A protocol for an observational cohort study of heat strain and its effect on fetal wellbeing in pregnant farmers in The Gambia

Introduction: Climate change predictions indicate that global temperatures are likely to exceed those seen in the last 200,000 years, rising by around 4°C above pre-industrial levels by 2100 (without effective mitigation of current emission rates). In regions of the world set to experience extreme temperatures, women often work outside in agriculture even during pregnancy. The implications of heat strain in pregnancy on maternal health and pregnancy outcome are not well understood. This protocol describes a study to assess the physiological response of pregnant women to environmental heat stress and the immediate effect this has on fetal wellbeing. Methods and analysis: The study will be performed in West Kiang district, The Gambia; a semi-arid zone in West Africa with daily maximum temperatures ranging from approximately 32 to 40°C. We will recruit 125 pregnant women of all ages who perform agricultural work during their pregnancy. Participants will be followed every two months until delivery. At each study visit fetal growth will be measured by ultrasound scan. During the course of their working day we will take the following measurements: continuous maternal physiological measurements (heart rate, respiratory rate, chest skin temperature and tri-axis accelerometer data); intermittent maternal tympanic core temperature, four point skin temperature, blood pressure; intermittent fetal heart rate and, if eligible, umbilical artery doppler; intermittent environmental measurements of air temperature, humidity, solar radiation and wind speed. Venous blood and urine will be collected at beginning and end of day for biomarkers of heat strain or fetal distress and hydration status.

This is a prospective observational cohort study of pregnant women who perform outdoor agricultural work during pregnancy, which has been recruiting since August 2019. This study will be conducted at Keneba field station, Medical Research Council Unit The Gambia at London School of Hygiene and Tropical Medicine (MRCG @ LSHTM). MRC Keneba is a rural field station based 2.5 hours inland from the coast, in Kiang West region where mostly subsistence agriculture is practiced. The climate in this area has two distinct seasons, the wet and dry season, which run from July to October and November to July, respectively. Farming of rice and groundnuts occur during the wet season and relies mostly on rainfall. In the dry season there are large “gardens”, which are used to grow a variety of vegetables. These tend to be watered by hand. Farming is a gender specific activity, with men growing the cash crops and women mainly growing food for household consumption or selling at the local markets. All agricultural work practiced by women in the region is done manually – planting, transplanting, weeding, harvesting, clearing, tilling and watering. Previous work in The Gambia has assessed the energy expenditure of pregnant women during different agricultural activities and also assessed the amount of time spent on these activities. These studies show that women will work between 50% to 83% of a 9-hour day on agricultural work, depending on the season, even when pregnant 45– 48. The mean monthly temperatures in 2017 varied from 25.4–30.3°C and maximum monthly temperatures from 31.5–39.5°C. The maximum monthly Wet Bulb Globe Temperature (WBGT) varied from 24.7–29.3°C. This gives the exposure during the hottest times of the day. This exposure is at a level that international guidelines would identify as at risk of heat illness. The annual average temperature rise since 1980 is just below 1°C. Most villages do not have electricity and therefore no access to air conditioning or electric fans. Water is mostly supplied through public bore holes, although some homes do have tapped water. Community sensitization and discussion will occur in each village prior to any visits. Once agreement and consent from the village elders has been obtained, we will recruit 125 participants. Pregnant women will be approached and informed about the study in their preferred language. Inclusion and exclusion criteria are set out below. Inclusion criteria: Exclusion criteria: Previous studies on physiological changes in pregnant women working in heat have not been completed. Based on published literature, we expect around 35% of agricultural workers to experience heat strain and assume this incidence risk remains at least as high in pregnancy 49, 50. Assuming an unexposed incidence risk of fetal distress to be 5% with an alpha of 0.05, we will need to recruit 99 participants to be powered to detect an exposed incidence risk of 30% with fetal distress. Taking into account drop-out rates due to fetal loss, we will recruit 125 participants. Pregnant women, of any gestation, identified by the demographic surveillance system (DSS), antenatal clinics or village assistants will be approached and consented if eligible. They