The effectiveness of healthy lifestyle interventions on weight gain in overweight pregnant women: A cluster‐randomized controlled trial

Abstract Aim Interventions based on adopting a healthy lifestyle have been less successful. The aim of this study was to investigate the effectiveness of healthy lifestyle interventions on weight gain in overweight pregnant women. Design A cluster randomized controlled trial. Methods Health centres were selected by simple random sampling; then, 66 overweight pregnant women were enrolled by convenience sampling and divided into intervention and comparison groups. Intervention group received individual nutritional counselling and physical activity training. The data were collected in several stages with the demographic and obstetric questionnaire, maternal weight record, food frequency and international physical activity questionnaire. Results Pregnancy weight gain‐4.75(CI 95%: −4.02, −5.48) was significantly lower in the intervention group (p < .001). Comparing between groups with adjustment for baseline values indicated that there was a statistically significant difference in terms of total calorie 95.46 (CI 95%: −22.37, 213.30), carbohydrate 23.45 (CI 95%: 2.12, 44.78), protein −7.16 (CI 95%: −12.85, −1.47) and fat 8.82 (CI 95%: 2.21, 15.67) intake. Despite the higher level of physical activity in the intervention group, there was no statistically significant difference between the two groups. Conclusion Counselling interventions for healthy living during pregnancy can lead to controlling weight gain, improving dietary habits and increasing the physical activity in overweight pregnant women.


| INTRODUC TI ON
Increased prevalence of obesity and overweight in different age groups, especially in women during their reproductive age, is one of the important health problems (Dodd et al., 2014;Jepson, Harris, Platt, & Tannahill, 2010), and disregard for the desired weight gain during pregnancy due to adverse effects on maternal and foetal health is one of the major concerns in prenatal care (Loh, Oen, Koo, Ng, & Yap, 2018). Reports indicate that more than 1.9 billion (39%) of the world's population are considered as overweight and obese, with 40% of them being women (O'Brien, Cramp, & Dodd, 2016).
Studies conducted in Iran also show a two-fold increase in the prevalence of obesity in women (Jafari-Adli et al., 2014;Kiadaliri, Jafari, Faghihzadeh, Kalantari, & Asadi-Lari, 2014) in such a way that the prevalence of overweight in women is 27.5%-38.5% and the prevalence of obesity is 12.6%-25% (Jafari- Adli et al., 2014). According to studies, 50% of women in developed countries have a body mass index above 25 at the time of pregnancy (Dodd et al., 2014;O'Brien et al., 2016). On the other hand, pregnancy is a critical time for women and their children to become overweight and obese (Thangaratinam et al., 2012;Tol, Tavassoli, Shariferad, & Shojaezadeh, 2011;Tovar, Chasan-Taber, Bermudez, Hyatt, & Must, 2010). Because pregnant women face a myth of "eating for two" and "more rest", are better during pregnancy (Fathnezhad- Warren, 2013). Therefore, a statistically significant percentage of pregnant women lose control over their eating habits and their emotional eating increases (Kimmel, Ferguson, Zerwas, Bulik, & Meltzer-Brody, 2016). According to studies, 20 to 40 per cent of women in Europe gain more than recommended weight during the pregnancy (Mourtakos et al., 2015). However, maternal overweight and obesity are associated with increased prevalence of chronic diseases, increased healthcare costs (Gebler, Charuvastra, & Silver, 2015), increased risk of maternal and neonatal mortality, increased risk of gestational diabetes, hypertension in pregnancy, greater use of induction of labour and caesarean section, neonatal macrosomia, hard labour (dystocia), increased probability of admission to intensive care unit and neonatal hyperbilirubinemia (Dodd et al., 2014;Mourtakos et al., 2015). Increased risk of developing type 2 diabetes (Liu, Ao, Yang, & Wang, 2013;O'Brien et al., 2016) and future cardiovascular disease in mother and child are some of the long-term complications of obesity and high weight gain during the pregnancy (Shah, Retnakaran, & Booth, 2008). Evidence suggests that the complications of excessive weight gain during the pregnancy are directly associated with increased body mass index (BMI) (Dodd et al., 2014).
In addition, due to inappropriate diets obese and overweight people face nutritional deficiencies more than people with normal weight, with studies suggesting unbalanced intake of macronutrients and micronutrients in such individuals (Groth & Morrison-Beedy, 2013;Hui et al., 2012). Proper diet and physical activity are important components of lifestyle; they not only affect pregnancy outcomes but also have intergenerational effects (Lindström, 2012). They also improve maternal and child physical and mental health (A. Hui et al., 2012;Thompson, Vamos, & Daley, 2015). However, interventions based on diet, physical activity and healthy lifestyle had no statistically significant effect on weight loss and improvement of pregnancy outcomes (Phelan et al., 2011). Studies carried out in Iran suggest that pregnant women have poor knowledge about physical activity (Noohi, Nazemzadeh, & Nakhei, 2010) and proper diet during pregnancy (Kamalifard, Charandabi, Mameghani, Jafarabadi, & Omidi, 2012).
And since the best interventions to get optimize outcomes for maternal and child are still unknown and research on adopting a healthy lifestyle has not been robust (Oteng-Ntim, Varma, Croker, Poston, & Doyle, 2012) and the findings are inconsistent (Baheiraei, Mirghafourvand, Mohammadi, & Charandabi, 2012;Groth & Morrison-Beedy, 2013). Therefore, we performed a cluster randomized trial on a lifestyle intervention for overweight pregnant women.
The research questions were as follows: 1. Is there the efficacy of healthy lifestyle interventions including individual nutritional counselling and physical activity on weight gain during pregnancy in overweight pregnant women? 2. Is there the effectiveness of these interventions on maternal and neonatal outcomes (gestational diabetes, neonatal weight, etc.)?

