74			0.42	0.40
Oils & fats			0.67		0.36
High energy beverages			0.58			0.45			0.54
Fish & seafood			0.46			0.44			0.61
Sweet foods			0.34			0.39			0.37
Poultry & meat			0.33			0.28			0.41
Bread, cereal & cereal products						0.66			0.64
Processed meat						0.41			0.42
Total variance	12.89%	12.58%	12.04%	11.93%	12.78%	13.86%	10.12%	12.67%	16.31%

Only food groups with absolute factor loadings > 0.30 were retained in each pattern for simplicity.

Table 3 shows the odds ratio (ORs) and 95% CI for the risk of GDM according to adherence tertiles to DPs before and during pregnancy. During pre-pregnancy, a similar trend was observed for all DPs, in which women with HA were at reduced risk of GDM. In the first trimester, women with HA to DP 4 and DP 5 were at reduced risk of GDM, and this trend remained similar in the second trimester (DP 7 and DP 8). Meanwhile, women with HA to DP 6 were at higher risk of GDM in the first trimester, but they were at lower risk for GDM in the second trimester (DP 9). The associations between adherence to a DP and the risk of GDM were generally not significant, except for DP 2 and DP 5, e.g. women with HA to DP 2 (adjusted OR: 0.45, 95% CI: 0.20–0.91) and DP 5 (adjusted OR: 0.28, 95% CI: 0.11–0.68) had significantly reduced risk of GDM compared to women with LA. The findings of the stratified analysis showed that the significant association between DP 5 and GDM was only observed among underweight or normal-weight women. Meanwhile, a significant association between DP 2 and GDM risk was found only among obese women (Table 4).

Table 3

Adjusted odd ratios and 95% confidence intervals for the association between dietary pattern and GDM (N = 452).

Dietary pattern b	Maternal glycemia a
	Adjusted OR	p-value
Pre-pregnancy
DP 1
• LA	1.00
• MA	0.83 [0.39–1.76]	0.62
• HA	0.82 [0.38–1.75]	0.60
DP 2
• LA	1.00
• MA	0.65 [0.31–1.37]	0.25
• HA	0.45 [0.20–0.91]	0.04*
DP 3
• LA	1.00
• MA	0.70 [0.33–1.49]	0.36
• HA	0.79 [0.38–1.64]	0.52
First trimester
DP 4
• LA	1.00
• MA	0.78 [0.37–1.63]	0.51
• HA	0.81 [0.38–1.71]	0.57
DP 5
• LA	1.00
• MA	0.70 [0.35–1.40]	0.31
• HA	0.28 [0.11–0.68]	0.01*
DP 6
• LA	1.00
• MA	1.27 [0.59–2.72]	0.54
• HA	1.15 [0.54–2.47]	0.72
Second trimester
DP 7
• LA	1.00
• MA	0.67 [0.33–1.38]	0.17
• HA	0.51 [0.24–1.11]	0.07
DP 8
• LA	1.00
• MA	0.69 [0.32–1.50]	0.35
• HA	0.73 [0.34–1.58]	0.42
DP 9
• LA	1.00
• MA	1.24 [0.60–2.55]	0.56
• HA	0.73 [0.33–1.63]	0.44

Note.

a The reference category is non GDM.

b Dietary patterns were classified in tertiles of adherence (1st tertile = low adherence (LA); 2nd tertile = moderate adherence (MA) & 3rd tertile = high adherence (HA)).

Adjusted for clinic, gestational week at OGTT performed, maternal age, ethnicity, medical history of GDM and family history of DM.

*p<0.05

Table 4

Adjusted odd ratios and 95% confidence intervals for the associations between DP 2 and DP 5 with GDM stratified by pre-pregnancy BMI (n = 452).

Pre-pregnancy BMI	Dietary pattern b	Maternal glycemia a
		Adjusted OR	p-value
Underweight and normal weight (n = 298)	DP 2
	• LA	1.00
	• MA	0.69 [0.25–1.90]	0.47
	• HA	0.56 [0.20–1.58]	0.27
	DP 5
	• LA	1.00
	• MA	0.71 [0.28–1.81]	0.47
	• HA	0.27 [0.08–0.87]	0.02*
Overweight (n = 103)	DP 2
	• LA	1.00
	• MA	0.61 [0.26–1.71]	0.34
	• HA	0.36 [0.06–1.38]	0.08
	DP 5
	• LA	1.00
	• MA	0.52 [0.20–1.36]	0.42
	• HA	0.41 [0.15–1.78]	0.24
Obese (n = 51)	DP 2
	• LA	1.00
	• MA	-
	• HA	0.42 [0.03–0.57]	0.04*
	DP 5
	• LA	1.00
	• MA	-
	• HA	0.24 [0.08–4.61]	0.34

Note.

a The reference category is non GDM.

b Dietary patterns were classified in tertiles of adherence (1st tertile = low adherence (LA); 2nd tertile = moderate adherence (MA) & 3rd tertile = high adherence (HA)).

