The objective of this study was to examine the associations between active and passive smoking in different periods of pregnancy and changing smoking habits during pregnancy, with low birthweight and preterm birth. The study was embedded in the Generation R Study, a population-based prospective cohort study from early fetal life onwards in Rotterdam, The Netherlands. Active and passive smoking were assessed by questionnaires in early, mid- and late pregnancy. Analyses were based on 7098 pregnant women and their children. Active smoking until pregnancy was ascertained and was not associated with low birthweight and preterm birth. Continued active smoking after pregnancy was also recorded and was associated with low birthweight (adjusted odds ratio 1.75 [95% CI 1.20, 2.56]) and preterm birth (adjusted odds ratio 1.36 [95% CI 1.04, 1.78]). The strongest associations were found for active maternal smoking in late pregnancy. Passive maternal smoking in late pregnancy was associated with continuously measured birthweight (P for trend <0.001). For all active smoking categories in early pregnancy, quitting smoking was associated with a higher birthweight than continuing to smoke. Tendencies towards smaller non-significant beneficial effects on mean birthweight were found for reducing the number of cigarettes without quitting completely. This study shows that active and passive smoking in late pregnancy are associated with adverse effects on weight and gestational age at birth. Smoking in early pregnancy only, seems not to affect fetal growth adversely. Health care strategies for pregnant women should be aimed at quitting smoking completely rather than reducing the number of cigarettes.

We examined ethnic differences in infant mortality and the contribution of several explanatory variables. Data of Statistics Netherlands from 1995 to 2000 were studied (1,178,949 live borns). Proportional hazard analysis was used to show ethnic differences in total and cause-specific infant mortality. Obstetric, demographic and -geographical variables, and socio-economic status were considered as possible determinants. The four major ethnic minority groups showed an elevated risk of infant mortality, ranging from 1.28 in Turkish infants to 1.50 in Antillean/Aruban infants. In the early neonatal period, risks were elevated for Surinamese (hazard ratio [HR] 1.48, 95% confidence intervals [CI] 1.23, 1.78) and Antilleans/Arubans (HR 1.43, 95% CI 1.06, 1.92). In the post-neonatal period, risks were only elevated for Turkish (HR 2.20, 95% CI 1.80, 2.69) and Moroccan infants (HR 2.06, 95% CI 1.67, 2.55). Surinamese and Antillean/Aruban infants had an elevated risk of dying from perinatal causes (HR 1.62, 95% CI 1.33, 1.98 and 1.69, 95% CI 1.24, 2.29 respectively), Turkish and Moroccan infants had an elevated risk of dying from congenital anomalies (HR 1.42, 95% CI 1.16, 1.73 and 1.46, 95% CI 1.20, 1.79 respectively). Inequalities as a result of socio-economic position and demographic factors, such as marital status and maternal age, partially explain the ethnic differences in infant mortality. We conclude that ethnic minority groups in The Netherlands have a higher infant mortality than the native population, which in part seems preventable by reducing inequalities in socio-economic status. Marital status and age of the mother are important other risk factors of infant mortality.

Patterns and correlates of maternal smoking could differ according to ethnic background, and these differences might have consequences for intervention strategies. In the Generation R study, we examined patterns of smoking during pregnancy and the associations of socioeconomic (educational level), demographic (maternal age, marital status, generational status, parity) and lifestyle (alcohol consumption, partner smoking) correlates with smoking during pregnancy in 5,748 women of Dutch, Turkish, Moroccan, Surinamese-Hindustani, Surinamese-Creole, Capeverdean and Antillean ethnic background. Smoking rates before pregnancy were highest in the Turkish group (43.7%) and lowest in the Moroccan group (7.0%). Compared with Dutch women (24.1%), Turkish and Moroccan women were less likely to quit smoking before pregnancy (17.0% and 5.9%, respectively; p<.001). Turkish and Moroccan women (72.0% and 70.6%, respectively) were more likely to continue smoking during pregnancy compared to Dutch women (58.6%, p<.001). Lower education was associated with smoking during pregnancy only in the Dutch group. No significant association of education with smoking was seen in the non-Dutch groups. Second-generation (i.e., foreign-born) Turkish and Capeverdean women were more likely to smoke during pregnancy compared with first-generation women. Partner smoking was associated with smoking during pregnancy in all ethnic groups except for Surinamese-Creole and Antillean. Maternal alcohol consumption was associated with smoking during pregnancy in all ethnic groups except for Capeverdean. Smoking rates and correlates of smoking during pregnancy varied by ethnic background. These observations should be considered when designing maternal smoking prevention and intervention strategies.