| Study design and
This open-label, prospective, cluster randomized controlled trial was conducted from 4 April 4-20 January 2018 in Tabriz, Iran. The study had two-arm parallel-group design. Tabriz is the fifth largest city of Iran and has 20 health centre in 12 clusters. Randomization was performed at the cluster level. Cluster randomization avoids dissemination bias, which would have occurred if individuals were randomized and treated within the same practice. First, by using Randomizer software, 6 health centres were selected through simple random sampling. Then, three Health centre was randomly selected as intervention centres and three health centre as comparison group. Then at each of them, list of all pregnant mothers meeting the inclusion criteria (first based on Nulliparity then by other criteria) were listed based on demographic data in the records of pregnant mothers and enrolled into study through convenience sampling.
Initially, they were called and invited to participate in the study. In case of not participating in the first study session, other eligible participants were requested to join the study to complete the final sample size. For determining the sample size, primary object of the study was used which was the weight gain during pregnancy. The sample size for comparing two groups was calculated by formula below with the confidence interval of 95% and power of 90%. And the effect size was concidred equivalent to similar study (Wolff, Legarth, Vangsgaard, Toubro, & Astrup, 2008), 22 individuals were assigned to each group. Due to three stages of follow-up and probability of sample loss, sample size increased by 50% and final sample size was 33 in each group and 66 in total. STATA13 software was used to determine sample size: Study parameter: alpha = 0.0500, power = 0.9000, delta = −6.7000, m1 = 13.3000, m2 = 6.6000, sd1 = 7.5000, sd2 = 5.5000, estimated sample sizes' = 44 so N per group = 22.

| Study participants
Pregnant women meeting all inclusion criteria including primiparous women with singleton pregnancy with 16-20 weeks of gestational age, BMI 25-29.9, age range 18-40 years and no underlying chronic diseases or risk factors were enrolled into study. Exclusion criteria were taking any type of chemical or herbal drug (other than dietary supplements recommended during pregnancy) that could interfere with normal weight gain during pregnancy, unwillingness of the mother to continue the study or to participate simultaneously. In another studies with similar interventions, emerging of any disorders or gynaecological problems requiring special care for the mother that could interfere with the interventions of the present study and to have an abortion. To reduce the dropout rate of the study, subjects initially were trained and briefed to attend the follow-up programmes; otherwise, telephone calls were made ( Figure 1).

| Socio-demographic and pregnancy history
It consists of the demographic variables of pregnant women containing age, educational level, occupational status of pregnant women and their spouses, self-assessment of household economic status, and midwifery variables, including the first day of the last menstruation, probable due date, gestational age based on first trimester ultrasound, number of pregnancies and childbirths, and height and weight before pregnancy.

| Physical activity questionnaire
In this study, the short form of the self-reported International

| Intervention protocol and its implementation steps
Most of the interventions in the intervention group included individual counselling as listed below: 1. Prescribing a balanced and flexible diet based on the maternal weight gain pattern during pregnancy, correcting nutritional misconceptions while respecting family cultural beliefs.
2. To help with optimal weight gain during pregnancy, it was recommended to exercise safely through in-person training, providing pamphlets and educational booklets. Trainings included aerobic exercise at least 3 times a week with moderate intensity and an exercise programme including 5-min stretching exercises.
A.fnK enrolled clusters and who assigned clusters to interventions, and S.H generated the random allocation sequence. To perform the intervention, after selecting healthcare centres as intervention site, pregnant women meeting the inclusion criteria who were 16-20 weeks pregnant were included in the study and were followed up in three time periods; 26-28 weeks, 35-37 weeks and 7 days postpartum. At the beginning of the study, after randomization and completing written informed consent, participants were provided with demographic, obstetric, food frequency and physical activity questionnaires which were completed in 45-60 min after explanations of the researchers. Then, in the intervention group, the researcher provided the necessary training and counselling on physical activity and nutritional information during pregnancy for the mother for 30 to 45 min depending on the mother's awareness and need. Nutrition and diet consultation was provided by nutritionist to the mothers. The researcher also provided the intervention group with educational booklets and pamphlets that included information on diet and physical activity. Afterwards, mothers were advised to attend follow-up and weight control sessions during pregnancy. In addition, the researcher shared their contact number with the mothers of the intervention group to respond their self-care questions any time during the 6-month intervention period. In each session, counselling was provided depending on the needs of the mothers in the intervention group. To avoid imposing additional costs and problems on commuting to the health centre, attempts were made to co-ordinate primary interventions and follow-up visits with prenatal care visits. Therefore, all scheduled visits were concomitant with one of the mother's visits to the health centre. The visits were as follows:

