Stepped-wedge cluster-randomised controlled trial to assess the cardiovascular health effects of a managed aquifer recharge initiative to reduce drinking water salinity in southwest coastal Bangladesh: study design and rationale

Introduction Saltwater intrusion and salinisation have contributed to drinking water scarcity in many coastal regions globally, leading to dependence on alternative sources for water supply. In southwest coastal Bangladesh, communities have few options but to drink brackish groundwater which has been associated with high blood pressure among the adult population, and pre-eclampsia and gestational hypertension among pregnant women. Managed aquifer recharge (MAR), the purposeful recharge of surface water or rainwater to aquifers to bring hydrological equilibrium, is a potential solution for salinity problem in southwest coastal Bangladesh by creating a freshwater lens within the brackish aquifer. Our study aims to evaluate whether consumption of MAR water improves human health, particularly by reducing blood pressure among communities in coastal Bangladesh. Methods and analysis The study employs a stepped-wedge cluster-randomised controlled community trial design in 16 communities over five monthly visits. During each visit, we will collect data on participants’ source of drinking and cooking water and measure the salinity level and electrical conductivity of household stored water. At each visit, we will also measure the blood pressure of participants ≥20 years of age and pregnant women and collect urine samples for urinary sodium and protein measurements. We will use generalised linear mixed models to determine the association of access to MAR water on blood pressure of the participants. Ethics and dissemination The study protocol has been reviewed and approved by the Institutional Review Boards of the International Centre for Diarrheal Disease Research, Bangladesh (icddr,b). Informed written consent will be taken from all the participants. This study is funded by Wellcome Trust, UK. The study findings will be disseminated to the government partners, at research conferences and in peer-reviewed journals. Trial registration number NCT02746003; Pre-results.

Introduction: Saltwater intrusion and salinization have contributed to drinking water scarcity in many coastal regions globally, leading to dependence on alternative sources for water supply. In southwest coastal Bangladesh, communities have few options but to drink brackish groundwater which has been associated with high blood pressure among the adult population, and preeclampsia and gestational hypertension among pregnant women. Managed aquifer recharge (MAR), the purposeful recharge of surface water or rainwater to aquifers to bring hydrologic equilibrium, is a potential solution for salinity problem in southwest coastal Bangladesh by creating a freshwater lens within the brackish aquifer. Our study aims to evaluate whether consumption of MAR water improves human health, particularly by reducing blood pressure among communities in coastal Bangladesh.

Methods and analysis:
The study employs a stepped-wedge cluster-randomized controlled community trial design to evaluate the impact of the programmatic implementation of MAR intervention in 16 communities over five monthly visits. During each visit, we will collect data on participants' source of drinking and cooking water and measure the salinity level and electrical conductivity of household stored water. At each visit, we will also measure blood pressure of participants' ≥ 20 years of age and pregnant women and collect urine samples for urinary sodium and protein measurements. We will use generalized estimating equations with robust standard errors to determine the association of access to MAR water on blood pressure of the participants.

Ethics and dissemination:
The study protocol has been reviewed and approved by the

Strengths and limitation of the study
• This is the first study to evaluate the health impact of managed aquifer recharge in southwest coastal Bangladesh.
• The stepped-wedge trial ensures we will have counterfactual data as well as gradual access to MAR water in all communities.
• Objective measurement of exposure (drinking water salinity) and outcomes (urinary sodium and blood pressure).
• The magnitude of exposure will vary geographically and across time period. Therefore, MAR water salinity will differ across communities at a single point of time, and also for the same community at different points of time.
• Compliance of the intervention may be different across sites and for individuals of different professions. Saltwater intrusion and salinization have increased groundwater salinity in many coastal aquifers and small islands across the world [1][2][3][4][5] . This is driven by a number of climatological and anthropogenic factors including global warming, increased cyclones and tidal surges, reduced river discharge, and increased groundwater abstraction in excess of recharge [6][7][8] . Communities in many coastal regions rely on groundwater as their main source of drinking water 9 as well as freshwater for domestic, agricultural and industrial purposes 10 . Nearly half of the world's population resides in coastal areas 3 and 10% of these live in low-lying coastal areas where surface elevation is generally <10 meters above mean sea level 11 . Salinization in coastal areas is expected to increase in the future because of increased groundwater withdrawal due to population and economic growth and sea-level rise 3 . As the world's population and economic activities continue to grow, groundwater supplies are progressively under threat of depletion, which increases the importance of monitoring, management and conservation of coastal freshwater aquifers 12 13 .
One approach to minimize the impact of groundwater salinization is to enhance groundwater recharge into coastal aquifers 14 . Managed aquifer recharge (MAR) is an approach to artificially promote freshwater recharge to increase storage. MAR involves infiltration of freshwater (e.g. rainwater and pond water) into aquifers to create a store of freshwater within the naturally brackish aquifer (Figure 1) [15][16][17] . MAR represents a promising adaptive strategy for increasing freshwater availability and sustaining a year-round drinking water supply that is protected from evaporation, and could be resilient to tidal storms, cyclones and surface water salinity since freshwater infiltration and storage occur under confined conditions 18 .
The brackishness of groundwater in southwest coastal Bangladesh is caused by a combination of climatic and anthropogenic factors including sea level rise, frequent cyclones and tidal storms, and shrimp cultivation [19][20][21][22] . In the future, climate change and sea-level rise are expected to cause more cyclones, tidal surges and flood in this region that will further affect surface water and groundwater salinity 7 . In many areas in southwest coastal Bangladesh, both shallow and deep aquifers contain naturally brackish water causing acute scarcity of drinking water (Figure 2A) 23 . Water salinity in southwest coastal Bangladesh follows a clear seasonal pattern ─ higher in the dry season than the wet season 24 . Salinity of water bodies builds up from October to May, peaking during February to early May. After May, salinity decreases sharply due to rainfall and increased upstream river flow. People collect and use rainwater for drinking and cooking purpose during the monsoon when precipitation is intense, but during the dry season, they generally rely on pond water or saline tube well water 25 .
The Geology Department of the University of Dhaka, in collaboration with UNICEF and the Department of Public Health Engineering (DPHE), Government of Bangladesh have piloted 20 small-scale MAR projects in three districts of southwest coastal Bangladesh to evaluate the feasibility of MAR for drinking water supply in rural communities 26 . The shallow brackish aquifer was the target storage zone that is overlain by Holocene clay aquitard of 3 to 15 meter thickness 16 . In the pilot phase, an average storage of approximately 900 m 3 of fresh water per year per site was established, sufficient to deliver 15 L of safe drinking water per day per household, which can fulfill the demand for drinking and cooking for approximately 300 people in 60-70 households during the dry season 27 28 . The second phase of the MAR project started in Epidemiological studies have demonstrated that high sodium intake is associated with elevated blood pressure [29][30][31] and other cardiovascular diseases 32 . A study conducted among the adult population residing in southwest coastal Bangladesh suggest drinking saline water was associated with high blood pressure after adjusting for personal, lifestyle and environmental factors 33 . Among study participants, the mean systolic blood pressure for those consuming water with sodium from the lowest quintile was 119.4 (SD 13.7) and from the highest sodium quintile was 126.7 mm Hg (18·0) 33 . Excess sodium intake from drinking brackish water has been also associated increased gestational hypertension and preeclampsia among pregnant women in southwest coastal Bangladesh 34 35 . Mean systolic and diastolic blood pressure were higher among the pregnant women who drank tube well or pond (saline sources) water compared to those who drank rain water 34 36 . The mean systolic blood pressure of pregnant women from these areas was 102.4 mm Hg among those who drank rainwater, 112.6 mm Hg among pond water users and 119.4 mm Hg among brackish groundwater users 34 . High blood pressure during pregnancy is associated with high maternal mortality, and adverse pregnancy and fetal outcomes 37 .
There are 37 million people living in the southwest coastal region and 20 million are currently affected by drinking water salinity 38 . The estimated mean global sodium consumption is 3.95 g per day (range 2.2 to 5.5 g per day) 39 , but people in southwest coastal Bangladesh consume up to 16 g of sodium per day through drinking brackish groundwater 40 . While MAR water can potentially reduce exposure to saline water, the health effects of providing access to

Objectives
Randomized controlled trials demonstrate that reduction in dietary sodium in adults decreases blood pressure among people both with and without hypertension [41][42][43] . Meta-analysis of randomized trials also suggest modest reduction in salt intake for four or more weeks causes significant reduction in blood pressure at the population-level 44 45 . We will assess whether access to low-salinity MAR water is associated with cardiovascular health benefits measured by reduction of blood pressure, urinary sodium and protein excretion among a population ≥20 years old and pregnant women in southwest coastal Bangladesh.