Adjusted for clinic, gestational week at OGTT performed, maternal age, ethnicity, medical history of GDM and family history of DM.

*p<0.05

DP trajectory from pre-pregnancy to second trimester of pregnancy for non-GDM and GDM women are depicted in Fig 2. The number of women with GDM (n = 48) were relatively small. There were significant differences in pattern 1, 2 and 3 between non-GDM and GDM women. While non-GDM women maintained high adherence to all three DPs before and during pregnancy, GDM women changed their diets during pregnancy. For pattern 1 (DP 1, DP 4 and DP 7 –prudent diet, high in fruits, vegetables, nuts, seeds, legumes, eggs, milk and dairy), GDM women had low DP score before pregnancy, decreased score at the first trimester and further decreased score at second trimester. For pattern 2 (DP 2, DP 5 and DP 8 –high in condiments & spices, and sugar, spread & creamer), a V-shaped trend was observed in GDM women, where DP score was at the lowest point in the first trimester and higher at the second trimester. For pattern 3 (DP 3, DP 6 and DP 9 –high protein, sugar and energy), GDM women showed an inverse V-shaped trend in that they had low DP score before pregnancy, the highest DP score at the first trimester and the lowest DP score at the second trimester.

Fig 2

DP trajectory before and during pregnancy by maternal glycemia adjusted for pre-pregnancy BMI and rate of gestational weight gain.

Discussion

In this prospective cohort of 452 pregnant women, three major DPs were identified before and during pregnancy. Only DP 2 and DP 5 showed significantly a lower risk of GDM after the adjustment for covariates. The study findings should be interpreted with caution as this pattern was relatively higher in loadings for sugar, spread & creamer, sauces, condiments & spices, and oils & fats. The percentage of under-reporting was 25.9–31.2%, calculated using the Goldberg method (rEI:BMR) with implausible reporters had rEI:BMR values that differed from physical activity levels by more than ± 2 standard deviations [47,48]. The food items observed in this pattern were mainly cooking ingredients or additions to food items consumed at main meals or snacks. However, it can be referred to as the common food pattern, as these foods are commonly consumed by the Malaysian population. This DP has not been previously identified in the Western dietary pattern [27-29,36] but was similar to the ‘less-healthy’ pattern (condiments, sugar, oils and fats, sweets and desserts, and tea and coffee) reported by pregnant women in Universiti Sains Malaysia (USM) Birth Cohort [49]. Thus, this pattern is believed to be the true DP of our study population. The finding also reflects specific differences in dietary patterns between the Malaysia and Western population.

Generally, women with diets high in sugar and fat show a higher risk of GDM [31,50]. Yet, this study showed that women with HA to DP 2 (characterized as high sugar and fat) in pre-pregnancy and DP 5 in the first trimester were in fact significantly at reduced risk of GDM compared to women with LA to such a relatively unhealthy diet. Subsequently, a stratified analysis was performed to determine the associations between DP 2 and DP 5 with GDM by pre-pregnancy BMI (Table 4). This analysis showed that the significant association between DP 5 and GDM was only observed among underweight or normal weight women. Meanwhile, a significant association between DP 2 and GDM risk was found only among obese women. In addition, the present study also found that among women with HA to DP 5, underweight or normal weight women had significantly lower total energy intake in the first trimester (1,541 kcal/day) than overweight or obese women (1,624 kcal/day) (S2 Table). The mean total energy intake was also below the requirement for pregnant women in first trimester (1,690 kcal/day) [51]. Based on these data we speculate that the lower total energy intake observed in women with this dietary pattern might be protective of insulin resistance and consequently hyperglycemia risk. Thus, a lower BMI status and energy intake of women with high HA to DP 5 may underlie the observed association with a lower GDM risk. We cannot examine the association between pre-pregnancy energy intake and DP 2 with risk of GDM as the pre-pregnancy energy intake cannot be assessed retrospectively. Nevertheless, further investigation is needed to confirm the observed associations for underweight or normal weight women and for overweight or obese women and understand the drivers for GDM risk.

The findings from the ANOVA repeated measures showed that non-GDM women tend to have consistently higher DP scores for all three patterns before pregnancy up to the second trimester while GDM women showed lower and/or inconsistent DP scores from pre-pregnancy up to the second trimester of pregnancy. These trends in DP scores were similar regardless of adjustments for covariates like gestational weight gain and pre-pregnancy BMI. Pattern 2 and pattern 3 can be considered as the “less healthy” patterns due to their higher energy, fat, and sugar content. Yet, pattern 2 represents the typical Malaysian DP in which additional cooking oils, condiments and sugary substances are used to prepare standard fresh food items. Previous studies have shown that genetic factors are important in determining susceptibility to GDM [52,53]. Inherited abnormalities of pancreatic islet β-cell function and/ or β-cell mass may be implicated in the etiology of GDM [54]. Thus, it is plausible that non-GDM women in the present study have genetic variants that make them less susceptible to the risk of GDM and can explain why they are not at risk of GDM despite having also high DP scores for less healthy DPs. In general, the percentage of women that developed GDM in this cohort was low, also compared to the reported prevalence of GDM in Malaysia [2,3,5]. This may indicate that the SECOST study was conducted among a relatively traditional and healthy population. Further study is needed to identify the factors including genetic that explain susceptibility for GDM and examine the associations between genes, DPs, and BMI with the risk of GDM.

