Tehran environmental and neurodevelopmental disorders (TEND) cohort study: Phase I, feasibility assessment

Abstract

Purpose

To advance knowledge about childhood neurodevelopmental disorders and study their environmental determinants, we conducted a study in Tehran, Iran to assess the feasibility of prospective birth cohort study.

Methods

We evaluated participation of pregnant women, feasibility of sampling biological material, and health care services availability in Tehran in four steps: (1) first trimester of pregnancy; (2) third trimester of pregnancy; (3) at delivery; and (4) two to three months after delivery. We collected related data through questionnaires, also various biological samples were obtained from mothers (blood, urine, milk and nails—hands and feet) and newborns (umbilical cord blood, meconium, and urine samples) from February 2016 to October 2017.

Results

overall 838 eligible pregnant women were approached. The participation rate was 206(25%) in our study and about 185(90%) of subjects were recruited in hospitals. Out of 206 participants in the first trimester, blood, urine, hand nail, and foot nail samples were collected from 206(100%),193(93%), 205(99%), and 205(99%), respectively. These values dropped to 65(54%), 83(69%), 84(70%), and 84(70%) for the remaining participants 120(58%) in the third trimester, respectively. Also, we gathered milk samples from 125(60%) of mothers at two to three months after delivery.

Conclusion

Our findings suggest that hospitals were better places for recruitment of subjects in a birth cohort in Tehran. We further concluded that birth cohort study recruitment can be improved by choosing appropriate gestational ages. Obtaining the newborn’s urine, meconium, and umbilical cord blood were challenging procedures and require good collaboration between hospital staff and researchers.

Introduction

Environmental exposures during pregnancy and childhood may permanently change the body’s structure, physiology, and metabolism that can promote the risk of illness long time after the environmental exposure has occurred, even across generations [1, 2]. Thus, pregnancy and birth cohort studies are designed to study the impacts of early exposures by collecting various biological samples at multiple time points during development of the child, and offer a promising approach to enhance etiologic knowledge of various diseases [3, 4]. On the other hand, the human genome sequencing and mapping provides basic information for elucidation of gene expression and protein function (biochemical pathways) implicated in the natural history of chronic diseases, which may consequently offer opportunities for a more effective treatment and improvement of patient management [5]. However, increasing evidence showed that genetic variants describe a limited proportion of variance of the risk of chronic diseases and environmental factors and its interaction with genetic has more pronounced role in the development of chronic diseases [6, 7]. For instance, common and highly complex chronic pathologies, such as neurodevelopmental disabilities, obesity, and asthma may be induced or promoted by environmental risk factors and the interaction with individual characteristics (e.g., genetics) [8–11].

Epidemiological studies have led to the hypothesis of the “developmental origins of health and disease” [12]. Based on this hypothesis, the development of chronic diseases are influenced not only by genetic and adult life-style factors, but also by environmental factors acting in early life (during the pre-conceptual, fetal, and infant phases of life) [1]. Although many birth cohort studies are being developed and implemented in Iran, the published results have not sufficiently investigated effects of environmental factors on pregnancy outcomes and children development [13–15].

For the first time in Iran, Tehran Environment and Neurodevelopmental Disorders (TEND) project, as a prospective population-based birth cohort study in Tehran, Iran, is established to study the potential associations between pre- and post-natal environmental exposures and neurodevelopmental changes.

Cohort studies are a long-term investment requiring solid designs and careful planning, implementation and monitoring of high quality standard procedures to ultimately provide insights into the causes of the development of chronic pathologies.

Sustainable implementation of cohort studies requires also optimal integration of its research infrastructure into processes of the community health system, which in turn calls for proper piloting. In this regard, the first phase of the TEND project as a feasibility study aimed to evaluate: (1) factors relating to willingness and acceptance of pregnant women to participate and cooperate in a pregnancy cohort study and; (2) the feasibility of sampling biological material from pregnant women from inception until delivery; (3) the accessibility and acceptability of different healthcare services in Tehran to integrate research related procedures in the daily work. Overall, we aimed to provide evidence-based recommendations for launching of a larger birth cohort study in Tehran to quantify the effect of environmental factors on children mental development.

Materials and methods

The present study was conducted in Tehran, the capital of Iran with the population of more than 8.7 million people (according to 2016 national census). Across this large city, spatial variation in environmental exposures, socioeconomic conditions, and health care services availability is substantial, and the number of newborn delivery is very large.

