The reduced risk of preeclampsia with maternal smoking during pregnancy is inconsistent with the many well-documented harmful effects of smoking, including increased risk of stillbirth and infant mortality.1,2 The reduction in risk with maternal smoking is estimated to be 30%–50%.3–5 However, arriving at a clear biologic mechanism to explain these findings has been complicated by the observation that maternal active smoking is also strongly associated with decrements in birthweight6–8 and modest increases in risk for preterm delivery,9,10 both of which also result from preeclampsia and should thus be decreased if preeclampsia is prevented by smoking.
Preeclampsia is not the only condition thought to be prevented by active smoking. Parkinson disease is another.11 It is possible that some therapeutic intervention could evolve from a better understanding of the biological mechanism of such protective effects. Even though the net harm from smoking remains unarguable, understanding how an exogenous, modifiable factor affects preeclampsia risk could also shed light on the etiology of preeclampsia.
Despite the robust literature examining the influence of passive/environmental smoke exposure on birthweight or prematurity, to our knowledge there have been no studies on the influence of passive smoke exposure on preeclampsia or gestational hypertension. Additionally, there is very limited literature on trimester-specific effects of active smoking on hypertensive disorders of pregnancy, with one recent study based in the Swedish Medical Birth Registry finding that reduced risk of preeclampsia was limited to women who continued smoking into their late pregnancy.5
We undertook a prospective investigation of the trimester-specific effects of active and passive smoking on preeclampsia and gestational hypertension in the Norwegian Mother and Child cohort (MoBa), a population-based cohort recruited between 1999 and 2008, comprising about 18% of all deliveries in Norway during the study period.12
Study Population and Design
The Norwegian Mother and Child cohort (MoBa) enrolled over 100,000 women between 1999 and 2008. Study design and selection characteristics have been described in detail elsewhere.12,13 Briefly, women were invited to participate by mail prior to their routine ultrasound examination at their local hospital, which generally occurs around 18 weeks’ gestation. Women consenting to participate completed two self-administered exposure questionnaires during their pregnancy. The first maternal and paternal questionnaires (Q1) were sent by mail with the consent form prior to the ultrasound appointment (95% completed). The mean gestational age at which the first questionnaire was completed was 17.6 weeks (standard deviation = 2.8). A second questionnaire covering maternal health status and habits since the first questionnaire was mailed to participants at approximately week 30 (Q3) (92% completed, mean gestational age 30.7 [1.6] weeks). In some cases (0.4%), the last questionnaire was received following delivery. Women were allowed to participate for more than one pregnancy within the study period (15% with repeat pregnancies). Participation rate varied by study year,12 but averaged 39%. The final MoBa population comprised approximately 18% of all deliveries in Norway during the study period. There were no differences between the enrolled MoBa population and the entire population with respect to prevalence of preeclampsia and chronic hypertension, although active smokers were under-represented among the MoBa participants (11% vs. 6%).12
Pregnancy Smoking Information
Women were queried twice during pregnancy regarding their personal smoking habits and their exposure to passive/environmental tobacco smoke at work or home. In the first questionnaire (~18 weeks’ gestation), mothers were asked whether they currently smoked (if so, how many cigarettes per day/week), or if they had stopped smoking after they became pregnant, and if so, at what gestational age. If a woman reported quitting smoking after she became pregnant but did not report the amount she smoked up until she quit (n= 10,073), we used the amount she reported in the 3 months before she became pregnant. If that was unavailable (n = 417), she was considered to be a smoker until the gestational week she reported quitting, but with a missing dose for that period. If a woman reported quitting but did not provide the gestational age at which she quit (n = 148), we assumed she quit at the midpoint of her current trimester. If a woman reported smoking during pregnancy, but did not report whether or not she quit (n = 5,852), we assumed she continued smoking. We used the dose information provided in the first questionnaire to cover the first- and second-trimester exposure, or up until a specified quit week. In the next questionnaire (~32 weeks’ gestation), women were asked whether they currently smoked and how much, and, if they had quit at some point during pregnancy, at what gestational age. We used the dose information provided in this questionnaire to cover the third trimester until delivery, or until a specified quit week. Dose information was categorized into four groups: “None” and then tertiles rounded to the nearest whole cigarette above (>0 to <2, 2 to <8, ≥8 cigarettes per day). In both questionnaires, passive smoke exposure at work and home was reported, along with the number of hours exposed per day.