Similar to previous studies [28,29,33], the present study supported that only DPs in the first trimester and before pregnancy were significantly associated with the risk of GDM. Maternal metabolism in early pregnancy is critical not only for maternal health but also gestational programming [55,56]. During the first few weeks of pregnancy, the presence of placenta causes a reduction in growth hormone levels, resulting in enhanced insulin sensitivity [57]. GDM occurs if β-cells are unable to compensate for the changes in insulin resistance [58]. In another word, women are exposed to metabolic dysregulation in early pregnancy, and this metabolic abnormality may increase the risk of adverse pregnancy outcomes. Oostvogels et al. (2017) showed that pre-pregnancy maternal weight status and early pregnancy maternal lipid profile are independently associated with offspring adiposity [59], which reinforces the important role of maternal metabolism before and in early pregnancy. The precise underlying mechanisms between pre-pregnancy and early pregnancy DPs with the risk of GDM, however, remains to be elucidated. Nevertheless, the available evidence seems to support that women planning for pregnancy or in early pregnancy should be the target of lifestyle intervention for preventing GDM [60-62].

This study is not without any limitation. The results of the PCA approach might be affected by several arbitrary decisions such as determination of a number of factors to extract, the components labeling, the method of rotation, and even their interpretation [63]. In this study, the total variance of 3 patterns explained in each time point is relatively low (<40%). However, this finding was comparable with other previous studies that reported a total variance of 16.6–32.9% [23,27,31]. Besides, there is also the possibility of recall bias and misreporting when using SFFQ for dietary assessment. Nevertheless, the use of well-trained enumerators in data collection could reduce these errors. Although the effects of potential covariates were adjusted in the statistical methods, residual confounding of unknown confounders cannot be excluded. Besides, nutrition knowledge for study population is unknown. Future studies could include extending confounders to the model as well as to determine the nutrition knowledge of pregnant women. Despite these limitations, the present study is worthwhile as it provides important insights into the association between DPs before and during pregnancy and the risk of GDM, which such information is lacking in the literature.

Conclusions

The present study found that diets rich in sugar, spread and creamer, spices and condiments (DP 2 and DP 5) were significantly associated with reduced risk of GDM. The unexpected findings could be due to factors such as, lower BMI, and reduced energy intake. Furthermore, non-GDM women maintained a high DP score for all DPs from pre-pregnancy until second trimester of pregnancy, meanwhile GDM women showed lower and/or inconsistent DP scores, suggesting that non-GDM might have a lower genetic susceptibility to GDM as they are not at risk of GDM despite having also high DP scores for less healthy DPs (pattern 2 and pattern 3). A future large-scale prospective study or a well-designed randomized controlled trial is warranted to further confirm the association between the dietary pattern and the GDM in the Malaysian population. Such findings could inform efforts to strengthen existing or develop appropriate health and nutrition strategies to address the increasing rate of GDM in Malaysia.

Adjusted odd ratios and 95% confidence intervals for dietary pattern 2 and GDM (n = 452) [Sensitivity analysis].

(DOCX)

Click here for additional data file.

Energy intake in first trimester of women by pre-pregnancy BMI.

(DOCX)

Click here for additional data file.

18 Sep 2019

PONE-D-19-22261

The association between dietary patterns before and in early pregnancy and the risk of gestational diabetes mellitus (GDM): data from the Malaysian SECOST cohort

PLOS ONE

Dear Dr Yong,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

• The justification for using the Principal Component Analysis is not clear because it is possible to assess the relationship between frequency of consumption of specific food groups (e.g. 17 food groups mentioned in the article) and risk of Gestational Diabetes Miletus (GDM) directely without extracting hidden DPs. The identification of hidden dietary patterns (principal factors/components) through the PCA simply complicated the interpretation of the findings.

• The authors presented conflicting findings about the significance of diet and dietary pattern for GDM. At some point they stated that DP2 and DP5 generated by PCA are significant predictors of GDM while latter they argued that dietary pattern as assessed by FFQ has no association with GDM.

• In the abstract it is stated that “Our FFQ analysis did not reveal any clear associations between (pre)pregnancy dietary patterns and the risk to develop GDM (or not) …..” however in the results section, no result is provided regarding the direct association between frequency of consumption of food groups and risk of GDM.