Study design:
We will implement a stepped-wedge cluster randomized controlled trial in 16 MAR communities. The stepped wedge design allows communities to gradually have access to MAR water; however, the point at which their access commences will be randomly assigned 46 .
In this way, each MAR site will contribute data for both the intervention and the control time periods. We will have five monthly steps in the stepped wedge trial (Figure 3). In the first step, none of the communities will have MAR water available for drinking and we will collect baseline information. During each subsequent month, four randomly selected communities will receive access to MAR water for drinking and cooking. In the last (fifth) monthly step, all the communities will have access to MAR drinking water.

Study settings:
The study will be conducted at the community-level in three districts in southwest coastal Bangladesh ─ Khulna, Satkhira and Bagerhat ─ where 75 MAR systems have been installed. Of these, 30 systems are already in use. We will select 16 of the remaining 45 MAR communities for the study based on consultation with Dhaka University and UNICEF using three criteria: communities have not started drinking MAR water, acceptable level of arsenic in MAR water, and electrical conductivity (a measure of salinity) of water below 2000 µS/cm, an indication that the MAR system has successfully reduced salinity in the aquifer.

Participant eligibility criteria:
The inclusion criteria will be households whose ≥ 20 year old members and pregnant women willingly agree to exclusively use MAR water during the dry season for drinking and cooking purposes. We will enroll post-adolescent ≥ 20 year population as an individual's response to salt intake (salt-sensitivity) increases with age 47 and adolescence is associated adrenal and nervous system maturation that may contribute to salt sensitivity 48 49 .
We will approach households living near the MAR water access point. We will enroll the 28 closest households surrounding each MAR systems that meet the inclusion criteria and consent to study participation. All household members > 20 year of age will be eligible and enrolled in the study from the selected 28 households. In addition, we will enroll households that include a pregnant woman in each selected MAR catchment community irrespective of household selection, from a list of pregnant women developed by a female promoter.
Intervention: We will randomly select four MAR communities to drink MAR water in each step following the first (baseline) step (Figure 3). The Dhaka University team have determined the catchment area under each MAR site based on geographic distance from MAR infrastructure and have developed a list of households who expressed willingness to drink MAR water. We will deploy a local trained promoter at each MAR site, who will visit these households, list members and identify pregnant women in these and other households under catchment areas of each MAR site. Promoters will encourage all household members to drink MAR water exclusively while they are at home, carry a bottle of MAR water while they go out for work and other activities, and to cook with MAR water.
Sample size: The sample size of the stepped wedge trial was calculated based on the primary objective: reduction of systolic blood pressure. Data from southwest coastal Bangladesh demonstrated a 7 mmHg lower mean systolic blood pressure among pregnant women who drank pond water compared to those who drank brackish water, and a 17 mmHg lower mean systolic blood pressure for pregnant women who drank rainwater compared to those who drank brackish water 34 . The mean reduction in systolic blood pressure following long term modest salt reduction is 4.2 mmHg 44 . For calculations we considered a mean systolic blood pressure reduction of at least 3 mmHg among the communities who will have access to MAR water for drinking and cooking purposes compared to those who will use brackish water. The entire 10 population under the MAR community catchment area will be considered as one cluster. Each MAR community serves approximately 300 people, constituting 60-70 households in the catchment area 27 . The mean household size in southwest coastal Bangladesh is 4.2, approximately 52% of household members are ≥ 20 years 50 51 . Since we will collect blood pressure for all household members ≥ 20 years, we estimated an average of 2.2 participants in this age group per household. We will select 28 households per cluster (28 *2.2 = 60 people per cluster). A previously published study in Bangladesh reported standard deviation (SD) for systolic blood pressure as 13.51 52 , but we considered standard deviation of systolic blood pressure 20 to capture a greater variation of blood pressure. We calculated the sample size for the

Data collection methods:
The outcomes of interest will be measured from selected participants of 16 communities irrespective of their access to MAR water during five monthly visits. The interval between two successive blood pressure measurements or two consecutive data collection visits will be at least one month for each participant. After taking informed written consent from household heads and each participant, research staff will administer a survey using a structured questionnaire programmed on handheld computers to collect information on drinking water  11 sources, household members' socio-demographic status, participants' medical history and cardiovascular risk factors including smoking history and alcohol intake during the first monthly visit. At this time they will also measure participants' weight, height, waist and hip circumference and blood pressure, and collect urine samples. During subsequent monthly visits, research staff will measure blood pressure, weight and collect urine samples. If a pregnant woman is identified, research staff will measure her blood pressure, and collect urine and blood samples at every monthly visits.
Exposure assessment: During each visit, research staff will inquire about household members' reported primary water sources used for drinking and cooking purposes in last 24 hours, collect reported information on whether participants exclusively used the primary water sources and explore whether any alternative water sources were used. They will also ask about the frequency of collection and cost of primary water sources for drinking and cooking, time required to collect water, amount of collected water and when the last water was collected. Research staff will observe the presence of stored water in households, ask the sources of each container of stored water and will collect water samples that have been stored for drinking and cooking (if any) to measure the salinity level, electrical conductivity, resistivity, total dissolve solutes (TDS) and temperature of water samples using Hanna Salinity™ meter. They will also collect the MAR water from the source MAR outlet to measure the salinity level, electrical conductivity, resistivity, total dissolve solutes (TDS) and temperature of water. Research staff will collect 24 hour urine from each participant to measure the urinary sodium excretion as a proxy for sodium intake 55 56 .
Outcomes assessment: Systolic blood pressure of the ≥ 20 year old population and pregnant women is the primary outcome of this study. Secondary outcomes will be diastolic blood pressure, mean arterial pressure, and pulse pressure ( Table 1). The different components of blood pressure are independent cardiovascular risk factors associated with sodium intake [57][58][59][60][61] , and indicate future risk for different cardiovascular diseases 62 . High systolic blood pressure puts more stress on the vascular wall and is strongly associated with risk for future intracerebral haemorrhage, subarachnoid haemorrhage, angina, myocardial infarction and peripheral vascular diseases 62 . Raised diastolic blood pressure is associated with aortic and thoracic aneurysm 62 .
Some effects of high sodium intake are independent of high systolic and diastolic blood pressure such as arterial stiffness and left ventricular mass ─ both of these are independent predictor of future cardiovascular diseases. High pulse pressure (difference between systolic and diastolic blood pressure) is associated with increased arterial stiffness 58 60 , and high mean arterial pressure (diastolic blood pressure plus one-third of systolic blood pressure) and high pulse pressure is independently associated with increased left ventricular mass 58 60 .
Reduced salt intake for four weeks has been associated with decreased proteinuria in blinded randomized controlled trial 79 . Preeclampsia is associated with high maternal mortality and adverse pregnancy and fetal outcomes 37 and is characterized by high blood pressure and proteinuria after 20 weeks of pregnancy 80 85 86 . Caffeine (tea, coffee, carbonated beverages), eating, heavy physical activities and smoking will be proscribed for 30 minutes prior to measuring blood pressure. Participants will rest for 5 minutes sitting on a chair keeping their arm supported. An appropriate sized cuff and calibrated instrument will be used for different age groups and the blood pressure instrument will be positioned at heart level. Blood pressure will be measured three times; first left arm, then right arm, then again left arm. Both systolic and diastolic blood pressure will be recorded from all measurements. The arithmetic mean of three systolic blood pressure measurements will be used as the primary outcome. However, if a systolic blood pressure measurement differs by 10% from the other measurements, that measurement will be excluded when calculating the arithmetic mean systolic blood pressure.
Biomarker measurements: Field research staff will instruct the participants' to collect 24 hours urine sample during each household visit. The volume of the urine samples will be noted at household level, and a sample of 25 ml urine will be collected and transported to the field laboratory at 2-8° C within 6 hours of collection for processing, analysis and storage. Aliquots of each participant's urine sample will be made for biochemical and electrolyte measurements.
Urinary creatinine concentration will be measured by a colorimetric method (Jaffe reaction) 14 using a semi-auto biochemistry analyzer (Evolution 3000, BSI, Italy). Urinary total protein will be estimated using a light sensitive colored reagent (Randox.UK). We will use the direct Ion Selective Electrode (ISE) method for urinary sodium measurements using a semi-auto electrolyte analyzer (Biolyte2000, Bio-care Corporation,Taiwan). We will measure the uric acid concentration in blood by an enzymatic colorimetric method (Evolution 3000, BSI, Italy). We will perform routine Quality Control for all tests using standard quality control reagents (Bio-Rad Laboratories, USA) and 5% of samples will be cross checked at the International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b) central laboratory, in Dhaka. We will use a chemifluorescent competitive enzyme-linked immunosorbent assay (ELISA) that quantifies marinobufagenin (MBG) levels in urine.