Total and cause-specific infant mortality seem to differ according to generational status and age at immigration of the mother. The direction of these trends however differs between ethnic populations. This may be related to acculturation and selective migration.

meeste van deze relaties komen naar voren uit genoombreed associatieonderzoek, waarbij honderdduizenden SNP's van een grote groep patiënten worden vergeleken met die van een controlegroep. 3 Het menselijk genoom is opgebouwd uit ruim drie miljard baseparen die samen de 46 chromosomen vormen. Slechts een klein deel (< 1%) van die baseparen codeert voor specifieke eiwitten. De coderende stukken DNA noemt men genen -naar schatting zijn dit er ongeveer 25.000. Een SNP is een variatie in één enkel basepaar. SNP's komen bij iedereen in grote aantallen voor en kunnen, als ze in of nabij een gen liggen, een deel van de variatie in de genactiviteit verklaren. Deze verschillen in genactiviteit kunnen op hun beurt leiden tot verschillen in fenotype, en eventueel ook tot ziekte. De voorspellende waarde van een genetisch profiel voor multifactoriële ziektenOmdat de meeste SNP's slechts een klein effect hebben en verschillende genen een rol spelen bij het ontstaan van multifactoriële ziekten, kan men deze ziekten niet voorspellen op basis van één enkel gen. Men doet dat op basis van meerdere genen, die men samenvoegt tot een genetisch profiel. In feite zijn genetische profielen vaak niet meer dan risicoscores op basis van de combinatie van een aantal SNP's. 4 Deze risicoscores zijn doorgaans statistisch significant geassocieerd met het risico op de ziekte, 5 maar dat betekent nog niet dat ze ook in staat zijn om de ziekte te voorspellen. Het voorspellende vermogen van genetische pro- Het voorspellen van ziektes wordt big businessDe afgelopen jaren is enorme vooruitgang geboekt in het identificeren van nieuwe genen voor veelvoorkomende ziekten, zoals type-2-diabetes en coronaire hartziekten. 1 Met name de genoombrede DNA-scans waarmee men in één keer honderdduizenden DNA-variaties, oftewel single nucleotide polymorphisms (SNP's), kan onderzoeken, hebben hieraan een belangrijke bijdrage geleverd. 1 De grote belofte is dat deze nieuwe kennis zal leiden tot 'personalized medicine', waarbij beslissingen over preventie en behandeling kunnen worden afgestemd op het individuele genetische profiel van de patiënt. 2 Sinds een paar jaar bieden diverse bedrijven zoals deCODEme, 23andMe en Navigenics 'personal genome scans' aan, die naar hun zeggen de consument inzicht geven in de risico's op diverse aandoeningen. Zij bieden deze commerciële tests niet aan via de reguliere zorgverleners, maar via internet. De risico's die deze scans aan het licht kunnen brengen, betreffen vooral multifactoriële ziekten zoals type-2-diabetes, hart-en vaatziekten, maculadegeneratie, sommige vormen van kanker en reuma (zie de tabel). Vanwege de enorme media-aandacht voor genenonderzoek en het groeiende aanbod van dergelijke commerciële tests zullen patiën-ten steeds vaker vragen gaan stellen over de rol van genen bij het ontstaan van ziekten. Consumenten die een persoonlijke genetisch profiel hebben laten maken, zullen vaak niet weten wat dat profiel precies betekent en wat ze met de voorspelde risico's moeten doen. In dit commentaar bespreken wij de stand ...