I recommend the authors to focus on the relationship between frequency of consumption of specific food groups (as measured by FFQ) and occurrence of GDM. Kind of multivariable analyses for assessing the association between the 17 food groups and GDM may provide practically meaningful findings.

• Table 3 presents the results of three different models. I recommend the authors to make interpretation based on one model which they consider as the best.

• The authors did not satisfactorily explain how DPs relatively higher in loadings for sugar, spread & creamer, sauces, condiments & spices, and oils & fats are associated with reduced risk of GDM. In the discussion section, they mentioned that the significant association was only observed among underweight and normal weight women; however the results of this stratified analysis had not been present in the results section.

We would appreciate receiving your revised manuscript by Nov 02 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Samson Gebremedhin Gebreselassie, PhD

Academic Editor

PLOS ONE

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating the following in the Acknowledgments Section of your manuscript:

"SECOST was supported by a research grant from Danone Dumex (Malaysia) Shd. Bhd. We are extremely grateful to all the women who took part in this study, the medical officers and nurses for their assistance during data collection, and the whole SECOST team, which includes interviewers, research assistants, and volunteers."

We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

"No - The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

Additionally, because some of your funding information pertains to commercial funding, we ask you to provide an updated Competing Interests statement, declaring all sources of commercial funding.

In your Competing Interests statement, please confirm that your commercial funding does not alter your adherence to PLOS ONE Editorial policies and criteria by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” as detailed online in our guide for authors  http://journals.plos.org/plosone/s/competing-interests.  If this statement is not true and your adherence to PLOS policies on sharing data and materials is altered, please explain how.

Please include the updated Competing Interests Statement and Funding Statement in your cover letter. We will change the online submission form on your behalf.

3. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

Additional Editor Comments (Section-by-section comments):

Background

• Please provide clear definitions for dietary pattern and “prudent diet”.

• Page 3, line 73-77: the sentences are not clear. Please revisit them.

• The background section should give a highlight of what other studies conducted in Western or Asian population have found about the relationship between dietary pattern and GDM.

Methods

• Please comment on the adequacy of the sample size (power) for addressing the research questions at hand.

• Line 184: please provide additional information on how the sensitivity analysis was made.

• Line 188-91: not clear. Please revisit it.

• Line 138-40: “The factor scores for each dietary pattern and each individual were calculated by summing intakes of food items weighted by their factor loadings”. Line 187-188: ANOVA was used to plot the DP trajectory with similar pattern by maternal glycemia categories. These sentences are not clear, please revisit them.

• “Three major patterns were identified before and during pregnancy: pattern 1 consisted of DP 1, DP 4, and DP 7, pattern 2 consisted of DP 2, DP 5, and DP 8, and pattern 3 included DP 3, DP 6 and DP 9.” The sentence is not clear.

Results

• Table 1: What was the justification for using age of 30 as a cut-off here? Use of 35 years as a cut-off would be practically meaningful b/c the risk of many obstetric complications tends to increase after the age of 35 years. Similarly, the cutoffs used for classifying height has not been justified.

• Table 1: STPM/ Matric??? Not clear

• If the information was collected during pregnancy, how did you manage to calculate pre-pregnancy BMI?

• Line 214-15: “Three specific DPs were observed at each time period before and during pregnancy”. It is not clear how you decided to extract three components at each round. Did you evaluate the scree plot or decide take all the components having Eigenvalues above 1? Or did you simply take the first three components with the highest Eigenvalues?

• Table 2: the combined total variance explained by the three principal components is relatively low (<40%) in all of the three rounds (pre-pregnancy, first trimester and second trimester). It is usually recommended that at least 60% of the variance should be explained by the extracted components. Do you think that the PCA has adequately explained the original 17 food groups?

• I could not get the results of the “sensitivity analysis” mentioned in the methods section.

• Line 259-271: the paragraph is not clear. Please revisit it again.

Discussion

• Line 297-9: “Subsequently, a stratified analysis was performed to determine the associations between DP 2 and DP 5 with GDM by pre-pregnancy BM”. Please provide the findings of the stratified analysis in the results section. I see nothing about this analysis in the methods and results section.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: N/A

Reviewer #2: Thank you for asking me to review this interesting paper which has results contrary to expectations. The findings suggest that an unhealthy diet is protective against the development of gestational diabetes in a Malaysian population. The work is overall well conducted and well put together but the conclusions need further explanation and discussion in order to reconcile the findings.

Women eating DP 2 and 5 had a diet rich in sugar, spread and creamer, spices and condiments, but this is unlikely to account for all calories. It would be helpful to know if these women were eating much else besides? Perhaps a 'typical day' idea of what this diet looked like would be useful. It would also be helpful to know what exactly the number of calories were for each group. If this group was eating only small amounts of other food stuffs then it is possible the calorie amount was very different.

What %carbohydrate was seen in each group? Is there information on dietary macronutrients? It is also possible that the difference in carb or fat content also drives this association.