Statistical methods:
We will conduct an intention-to-treat analysis. Since we will have longitudinal data on individual participants, generalized estimating equations (GEE) with an appropriate autoregressive correlation structure, to account for repeated measures within individual study participants will be used as primary analysis. The systolic blood pressure will be regressed on an indicator variable for access to MAR water, and dummy variables for the study sites and time periods. The interpretation of GEE model output will be the population average association of MAR water exposure with blood pressure, conditional on study site. In secondary analyses, we will adjust for age, sex, weight and height, personal, dietary and environmental factors ( Table 2). We will use weighted GEE or multiple imputations methods to account for missing observations. In secondary analyses, we will consider the association of drinking MAR water with diastolic blood pressure, pulse pressure, mean arterial pressure, urinary sodium and protein concentrations. For the primary analysis of pregnant women for systolic blood pressure outcome, we will a run random effect model considering MAR (dummy variable) and gestational months as predictor variables.

Ethics
Informed written consent will be taken from all participants and household heads. This study protocol has been reviewed and approved by the icddr,b Ethical Review Committee. If research staff members identify hypertensive patients or cases of hypertension during pregnancy, they will refer patients to the local government health facilities for further management. In addition, research staff will train pregnant women and family members to recognize the danger signs during pregnancy, and will also instruct them where to seek medical care if such danger signs appear. During monthly household visits research staff will encourage pregnant women to attend prenatal visits.

Dissemination
Study findings will be shared with the Department of Public Health Engineering (DPHE), Department of Environment of the government of Bangladesh and other partner NGOs working in southwest coastal Bangladesh for safe access to drinking water. We will discuss the scope and limitations of the MAR system to address the demand of safe drinking water based on findings from this study. We will be submitting abstracts to international conferences for dissemination to the international audience working on safe water. We will develop manuscripts and submit to peer-reviewed journals to publish our research.

Discussion
This will be the first study to assess the health impact of an environmental intervention to reduce groundwater salinity in southwest coastal Bangladesh. Our study has several strengths. The stepped-wedge trial ensures that we will have counterfactual data as well as gradual access to MAR water in all communities. In a stepped-wedge design, treatment effect of an intervention can be estimated from between-and within-cluster comparisons as participants will act as their own control, compared to only between-cluster comparison in a parallel cluster randomized design 87 . We will use the same instruments for data collection and outcome measurements for all steps for exposed and unexposed communities, which will mitigate bias in data collection.
Measurement of water source salinity and urinary sodium concentration will explain the biologically plausible effect of drinking MAR water on blood pressure. The primary, secondary and tertiary outcomes are objective and would reduce the risk of reporting bias. Collection of detailed exposure and co-variate data will help in determining a valid association between drinking MAR water and health benefits. We will have several outcome variables and biomarkers that will ensure a comprehensive health benefit evaluation of access to MAR water.
The study has several limitations. We will be unable to control salinity level of drinking  88 89 . To account this, we will specify the versions of exposure by measuring the salinity level of MAR water available in participants' households and interpret the response as unit change of salinity at different MAR water salinity-levels. As compliance of drinking MAR water may be different across communities despite active encouragement by promoters, one problem of the intention to treat analysis is that if the proportions of participants who will always drink MAR water is low compared to those who will not, the potentially greater effect of MAR intervention on blood pressure of fully participating individuals may be washed out by the smaller effect of those who will not comply. To account for this, as a secondary analysis, we will conduct "instrumental variable" (IV) analyses by jointly running two regression models: a regression model for predicting urinary sodium excretion by drinking water salinity, and a regression model for blood pressure prediction given the urinary sodium excretion.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59

Methods and analysis:
The study employs a stepped-wedge cluster-randomized controlled 41 community trial design in 16 communities over five monthly visits. During each visit, we will 42 collect data on participants' source of drinking and cooking water and measure the salinity level 43 and electrical conductivity of household stored water. At each visit, we will also measure the 44 blood pressure of participants' ≥ 20 years of age and pregnant women and collect urine samples 45 for urinary sodium and protein measurements. We will use generalized linear mixed models to 46 determine the association of access to MAR water on blood pressure of the participants.

47
Ethics and dissemination: The study protocol has been reviewed and approved by the 48 Institutional Review Boards of the International Centre for Diarrheal Disease Research, 49 Bangladesh (icddr.b). Informed written consent will be taken from all the participants. This study 50 is funded by Wellcome Trust, UK. The study findings will be disseminated to the government 51 partners, at research conferences, and in peer-reviewed journals.  53 54 Strengths and limitation of the study 55 56 • This is the first study to evaluate the health impact of managed aquifer recharge in 57 southwest coastal Bangladesh. 58 • The stepped-wedge trial ensures we will have counterfactual data as well as gradual 59 access to MAR water in all communities. 60 • Objective measurement of exposure (drinking water salinity) and outcomes (urinary 61 sodium and blood pressure).

62
• The magnitude of exposure will vary geographically and across time period. Therefore, 63 MAR water salinity will differ across communities at a single point of time, and also for 64 the same community at different points of time.

65
• Compliance of the intervention may be different across sites and for individuals of 66 different professions. 67 68 Background and rationale: Saltwater intrusion and salinization have increased 70 groundwater salinity in many coastal aquifers and small islands across the world. [1][2][3][4][5] This is 71 driven by a number of climatological and anthropogenic factors including global warming, 72 increased cyclones and tidal surges, reduced river discharge, and increased groundwater 73 abstraction in excess of recharge. [6][7][8] Communities in many coastal regions rely on groundwater as 74 their main source of drinking water 9 as well as freshwater for domestic, agricultural and 75 industrial purposes. 10 Nearly half of the world's population resides in coastal areas 3 and 10% of 76 these live in low-lying coastal areas where surface elevation is generally <10 meters above mean 77 sea level. 11 Salinization in coastal areas is expected to increase in the future because of increased 78 groundwater withdrawal due to population and economic growth and sea-level rise. 3 As the 79 world's population and economic activities continue to grow, groundwater supplies are 80 progressively under threat of depletion, which increases the importance of monitoring, 81 management and conservation of coastal freshwater aquifers. 12 13 82 One approach to minimize the impact of groundwater salinization is to enhance 83 groundwater recharge into coastal aquifers. 14 Managed aquifer recharge (MAR) is an approach to 84 artificially promote freshwater recharge to increase storage. MAR involves infiltration of 85 freshwater (e.g. rainwater and pond water) into aquifers to create a store of freshwater within the 86 naturally brackish aquifer (Figure 1). [15][16][17] MAR represents a promising adaptive strategy for 87 increasing freshwater availability and sustaining a year-round drinking water supply that is 88 protected from evaporation, and could be resilient to tidal storms, cyclones and surface water 89 salinity since freshwater infiltration and storage occur under confined conditions. 18 90 tidal storms, and shrimp cultivation. [19][20][21][22] In the future, climate change and sea-level rise are 93 expected to cause more cyclones, tidal surges and flood in this region that will further affect 94 surface water and groundwater salinity. 7 In many areas in southwest coastal Bangladesh, both 95 shallow and deep aquifers contain naturally brackish water causing acute scarcity of drinking 96 water (Figure 2A). 23

115
Epidemiological studies have demonstrated that high sodium intake is associated with 116 elevated blood pressure [30][31][32] and other cardiovascular diseases. 33 A study conducted among the 117 adult population residing in southwest coastal Bangladesh suggest drinking saline water was 118 associated with high blood pressure after adjusting for personal, lifestyle and environmental 119 factors. 34 Among study participants, the mean systolic blood pressure for those consuming water 120 with sodium from the lowest quintile was 119.4 (SD 13.7) and from the highest sodium quintile 121 was 126.7 mm Hg (18·0). 34 Excess sodium intake from drinking brackish water has been also 122 associated increased gestational hypertension and preeclampsia among pregnant women in 123 southwest coastal Bangladesh. 35 36 Mean systolic and diastolic blood pressure were higher among 124 the pregnant women who drank tube well or pond (saline sources) water compared to those who 125 drank rain water. 35 37 The mean systolic blood pressure of pregnant women from these areas was 126 102.4 mm Hg among those who drank rainwater, 112.6 mm Hg among pond water users and 127 119.4 mm Hg among brackish groundwater users. 35 High blood pressure during pregnancy is 128 associated with high maternal mortality, and adverse pregnancy and fetal outcomes. 38

130
There are 37 million people living in the southwest coastal region and 20 million are 131 currently affected by drinking water salinity. 39 The estimated mean global sodium consumption 132 is 3.95 g per day (range 2.2 to 5.5 g per day), 40  because people don't use them. There is also considerable controversy on whether reducing 139 sodium intake improves health. 42 In addition the MAR system will also alter the intake of other 140 cations such as calcium and magnesium that may have health impact. [43][44][45][46] Because the cost of analysis of randomized trials also suggest modest reduction in salt intake for four or more weeks 151 causes significant reduction in blood pressure at the population-level. 50 51 The primary objective 152 of the study is to assess whether access to low-salinity MAR water can reduce blood pressure of 153 community members ≥20 years of age. Secondary objectives include whether access to MAR 154 water can reduces urinary sodium and total protein excretion. We will evaluate water salinity, 155 urinary sodium excretion and blood pressure so that we understand whether or not we achieved 156 our immediate targets along the causal pathway.  implemented and randomized at individual-or household-levels. Once a community will have 164 access to MAR water, it is difficult to withhold the access of MAR water for some households.