will attend the Keneba antenatal clinic where socio-economic, demographic, medical and obstetric details will be collected. These will include any past medical history, past obstetric history including previous miscarriages, stillbirths, premature births or low birth weight infants. A baseline ultrasound will be performed by a trained member of staff. Gestational age will be determined based on an early ultrasound scan (under 28 weeks gestation) using biparietal diameter. If an early ultrasound scan has not been performed then biparietal diameter will still be used but with the expectation of reduced accuracy in the estimation of gestation. In women between 28–34 weeks we will perform an UmbiFlow™ scan. The UmbiFlow™ device was designed in South Africa for use in low-resource settings to identify women at risk of poor birth outcomes due to placental pathology. It measures the resistance index (RI) in the umbilical artery and plots this on a normogram based on gestational age (see Figure 5 for an example). This device is designed for use by unskilled practitioners and requires minimal training. It has been validated for gestational ages 28–34 weeks 51, 52. Within the next 2 weeks, on the day they are working outside, they will attend Keneba field station where they will have baseline anthropometry, physiology readings and blood and urine collection. They will be fitted with an Equivital™ LifeMonitor device. This is a portable, multi-parameter telemetry device that sits within a Lycra chest belt with inbuilt fabric sensors 53. Once wearing the LifeMonitor device they will complete a 6-minute walk test to determine cardiovascular reserve and calibrate the device 54. During the working day (duration recorded) we will record their tympanic temperature and the ambient conditions every hour. At middle and end of day we will assess fetal heart rate ± umbilical artery flow. Maternal measurements will include a four-point skin temperature using an infrared, non-contact thermometer. Measurements are taken from four-left-hand sided points from 20 cm away; chest, mid-tricep, mid-thigh and mid-calf 55. A Perfect ®Prime thermal imaging camera IR10019 with a resolution of 320 x 240 and pixels of 76,800 will be used to take two pictures per time point; from the waist up (with head-dress removed), and from the waist down (with legs revealed). Heart rate and blood pressure will be measured with an automatic OMRON M3 Intellisense device. Standardised ratings for thermal sensation and comfort will be recorded. At the end of the participant’s normal working day we will collect end line blood and urine and take a final measurement of weight and bioimpedence. Participants will be followed every 2 months during the course of their pregnancy. Figure 6 and Figure 7 give an overview of study processes and timing. DSS, demographic surveillance survey; PMH, past medical history; USS, ultrasound scan. MUAC, mid-upper arm circumference; BP, blood pressure; USS, ultrasound scan. After delivery, data will be collected on birth outcome, birth weight, gestational age, infant sex and maternal and newborn status. Recruitment will be over a 12-month period to ensure different seasonal exposure to work and heat. By recruiting over the course of a year and repeating measures every two months we will capture different trimesters for the same women. This will give us an estimate of the physiological changes that occur at different heat exposures and by different trimesters and identify if these alterations lead to altered fetal wellbeing. The primary outcome is a measurement of fetal distress. We define compromised fetal wellbeing as either: (i) a baseline fetal heart rate above 160 bpm or below 115 bpm; and/or (ii) if the fetus is 28–34 weeks gestation, then UmbiFlow™ above the 75th percentile of established resistance index graphs, or absent end diastolic flow, in keeping with the findings from South Africa and the developers of UmbiFlow™ 51, 52. Study staff will collect a venous blood and urine sample for each participant for use in study laboratory procedures. All samples aim to identify maternal heat strain or fetal wellbeing. Table 1 gives the laboratory sampling and justification. Whole blood samples from each participant will be used to prepare six dried blood spots of 10 µl each on filter paper and stored for biomarker testing. Serum samples will be separated and stored at -80°C for future analysis. CRP- C-reactive protein; IL-6 interleukin-6; TNF, tumor necrosis factor. Statistical analysis will be performed using R. Appropriate descriptive analysis will be used to present maternal characteristics and environmental heat stress exposures. Data will be assessed for normality and skewed data will be appropriately transformed. Metabolic rate and energy expenditure will be determined from the raw accelerometer and heart rate data using complex non-linear modelling. The 6-minute calibration test will allow development of individual and trimester specific estimates of metabolic rate. These will be cross-checked