I personally would urge caution in interpreting these findings and would suggest that the relevance of this to public health remains unclear. This is not reason to reject this paper, but there is a risk the findings could be misinterpreted.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

20 Nov 2019

Response to the decision letter

We are grateful to the editor and reviewers for their constructive comments on our manuscript. We have addressed all comments and suggestions accordingly and submitted a revised version of the manuscript for further consideration by the journal. Specifically, we have provided point-by-point response to comments / suggestion put forth by the editor and reviewers.

1. The justification for using the Principal Component Analysis is not clear because it is possible to assess the relationship between frequency of consumption of specific food groups (e.g. 17 food groups mentioned in the article) and risk of Gestational Diabetes Miletus (GDM) directly without extracting hidden DPs. The identification of hidden dietary patterns (principal factors/components) through the PCA simply complicated the interpretation of the findings.

Response:

A paragraph to justify the use of Principal Component Analysis (PCA) to identify dietary patterns has been included in the introductory section (Line 69-81, text in red).

“Dietary intake during pregnancy is commonly assessed through the intakes of energy, macronutrients, micronutrients or food groups [12,13], which may then be examined in relation to an imbalanced maternal diet and poor maternal nutritional status. Dietary pattern is a relatively new approach that describes a combination of commonly consumed foods [14] that allows for the diet to be described as a whole [15]. Nutritional health outcomes are often the result of multiple synergies among nutrients and foods rather than just the sum of the individual food [16]. Although there are several approaches to identifying dietary patterns of pregnant women, the posterior-approach derived from principal component analysis (PCA) is the most commonly used for deriving dietary patterns during pregnancy [10-12]. PCA is a technique to reduce a large of correlated variables into a smaller number of components [21,22], revealing the underlying structure within diets of the population. Numerous studies examining the role of PCA derived DPs in adverse pregnancy outcomes, such as GDM, pre-eclampsia, and pre-term birth [23-25] have been published.”

2. The authors presented conflicting findings about the significance of diet and dietary pattern for GDM. At some point they stated that DP2 and DP5 generated by PCA are significant predictors of GDM while latter they argued that dietary pattern as assessed by FFQ has no association with GDM.

Response:

The association between dietary pattern and GDM was examined using logistic regression. Three models were examined. The first model adjusts for clinic and gestational week at OGTT. Model 2 adjusts for covariates of model 1 plus maternal age, and ethnicity. Model 3 then adjusts for covariates in model 2, plus income, pre-pregnancy BMI, medical history of GDM and family history of DM. This study found a significant association between DP and GDM in model 1 and model 2. However, after adjusting for covariates in model 3, the association becomes non-significant. The models are intended to portray the influence of different types of covariates (model 1 – clinic factors; model 2 – biological factors; model 3 – familial / environmental factors) on the association between DP and GDM. However, as suggested by reviewer, we will only retain 1 model, but which include the following covariates – clinic + biological factors + family history of DM + previous GDM.

3. In the abstract it is stated that “Our FFQ analysis did not reveal any clear associations between (pre)pregnancy dietary patterns and the risk to develop GDM (or not) …..” however in the results section, no result is provided regarding the direct association between frequency of consumption of food groups and risk of GDM.

Response:

We have revised the conclusion of the abstract accordingly (Line 47-53, text in red).

Abstract

Overall, a relative low percentage of women with GDM was found in this cohort. The risk was lower in women with high adherence to a relatively unhealthy dietary pattern, i.e. DP 2 and DP 5. The lower body mass index (BMI) status and energy intake of women showing a high adherence to DP 2 in the first trimester may underlie the observed association with a lower GDM risk. Additionally, genetic variance might explain the less susceptibility to GDM despite HA to unhealthy DPs among non-GDM women.

4. I recommend the authors to focus on the relationship between frequency of consumption of specific food groups (as measured by FFQ) and occurrence of GDM. Kind of multivariable analyses for assessing the association between the 17 food groups and GDM may provide practically meaningful findings.

Response:

We would like to retain the dietary pattern analysis as to emphasize the importance of food / nutrient interaction that represents usual diet. A paragraph to justify the usage of dietary pattern has been included in the Introduction section (refer to Response 1).

5. Table 3 presents the results of three different models. I recommend the authors to make interpretation based on one model which they consider as the best.

Response:

We have revised Table 3 accordingly.

6. The authors did not satisfactorily explain how DPs relatively higher in loadings for sugar, spread & creamer, sauces, condiments & spices, and oils & fats are associated with reduced risk of GDM. In the discussion section, they mentioned that the significant association was only observed among underweight and normal weight women; however, the results of this stratified analysis had not been present in the results section.

Response:

A lower BMI status and energy intake of women with high HA to DP 5 may underlie the observed association with a lower GDM risk. The findings of the stratified analysis have been included in the methods and results section. Please refer to Table 4 for the stratified analysis.