Birth Outcome Information
Birth outcome information was obtained by linkage with the Medical Birth Registry of Norway. The registry receives delivery information from hospitals using a standardized birth notification form.14 This form provides demographic information on the mother and father, mother’s health before and during pregnancy including chronic diseases and pregnancy complications, as well as infant birth defects and delivery characteristics. A designation of gestational hypertension alone, or preeclampsia/eclampsia, is completed by a hospital midwife using information from the mother’s antenatal medical record and additional data recorded at the maternity department upon admission. Preliminary analyses of the positive predictive value of preeclampsia registrations as compared with blood pressure and protein measures on personal antenatal records and hospital discharge codes indicate that the preeclampsia registrations have strong validity (Klungsøyr K, unpublished data). Information on maternal chronic hypertension and diabetes is also contained in the registry. Preeclampsia and gestational hypertension were considered separately in the analysis. Women with preexisting/chronic hypertension, including preeclampsia superimposed on chronic hypertension, were excluded from all analyses (n= 372). Overall, the risk of preeclampsia in this population was 3.5% (n = 2,564) and the risk of gestational hypertension was 1.9% (n = 1,935).
MoBa data were from the version 4 release of quality-assured data. Of the 90,190 MoBa pregnancies in which Q1 was completed, 77,041 also had Q3 completed, of which 74,439 were singletons. Additionally, women who delivered prior to the onset of their third trimester (gestational age at delivery <27 weeks) were excluded (n = 2,378). There were 8,209 women who contributed two or more pregnancies to this analysis (7,890 with two pregnancies, 311 with three pregnancies, and eight with four pregnancies).
The following covariates were considered as potential confounders or effect modifiers: maternal age (linear and quadratic term), prepregnancy body mass index, parity, marital status, maternal education, educational disparity between mother and father, household income, and any diabetes prior to or during pregnancy. We constructed directed acyclic graphs to explore the possible relations between active and passive smoking, the above covariates, and preeclampsia or gestational hypertension. We adjusted for covariates indicated by an inspection of directed acyclic graphs, although we also included highly predictive covariates that improved the overall precision of the model. We evaluated the final model for parsimony after inspecting changes in the main effect estimate. Women who were missing active or passive smoking exposure information were dropped from the models. For the main analysis of active and passive smoking in relation to preeclampsia and gestational hypertension, multiple imputation (five replications) was used to impute missing values for body mass index (BMI) and education. Specifically, in SAS, we first used the Markov Chain Monte Carlo method to construct five data sets, just enough missing values to have a monotone missing value, and then used the discriminant function on each of these data sets to impute education. The imputed continuous values for BMI were placed in discrete categories. We evaluated the interaction between smoking and maternal BMI, smoking and maternal age, and smoking and parity, including any interactions in the model if the P value for the product term was less than 0.10. Multivariable logistic regression using Proc Genmod was implemented in SAS 9.2 (Cary, NC), using the repeated statement to account for multiple observations per woman.
There were 74,067 singleton pregnancies with complete pregnancy questionnaire data (Table 1). Among these, 17,014 (23%) reported active smoking at some point during pregnancy. Among women who reported any smoking during pregnancy, 47% reported quitting in the first trimester, 6% reported quitting in the second trimester, and 1% reported quitting in the third trimester. In general, smokers tended to be younger, primiparous, less educated, more likely to be single at the time of pregnancy, and have smaller personal incomes. Smokers were also much more likely to report passive smoke exposure at home or work.