In Tehran, several governmental and private healthcare centers, hospitals, and clinics provide prenatal care services. Divided into three governmental sectors, the city has three medical universities, namely Tehran University of Medical Sciences (TUMS), Shahid Beheshti University of Medical Sciences (SBUMS), and Iran University of Medical Sciences (IUMS) that supervise all healthcare duties within the city. These universities provide health care services to inhabitants of their catchment areas. TUMS is responsible for healthcare services for about 3 million people in the southern areas of Tehran [16]. TUMS runs 14 hospitals, including six providing prenatal care and delivery facilities with nearly 3′000 deliveries per year. Also, TUMS has 30 comprehensive health care center (2016 data) [17].

The inclusion criteria for subjects of this study were: (1) Iranian nationality; (2) living in Tehran; (3) gestational age less than 13 weeks at the recruitment day; and (4) deciding to give birth at one of our research hospitals. The TUMS Ethical Committees reviewed and approved the study design, procedures, and informed consent format (IR.TUMS.REC.1394.1180). For the feasibility study, we recruited pregnant women who referred to three hospitals and two healthcare centers with prenatal care unit and sampling facilities, located in different areas of Tehran metropolis.

The study was not designed or powered to detect differences by subgroups, we aimed for a sample size of 100 participants per health care services (hospitals and health care center) to estimate of mother participation and follow up, and sampling biological material collection with reasonable precision.

Data collection was not started simultaneously in all hospitals and centers. Two hospitals recruited mothers from February 2016 to October 2017 (21 months), one from September 2016 to September 2017 (12 months), one health care center from November 2016 to April 2017 (6 months), and the next health care center from November 2016 to September 2017 (11 months). Figure 1 shows the location of the research sites and participants.

Fig. 1.

The location of hospitals, healthcare centers, and participants

Data and biospecimen collection

Women who agreed to participate were asked to sign the informed consent. Data and bio samples were collected on four occasions; (1) in the first trimester of pregnancy; (2) in the third trimester of pregnancy; (3) at delivery, and (4) two to three months after delivery.

To formulate questionnaires and data collection tools, we contacted many researchers who had done similar studies around the world in addition to local experts. Using their experiences, we developed a structured questionnaire to evaluate the effects of environmental pollutants on pregnancy outcomes and child’s developmental abnormalities. Data were collected on socio-demographic, life style (e.g. smoking and alcohol), medical history, history of previous pregnancy (if any), diet, occupation and work environment, and potential sources of exposure data along with blood and spot urine collection, blood pressure, and anthropometric measurements. The details of characteristics and exposures and bio samples that were collected in each visit of study were presented in Table 1.

Table 1.

Baseline characteristics and demographic data of mothers and their newborns

Infant characteristics from (180 baby deliveries)
Sex
Male	74 (41%)
Female	106(59%)
Missing	0
Gestational age at birth, weeks
<37	60 (33.3%)
> = 37	95 (52.7%)
Missing	15(8.3%)
Birth weight, grams
<2500	9 (5%)
> = 2500	168 (93.3%)
Missing	3(1.6%)
Mean (SD)	3197 (539)
Mothers characteristics
Age mean (SD)	31.90 (5.9)
Range	19–47
Missing	0
Gestational age at participation week (mean (SD)	10.61 (3.13)
Mode of birth
Cesarean section   Normal vaginal delivery	126 (72.4%) 48 (26.6%)
Missing	6 (3.3%)

The first and third trimester questionnaires were administered through a paper based questioner during a face-to-face interview by trained research staff. The information about pregnancy outcomes, including mother and newborn, were collected by research collaborators in the hospitals’ delivery and operation rooms.

If the study participants did not deliver in our research hospitals, we called the women for a telephone interview. Two to three months after delivery, questionnaires about maternal depression (Beck’s Depression Inventory), child’s feeding and sleep habits, and a milk collection container were sent to participants’ home. A few days later, in the presence of a female questioner, milk samples were collected at home.

The present study obtained various biological samples from mothers (blood, urine, milk and nails (hands and feet)) and newborns (umbilical cord blood, meconium and urine samples). Using a cool box with the cold chain condition, collected blood and urine samples were transferred to the TUMS Central Laboratory Complex. The samples stored in the deep-freezer (−80 °C).