Methods:

A stratified analysis was performed to determine the associations between DP 2 and DP 5 with GDM by pre-pregnancy BMI (Table 4).

Results:

The findings of the stratified analysis showed that the significant association between DP 5 and GDM was only observed among underweight or normal-weight women. Meanwhile, a significant association between DP 2 and GDM risk was found only among obese women.

Background

1. Please provide clear definitions for dietary pattern and “prudent diet”.

Response:

The definitions for the two terms have been incorporated:

Dietary pattern – (Line 72-73, text in red)

Prudent dietary pattern – (Line 96-100, text in red)

2. Page 3, line 73-77: the sentences are not clear. Please revisit them.

It could also be attributed to the different food items reported for the respective DPs. For example, the “prudent” DP in a prospective cohort study of women in Northern China was characterized by higher intakes of dark-coloured vegetables and deep-sea fish [12], whereas the “prudent” DP of women in the Born in Guangzhou Cohort Study (BIGCS) was predominantly dairy products, nuts, eggs, fish, soups, and fruits [13].

Response:

We have revised the sentences. (Line 96-100, text in red)

It could also be attributed to the different food items reported for the respective DPs. For example, the “prudent” DP, which is a diet delivers health benefits, observed in a prospective cohort study of women in Northern China was characterized by higher intakes of dark-colored vegetables and deep-sea fish [26], while the “prudent” DP of women in the Born in Guangzhou Cohort Study (BIGCS) was predominantly dairy products, nuts, eggs, fish, soups, and fruits [27].

2. The background section should give a highlight of what other studies conducted in Western or Asian population have found about the relationship between dietary pattern and GDM.

Response:

A paragraph on the relationship between dietary pattern and GDM has been included in the Introduction section. (line 83-89, text in red)

Associations between DP before and during pregnancy with risk of GDM has been examined [26-30]. Most studies showed that an adoption of the prudent pattern, diet that was rich in vegetables, fruits, whole grains and legumes showed significant lower risk of GDM [26-28], whereas adhering to the Western pattern, which is characterized by high intakes of red meat, processed meats and refined foods, was associated with increased risk of GDM [23,26,31]. There are several suggested mechanisms by which Western pattern may increase the risk of GDM through inflammation, and placental dysfunction [26,29,30].

Methods

3. Please comment on the adequacy of the sample size (power) for addressing the research questions at hand.

Response:

The adequacy of the sample size of pregnant women was checked to ensure a 5% statistical significance level and 80% power. (Line 126-128, text in red)

For PCA analysis, the Kaiser-Meyer-Olkin measure of sampling adequacy was used to ensure that the sample size was adequate for these analyses. The Kaiser-Meyer-Oklin values were 0.68 (pre-pregnancy), 0.71 (first trimester), and 0.65 (second trimester). (line 160-163, text in red)

4. Line 184: please provide additional information on how the sensitivity analysis was made.

Response:

We have revised the sensitivity analysis. Sensitivity analyses were run between acceptable reporters and under-reporters of energy intake. Parity and total energy intake have been taken out from the sensitivity analysis model, as parity was not significant associated with GDM risk. For energy intake, the relationships between DP and GDM are likely to be at least partially mediated by total energy intake. As energy intake is likely to be on the causal pathway between dietary variables and the outcome of interest, it may be inappropriate to adjust for total energy when modelling DPs and GDM. However, we have to some degree, accounted for total energy intakes by identifying the dietary patterns: a higher z-score for the observed dietary pattern indicates that an individual has a high intake of food relevant to the specific dietary pattern.

5. Line 188-91: not clear. Please revisit it.

Multivariable-adjusted odds ratio (OR) with 95% confidence interval (CI) were presented. Potential covariates included in the multivariable model were age (continuous), ethnicity (Malays/ Non-Malays), household income (continuous), pre-pre-pregnancy BMI (continuous), history of GDM (yes or no), and family history of DM (yes or no). In sensitivity analyses, additional adjustment for parity (yes or no), energy intake at that trimester (continuous), and omitting under-reporters (yes or no) was investigated. Repeated measures ANOVA was used to plot the DP trajectory with similar pattern by maternal glycemia categories. Three major patterns were identified before and during pregnancy: pattern 1 consisted of DP 1, DP 4, and DP 7, pattern 2 consisted of DP 2, DP 5, and DP 8, and pattern 3 included DP 3, DP 6 and DP 9. Significant level of the statistical analysis was set at the cut-off point of p< 0.05.

Response:

We have revised lines 188-91 accordingly. (Line 211-219, text in red)

Adjusted odds ratio (OR) with 95% confidence interval (CI) of the association between DPs and GDM were presented. A further stratified analysis was performed to determine the associations between DP 2 and DP 5 with GDM by pre-pregnancy BMI (Table 4). Repeated measures ANOVA was used to plot the DP trajectory for GDM and non-GDM groups. Three major patterns were identified at each time point: pattern 1 (DP 1, DP 4, and DP 7), pattern 2 (DP 2, DP 5, and DP 8), and pattern 3 (DP 3, DP 6 and DP 9). Sensitivity analyses were run between acceptable reporters and under-reporters of energy intake. The findings did not change the main findings (S1 Table). The cut off point for statistical significance was set at p<0.05.