Rates of smoking declined dramatically during pregnancy. Approximately 23% of women reported active smoking in trimester 1, which declined to 9% in trimester 2 and 8% in trimester 3. Patterns of smoking behavior and their associated risks for preeclampsia and gestational hypertension are described in Table 2. Women who quit smoking prior to the onset of their third trimester had roughly the same risk of preeclampsia and gestational hypertension as nonsmokers. In general, smoking during the third trimester was required to produce a substantial decrement in risk of disease. Because few third-trimester smokers take up the habit in the third trimester, we cannot separate the effect of continuous exposure from exposure during the third trimester alone. For both preeclampsia and gestational hypertension, women with continuous exposure throughout pregnancy had the strongest reduction in risk of disease (Table 2).
We estimated the associations of trimester 1, trimester 2, and trimester 3 active and passive smoking with preeclampsia and gestational hypertension (Table 3). In models considering each trimester separately (model 1), active smoking only (ever vs. never) was associated with lower odds of preeclampsia in all trimesters, with the strongest associations in trimesters 2 and 3. Passive smoke exposure alone was also associated with minor decrements in risk across pregnancy. Although consistent in direction with active smoking, the passive smoking estimates were weak. Active and passive smoke exposure together was also associated with reduced odds of preeclampsia in every trimester—again most strongly in trimesters 2 and 3. Women who smoked throughout the pregnancy also tended to smoke more heavily (40% of women who quit smoking before their second trimester reported smoking less than 1 cigarette per day, whereas only 13% of women who smoked throughout pregnancy reported smoking so little in the first trimester). We conducted a sensitivity analysis in an attempt to isolate the effect of timing from that of dose in early pregnancy. We reran the third-trimester model 1 (Table 3), including the additional adjustment variable of dose in trimester 1, and found no appreciable change in the estimate of relative risk or the precision of the overall association. This suggests that there is some independent effect of exposure timing on the overall association of smoking with preeclampsia.
To shed additional light on the association with exposure in trimester 3, we performed additional analyses (model2) that adjusts the trimester 3 association for smoking in trimester 2. (We limited ourselves to adjusting trimester 3 results because, in general, smoking status in trimesters 1 and 2 were correlated in the sense that exposure history was ascertained during the same interview, and there were no nonsmokers at trimester 1 who reported starting smoking at trimester 2.) The analysis suggests that adjusting for prior trimesters did not substantially change the trimester 3 effects, although the association became much less precise (eg, for the association of active smoking with preeclampsia, odds ratio [OR] = 0.56 unadjusted for prior history vs. 0.63 adjusting for trimester 2 smoking, which is a 13% change in OR).
A similar pattern was found for gestational hypertension (approximately 11% change in estimated effect) (Table 3). We found no evidence of effect modification of these associations by parity, maternal age, or body mass index. Moreover, results were unchanged when adjusted for the length of the third trimester (in order to account for the possibility that smokers delivered prior to development symptoms of preeclampsia). Overall, active smoking in trimester 2 and trimester 3 appeared to be associated with the strongest reductions in risk for both preeclampsia and gestational hypertension.
We used multiple imputation to account for missing data on key covariates in Table 3. When observations with missing data for key covariates were instead excluded, in general, the point estimates and confidence limits of the smoking effects were consistent to the second decimal place (data not shown).
For both preeclampsia and gestational hypertension, odds of disease dropped markedly from the lightest to the heaviest smokers (Table 4). However, overall there were few “light” smokers during this period, and a substantial number reporting no daily cigarette usage.