165
Therefore, we will conduct a cluster randomized trial where each community will be considered 166 as a cluster. The stepped wedge design allows communities to gradually have access to MAR 167 water; however, the point at which their access commences will be randomly assigned. 52 In this 168 way, each MAR site will contribute data for both the intervention and the control time periods. 169 We will have five monthly steps in the stepped wedge trial (Figure 3). In the first step, none of 170 the communities will have MAR water available for drinking and we will collect baseline 171 information. During each subsequent month, four randomly selected communities will receive 172 access to MAR water for drinking and cooking. In the last (fifth) monthly step, all the 173 communities will have access to MAR drinking water. year old hypertensive household members will be also eligible for participation but research staff and adolescence is associated adrenal and nervous system maturation that may contribute to salt 191 sensitivity. 56 57 They will approach households living near the MAR water access point. They computer-generated random numbers. Randomization will be conducted by an investigator who 203 will not be directly involved in implementation of the stepped wedge study and this will be done 204 before commencement of the study. The study could not be blinded therefore there was no 205  the brackish aquifer through the MAR systems will be ongoing for 1-2 years in each community, 207 people will not have access to MAR water until a formal handover of MAR systems to 208 community members. The implementers form a community management team who are 209 responsible for maintenance of each MAR system following handover. We have synchronized 210 the community handover with the randomization schedule in the 16 communities and our 211 community health promoters informed participants when the MAR water will be available for 212 consumption.

213
We will deploy a local trained promoter at each MAR site, who will visit these 214 households, list members and identify pregnant women in these and other households in 215 catchment areas of each MAR site. Promoters will encourage all household members to drink 216 MAR water exclusively while they are at home, carry a bottle of MAR water while they go out 217 for work and other activities, and to cook with MAR water. The community health promoters 218 will visit households with promotional materials (e.g. flip charts) from the beginning of the study 219 to inform household members about adverse health effects of drinking brackish water and 220 potential benefits of drinking low-salinity MAR water. As per the randomization schedule, they 221 will inform households when MAR water will be available for consumption. drank rainwater compared to those who drank brackish water 35 . The mean reduction in systolic 228 Hughes therefore sample size was 1396 x 3.04. We also inflated the total sample size considering 254 10% loss to follow-up. We then followed the approaches of determining the number of clusters 255 required for total sample size considering a fixed cluster size of 60. 62 We then calculated the 256 number of steps that we need to randomize by dividing total cluster by steps. We calculated that 257 16 MAR communities will be required for the study with four communities randomized to access 258 MAR water in each step ( Figure 3).

259
Data collection methods: The outcomes of interest will be measured from selected 260 participants of 16 communities irrespective of their access to MAR water during five monthly 261 visits. The interval between two successive blood pressure measurements or two consecutive 262 data collection visits will be at least one month for each participant. After taking informed 263 written consent from household heads and each participant, research staff will administer a 264 survey using a structured questionnaire to collect information on drinking water sources, 265 household members' socio-demographic status, medical history, family history of cardiovascular 266 diseases, medication intake and cardiovascular risk factors including smoking history and 267 alcohol intake during the first monthly visit. At this time they will also measure participants' 268 weight, height, waist and hip circumference and blood pressure, and collect urine samples.

304
(diastolic blood pressure plus one-third of systolic blood pressure) and high pulse pressure is 305 independently associated with increased left ventricular mass. 66 68 306 We will also measure the creatinine adjusted protein excretion of all participants as 307 tertiary outcomes ( Table 1). Proteinuria is a biomarker for future risk of cardiovascular 308 diseases, [71][72][73][74] and is associated with the pathogenesis of cardiovascular diseases, including 309 hypertension, 75 76 chronic kidney disease, 77 myocardial ischemia, 78 carotid artery thickness, 79 80 310 left ventricular hypertrophy, 81 82 hyperlipidemia, 83 atherosclerosis, 84 and coronary artery 311 calcification. 72 80 85 86 Reduced salt intake for four weeks has been associated with decreased 312 proteinuria in blinded randomized controlled trial. 87 Preeclampsia is associated with high 313 maternal mortality and adverse pregnancy and fetal outcomes 38  to measuring blood pressure. Participants will rest for 5 minutes sitting on a chair keeping their 323 arm supported. An appropriate sized cuff and calibrated instrument will be used for different age 324 groups and the blood pressure instrument will be positioned at heart level. Blood pressure will be 325 measured three times; first left arm, then right arm, then again left arm. Both systolic and 326 diastolic blood pressure will be recorded from all measurements. The arithmetic mean of three 327 systolic blood pressure measurements will be used as the primary outcome. However, if a 328 systolic blood pressure measurement differs by 10% from the other measurements, that 329 measurement will be excluded when calculating the arithmetic mean systolic blood pressure.

330
Biomarker measurements: Field research staff will instruct the participants' to collect 331 24 hours urine sample during each household visit. The volume of the urine samples will be 332 noted at household level, and a sample of 25 ml urine will be collected and transported to the 333 field laboratory at 2-8° C within 6 hours of collection for processing, analysis and storage.

334
Aliquots of each participant's urine sample will be made for biochemical and electrolyte 335 measurements. Urinary creatinine concentration will be measured by a colorimetric method 336 (Jaffe reaction) using a semi-auto biochemistry analyzer (Evolution 3000, BSI, Italy). Urinary 337 total protein will be estimated using a light sensitive colored reagent (Randox.UK). We will use 338 the direct Ion Selective Electrode (ISE) method for urinary sodium measurements using a semi-339 auto electrolyte analyzer (Biolyte2000, Bio-care Corporation, Taiwan). We will measure the uric 340 acid concentration in blood by an enzymatic colorimetric method (Evolution 3000, BSI, Italy). 341 We will perform routine Quality Control for all tests using standard quality control reagents 342 and data values will be programmed to minimize data entry errors. The dataset downloaded from 347 handheld devices will be cleaned and checked by the site investigators. All laboratory data will 348 be double entered. Data will be stored in icddr,b's data repository system, in compliance with the 349 system's requirements and will be publicly available after analyzing the primary result.

350
The research staff will be trained for identifying adverse events such as hypertension and 351 hypertensive disorders in pregnancy. They will report to the investigators following 352 identification of these patients and the investigators will assess whether these adverse events 353 need to be reported to icddr,b's Ethical Review Committee.

354
Statistical methods: We will conduct an intention-to-treat analysis for the primary 355 analysis. For the primary analysis we will assess whether access to MAR water reduce the 356 systolic blood pressure (continuous outcome) of the >20 years old community member. We will 357 use generalized linear mixed models with appropriate links for the primary analysis considering 358 random effects for community, households and participants, and a fixed effects of steps or visits. 359 We will adjust the effect of MAR systems on blood pressure for age, sex, weight and height, 360 personal, and socioeconomic factors ( Table 2). We will inspect the missing data patterns and 361 use multiple imputation with chained equations to jointly impute data on missing exposure and 362 confounders to preserve an unbiased association estimate if the data are missing at random 363 conditional on measured variables. We will also conduct subgroup analyses among the 364 households that adhere with the MAR intervention and that exclusively use MAR water. In gestation, we will include them in pregnant women analysis from the eighth week of gestation.

376
Ethics: Informed written consent will be taken from all participants and household heads.

377
Consent will be also taken for ancillary studies and future use of specimens collected from study 378 participants. This study protocol has been reviewed and approved by the icddr,b Ethical Review

379
Committee. Approval will be taken for any addition or modification of the protocol from icddr,b 380 Ethical Review Committee. If research staff members identify hypertensive patients or cases of 381 hypertension during pregnancy, they will refer patients to the local government health facilities 382 for further management. In addition, research staff will train pregnant women and family 383 members to recognize the danger signs during pregnancy, and will also instruct them where to 384 seek medical care if such danger signs appear. During monthly household visits research staff 385 will encourage pregnant women to attend prenatal visits. All dataset will be anonymous without 386 the personal identifiers and participants' privacy will be maintained during data storage, analysis 387 and dissemination. partner NGOs working in southwest coastal Bangladesh for access to safe drinking water. We 391 will discuss the scope and limitations of the MAR system to address the demand of safe drinking 392 water based on findings from this study. We will be submitting abstracts to international 393 conferences for dissemination to the international audience working on safe water. We will 394 develop manuscripts and submit to peer-reviewed journals to publish our research. This will be the first study to assess the health impact of an environmental intervention to 397 reduce groundwater salinity in southwest coastal Bangladesh. Our study has several strengths.