6. Line 138-40: “The factor scores for each dietary pattern and each individual were calculated by summing intakes of food items weighted by their factor loadings”. Line 187-188: ANOVA was used to plot the DP trajectory with similar pattern by maternal glycemia categories. These sentences are not clear, please revisit them.

Response:

Line 164-166: The factor scores for each DP and for each individual item were calculated by summing the food item intakes weighted by their factor loadings.

Line 214-215: Repeated measures ANOVA was used to plot the DP trajectory for GDM and non-GDM groups.

7.“Three major patterns were identified before and during pregnancy: pattern 1 consisted of DP 1, DP 4, and DP 7, pattern 2 consisted of DP 2, DP 5, and DP 8, and pattern 3 included DP 3, DP 6 and DP 9.” The sentence is not clear.

Response:

Line 215-216: Three major DPs were identified at each time point: pattern 1 (DP 1, DP 4, and DP 7) pattern 2 (DP 2, DP 5, and DP 8) and pattern 3 (DP 3, DP 6 and DP 9).

Results

8. Table 1: What was the justification for using age of 30 as a cut-off here? Use of 35 years as a cut-off would be practically meaningful b/c the risk of many obstetric complications tends to increase after the age of 35 years. Similarly, the cutoffs used for classifying height has not been justified.

Response:

We have revised the age group to 35 years. The classification of height was based on the tertile of height.

9.Table 1: STPM/ Matric??? Not clear

Response:

STPM refers to the Malaysian Higher School Certificate and Matric refers to Matriculation.

Footnote has been added for the abbreviation of STPM/Matriculation (Table 1).

10. If the information was collected during pregnancy, how did you manage to calculate pre-pregnancy BMI?

Response:

Women were requested to recall pre-pregnancy body weight. Pre-pregnancy body mass index (BMI) (kg/m2) was calculated as pre-pregnancy weight divided by the square of height and classified according to the recommendation of World Health Organization [14]. (Line 191-194, text in red)

11. Line 214-15: “Three specific DPs were observed at each time period before and during pregnancy”. It is not clear how you decided to extract three components at each round. Did you evaluate the scree plot or decide take all the components having Eigenvalues above 1? Or did you simply take the first three components with the highest Eigenvalues?

Response:

The number of factors was determined by eigen values greater than 1.5 [15]. (line 156-157, text in red).

12. Table 2: the combined total variance explained by the three principal components is relatively low (<40%) in all of the three rounds (pre-pregnancy, first trimester and second trimester). It is usually recommended that at least 60% of the variance should be explained by the extracted components. Do you think that the PCA has adequately explained the original 17 food groups?

Response:

We agree that the total variance explained by the 3 patterns at each time point is relatively low. However, several studies that used PCA to derive DP during pregnancy reported a comparable percentage of total variance explained (<30%) % [23,27,31]. We believe that it is reasonable to expect a total variance of < 40% in this study. We have addressed this issue as a study limitation.

Limitation:

In this study, the total variance of 3 patterns explained in each time point is relatively low (<40%). However, this finding was comparable with other previous studies that reported a total variance of 16.6 – 32.9% [23,27,31].

13. I could not get the results of the “sensitivity analysis” mentioned in the methods section.

Response:

The result of the sensitivity analysis has been added. (S1 Table)

14. Line 259-271: the paragraph is not clear. Please revisit it again.

Differences in DP trajectory from pre-pregnancy until the second trimester of pregnancy between non-GDM and GDM women are depicted in Fig 2. While non-GDM women maintained high adherence to all three DPs before and during pregnancy, GDM women clearly seemed to change their diets over the course of pregnancy. For pattern 1 (DP 1, DP 4 and DP 7 – prudent diet, high in fruits, vegetables, nuts, seeds, legumes, eggs, milk and dairy), GDM positive women had low DP scores before pregnancy, decreased scores at the first trimester further decreasing at second trimester time point. For pattern 2 (DP 2, DP 5 and DP 8 – high in condiments & spices, and sugar, spread & creamer), a V-shaped trend was observed in GDM women, where DP score was at the lowest point in the first trimester and higher at the second trimester. For pattern 3 (DP 3, DP 6 and DP 9 – high protein, sugar and energy), GDM women showed an inverse V-shaped trend in that they had low DP score before pregnancy, the highest DP score at the first trimester and the lowest DP score at the second trimester.