against historic data on energy expenditure of pregnant women in West Kiang per activity type. Heat strain will be determined by either the physiological strain index (PSI) or the Center for Disease Control (CDC) recommended signs and symptoms score. The PSI model is based on changes in heart rate and core body temperature and therefore gives an indication of the combined thermal and cardiovascular load: Where 0 indicates baseline and 1 indicates rate or value during exposure 20. This has been used in multiple studies on physiological changes in exercise and/or heat but not in pregnancy 56, 57. The CDC method is based on a series of symptoms related to heat illness, which vary from heat rash to heat stroke 58. We will include those related to heat stroke, heat exhaustion and heat cramps or a core temperature above 38°C, but will not include symptoms of heat rash or sunburn as these are not related to the physiological changes we are interested in. There are several heat stress indices we will calculate based on the direct field measurement we will take. These will include the WBGT, the Universal Thermal Climate Index, the apparent temperature and the heat index. A mixed-effect linear model will be run, using lme4 package in R to allow fixed and random effects to be incorporated appropriately. The expected final model will be of the form: Fetal distress(ij) = fixed part [heat stress index + PSI/heat strain + maternal age + gestational month (or trimester) + nutritional status + metabolic rate + cardiac reserve + heat illness symptoms + ΔHct + Δbioimpedence] + random term [individual participant] Fetal compromise(ij) = presence or absence of fetal distress as defined above for individual i at gestational month j (1…9). Different commonly used heat stress indices as described above will be validated against heat strain data for clinical correlation. Changes in fetal heart rate from baseline, stratified by trimester will be explored. Heart rates > 170 and > 180 will also be used as cut offs for fetal distress, although the numbers may be small. Changes in biomarkers of heat strain or feto-placental function will be analysed by ANOVA stratified by trimester and heat stress exposure. This study has been approved by the Gambia government/MRC Joint ethics committee (ref: 16405) and the London School of Hygiene and Tropical Medicine Ethics Advisory Board. Written informed consent and information sheets will be provided to all participants. A trained study staff member will conduct individual screening interviews and informed consent procedures in the preferred language of the participant. If the participant is unable to write, her fingerprint will be used as substitute for a signature, and an impartial adult witness to the entire consent procedure will provide their signature. Potential participants will be able to ask questions and discuss the study with study staff at any time during and after study activities. Participants are free to withdraw consent at any time during the course of the study and this will not impact on future care provision. Risks associated with participating in this study are minimal. Participants will be screened at the start of the day and should they demonstrate any signs or symptoms of illness or concern they will be advised to seek the attention of the Keneba health clinic and participation in the study will be delayed until they are well. Should a participant be hypertensive but not pre-eclamptic then may still enter the study, but we will refer them to antenatal services for treatment of their hypertension. Should a participant develop pre-eclampsia after recruitment, diagnosed at antenatal clinic or on subsequent visits, they will be referred to Keneba antenatal services and not included in the daily assessment of maternal heat strain and fetal wellbeing. However, if they are willing to remain in the study, pregnancy outcome data will still be collected. Participants will have additional venous blood samples taken using aseptic technique with universal precautions to minimize the risk of infection, personal discomfort, transient bleeding and bruising that may result. An ultrasound scan is part of routine antenatal care and an additional scan adds no harm to maternal or fetal health. The risks of wearing the portable recording devices include chaffing of the skin and discomfort, which we will minimize by ensuring a good fit at the beginning of the day and checking for any skin irritation at the end of the day. During the 6-minute walk test the participant can stop at any point during the course of the exercise and standardised feedback was collected. Additionally, this test will be performed at Keneba field station, close to the clinic area and if any untoward symptoms are experienced, they will assess and treat the participant as required. During the course of the day, if a significant heat load is experienced and the maternal core temperature increases beyond 38.5 oC (see below) the guidelines on treatment of heat strain will be followed