398
The stepped-wedge trial ensures that we will have counterfactual data as well as gradual access 399 to MAR water in all communities. In a stepped-wedge design, treatment effect of an intervention 400 can be estimated from between-and within-cluster comparisons as participants will act as their 401 own control, compared to only between-cluster comparison in a parallel cluster randomized 402 design. 61

403
We explored other study designs such as non-randomized study. the same season after satisfactory reduction in salinity. Therefore, a parallel group cluster 427 randomized trial was not suitable as we could not randomly select control sites and postpone the 428 community to drink water for the dry season. We believe that the stepped-wedge trial design was 429 appropriate to obtain counterfactual data as well as fulfilling the programmatic need. 430 We will use the same instruments for data collection and outcome measurements for all 431 steps for exposed and unexposed communities, which will mitigate bias in data collection.

432
Measurement of water source salinity and urinary sodium concentration will explain the 433 biologically plausible effect of drinking MAR water on blood pressure. All outcomes planned to 434 measure in the study are objective outcomes that will reduce the risk of reporting bias. Collection 435 of detailed exposure and co-variate data will help in determining a valid association between 436 drinking MAR water and health benefits. We will have several outcome variables and 437 biomarkers that will ensure a comprehensive health benefit evaluation of access to MAR water.

Purpose of the research
Hello (Assalamualaikum/Nomoshkar). My name is ________ and I work with the ICDDR,B (Cholera Hospital) in Dhaka. You are aware that a Managed aquifer recharge scheme has already been installed in your community that would like reduce the salinity problem in your drinking water. You may also know that drinking saline water is associated with hypertension. We will conduct a pilot study to understand the mean blood pressure of the household members who are drinking water from MAR.
Background (brief introduction of the issue and the need for/ importance of the research) Salinity in drinking water is a big problem in your community and in many other communities in the south-western coastal region of Bangladesh. Previous research has demonstrated that drinking saline water caused high blood pressure among the adult population and gestational hypertension among the pregnant women. Hypertension risk many other cardiovascular diseases and gestational hypertension also causes adverse pregnancy and fetal outcomes. Managed aquifer recharge (MAR) is a promising sustainable intervention that can reduce salinity in groundwater. Moreover, MAR will not be affected by the cyclone and tidal surges. There is a potential to scale-up MAR across the south-western coastal region if we can demonstrate that drinking MAR water provide health benefits and do not pose any risk to human health. Now we would like to conduct a pilot study where we will measure the mean blood pressure of your household members and the salt concentration of their urine.

Why invited to participate in the study?
You're invited to participate in our study because your household is located in such an area where a new MAR site has been installed. Rain water and pond water is being infiltrated into the groundwater aquifers in this MAR scheme to lower the groundwater salinity. Currently, your household may be using MAR water or other water sources for drinking and cooking purposes. Since your household is among one of the households under the MAR catchment area, we are inviting you to participate in this study.

Methods and procedures
If your household decides to participate, we make two visits to your households. The study duration will be around four months. We will collect information from all the available household members >20 years old. In each visit, we will measure the blood pressure of all the >20 years household members and may also measure their 24 hour ambulatory arterial blood pressure by strapping a device in their arms for next 24 hours. We will collect about 20 ml urine from all the household members as spot urine and will provide a large bag to collect the total amount of urine during the next 24 hours time. We will measure the total protein, creatinine, in urine that will provide important information about your kidney health. In the first visit, the research staff will collect the demographic and socio-economic information, smoking and diseases history, and measure height of available household members. We will also collect 5 ml blood from the participant to assess the genetic marker to understand whether their gene indicates that reduction of salt intake can reduce blood pressure of your family members. Samples will be stored for up to 5 years after the end of the study because we want to take advantage of expected future advances in characterizing genetic markers and other parameters relevant to the salt sensitivity of hypertension. At the end of this time, the samples will be destroyed and a certificate of their destruction will be provided to iccdr,b. If you wish to withdraw your samples from the study earlier you may contact the investigators.
In addition, we will collect information about your drinking water sources during each visit and will collect your stored drinking water samples for measuring salinity and sodium concentration in water.

Risk and benefits
There is no more than minimal risk involved in this study. Your household members may feel bit discomfort during blood pressure measurement, blood and urine sample collection, but these will provide important health messages about your household member's health. Hypertension is very prevalent in Bangladesh and it remains undetected to many people. By participating in this study, your household members will have a screening for hypertension.
There will be also no social harm to you and the rest of the households in this village by your participation in this study. During the visit day, your household need to spend time (~90 minutes) from your daily life for study purposes. There is no direct monetary compensation benefit for participating in this study. However, if drinking MAR water seems to provide health benefits, the study result will help to scale MAR across coastal Bangladesh that many people will have access to safe water with low salinity.

Privacy, anonymity and confidentiality
We will maintain the confidentiality of what information you give us and we will not disclose your identity when we write our reports. We will only use the collective information for the purpose of this study, and we will not use your name in sharing and publishing the results of this study. We expect the steps we take will keep all of your information confidential.

Future use of information
The information collected from this study may be shared with other researchers if needed, but we will strictly maintain your confidentiality and privacy. In the future, we may wish to perform additional tests on the urine and blood sample that will be collected. The samples will be stored at icddr,b.

Right not to participate and withdraw
Taking part in the study is completely voluntary. You may choose not to answer any or all of the questions that will be asked about your household or your behavior. You can drop out of this

Principle of compensation
You need not to pay us to take part in this study, and similarly we will not pay you money for attending in the study. You will also be provided best possible, free treatment, for research related injuries.
[Note: Payment for loss of earning of the study participants may be considered in case the participants require extended hospitalization or confinement only for the purpose of the research and/or reimbursement of cost of transportation for participating in the study. However, the amount should be equivalent to the loss and not so high that the offer might induce (influence/bias judgment) participation in the study.

Purpose of the research
Hello (Assalamualaikum/Nomoshkar). My name is ________ and I work with the ICDDR,B (Cholera Hospital) in Dhaka. You are aware that a Managed aquifer recharge scheme has already been installed in your community that would like reduce the salinity problem in your drinking water. You may also know that drinking saline water is associated with hypertension. We will conduct a pilot study to understand the mean blood pressure of the household members who are drinking water from MAR .
Background (brief introduction of the issue and the need for/ importance of the research) Salinity in drinking water is a big problem in your community and in many other communities in the south-western coastal region of Bangladesh. Previous research has demonstrated that drinking saline water caused high blood pressure among the adult population and gestational hypertension among the pregnant women. Hypertension risk many other cardiovascular diseases and gestational hypertension also causes adverse pregnancy and fetal outcomes. Managed aquifer recharge (MAR) is a promising sustainable intervention that can reduce salinity in groundwater. Moreover, MAR will not be affected by the cyclone and tidal surges. There is a potential to scale-up MAR across the south-western coastal region if we can demonstrate that drinking MAR water provide health benefits and do not pose any risk to human health. Now we would like to conduct a pilot study where we will measure your mean blood pressure and the salt concentration of urine.

Why invited to participate in the study?
You're invited to participate in our study because your household is located in such an area where a new MAR site has been installed. Rain water and pond water is being infiltrated into the groundwater aquifers in this MAR scheme to lower the groundwater salinity. Currently, you may be using MAR water or other water sources for drinking and cooking purposes. We are inviting all the household members >20 years old in this MAR scheme catchment area to routinely measure blood pressure and collect urine samples. Since your age is >20 (or 20) years, your participation is very important for the study.

Methods and procedures
If you decide to participate, we will make two visits to measure your blood pressure, collect urine and blood. We may also measure your 24 hour ambulatory arterial blood pressure by strapping a device in your arm for 24 hours. We will collect about 20 ml urine as spot urine and will provide a large bag to collect the total amount of urine during the 24 hours time. We will measure the total protein, creatinine, in urine that will provide important information about your kidney health. In the first visit, the research staff will collect the demographic and socioeconomic information, smoking and diseases history, and measure height of available household members. We will also collect 5 ml of your blood to assess the genetic marker to understand whether your gene indicates that reduction of salt intake can reduce blood pressure of your family members.
So, if you agree, we may also invite you for some further discussions. We would like to record this sessions in a tape recorder as it is very difficult to write all the conversation in a note pad. We would also like to take some notes from this session. This session will last about an hour and half.