We used Moraes et al. (2015) [18] method for collecting newborns’ meconium samples. Trained staff sealed a waterproof dressing on the inner surface of the nappy cloth with a few sterile cotton cloths in the delivery room. One diaper for the newborn urine and another one for meconium sample were used. After collection of urine and meconium samples, we packed them with plastic zip-lock, put them in a cool-box and transferred it to the TUMS Central Laboratory Complex. In the laboratory, meconium was separated from cotton by a spatula and stored inside 3 cc Cryovials. Urine samples were extracted from cotton diaper by a 20 cc piston pressure syringe and then stored in 2 cc Cryovials. This procedure was done by a trained person under Cryovials laminar hood. After Cryovials was labeled by a coding device, samples were stored in a deep-freezer. Table 1 shows data and sampling collection time-tables. All blood, urine and breast milk stored at −80 °C as required.

Tracking and adherence strategy

We pursued various strategies to keep participation rates high. First, we have given an ID card to pregnant women at the time of survey recruitment. This ID card helped the staff of hospitals to identify the study participants from others in prenatal visits, laboratory administration, and in the delivery room. Second, we presented some gifts to the participants to encourage them to continue collaborations with the survey. Third, we gave the research supervisor contact information to all women for their reference whenever they wanted. In addition, we set up a social media channel for offering pregnancy, and infancy health information for all participants. In this channel we also provided a private room for mothers who needed to ask more questions from the specialists. In case of needed more health information and consultation, mothers and children could visit our collaborator physicians free of charge. Fourth, to abate loss to follow up due to residential mobility of participants, and to increase our knowledge of their health conditions, we had regular phone calls to the participants.

Results

During the study period, a total of 838 eligible pregnant women were approached, of which 206 (25%) agreed to participate. The time needed for admission tasks, study objective presentation, obtaining inform consent, essential form documentation, bio specimen sampling, and completion of the questionnaire for any case was in average 90 and 60 min in first and second trimester, respectively. Table 1 gives an overview of the basic characteristics of the participating mothers and children.

Figure 2 shows the participation rate during different stages of the study. As presented in this Fig. 2, we collected blood sample from 206 (100%), urine sample from 193 (93%), hand nail sample from 204 (99%), and foot nails sample from 205 (99%) women in the first trimester of pregnancy. In the third trimester of pregnancy, 120 (58%) pregnant women referred to the research sites for giving biological sampling and filling the questionnaires. Among these 120 women, 65 (54%) blood samples, (70%) hand and foot nail, and 83 (69%) urine samples was collected. Totally, blood samples were obtained for 31% and urine samples were obtained for 40% of pregnant women in the third trimester of pregnancy (out of 206 subjects). Out of all registered mothers in the first trimester, 23 (11%) had abortion and 3 (1%) had stillbirth.

Fig. 2.

The participation rate during different study stages. The percentages are out of the 206 subjects that initially entered the study in first trimester

Among 180 baby deliveries and 3 stillbirths, 133 (72.6%) were born in the research hospitals. This percentage for women recruited in hospitals (n = 164) was 78.4%, 78% and 96%, for hospital A, B, and C, respectively. The survey obtained 44 (33%) cord blood samples, 25 (18%) meconium, and 4 (3%) newborns’ urine sample from all 133 births in the research hospitals. In short, the success rate was 21% for collecting umbilical cord blood samples and 1% for neonate urine sample in all recruited participants (n = 206). We obtained the delivery questionnaire for 193 (93%) of cases up to one week after birth [44 (23%) by the staff of hospitals and 146 (77%) through telephone interviews].

We gathered depression questionnaires (the Beck Depression Inventory) from 140 (67% of 206) and milk samples from 125 (60%) mothers at two to three months after delivery. Thirteen questionnaires and samples were obtained at the hospital and healthcare centers and for other participants the depression inventory and milk containers had been collected at the mothers’ home.

Details of participation rate and feasibility of collection of bio samples based on study visit are reported and compared between hospitals and health care centers in Fig. 3. As illustrated in Fig. 3, a total of 185 (90%) samples were taken from the hospitals and 21 (10%) from healthcare centers. In the third trimester of pregnancy, 113 (61%) of pregnant women that recruited in first trimester in hospitals referred to the hospitals for filling the questionnaires. This participation rate was 7 (21%) in health care centers.

Fig. 3.

The comparison of participation rate and feasibility of collection of bio samples between hospitals and health care centers during different study visits

Figure 4 shows the pregnant women gestational age distribution at the time of recruitment in the study. The highest number of mothers registered during gestational age 8 (16%) and 12 (21%) weeks. After modifying the inclusion criteria, from 13 to 16 weeks, 14 (7%) pregnant women were registered at gestational age of 14 to 16 week.

Fig. 4.