In this population-based study of Norwegians, we observed a strong reduction in risk of preeclampsia and, to a lesser degree, gestational hypertension, among women who reported smoking in their third trimester of pregnancy. Moreover, the associations were stronger among women who reported heavier smoking in this period, both for preeclampsia and for gestational hypertension. It is notable that women who quit smoking prior to the onset of their third trimester experienced roughly the same risk of preeclampsia as nonsmokers. This is somewhat inconsistent with a recent study indicating that women who smoke in the first trimester have a steeper increase in mean systolic blood pressure throughout pregnancy, and a steeper increase in mean diastolic blood pressure in later pregnancy, compared with nonsmokers.15 However, their study was much smaller than ours, with only 49 exposed gestational hypertension cases and 20 exposed preeclampsia cases.
Trimester-specific associations with smoking have not been widely studied. In a small hospital-based case-control study, Xiong et al16 found smoking in pregnancy to be inversely related to risk, but there were too few exposed cases to distinguish early versus late effects. Marcoux et al,17 also in a hospital-based case-control study, reported that smoking cessation prior to 20 weeks’ gestation resulted in an attenuation of the apparent protection provided by maternal smoking. England et al18 analyzed smoking data collected from a randomized clinical trial, and found that prepregnancy smoking was not protective, while smoking at the enrollment visit (which occurred at some point between 13 and 21 weeks) was. Very similar results were found by Zhang and colleagues,19 in historical data from the Collaborative Perinatal Project (CPP) (smoking information collected at ~18 weeks), and also by Sibai et al20 in an analysis of a randomized clinical trial with personal smoking history: prepregnancy smoking was not associated with reduced risk, but current smoking at enrollment was. However, none of these studies was able to describe smoking patterns continuously across trimesters. It is also likely that in each of these studies, except perhaps the CPP, a large proportion of first-trimester smokers were misclassified because the largest drop in smoking rates occurred between the second and third trimesters. Few women appear to quit smoking later in pregnancy, making it difficult to disentangle second and third trimester-specific effects.
Wikström and colleagues5 recently reported results from the Swedish Birth Register, using all singleton births in Sweden during the years 1999–2006. Information on smoking and snuff use was collected at two time points: at the first antenatal visit (approximately 15 weeks’ gestation) and at approximately week 30–32. Information on first-trimester smoking was not available (eg, if a woman was not a current smoker at 15 weeks). Wikström et al5 analyzed smoking in discrete categories: smoking or smokeless tobacco use at the first interview and smoking or smokeless tobacco use at the second interview. In that study, continuous smokers and smokers at the second interview had the greatest reduction in risk, both for preeclampsia and gestational hypertension, with no reduced risk associated with snuff use (suggesting that the causal agent is not nicotine, but rather some product of combustion).
It remains plausible that residual or unmeasured confounding partly explains the consistently inverse associations of prenatal smoking with preeclampsia. One possible pathway is that genetic traits associated with addiction to smoking are also directly associated with preeclampsia through another unblocked pathway. A similar concern has arisen in the literature linking prenatal smoking with childhood attention deficit hyperactivity disorder (ADHD). Although maternal prenatal smoking has been consistently associated with increased risks of ADHD in offspring in a number of human21,22 and animal23,24 studies, the use of a sib-pair design to control for unmeasured confounding reduced the magnitude of the association dramatically.25 Likewise, the CPP (conducted when smoking was far more prevalent during pregnancy) found no association between maternal smoking and ADHD.26 However, a protective and dose-response association between antenatal smoking and preeclampsia was reported in the CPP,19 dispelling some of the concerns that residual confounding might entirely explain the association of smoking with preeclampsia. Preeclampsia studies using designs that are more robust to genetic or other unmeasured confounders, such as the sib-pair design, should be explored, but may be challenging to implement given the strong association between parity and preeclampsia.