Risk and benefits
There is no more than minimal risk involved in this study. You may feel bit discomfort during blood pressure measurement, and urine sample collection, but these will provide important health messages about your health. Hypertension is very prevalent in Bangladesh and it remains undetected to many people. By participating in this study, you will be screened for hypertension.
There will be also no social harm to you and the rest of the households in this village by your participation in this study. During the visit day, you may need to spend time (~90 minutes) from your daily life for study purposes. There is no direct monetary compensation benefit for participating in this study. However, if drinking MAR water seems to provide health benefits, the study result will help to scale MAR across coastal Bangladesh that many people will have access to safe water with low salinity.

Privacy, anonymity and confidentiality
We will maintain the confidentiality of what information you give us and we will not disclose your identity when we write our reports. We will only use the collective information for the purpose of this study, and we will not use your name in sharing and publishing the results of this study. We expect the steps we take will keep all of your information confidential.

Future use of information
The information collected from this study may be shared with other researchers if needed, but we will strictly maintain your confidentiality and privacy. In the future, we may wish to perform additional tests on the urine sample that will be collected. The samples will be stored at icddr,b.

Right not to participate and withdraw
Taking part in the study is completely voluntary. You may choose not to answer any or all of the questions that will be asked about your household or your behavior. You can drop out of this study at any time during the interview or anytime during the study period. You have the right to refuse participation in this study. You need not to pay us to take part in this study, and similarly we will not pay you money for attending in the study. You will also be provided best possible, free treatment, for research related injuries.

Principle of compensation
[Note: Payment for loss of earning of the study participants may be considered in case the participants require extended hospitalization or confinement only for the purpose of the research and/or reimbursement of cost of transportation for participating in the study. However, the amount should be equivalent to the loss and not so high that the offer might induce (influence/bias judgment) participation in the study.

Methods: Assignment of interventions (for controlled trials)
Allocation: Sequence generation 16a Method of generating the allocation sequence (eg, computer-generated random numbers), and list of any factors for stratification. To reduce predictability of a random sequence, details of any planned restriction (eg, blocking) should be provided in a separate document that is unavailable to those who enrol participants or assign interventions Page 9 Allocation concealment mechanism 16b Mechanism of implementing the allocation sequence (eg, central telephone; sequentially numbered, opaque, sealed envelopes), describing any steps to conceal the sequence until interventions are assigned NA Implementation 16c Who will generate the allocation sequence, who will enrol participants, and who will assign participants to interventions Page 9 Blinding (masking) 17a Who will be blinded after assignment to interventions (eg, trial participants, care providers, outcome assessors, data analysts), and how NA 17b If blinded, circumstances under which unblinding is permissible, and procedure for revealing a participant's allocated intervention during the trial NA

Methods and analysis:
The study employs a stepped-wedge cluster-randomized controlled 41 community trial design in 16 communities over five monthly visits. During each visit, we will 42 collect data on participants' source of drinking and cooking water and measure the salinity level 43 and electrical conductivity of household stored water. At each visit, we will also measure the 44 blood pressure of participants' ≥ 20 years of age and pregnant women and collect urine samples 45 for urinary sodium and protein measurements. We will use generalized linear mixed models to 46 determine the association of access to MAR water on blood pressure of the participants.

Ethics and dissemination:
The study protocol has been reviewed and approved by the 48 Institutional Review Boards of the International Centre for Diarrheal Disease Research, 49 Bangladesh (icddr.b). Informed written consent will be taken from all the participants. This study 50 is funded by Wellcome Trust, UK. The study findings will be disseminated to the government 51 partners, at research conferences, and in peer-reviewed journals.  53 54 Strengths and limitation of the study 55 56 • This is the first study to evaluate the health impact of managed aquifer recharge in 57 southwest coastal Bangladesh. 58 • The stepped-wedge trial ensures we will have counterfactual data as well as gradual 59 access to MAR water in all communities. 60 • Objective measurement of exposure (drinking water salinity) and outcomes (urinary 61 sodium and blood pressure).

62
• The magnitude of exposure will vary geographically and across time period. Therefore, 63 MAR water salinity will differ across communities at a single point of time, and also for 64 the same community at different points of time. 65 • Compliance of the intervention may be different across sites and for individuals of 66 different socioeconomic status. 67 68 Background and rationale: Saltwater intrusion and salinization have increased 70 groundwater salinity in many coastal aquifers and small islands across the world. [1][2][3][4][5] This is 71 driven by a number of climatological and anthropogenic factors including global warming, 72 increased cyclones and tidal surges, reduced river discharge, and increased groundwater 73 abstraction in excess of recharge. [6][7][8] Communities in many coastal regions rely on groundwater as 74 their main source of drinking water 9 as well as freshwater for domestic, agricultural and 75 industrial purposes. 10 Nearly half of the world's population resides in coastal areas 3 and 10% of 76 these live in low-lying coastal areas where surface elevation is generally <10 meters above mean 77 sea level. 11 Salinization in coastal areas is expected to increase in the future because of increased 78 groundwater withdrawal due to population and economic growth and sea-level rise. 3 As the 79 world's population and economic activities continue to grow, groundwater supplies are 80 progressively under threat of depletion, which increases the importance of monitoring, 81 management and conservation of coastal freshwater aquifers. 12 13 82 One approach to minimize the impact of groundwater salinization is to enhance 83 groundwater recharge into coastal aquifers. 14 Managed aquifer recharge (MAR) is an approach to 84 artificially promote freshwater recharge to increase storage. MAR involves infiltration of 85 freshwater (e.g. rainwater and pond water) into aquifers to create a store of freshwater within the 86 naturally brackish aquifer (Figure 1). [15][16][17] MAR represents a promising adaptive strategy for 87 increasing freshwater availability and sustaining a year-round drinking water supply that is 88 protected from evaporation, and could be resilient to tidal storms, cyclones and surface water 89 salinity since freshwater infiltration and storage occur under confined conditions. 18 90  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  tidal storms, and shrimp cultivation. [19][20][21][22] In the future, climate change and sea-level rise are 93 expected to cause more cyclones, tidal surges and flood in this region that will further affect 94 surface water and groundwater salinity. 7 In many areas in southwest coastal Bangladesh, both 95 shallow and deep aquifers contain naturally brackish water causing acute scarcity of drinking 96 water (Figure 2A). 23 24 Water salinity in southwest coastal Bangladesh follows a clear seasonal  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59

115
Epidemiological studies have demonstrated that high sodium intake is associated with 116 elevated blood pressure [30][31][32] and other cardiovascular diseases. 33 A study conducted among the 117 adult population residing in southwest coastal Bangladesh suggest drinking saline water was 118 associated with high blood pressure after adjusting for personal, lifestyle and environmental 119 factors. 34 Among study participants, the mean systolic blood pressure for those consuming water 120 with sodium from the lowest quintile was 119.4 (SD 13.7) and from the highest sodium quintile 121 was 126.7 mm Hg (18·0). 34 Excess sodium intake from drinking brackish water has been also 122 associated increased gestational hypertension and preeclampsia among pregnant women in 123 southwest coastal Bangladesh. 35 36 Mean systolic and diastolic blood pressure were higher among 124 the pregnant women who drank tube well or pond (saline sources) water compared to those who 125 drank rain water. 35 37 The mean systolic blood pressure of pregnant women from these areas was 126 102.4 mm Hg among those who drank rainwater, 112.6 mm Hg among pond water users and 127 119.4 mm Hg among brackish groundwater users. 35 High blood pressure during pregnancy is 128 associated with high maternal mortality, and adverse pregnancy and fetal outcomes. 38

130
There are 37 million people living in the southwest coastal region and 20 million are 131 currently affected by drinking water salinity. 39 The estimated mean global sodium consumption 132 is 3.95 g per day (range 2.2 to 5.5 g per day), 40  because people do not use them. There is also considerable controversy on whether reducing 139 sodium intake improves health. 42 In addition the MAR system will also alter the intake of other 140 cations such as calcium and magnesium that may have health impact. [43][44][45][46] Because the cost of analysis of randomized trials also suggest modest reduction in salt intake for four or more weeks 151 causes significant reduction in blood pressure at the population-level. 50 51 The primary objective 152 of the study is to assess whether access to low-salinity MAR water can reduce blood pressure of 153 community members ≥20 years of age. Secondary objectives include whether access to MAR 154 water can reduces urinary sodium and total protein excretion. We will evaluate water salinity, 155 urinary sodium excretion and blood pressure so that we understand whether or not we achieved 156 our immediate targets along the causal pathway. implemented and randomized at individual-or household-levels. Once a community will have 164 access to MAR water, it is difficult to withhold the access of MAR water for some households.