Response:

We have revised the paragraph accordingly. (Line 298-309, text in red)

DP trajectory from pre-pregnancy to second trimester of pregnancy for non-GDM and GDM women are depicted in Figure 2. While non-GDM women maintained high adherence to all three DPs before and during pregnancy, GDM women changed their diets during pregnancy. For pattern 1 (DP 1, DP 4 and DP 7 – prudent diet, high in fruits, vegetables, nuts, seeds, legumes, eggs, milk and dairy), GDM women had low DP score before pregnancy, decreased score at the first trimester and further decreased score at second trimester. For pattern 2 (DP 2, DP 5 and DP 8 – high in condiments & spices, and sugar, spread & creamer), a V-shaped trend was observed in GDM women, where DP score was at the lowest point in the first trimester and higher at the second trimester. For pattern 3 (DP 3, DP 6 and DP 9 – high protein, sugar and energy), GDM women showed an inverse V-shaped trend in that they had low DP score before pregnancy, the highest DP score at the first trimester and the lowest DP score at the second trimester.

Discussion

15. Line 297-9: “Subsequently, a stratified analysis was performed to determine the associations between DP 2 and DP 5 with GDM by pre-pregnancy BM”. Please provide the findings of the stratified analysis in the results section. I see nothing about this analysis in the methods and results section.

Response:

Please refer to Table 4 for the stratified analysis. The findings of the stratified analysis have been included in the methods and results section as below.

Thank you for asking me to review this interesting paper which has results contrary to expectations. The findings suggest that an unhealthy diet is protective against the development of gestational diabetes in a Malaysian population. The work is overall well conducted and well put together but the conclusions need further explanation and discussion in order to reconcile the findings.

Response:

We have revised the conclusion accordingly. (Line 404-415, text in red)

Conclusion:

The present study found that diets rich in sugar, spread and creamer, spices and condiments (DP 2 and DP 5) were significantly associated with reduced risk of GDM. The unexpected findings could be due to factors such as, lower BMI, and reduced energy intake. Furthermore, non-GDM women maintained a high DP score for all DPs from pre-pregnancy until second trimester of pregnancy, meanwhile GDM women showed lower and/or inconsistent DP scores, suggesting that non-GDM might have a lower genetic susceptibility to GDM as they are not at risk of GDM despite having also high DP scores for less healthy DPs (pattern 2 and pattern 3). A future large-scale prospective study or a well-designed randomized controlled trial is warranted to further confirm the association between the dietary pattern and the GDM in the Malaysian population. Such findings could inform efforts to strengthen existing or develop appropriate health and nutrition strategies to address the increasing rate of GDM in Malaysia.

Women eating DP 2 and 5 had a diet rich in sugar, spread and creamer, spices and condiments, but this is unlikely to account for all calories. It would be helpful to know if these women were eating much else besides? Perhaps a 'typical day' idea of what this diet looked like would be useful.

Response:

The detail of dietary pattern before and during pregnancy has been published elsewhere (DOI: 10.4162/nrp.2019.13.3.230). (Line 168-170, text in red)

It would also be helpful to know what exactly the number of calories were for each group. If this group was eating only small amounts of other food stuffs then it is possible the calorie amount was very different. What %carbohydrate was seen in each group? Is there information on dietary macronutrients? It is also possible that the difference in carb or fat content also drives this association.

Response:

Yes, we have run the analysis as indicated (energy and macronutrients with GDM). However, only total energy showed a significant association. Please refer to supplementary tables 2.

I personally would urge caution in interpreting these findings and would suggest that the relevance of this to public health remains unclear. This is not reason to reject this paper, but there is a risk the findings could be misinterpreted.

Response:

Thank you for your valuable comments. We have revised the manuscript accordingly. I hope the revised manuscript will meet the requirements of PlosOne.

The study findings should be interpreted with caution as this pattern was relatively higher in loadings for sugar, spread & creamer, sauces, condiments & spices, and oils & fats. (Line 316-318, text in red).

17 Dec 2019

The association between dietary patterns before and in early pregnancy and the risk of gestational diabetes mellitus (GDM): data from the Malaysian SECOST cohort

PONE-D-19-22261R1

Dear Dr. Yong,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Samson Gebremedhin, PhD

Academic Editor

PLOS ONE

Additional Editor Comments:

Please accommodate the additional comments provided by reviewer 2.

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: No

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: This version is overall much improved. A few minor points:

Abstract:

HA should be defined (last line)

Methods:

line 126. Why were women excluded?

Can you be sure that women with pre-existing glucose intolerance were excluded? If not, then you need to acknowledge this as a limitation.

lines 299-310 - good to remember that these are based on relatively small numbers of women. What are the related confidence intervals? Are these changes significant?

Is there any information on total calories consumed? If so this should be added to table 2.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

23 Dec 2019

PONE-D-19-22261R1

The association between dietary patterns before and in early pregnancy and the risk of gestational diabetes mellitus (GDM): data from the Malaysian SECOST cohort

Dear Dr. Yong:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Samson Gebremedhin

Academic Editor

PLOS ONE