Preeclampsia is considered a two-stage disease.27 The first stage is caused by abnormal placentation due to insufficient invasion of extravillous cytotrophoblasts, leading to insufficient remodeling of the maternal spiral arteries and a reduced placental perfusion.28,29 The second stage of preeclampsia (the maternal systemic disorder) is likely to develop as a consequence of placental ischemia. Because sufficient uteroplacental blood flow is a key element in successful implantation and reproductive outcome, increased attention has been devoted to altered angiogenesis in preeclampsia. Recent research has focused on angiogenic factors and their role in the development of preeclampsia. Placental ischemia is believed to release endothelial damaging factors to the maternal circulation,27 and antiangiogenic factors are likely to play a role in the development of preeclampsia. Decreased concentrations of circulating angiogenic factors, free placental growth factor and free vascular endothelial growth factor, have been reported in preeclamptic pregnancies,30 as have increased concentrations of the antiangiogenic factors soluble fms-like tyrosine kinase 1 (sFlt-1) (which is an inhibitor of placental growth factor and vascular endothelial growth factor) and soluble endoglin (sENG).31,32
The inverse association between smoking during pregnancy and risk of preeclampsia could be causal. Various biologic mechanisms have been proposed to explain this association, including carbon monoxide–mediated inhibition of inflammation,33 enhanced vasodilation,34 and suppression of platelet aggregation,35 plasminogen activation,35 apoptosis,36,37 reactive oxygen species formation38, and sFlt-1, an antiangiogenic factor.39,40 Altered function of sFlt-1 has been associated with well-established risk factors for preeclampsia. Primiparity and twin pregnancy are major risk factors, separately2 and together.41 Wolf et al42 reported a decrease in circulating sFlt-1 from first to second pregnancy, suggesting that higher levels of sFlt-1 may predispose to preeclampsia. Higher levels of sFlt-1 have also been found in twin pregnancies,43 and lower sFlt-1 levels have been found in pregnancies of smokers,44 further supporting this hypothesis. A recent nested case-control study demonstrated higher gestational age–specific birthweights and lower sFlt-1 levels in smokers with preeclampsia as compared with nonsmokers with preeclampsia.45 The authors suggest smoking may exert this effect by reducing levels of sFlt-1.
Heterogeneity in associations of smoking with preeclampsia risk according to trimester of exposure may provide some clues as to the underlying mechanism. Clinical preeclampsia usually emerges after the middle of the second trimester, coincident with substantial rises in sFlt-1 and sENG, and may explain why smoking during the latter half of the pregnancy as demonstrated in our study and others seems more beneficial.46
There is likely to be some misclassification of outcome in our study (eg, gestational hypertension labeled preeclampsia or the reverse). In general, the patterns of association for these outcomes were similar, and thus it is difficult to argue that the smoking effect is specific to preeclampsia, although the magnitude of associations was stronger for this outcome. Differential recall of smoking behavior is made less likely by the prospective design of exposure ascertainment, with the first-trimester smoking behaviors reported prior to 20 weeks’ gestation. Smoking behaviors during late pregnancy may have been reported coincident with the symptoms or diagnosis of disease. Biomarkers of smoking exposure (such as cotinine levels)—unless they were collected at multiple times during pregnancy—would be unlikely to substantially improve estimates of exposure, given the significant variability in smoking patterns across pregnancy in this population and others.47 Additionally, the analytic approach would be improved by time-to-event modeling, in which competing risks for delivery (eg, prematurity) could be explicitly accounted for. Unfortunately, this approach requires the availability of gestational age at onset of symptoms, which is rarely available.
In conclusion, our results add to the weight of evidence implicating smoking exposure during late pregnancy as the critical determinant of the inverse association between antenatal smoking and preeclampsia. Although the pathways underlying the association remain unknown, eventual elucidation may help to identify treatment options for this dangerous and enigmatic disorder.