165
Therefore, we will conduct a cluster randomized trial where each community will be considered 166 as a cluster. The stepped wedge design allows communities to gradually have access to MAR 167 water; however, the point at which their access commences will be randomly assigned. 52 In this 168 way, each MAR site will contribute data for both the intervention and the control time periods. 169 We will have five monthly steps in the stepped wedge trial (Figure 3). In the first step, none of 170 the communities will have MAR water available for drinking and we will collect baseline 171 information. During each subsequent month, four randomly selected communities will receive 172 access to MAR water for drinking and cooking. In the last (fifth) monthly step, all the 173 communities will have access to MAR drinking water.

216
Agreement of the implementers and gatekeepers was sought at the beginning of the study during 217 site selection for inclusion of any site in trial and access of MAR water as per the randomization. 218 We will deploy a local trained promoter at each MAR site, who will visit these 219 households, list members and identify pregnant women in these and other households in 220 catchment areas of each MAR site. Promoters will encourage all household members to drink 221 MAR water exclusively while they are at home, carry a bottle of MAR water while they go out 222 for work and other activities, and to cook with MAR water. The community health promoters 223 will visit households with promotional materials (e.g. flip charts) from the beginning of the study 224 to inform household members about adverse health effects of drinking brackish water and 225 potential benefits of drinking low-salinity MAR water. As per the randomization schedule, they 226 will inform households when MAR water will be available for consumption. Hughes therefore sample size was 1396 x 3.04. We also inflated the total sample size considering 259 10% loss to follow-up. We then followed the approaches of determining the number of clusters 260 required for total sample size considering a fixed cluster size of 60. 62 We then calculated the 261 number of steps that we need to randomize by dividing total cluster by steps. We calculated that 262 16 MAR communities will be required for the study with four communities randomized to access 263 MAR water in each step (Figure 3). proxy for sodium intake. 63 64 295 Raised diastolic blood pressure is associated with aortic and thoracic aneurysm. 70 Some effects 303 of high sodium intake are independent of high systolic and diastolic blood pressure such as 304 arterial stiffness and left ventricular mass ─ both of these are independent predictor of future 305 cardiovascular diseases. High pulse pressure (difference between systolic and diastolic blood 306 pressure) is associated with increased arterial stiffness, 66 68 and high mean arterial pressure 307 (diastolic blood pressure plus one-third of systolic blood pressure) and high pulse pressure is 308 independently associated with increased left ventricular mass. 66 68

309
We will also measure the creatinine adjusted protein excretion of all participants as 310 tertiary outcomes ( Table 1). Proteinuria is a biomarker for future risk of cardiovascular 311 diseases, [71][72][73][74] and is associated with the pathogenesis of cardiovascular diseases, including 312 hypertension, 75 76 chronic kidney disease, 77 myocardial ischemia, 78 carotid artery thickness, 79 80 313 left ventricular hypertrophy, 81 82 hyperlipidemia, 83 atherosclerosis, 84 and coronary artery 314 calcification. 72 80 85 86 Reduced salt intake for four weeks has been associated with decreased 315 proteinuria in blinded randomized controlled trial. 87 Preeclampsia is associated with high 316 maternal mortality and adverse pregnancy and fetal outcomes 38  to measuring blood pressure. Participants will rest for 5 minutes sitting on a chair keeping their 326 arm supported. An appropriate sized cuff and calibrated instrument will be used for different age 327 groups and the blood pressure instrument will be positioned at heart level. Blood pressure will be 328 measured three times; first left arm, then right arm, then again left arm. Both systolic and 329 diastolic blood pressure will be recorded from all measurements. The arithmetic mean of three 330 systolic blood pressure measurements will be used as the primary outcome. However, if a 331 systolic blood pressure measurement differs by 10% from the other measurements, that 332 measurement will be excluded when calculating the arithmetic mean systolic blood pressure.

333
Biomarker measurements: Field research staff will instruct the participants' to collect 334 24 hours urine sample during each household visit. The volume of the urine samples will be 335 noted at household level, and a sample of 25 ml urine will be collected and transported to the 336 field laboratory at 2-8° C within 6 hours of collection for processing, analysis and storage.

337
Aliquots of each participant's urine sample will be made for biochemical and electrolyte 338 measurements. Urinary creatinine concentration will be measured by a colorimetric method 339 (Jaffe reaction) using a semi-auto biochemistry analyzer (Evolution 3000, BSI, Italy). Urinary 340 total protein will be estimated using a light sensitive colored reagent (Randox.UK). We will use 341 and data values will be programmed to minimize data entry errors. The dataset downloaded from 350 handheld devices will be cleaned and checked by the site investigators. All laboratory data will 351 be double entered. Data will be stored in icddr,b's data repository system, in compliance with the 352 system's requirements and will be publicly available after analyzing the primary result.

353
The research staff will be trained for identifying adverse events such as hypertension and 354 hypertensive disorders in pregnancy. They will report to the investigators following 355 identification of these patients and the investigators will assess whether these adverse events 356 need to be reported to icddr,b's Ethical Review Committee.

357
Statistical methods: We will conduct an intention-to-treat analysis for the primary 358 analysis. For the primary analysis we will assess whether access to MAR water reduce the 359 systolic blood pressure (continuous outcome) of the ≥20 years old participants. Pregnant women 360 will be included in the primary analysis because it is likely few pregnant women will be 361 identified in the 16 communities and separate analysis of pregnant women will be underpowered. 362 We will use generalized linear mixed models with appropriate links for the primary analysis 363 considering random effects for community, households and participants, and a fixed effects of 364 for health effects following receiving MAR intervention in secondary analysis.

372
Ethics: Informed written consent will be taken from all participants and household heads.

373
Consent will be also taken for ancillary studies and future use of specimens collected from study 374 participants. This study protocol has been reviewed and approved by the icddr,b Ethical Review

375
Committee. Approval will be taken for any addition or modification of the protocol from icddr,b 376 Ethical Review Committee. If research staff members identify hypertensive patients or cases of 377 hypertension during pregnancy, they will refer patients to the local government health facilities 378 for further management. In addition, research staff will train pregnant women and family 379 members to recognize the danger signs during pregnancy, and will also instruct them where to 380 seek medical care if such danger signs appear. During monthly household visits research staff 381 will encourage pregnant women to attend prenatal visits. All dataset will be anonymous without 382 the personal identifiers and participants' privacy will be maintained during data storage, analysis 383 and dissemination. water based on findings from this study. We will be submitting abstracts to international 389 conferences for dissemination to the international audience working on safe water. We will 390 develop manuscripts and submit to peer-reviewed journals to publish our research. This will be the first study to assess the health impact of an environmental intervention to 393 reduce groundwater salinity in southwest coastal Bangladesh. Our study has several strengths.

394
The stepped-wedge trial ensures that we will have counterfactual data as well as gradual access 395 to MAR water in all communities. In a stepped-wedge design, treatment effect of an intervention 396 can be estimated from between-and within-cluster comparisons as participants will act as their 397 own control, compared to only between-cluster comparison in a parallel cluster randomized 398 design. 61

399
We explored other study designs such as non-randomized study. the same season after satisfactory reduction in salinity. Therefore, a parallel group cluster 423 randomized trial was not suitable as we could not randomly select control sites and postpone the 424 community to drink water for the dry season. We believe that the stepped-wedge trial design was 425 appropriate to obtain counterfactual data as well as fulfilling the programmatic need. 426 We will use the same instruments for data collection and outcome measurements for all 427 steps for exposed and unexposed communities, which will mitigate bias in data collection.

428
Measurement of water source salinity and urinary sodium concentration will explain the 429 biologically plausible effect of drinking MAR water on blood pressure. All outcomes planned to 430 measure in the study are objective outcomes that will reduce the risk of reporting bias. Collection 431 of detailed exposure and co-variate data will help in determining a valid association between 432 drinking MAR water and health benefits. We will have several outcome variables and 433 biomarkers that will ensure a comprehensive health benefit evaluation of access to MAR water.

Methods: Assignment of interventions (for controlled trials)
Allocation: Sequence generation 16a Method of generating the allocation sequence (eg, computer-generated random numbers), and list of any factors for stratification. To reduce predictability of a random sequence, details of any planned restriction (eg, blocking) should be provided in a separate document that is unavailable to those who enrol participants or assign interventions Page 9 Allocation concealment mechanism 16b Mechanism of implementing the allocation sequence (eg, central telephone; sequentially numbered, opaque, sealed envelopes), describing any steps to conceal the sequence until interventions are assigned NA Implementation 16c Who will generate the allocation sequence, who will enrol participants, and who will assign participants to interventions  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  Biological  specimens   33 Plans for collection, laboratory evaluation, and storage of biological specimens for genetic or molecular analysis in the current trial and for future use in ancillary studies, if applicable NA *It is strongly recommended that this checklist be read in conjunction with the SPIRIT 2013 Explanation & Elaboration for important clarification on the items. Amendments to the protocol should be tracked and dated. The SPIRIT checklist is copyrighted by the SPIRIT Group under the Creative Commons "Attribution-NonCommercial-NoDerivs 3.0 Unported" license.

