The last century witnessed an important shift in the pattern of diseases worldwide. While developed countries have managed to bring about a signi- ficant reduction in incidence and mortality associated with infectious agents, chronic disorders have be- come the leading cause of illness, though for some of them their causes are still largely unknown at present. Modern lifestyles and industrial development have im- proved the quality of life but have also exposed the general population to a variety of air-, water- and food-borne toxic substances. Within the complex causal frame pro- posed for chronic disorders, there is increasing evidence of environmental pollutants being involved in the occu- rrence of these diseases1–6.

Although ambient pollutant levels are generally low -and indeed much lower than those found in specific work environments- their health impact can nonethe- less be quite relevant from a public health perspective, since exposure may affect the whole population during long periods of life. The increasing importance of en- vironmental contaminants has led to a great number of studies targeted at improving current knowledge on pollutant exposure levels among humans and their health effects. Since the 1990s, some countries have imple- mented surveillance systems using biomarkers to me- asure exposure to different pollutants in the general population, with the National Health and Nutrition Examination Surveys (NHANES), conducted by the National Center for Health Statistics7 in the USA, and the German Environmental Survey (GerES)8 in Germany being considered paradigms of this type of studies.

Currently, human biomonitoring is becoming one of the main priorities of the European Union. The deve- lopment of a coherent approach to biomonitoring in Eu- rope is specifically included within the actions of the European Environment and Health Action Plan for 2004109. Both European an American Health Authori- ties10,11 have stressed the importance of paying spe- cial attention to pollution levels in susceptible populations, and more specifically, in pregnant women and children. According to the US Agency for Toxic Substances and Disease Registry12, there are critical periods of structural and functional development during both pre- and post-natal life which make children more susceptible than adults to hazardous substances, though for most pollutants these periods are still unknown. Children may also differ from adults in their ca- pacity to repair damage from chemical insults and have a longer lifetime in which to express the deleterious con- sequences of such exposures.

In Spain, environmental surveillance has mainly re- lied on measures of selected pollutants in air, water, food and soil. In the last decade, however, several initiatives have been implemented with the aim of including human data in the framework of environmental research in different areas of the country13–18. One of the most recent examples of studies including biological measures is the Spanish Childhood and Environment (Infancia y Medio Ambiente [INMA]) study, a multicenter cohort project with 3100 pregnant women, designed to investigate the ef- fect of the most important pollutants during pregnancy and the first years of life19.

Madrid city and its metropolitan belt constitute the biggest urban area in Spain, with similar environmen- tal problems than other big western cities. Vallecas Health District presents some of the worst socioeconomic and health indicators within the region. In 2000 the Regional Government approved an important investment plan to promote the development of this District, which included specific Public Health objectives.

In this context, the Public Health Authority of the Madrid Autonomous Region, which had decided to promote biomarker-based assessment of exposure to specific en- vironmental pollutants in the general population, selected Vallecas Health District as one of the pilot study areas. The primary objective of this initiative was to design a surveillance system focused on newborns and their environment. Accordingly, a study was designed to study the feasibility of implementing a biomonitoring surveillance system in the region. The aims of that explora- tory research were to ascertain the population's exposure status using biomarkers; to test the selected participant enrolment strategy; to measure a wide range of pollutants, for the purpose of selecting a subgroup to be monitored in the proposed surveillance system; to obtain samples of several substrates, in order to choo- se those most suitable for use at a population level; and to obtain reference levels in the study areas. This paper reports the design, response rates and field work re- sults, and provides a description of the study population.

This is a descriptive study intended to enhance current knowledge of the pollution in the environment su- rrounding newborns. For that reason, it was decided that pregnant women at approximately 8 months' gestation and their partners would be included in the study population. In order to participate, subjects had to agree to answer an epidemiologic questionnaire and provide biologic specimens taken from themselves and the child, thereby constituting so-called «mother-father-newborn triplets».

Initially, this study was set out to recruit 100 complete mother-father-newborn triplets. To fulfill this ob- jective, however, the number of couples to be enrolled was increased by 50% (150), since there was a certain risk of losing mothers at delivery and, by extension, new- born participants.

Furthermore, we decided to include the two main types of urban environment in which most of the regional population lived. To this end, half the sample was thus recruited from the Madrid City district of Vallecas and the other half was drawn from the Greater Madrid metropolitan belt (Parla and Getafe health districts). Some basic sociodemographic features of these two areas are shown in table 1. In this paper, most data are reported for the study sample overall.