1. Salihu HM, Wilson RE. Epidemiology of prenatal smoking and perinatal outcomes. Early Hum Dev. 2007;83:713–720
2. Sibai BM, Hauth J, Caritis S, et al. Hypertensive disorders in twin versus singleton gestations. National Institute of Child Health and Human Development Network of Maternal-Fetal Medicine Units. Am J Obstet Gynecol. 2000;182:938–942
3. Conde-Agudelo A, Althabe F, Belizán JM, Kafury-Goeta AC. Cigarette smoking during pregnancy and risk of preeclampsia: a systematic review. Am J Obstet Gynecol. 1999;181:1026–1035
4. Engel SM, Janevic TM, Stein CR, Savitz DA. Maternal smoking, preeclampsia, and infant health outcomes in New York City, 1995-2003. Am J Epidemiol. 2009;169:33–40
5. Wikström AK, Stephansson O, Cnattingius S. Tobacco use during pregnancy and preeclampsia risk: effects of cigarette smoking and snuff. Hypertension. 2010;55:1254–1259
6. Wang X, Zuckerman B, Pearson C, et al. Maternal cigarette smoking, metabolic gene polymorphism, and infant birth weight. JAMA. 2002;287:195–202
7. Jaddoe VW, Troe EJ, Hofman A, et al. Active and passive maternal smoking during pregnancy and the risks of low birthweight and preterm birth: the Generation R Study. Paediatr Perinat Epidemiol. 2008;22:162–171
8. Reeves S, Bernstein I. Effects of maternal tobacco-smoke exposure on fetal growth and neonatal size. Expert Rev Obstet Gynecol. 2008;3:719–730
9. Haskins A, Mukhopadhyay S, Pekow P, et al. Smoking and risk of preterm birth among predominantly Puerto Rican women. Ann Epidemiol. 2008;18:440–446
10. Tsai HJ, Liu X, Mestan K, et al. Maternal cigarette smoking, metabolic gene polymorphisms, and preterm delivery: new insights on GxE interactions and pathogenic pathways. Hum Genet. 2008;123:359–369
11. Chen H, Huang X, Guo X, et al. Smoking duration, intensity, and risk of Parkinson disease. Neurology. 2010;74:878–884
12. Nilsen RM, Vollset SE, Gjessing HK, et al. Self-selection and bias in a large prospective pregnancy cohort in Norway. Paediatr Perinat Epidemiol. 2009;23:597–608
13. Magnus P, Irgens LM, Haug K, Nystad W, Skjaerven R, Stoltenberg CMoBa Study Group. . Cohort profile: the Norwegian Mother and Child Cohort Study (MoBa). Int J Epidemiol. 2006;35:1146–1150
14. Irgens LM. The Medical Birth Registry of Norway. Epidemiological research and surveillance throughout 30 years. Acta Obstet Gynecol Scand. 2000;79:435–439
15. Bakker R, Steegers EA, Mackenbach JP, Hofman A, Jaddoe VW. Maternal smoking and blood pressure in different trimesters of pregnancy: the Generation R study. J Hypertens. 2010;28:2210–2218
16. Xiong X, Zhang J, Fraser WD. Quitting smoking during early versus late pregnancy: the risk of preeclampsia and adverse birth outcomes. J Obstet Gynaecol Can. 2009;31:702–707
17. Marcoux S, Brisson J, Fabia J. The effect of cigarette smoking on the risk of preeclampsia and gestational hypertension. Am J Epidemiol. 1989;130:950–957
18. England LJ, Levine RJ, Qian C, et al. Smoking before pregnancy and risk of gestational hypertension and preeclampsia. Am J Obstet Gynecol. 2002;186:1035–1040
19. Zhang J, Klebanoff MA, Levine RJ, Puri M, Moyer P. The puzzling association between smoking and hypertension during pregnancy. Am J Obstet Gynecol. 1999;181:1407–1413