| Data analysis
After collecting data from all units, results were analysed using STATA 13 software. To describe the symmetric quantitative data, mean and standard deviation were used. Qualitative data were also reported with frequency and percentage. To compare the qualitative data, chi-square test was used. Also, considering the design of the study (cluster allocation) generalized linear models with random effects (random-effects model) were used to compare the mean of macronutrients at two time and physical activity at three-time points and interaction between time and group was investigated. Pairwise comparisons were performed with the follow-up test including Tukey. Independent t test was also used to compare fasting blood glucose and glucose tolerance test in 24-28 weeks of gestation.

| RE SULT
According to the demographic and obstetric characteristics of the studied units, the mean (SD) age of mothers was 25.50 (3.84) and most (50%) was in the age group of 25-30 years old. Fifty-five women (78.2%) had diploma and higher education and 53 women (80.3%) were housewives. Data analysis indicated no statistically significant difference between the demographic characteristics of the two groups, except for height and husband's education (Table 1).
Based on results, mean (SD) total calories consumed by pregnant women at the beginning and the end of the study were 2,100 (236) and 2,337 (238), respectively. Comparison of mean calorie, carbohydrate and protein intake before the intervention in the two groups showed no statistically significant difference; however, the intervention group had higher fat intake and the difference between two groups was statistically significant (p = .009). In spite of the fact that the amount of calories consumed at the end of the study was lower in the intervention group, there was no statistically significant difference between two groups (p = .112) but after adjustment baseline values comparison between two groups was statistically significant (p < .001). Also, there was a statistically significant difference in carbohydrate, protein and fat intake, indicating an increase in protein and a decrease in fat intake in the intervention group. Also, the amount of dietary fibre intake increased in the intervention group compared with the beginning of the study which was significantly (p < .001) higher than control group. Data analysis indicated that there was a statistically significant difference between the two groups in terms of dietary protein intake at the end of the study (p < .001). This difference was not significant at the start of the study indicating an increase in protein intake in the intervention group. However, the ratio of carbohydrate to total consumed calories at the end of the study was not statistically significant (p = .617). However, consumption in the intervention group was lower and had more fibre. Also, the intervention group had more calcium intake than the control group (p < .001) ( Table 2).
Although the level of activity was higher in the intervention group, especially during the study (second trimester of pregnancy), there was no statistically significant difference (p = .7) between the two groups at three different times (Table 3).
Based on the data related to weight gain in both groups, overall mean weight gain in the intervention group was statistically significant compared with the control group and the weight gain showed less growth (p < .001). Also differences in weight gain averages were statistically significant in the second and third trimesters (except first trimester); in spite of the fact that the weight gain of 5 women in the intervention group was in recommended range of IOM, there was no statistically significant difference between the two groups (p = .057) (Table 4).
Also, according to the results of Table 5, there was no statistically significant difference between the two groups in terms of mean fasting blood glucose and glucose tolerance test. Fisher test showed no statistically significant difference between two groups in terms of gestational diabetes (p = .56) ( Table 5).
Based on the independent samples t test, there was no statistically significant difference between birth weight, height and head circumference of neonates. In addition, caesarean delivery was the same in both groups and the most common cause was elective or planned caesarean section, although this variable was higher in the control group but there was no statistically significant difference in two groups (p = .81). Analysis of data indicated that there was no statistically significant difference in neonatal jaundice up to 7 days after delivery (p = .32). Other outcomes such as preterm labour, uterine atony and severe postpartum haemorrhage were not observed in any of the groups (Table 6). normal weight and has no effect on obese and overweight women.