with some additional considerations. In non-pregnant individuals, heat strain is determined to be a life-threatening emergency requiring immediate treatment when core temperature reaches 40.5°C 31. In pregnancy, heat is known to be teratogenic in the embryonic period, and throughout pregnancy compensatory mechanisms may be compromised. Hence we do not think it is ethical to allow the temperature to reach such a high level. Therefore, should maternal core temperature reach 38.5°C, this would result in an immediate review and treatment of heat strain would be commenced. This would include an overall clinical assessment of the women and fetus, immediate measures to treat the women and if these did not result in improvement within 30 minutes, consideration of transfer to the health facility. If during the intermittent measurements of fetal wellbeing, there are any concerns with fetal heart rate (either >160 or <115) 66, or regarding the Doppler results, then the participant will be assessed, encouraged to rest in the left lateral position, consume water, and have observations of fetal movement and maternal blood pressure taken. If after 30 minutes, the baseline heart rate has not returned to the normal range, the Doppler remains abnormal, or there are any clinical concerns, the women will be offered transport to the health clinic for further review and treatment. Any such events will be recorded as an incident case of fetal distress as per the primary outcome of this study. This is a field-based study in rural West Africa and therefore there are several limitations when comparing it to a laboratory based heat chamber study. We aim to characterize the physiological response of pregnant women to heat stress. Due to ethical considerations, this is an observational study only. Therefore if the women are not exposed to extreme heat in their usual work, we will not be able to measure this effect. However, from previous work in this setting, women were working in the heat during pregnancy and therefore we are confident that the exposure will occur. Since we are recruiting women who work in the gardens or farms during pregnancy, we are aware that our sample may be biased towards those in the lowest socio-economic group, which may affect the generalisability of the result. However, we consider this group to be of particular importance as they are likely to have little options for modification of behaviour or development of adaptation strategies. We will use Demographic Health Surveillance Data to compare our sample with the wider pregnant community over the time period of our study to determine the representation in the sample and any significant sources of bias. Our physiology measurements do not include continuous core temperature monitoring as would be the gold standard, due to practical constraints of field work. We therefore use tympanic temperature as a measure of core temperature, recognising the deficiencies in this measurement. Due to well documented measurement errors in this method, we are likely to underestimate the true temperature rise. Additionally the Equivital device measurements, heart rate, respiratory rate, skin temperature and tri-axis accelerometer will all have measurement errors. We will attempt to minimise these by removing impossible values, cleaning the signal and using the inbuilt algorithms for percentage confidence in the results. Due to the practicalities of accessing remote rice fields, and transporting all necessary equipment we will be unable to accurately assess water intake. Our assessment of hydration therefore is at the physiological level, but this is a limitation in the methodology especially considering that hydration is an important aspect of thermoregulation. All participants will be allocated a unique identifying number (UIN) at recruitment. Data generated by the wearable sensors will be downloaded from the devices at the end of the day, linked to the UIN and then wiped. During the study day, data will be collected on tablets using the REDCap application. On return to Keneba field station the tablets are synced, allowing transfer of encrypted data to the designated server. All data will be backed up regularly by the IT department in accordance with MRC SOP-INT-001. The database is centrally stored, data is secure and encrypted and held by MRC/LSHTM. No personal identifiable information will be available in any shared or published document. Primary data outputs will be in XML format. All study documents will be filed and stored for at least 10 years. The results of the study will be analysed and prepared for publication in open-access peer-reviewed international journals, staggered over time. At the end of the project a community event will be held to disseminate results to all those communities that participated in the study. We will comply with international standards and guidelines regarding open access of research data.

N/A

• bonell, ana
• hirst, jane
• vicedo-cabrera, ana m.
• haines, andy
• prentice, andrew m.
• maxwell, neil s.