The research team presented the study to hospital managers, pediatricians, gynecologists, and midwives at public hospitals in the selected areas, as well as to the primary-care departments of local health authori- ties and primary-care midwives.

It was decided that women were to be invited to participate in the study by the primary-care midwives in char- ge of pregnancy classes. The main reasons for opting for such an approach were: a) the desire to ensure that the research project was incorporated within the insti- tutional health framework, so that the data-extraction schedule for this study could be coordinated with the standard follow-up of pregnant women, and b) the im- portance of having a person who was trusted by the women, involved in the enrollment process.

Training sessions were held at which midwives le- arned about the different steps of the designated field- work protocol and were provided with the basic material needed to fulfill the role that they were expected to play.

All pregnant women in their eighth month of pregnancy, residing in the selected areas and attending child- birth preparation classes in the public health care system were invited to participate, until the designated sample size had been attained. Midwives outlined the study goals and requirements and provided informative brochures. All couples willing to participate and mee- ting the inclusion criteria were recruited. Both parents were required to give their prior written informed con- sent to participate. No financial rewards were offered, but the social utility of the study was explained.

No exclusions were made in terms of parents' or newborns' disease, race or place of origin. Inclusion criteria were: to have resided in one of the study areas for a minimum of one year; and not to have reported re- ceiving a blood transfusion during the previous year. For logistic reasons, we also required women to be aged over 15 years, had single pregnancies, and intended to deliver their babies at the local public hospital.

After a small pilot trial, recruitment started in Octo- ber 2003 and lasted until May 2004. In all, 630 women in 35 sessions of delivery-preparation courses were in- formed about the project and invited to participate. Of those eligible, 52% stated that they were willing to take part in the study and invite their partners to collabor- ate. Finally, 149 couples (29%) were enrolled (table 2). Two of these women gave birth before the first sche- duled date. Additionally, two couples and one father fail- ed to attend any appointment. All of them declined any subsequent call from the research group and withdrew from the study at an early stage.

Approximately 70% of the pregnant women attending the labor classes completed a brief questionnaire, designed to collect basic sociodemographic data so as to enable the representativeness of the final sample to be evalua- ted. The characteristics of participants and non-participants who completed this form are shown in table 3.

Table 4 summarizes the schedule for the field work in which parents were intended to play an active role. At enrolment, midwives assigned an identification num- ber to the future triplet, and presented them with a small bag. It contained general information about the project, an instruction leaflet, and the logistic material for the study, including a breast-milk pump to collect milk samples, which was also handed out as a gift. All the material in the bag was pre-labeled with the trio number plus an identification character that would allow each of its members to be distinguished («M» [madre]for mo- thers, «P» [padre]for fathers, and «H» [hijo/a]for children). The necessary material for collecting biologic spe- cimens was suitably labeled and wrapped in three packs, one for each member of the triplet.

All specimens were collected in accordance with specific protocols, in containers labeled with the subject identification number and date. In brief, five different types of specimens were collected (table 4). Adult blood specimens (30 ml) and cord blood (12.5 ml) were taken, using the vacutainer system with different tubes, namely: he- parine tubes for heavy metal and micronuclei analysis; tubes containing gelose for routine biochemical analyses; tubes with EDTA for routine hemogram; and empty, simple glass standard tubes for organochlorines and dioxins. The recipient laboratory identified the specimens with the case codes, performed the systematic hemogram and biochemical determinations, extracted the serum and then sent the heparinized tubes to the Genetic Department of the Ramón y Cajal Hospital for micronuclei analysis, and all remaining specimens to the Madrid Regional Public Health Laboratory (RPHL), where they were centra- lized, classified, frozen and stored at –70 oC.

Urine specimens (100 ml) corresponded to the first morning urine passed on the date appointed for blood extraction. Receptacles had to be wrapped in aluminum foil to avoid light exposure, and were kept refrigerated at 2-8 oC by the recipient laboratory until submission to the RPHL.

Placentas were frozen at –20 °C immediately after birth at the maternity ward.

Milk samples (20 ml) were collected by mothers at the end of third week post-delivery, after having fed their babies. Milk receptacles were kept refrigerated (2-8 oC) until collected by study personnel, and frozen and sto- red at the RPHL. Three weeks postpartum and after the baby had been bathed, a lock of newborn hair was cut with scissors by the parents and kept at room tempe- rature until analysis.