| D ISCUSS I ON
According to these researchers, women with high body mass index tend to be more resistant to these interventions (Polley, Wing, & Sims, 2002), despite of the fact that some studies have suggested the positive effects of education and motivational programmes on weight control in obese women. In a case-control randomized experimental study performed in Sweden, in obese pregnant women with additional visits, motivational programmes had a positive effect of on encouraging regular aerobic exercise, developing motivational habits and behaviours and balancing the energy intake and appropriate weight gain during pregnancy (Claesson et al., 2008 TA B L E 1 Demographic and obstetric characteristics of overweight pregnant women in two groups of intervention and comparison and obese women, and the recommended amount is 300 calories a day (Moore Simas et al., 2013), and in the present study, this amount was considered in dietary recommendations, and the goal was to promote healthy eating culture and to control optimal weight gain.
Additionally, in given consultation it was advised to use reduce consumption of saturated fats and carbohydrates and to increase the unsaturated fats, protein intake. Also, it was recommended to use fruits, vegetables and fibres more commonly so that most of the calories would be provided by these nutrients. Hui et al. conducted a study for investigating the effectiveness of healthy lifestyle interventions, including home exercise, walking 3 to 5 times per week and nutritional counselling in primiparous and non-diabetic women.
They concluded that calorie, triglycerides, cholesterol and saturated intake only decrease in women with normal weight (Hui et al., 2014).
Only in two studies conducted by Wolf and Mottella, the calorie amount was decreased which was due to limited calorie intake (Mottola et al., 2010;Wolff et al., 2008), despite of the fact that the gained weight in our study was lower in intervention group; only 5 pregnant women have reached the recommended weight for overweight women by IOM (7-11.5 kg) and the rest has gained more than the recommended range. Similar diet programmes for total calorie intake in obese, overweight and normal weighted women might be a cause for improper weight gain. Because IOM has the same advise about calorie intake in obese and overweight women (Mottola et al., 2010). It seems that following a treatment and having a regular physical activity might affect the weight gain pattern.
According to the results of this study, despite of the statistically significant increase of physical activity in intervention group in second trimester, there was no change in third trimester. In general, studies conducted for evaluating the effectiveness of counselling and educating the pregnant women to encourage exercising and physical activity had inconsistent results (Althuizen et al., 2013;Hui et al., 2014;Wolff et al., 2008). In the study conducted by Tomoda et al investigating the effect of educational intervention on physical activity during the pregnancy, however, there was no significant change in physical activity (moderate to vigorous level) in the intervention group (Tomoda, Ogita, & Tamura, 1996). Unlike Tomoda, Hui et al. found that physical activity increased significantly two months after the intervention (Hui et al., 2014). Also, Shakeri and colleagues found a statistically significant difference between the two groups by forming 8 training sessions (Shakeri, Fekri, Shahnavaz, & Shakibazadeh, 2012 in primiparous women showed that after education of relaxation training and safe exercise during pregnancy, there was statistically significant difference in mean score of physical activity before and after training (Heffernan, 2000). In a similar study by Mirmulaei and his colleagues, the mean scores of physical activity before training in the two groups showed no statistically significant difference, but after training, the difference was statistically significant (Mirmolaei,  To compare between groups with adjustment for baseline values.
type of educated activity, the right time to encourage and decision to change behaviour are also effective in the intervention results, so that specific training for each pregnancy trimester that is easy, more accessible and at the same time more acceptable could be more effective than general recommended activities. It seems that the type of training activity and the right time to motivate and decide to change behaviour also influence the outcome of the intervention so that training for any pregnancy course that is easier, more accessible and at the same time more acceptable. It can be more effective than recommended public activities especially in late pregnancy, the time According to systematic reviews and meta-analysis of studies related to healthy lifestyle-promoting interventions, these interventions were associated with appropriate weight gain pattern of mothers in pregnancy; however, there was no statistically significant difference in pregnancy outcomes between the study groups.

| Limitation
One limitation of the study was assuming that pregnant women will not be able to follow regular dietary or exercise regimes in the first trimester due to some common complaints such as nausea, vomiting, feeling tired and requiring more sleep. Participants were selected from mothers with 16-20 weeks of gestational age. However, it is recommended to begin interventions earlier and if possible before the pregnancy.

| CON CLUS ION
Pregnancy is an opportunity to identify healthy behaviours and to reinforce or motivate them to change health-related misconducts.
However, the results of the third sub-study suggest that despite some differences observed between women who received counselling and training to improve nutrition and physical activity with those who did not receive such trainings, the difference in optimal weight gain, based on world-class recommendations, occurred in a handful of pregnant women. However, statistically significant differences in the pattern of healthy eating and preventing overweight in the intervention group were promising, and it improves the efficacy of these interventions.

ACK N OWLED G EM ENTS
We are extremely grateful to Mrs. Anita Kazem Shayan as a Nutrition Consultant. We thank the women who participated in the study. We also appreciate the funding support from Shahid Beheshti University of Medical Sciences (SBMU).

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R S' CO NTR I B UTI O N
A.fnK contributed to developed of the concept, collected data,  TA B L E 6 Pregnancy and neonatal outcomes in two intervention and comparison groups