Purpose of the research
Hello (Assalamualaikum/Nomoshkar). My name is ________ and I work with the ICDDR,B (Cholera Hospital) in Dhaka. You are aware that a Managed aquifer recharge scheme has already been installed in your community that would like reduce the salinity problem in your drinking water. You may also know that drinking saline water is associated with hypertension. We will conduct a pilot study to understand the mean blood pressure of the household members who are drinking water from MAR.
Background (brief introduction of the issue and the need for/ importance of the research) Salinity in drinking water is a big problem in your community and in many other communities in the south-western coastal region of Bangladesh. Previous research has demonstrated that drinking saline water caused high blood pressure among the adult population and gestational hypertension among the pregnant women. Hypertension risk many other cardiovascular diseases and gestational hypertension also causes adverse pregnancy and fetal outcomes. Managed aquifer recharge (MAR) is a promising sustainable intervention that can reduce salinity in groundwater. Moreover, MAR will not be affected by the cyclone and tidal surges. There is a potential to scale-up MAR across the south-western coastal region if we can demonstrate that drinking MAR water provide health benefits and do not pose any risk to human health. Now we would like to conduct a pilot study where we will measure the mean blood pressure of your household members and the salt concentration of their urine.

Why invited to participate in the study?
You're invited to participate in our study because your household is located in such an area where a new MAR site has been installed. Rain water and pond water is being infiltrated into the groundwater aquifers in this MAR scheme to lower the groundwater salinity. Currently, your household may be using MAR water or other water sources for drinking and cooking purposes. Since your household is among one of the households under the MAR catchment area, we are inviting you to participate in this study.

Methods and procedures
If your household decides to participate, we make two visits to your households. The study duration will be around four months. We will collect information from all the available household members >20 years old. In each visit, we will measure the blood pressure of all the >20 years household members and may also measure their 24 hour ambulatory arterial blood pressure by strapping a device in their arms for next 24 hours. We will collect about 20 ml urine  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  from all the household members as spot urine and will provide a large bag to collect the total amount of urine during the next 24 hours time. We will measure the total protein, creatinine, in urine that will provide important information about your kidney health. In the first visit, the research staff will collect the demographic and socio-economic information, smoking and diseases history, and measure height of available household members. We will also collect 5 ml blood from the participant to assess the genetic marker to understand whether their gene indicates that reduction of salt intake can reduce blood pressure of your family members. Samples will be stored for up to 5 years after the end of the study because we want to take advantage of expected future advances in characterizing genetic markers and other parameters relevant to the salt sensitivity of hypertension. At the end of this time, the samples will be destroyed and a certificate of their destruction will be provided to iccdr,b. If you wish to withdraw your samples from the study earlier you may contact the investigators.
In addition, we will collect information about your drinking water sources during each visit and will collect your stored drinking water samples for measuring salinity and sodium concentration in water.

Risk and benefits
There is no more than minimal risk involved in this study. Your household members may feel bit discomfort during blood pressure measurement, blood and urine sample collection, but these will provide important health messages about your household member's health. Hypertension is very prevalent in Bangladesh and it remains undetected to many people. By participating in this study, your household members will have a screening for hypertension.
There will be also no social harm to you and the rest of the households in this village by your participation in this study. During the visit day, your household need to spend time (~90 minutes) from your daily life for study purposes. There is no direct monetary compensation benefit for participating in this study. However, if drinking MAR water seems to provide health benefits, the study result will help to scale MAR across coastal Bangladesh that many people will have access to safe water with low salinity.

Privacy, anonymity and confidentiality
We will maintain the confidentiality of what information you give us and we will not disclose your identity when we write our reports. We will only use the collective information for the purpose of this study, and we will not use your name in sharing and publishing the results of this study. We expect the steps we take will keep all of your information confidential.

Future use of information
The information collected from this study may be shared with other researchers if needed, but we will strictly maintain your confidentiality and privacy. In the future, we may wish to perform additional tests on the urine and blood sample that will be collected. The samples will be stored at icddr,b.

Right not to participate and withdraw
Taking part in the study is completely voluntary. You may choose not to answer any or all of the questions that will be asked about your household or your behavior. You can drop out of this

Principle of compensation
You need not to pay us to take part in this study, and similarly we will not pay you money for attending in the study. You will also be provided best possible, free treatment, for research related injuries.
[Note: Payment for loss of earning of the study participants may be considered in case the participants require extended hospitalization or confinement only for the purpose of the research and/or reimbursement of cost of transportation for participating in the study. However, the amount should be equivalent to the loss and not so high that the offer might induce (influence/bias judgment) participation in the study.

Purpose of the research
Hello (Assalamualaikum/Nomoshkar). My name is ________ and I work with the ICDDR,B (Cholera Hospital) in Dhaka. You are aware that a Managed aquifer recharge scheme has already been installed in your community that would like reduce the salinity problem in your drinking water. You may also know that drinking saline water is associated with hypertension. We will conduct a pilot study to understand the mean blood pressure of the household members who are drinking water from MAR .
Background (brief introduction of the issue and the need for/ importance of the research) Salinity in drinking water is a big problem in your community and in many other communities in the south-western coastal region of Bangladesh. Previous research has demonstrated that drinking saline water caused high blood pressure among the adult population and gestational hypertension among the pregnant women. Hypertension risk many other cardiovascular diseases and gestational hypertension also causes adverse pregnancy and fetal outcomes. Managed aquifer recharge (MAR) is a promising sustainable intervention that can reduce salinity in groundwater. Moreover, MAR will not be affected by the cyclone and tidal surges. There is a potential to scale-up MAR across the south-western coastal region if we can demonstrate that drinking MAR water provide health benefits and do not pose any risk to human health. Now we would like to conduct a pilot study where we will measure your mean blood pressure and the salt concentration of urine.

Why invited to participate in the study?
You're invited to participate in our study because your household is located in such an area where a new MAR site has been installed. Rain water and pond water is being infiltrated into the groundwater aquifers in this MAR scheme to lower the groundwater salinity. Currently, you may be using MAR water or other water sources for drinking and cooking purposes. We are inviting all the household members >20 years old in this MAR scheme catchment area to routinely measure blood pressure and collect urine samples. Since your age is >20 (or 20) years, your participation is very important for the study.

Methods and procedures
If you decide to participate, we will make two visits to measure your blood pressure, collect urine and blood. We may also measure your 24 hour ambulatory arterial blood pressure by strapping a device in your arm for 24 hours. We will collect about 20 ml urine as spot urine and  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  will provide a large bag to collect the total amount of urine during the 24 hours time. We will measure the total protein, creatinine, in urine that will provide important information about your kidney health. In the first visit, the research staff will collect the demographic and socioeconomic information, smoking and diseases history, and measure height of available household members. We will also collect 5 ml of your blood to assess the genetic marker to understand whether your gene indicates that reduction of salt intake can reduce blood pressure of your family members.
So, if you agree, we may also invite you for some further discussions. We would like to record this sessions in a tape recorder as it is very difficult to write all the conversation in a note pad. We would also like to take some notes from this session. This session will last about an hour and half.

Risk and benefits
There is no more than minimal risk involved in this study. You may feel bit discomfort during blood pressure measurement, and urine sample collection, but these will provide important health messages about your health. Hypertension is very prevalent in Bangladesh and it remains undetected to many people. By participating in this study, you will be screened for hypertension.
There will be also no social harm to you and the rest of the households in this village by your participation in this study. During the visit day, you may need to spend time (~90 minutes) from your daily life for study purposes. There is no direct monetary compensation benefit for participating in this study. However, if drinking MAR water seems to provide health benefits, the study result will help to scale MAR across coastal Bangladesh that many people will have access to safe water with low salinity.

Privacy, anonymity and confidentiality
We will maintain the confidentiality of what information you give us and we will not disclose your identity when we write our reports. We will only use the collective information for the purpose of this study, and we will not use your name in sharing and publishing the results of this study. We expect the steps we take will keep all of your information confidential.

Future use of information
The information collected from this study may be shared with other researchers if needed, but we will strictly maintain your confidentiality and privacy. In the future, we may wish to perform additional tests on the urine sample that will be collected. The samples will be stored at icddr,b.

Right not to participate and withdraw
Taking part in the study is completely voluntary. You may choose not to answer any or all of the questions that will be asked about your household or your behavior. You can drop out of this study at any time during the interview or anytime during the study period. You have the right to refuse participation in this study.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  You need not to pay us to take part in this study, and similarly we will not pay you money for attending in the study. You will also be provided best possible, free treatment, for research related injuries.