Pregnant women gestational age distribution at the time of recruitment in the study

The main reasons of women who rejected participation in first visit were:

1. inability to come for regular prenatal care and/or delivery at the research sites (106); (b) decision for elective abortion; (c) pregnancy complications that had changed the regular schedule of prenatal care; (d) disliked to give blood sample; (e) distrust in the research procedure safety; (f) lack of time needed for filling of questionnaires; (g) inappropriate interview room and crowded areas; (h) sensitive questions in questionnaires, and (l) disagreement of their spouse.

2. The most frequent reasons for lost to follow-up of mothers in the third trimester were: (a) not satisfaction with taking a relatively high volume of blood sample; (b) women’s difficulty to reach to the research locations; (c) insufficient information about the survey procedure and arrangements; (d) time limitation due to daily life (specially with other small children at home); (e) need for bed rest; (f) forgetting the schedule of study; (g) inappropriate interview room and crowded areas; and (h) spouse disagreement.

Discussion

Designing a large-scale birth cohort study requires adequate knowledge and assessment of various processes in the community health system to optimize the use of all resources. The current study aimed to evaluate feasibility and challenges of running a large-scale birth cohort in Tehran megacity.

We were able to recruit 24.5% of the pregnant women who admitted in the hospitals and healthcare centers in our study. This participation rate was lower than some previous reports in MOBA (45% [19]) and in SNIP (75% [20]). However, the success rate in a study in Bremen, Germany was similar to ours (24%) as they added umbilical cord blood samples to their study protocol [3]. Noteworthy, the participation rate was more than 50% in one hospital in our study. So that, one hundred and twenty mothers were enrolled in mentioned hospital due to good communication ability of our questioner, higher number of patient admission at the prenatal clinic, and good support and collaboration from the staff of hospital. Good participation of pregnant women due to suitable communicability of researcher was reported in a previous birth cohort study [21].

Therefore, attention to recruitment sites is crucial in this regard. Also, it should be considered that lower loss to follow up might be more important than lower participation at the beginning of the study.

In the present study, we decided to exclude the fetal screening centers and private clinics due to referral status of such centers and many of their patients may come from outside of Tehran. In addition, it may cause some difficulties for following-up in the subsequent phases, especially at time of delivery.

We chose four of the thirty healthcare centers that had laboratory facilities, necessary for collecting biological samples. Two centers did not cooperate with the study due to lack of human resources. Considering the importance of access to the mother and infant at the time of delivery, we accessed to 72.6% of participants at delivery and/or operation room. The rate was relatively high (81%) for those participants that were recruited to the survey at the research hospitals. Therefore, this finding suggests that hospitals are better places for birth cohort studies in terms of delivering prenatal care, collecting biological samples, and giving birth in Tehran.

The number of recruited pregnant women for each research site varied from 3 to 120 participants, and 90% of cases were enrolled in the hospitals. Finally, the study results showed that, among other potential centers, hospitals in Tehran might be a better place for conducting birth cohort studies due to several reasons including higher number of potential participants, better sampling facilities, experienced staff, and possibility of delivery at the hospital.

The reasons of reluctance for participation in the present study can be classified into three categories: (1) Personal reasons (such as unwillingness, stress, fear of blood sampling, etc.); (2) The survey location problems (crowded hospitals, inappropriate access rout, etc.); and (3) The requirements of study (a relatively high volume blood sample, lengthy questionnaire, etc.). Thus, the following suggestions may increase the participation rate: arrangement for physician visit in advance; interview room preparation; providing quiet environment for interview and sampling; decreasing the number of visits; reducing the blood sample volume; and facilitating the overall study process.

In the present study, the maternal participation rate for sampling and the questionnaire interviewing was low in the third trimester of pregnancy. However, by changing the strategy and offering a gift to mothers in this trimester rather than at delivery time participation rate increased (e.g. the participation rate increased from 34% to 76% after this procedure modification in the hospital with the highest pregnant women recruitment).

In addition, the most important reason for not cooperating in the third trimester was not willing to give blood sampling again. One of the main objective of the TEND phase I was to find the best biological sample This change can limit blood sample volume, or waive blood sampling from the study protocol and replace it with non-invasive sampling. In addition, for reducing fear of blood sampling, merging research blood sampling time with routine blood tests during the antenatal care could be effective. It can decrease the total blood sample volume and keep the number of needle seek at minimum.