As part of this research project, passive air samplers were positioned around study participants' place of re- sidence, in order to take outdoor measurements of air pollutants. The aim was to ascertain exposure to nitrogen dioxide and volatile organic compounds (benzene, to- luene, ethylbenzene, m+p-Xylene and o-xylene) in the two study areas. Passive air samplers, analytical determinations and methodology were in accordance with the INMA protocols20, though in our case a single sam- pling campaign was completed. A total of 80 passive air samplers were deployed, 29 in the urban area (Vallecas district) and 51 in the outlying area (Getafe and Parla districts). These measures may help to have a bet- ter knowledge of the study areas pollution profile.

Questionnaires were used in all three phases of the field work. The design of the principal epidemiologic questionnaire, completed at study entry, took the INMA protocols into account so as to allow for future comparisons with INMA results. Sex-specific versions were initially tes- ted on a small sample of mothers and fathers in each area, and some questions were reworded in line with the results of this trial. The interview lasted around 40 minutes and included questionnaires with several modules: sociodemographic information; tobacco exposure (lifelong use, recent and current smoking, and second- hand exposure); occupational history; socioeconomic data; environmental exposures at home; and, reproductive and general medical history (diseases, drug consump- tion, radiological exposure, and dental fillings). For women, an additional module registered characteristics and cli- nical events during the pregnancy. Dietary information was recorded, using a previously developed and validated food-frequency questionnaire21–23, which enables nutrient intake to be estimated.

At the hospital, health staff completed a sheet re- cording delivery and newborn health status and an- thropometric data. Finally, a lactation questionnaire re- gistered information on the lactation process and sample collection.

Both the study protocol and informed consent form complied with the principles of the Helsinki Declaration and were formally approved by the Ethical Committee of the Carlos III Institute of Health. As children were in- volved in this project, the final version of the protocol was also submitted to the official Children's Ombuds- man for the Madrid Autonomous Region. Personal data and biologic samples were stored in accordance with prevailing Spanish legal and ethical requirements. Confidentiality was guaranteed. Participants were pro- vided with a contact telephone number in the event that they might have any doubts or queries, or wish to with- draw their consent to participate at any time.

Table 5 summarizes the main results. Of the origi- nally planned 150 triplets, we finally recruited 145 mothers, 144 fathers and 135 newborns. It has to be taken into account that blood cord collection, which took place at approximately one month after initial parents' sample collection, had additional difficulties as it required that: a) mothers provide the material included in the kit to the hospital personnel, and inform that they were BioMadrid participants, and b) delivery circumstances per- mit the normal collection of samples.

Among adults, blood and urine specimens were ob- tained in over 97% of subjects, the general questionnaire attained similar figures, and dietary record she- ets were completed in 92% of cases.

In newborns, the most frequently obtained specimen was hair (96%) followed by cord blood (87%) (table 5). In one case lost to follow up was due to miscarriage. Other nine children with no biological samples were born alive, though were lost due to different reasons: one woman forgot to take the kit including instructions and material needed to collect the biological samples; four mothers were sent to a non-participant hospital to give birth due to overcrowded maternal wards in the refe- rence centers, and in four other cases complications in delivery distorted the logistical arranged procedures.

The lowest figures corresponded to placentas (81%) and milk samples (74%). Fourteen of the lactating mothers answered the lactation questionnaire but failed to collect a breast milk specimen or provided an insuffi- cient volume of milk. In general, participants reported that the instructions supplied were clearly understan- dable. All BioMadrid protocols are available upon request.

This project is the first comprehensive study of po- llutant levels in the population of the Madrid Region. It will provide an estimate of environmental exposure to different pollutants in several biologic specimens from fathers, mothers and newborns in the study areas, and will help design a feasible environmental surveillance system based on human biomarkers. In spite of the study's complex design, field work results can be con- sidered as satisfactory.

This project has enabled the complexity of imple- menting a surveillance program using biologic samples to be estimated. In the next future, we have still several questions to answer when laboratory analyses are completed, including the discussion about the most suit- able toxicants and biological samples to be considered in a surveillance system. Nevertheless, the financial and human effort needed to put a system like this into ope- ration calls for firm political backing. The results of this study will serve to tailor the design of any future surveillance system to the reality of the Madrid Region.

Financial support was obtained from the Madrid Regional Health & Consumer Affairs Authority and the Spanish Health Research Fund (Fondo de Investigación Sanitaria [FIS]) grant PI040777.

The authors are deeply indebted to the primary-care midwives and health staff of the Gregorio Marañón and Getafe Hospital maternity wards and laboratories, and, in particular, to the study participants themselves. Their unstinting and al- truistic collaboration made this study possible. We should also like to thank: the INMA study research team for their encour- agement and assistance; and Michael Benedict for his help with the English.