We added the study participants to a Telegram channel to keep connection with mothers, for the subsequent steps, and to provide useful health information. However, only 50 (24%) women joined this channel because some mothers did not have access to the Internet and/or smartphones. Additionally, some mothers believed that electromagnetic waves of mobile phones may have adverse health effects on their fetus. In the last trimester of pregnancy, particularly, many women reported that they turned off their smartphones to protect their baby. An Italian birth cohort feasibility study has reported 66% Internet access during pregnancy [22]. It seems that providing free internet access, offering smartphone, and promoting women’s knowledge about electromagnetic waves exposure may increase using social media Telegram channels among the study participants.

Although many of birth cohort studies have been launched at the time of delivery [23–25], assessment of effects of exposure to environmental factors during fetal life on the pregnancy outcomes will not be possible in such cohorts. In order to investigate effects of environmental exposures pregnancy cohorts should be launched instead, including at least the first trimester of pregnancy if not the time before conception [3, 26]. There are also some large studies that enrolled women in the more early stages of pregnancy [27–29]. We did not find any birth cohort study in Iran to evaluate appropriate gestational age for entering into the study. In the first step, we enrolled women up to 13 week of gestational age. During the last six months of the study, we changed the gestational age criteria from 13th to 16th weeks. Subsequently, the number of enrolled mothers increased dramatically. In this six-month period, a total of 86 mothers (41% of total) were enrolled (23 women were at 15 and 16 weeks of gestation).

The success rate for collecting umbilical cord blood samples was 21% of the total participated women and 32% for mothers who gave birth at our hospitals. In a similar study in Bremen, Germany, the percentage of umbilical cord blood samples was 53.7% [3]. In addition, three other studies have reported the success rate of umbilical cord blood sampling of 54%, 79%, and 82% [30–32]. One of the causes for the low percentage of umbilical blood samples was unavailability of permanent researcher in the delivery room. Another reason was emergency deliveries, which decreases sampling success rates. In some cases, the mothers did not allow the umbilical cord blood sampling because they believed it can reduce the blood volume of newborns. To increase the success rate in this procedure, we suggest employing some trained persons that cover the whole 24 h of delivery rooms to collect umbilical cord blood samples.

According to the study results, the urine and meconium collection from newborns were the most challenging sampling steps. When meconium was mixed with sampling cotton, separation of sticky meconium and transferring to Cryovials got a very hard procedure. The problem was solved when we withdraw cotton from the meconium sampling set. There were also some difficulties to obtain the newborns’ urine samples. First, in some cases the urine volume was very low that dried up before arriving to the laboratory. Second, extraction of urine from newborns’ nappies by syringe was very difficult. Third, sometimes urine and meconium samples were mixed. Finally, unavailability of refrigerator to hold the nappies in the research sites complicated the procedure more.

Two months after delivery, 82% of the mothers tended to continue their collaboration with the study, which could be an acceptable percentage for the further evaluations of children mental development. Although agreement of pregnant women to participate in the study was low at the time of recruitment, a large proportion of mothers agreed to remain in the survey at the postpartum phases of follow-up.

It could be said that the population recruited within this feasibility study was relatively selective because, only 25% agreed to participate. So, the results cannot be generalized to the general population. While this is a concerning issue when assessing prevalence and incidence measures, but lack of external validity in cohort studies causes less problems since the main objective of cohort studies is the investigation of associations within the study population and effect estimates for these associations are usually unbiased in this situation.

We tried to evaluate birth cohort study feasibility in the real model for Iranian health system with minimal interventions. It can make a reasonable estimation of the required logistics for conducting the main phase of study and it also provides a good picture regarding feasibility of the main phase.

Conclusion

In summary, the results of our study showed that the experience of researchers and their enthusiasm is one of the most important and effective factors in setting up a birth cohort study in Tehran health system. Although the level of participation in the study was not optimal, the rate of participation can be increased if suitable strategies be implemented, such as choosing appropriate gestational ages at the time of first recruitment as the numbers of available pregnant women at early gestational ages (e.g. less than 13 weeks) are limited. The tracking of mothers and gathering of biological samples in the third trimester of pregnancy was more difficult than other trimesters or postpartum in Tehran. Obtaining the newborn’s urine, meconium, and umbilical cord blood were challenging procedures and require good collaboration between hospital staff and researchers. We suggest choosing researchers from the hospital staff as they are more available for the study participants and sampling, and they can have better logistic support. More than 70% of the subjects that were included in the study had delivered babies in the research hospitals, which seem to be an acceptable percentage. Therefore, in the main phase of TEND, to improve efficiency and reduce costs, recruitment will be restricted to hospitals.