20. Sibai BM, Gordon T, Thom E, et al. Risk factors for preeclampsia in healthy nulliparous women: a prospective multicenter study. The National Institute of Child Health and Human Development Network of Maternal-Fetal Medicine Units. Am J Obstet Gynecol. 1995;172(2 Pt 1):642–648
21. Obel C, Linnet KM, Henriksen TB, et al. Smoking during pregnancy and hyperactivity-inattention in the offspring–comparing results from three Nordic cohorts. Int J Epidemiol. 2009;38:698–705
22. Braun JM, Kahn RS, Froehlich T, Auinger P, Lanphear BP. Exposures to environmental toxicants and attention deficit hyperactivity disorder in U.S. children. Environ Health Perspect. 2006;114:1904–1909
23. Schneider T, Ilott N, Brolese G, Bizarro L, Asherson PJ, Stolerman IP. Prenatal exposure to nicotine impairs performance of the 5-choice serial reaction time task in adult rats. Neuropsychopharmacology. 2011;36:1114–1125
24. Garcia-Rill E, Buchanan R, McKeon K, Skinner RD, Wallace T. Smoking during pregnancy: postnatal effects on arousal and attentional brain systems. Neurotoxicology. 2007;28:915–923
25. Obel C, Olsen J, Henriksen TB, et al. Is maternal smoking during pregnancy a risk factor for hyperkinetic disorder?–Findings from a sibling design. Int J Epidemiol. 2011;40:338–345
26. Ball SW, Gilman SE, Mick E, et al. Revisiting the association between maternal smoking during pregnancy and ADHD. J Psychiatr Res. 2010;44:1058–1062
27. Redman CW, Sargent IL. Latest advances in understanding preeclampsia. Science. 2005;308:1592–1594
28. Lain KY, Roberts JM. Contemporary concepts of the pathogenesis and management of preeclampsia. JAMA. 2002;287:3183–3186
29. Sargent IL, Germain SJ, Sacks GP, Kumar S, Redman CW. Trophoblast deportation and the maternal inflammatory response in pre-eclampsia. J Reprod Immunol. 2003;59:153–160
30. Tjoa ML, Levine RJ, Karumanchi SA. Angiogenic factors and preeclampsia. Front Biosci. 2007;12:2395–2402
31. Maynard SE, Min JY, Merchan J, et al. Excess placental soluble fms-like tyrosine kinase 1 (sFlt1) may contribute to endothelial dysfunction, hypertension, and proteinuria in preeclampsia. J Clin Invest. 2003;111:649–658
32. Venkatesha S, Toporsian M, Lam C, et al. Soluble endoglin contributes to the pathogenesis of preeclampsia. Nat Med. 2006;12:642–649
33. Otterbein LE, Bach FH, Alam J, et al. Carbon monoxide has anti-inflammatory effects involving the mitogen-activated protein kinase pathway. Nat Med. 2000;6:422–428
34. Zhang F, Kaide JI, Rodriguez-Mulero F, Abraham NG, Nasjletti A. Vasoregulatory function of the heme-heme oxygenase-carbon monoxide system. Am J Hypertens. 2001;14(6 Pt 2):62S–67S
35. Fujita T, Toda K, Karimova A, et al. Paradoxical rescue from ischemic lung injury by inhaled carbon monoxide driven by derepression of fibrinolysis. Nat Med. 2001;7:598–604
36. Brouard S, Otterbein LE, Anrather J, et al. Carbon monoxide generated by heme oxygenase 1 suppresses endothelial cell apoptosis. J Exp Med. 2000;192:1015–1026
37. Liu XM, Chapman GB, Peyton KJ, Schafer AI, Durante W. Carbon monoxide inhibits apoptosis in vascular smooth muscle cells. Cardiovasc Res. 2002;55:396–405
38. Wang X, Wang Y, Kim HP, Nakahira K, Ryter SW, Choi AM. Carbon monoxide protects against hyperoxia-induced endothelial cell apoptosis by inhibiting reactive oxygen species formation. J Biol Chem. 2007;282:1718–1726
39. Powers RW, Roberts JM, Cooper KM, et al. Maternal serum soluble fms-like tyrosine kinase 1 concentrations are not increased in early pregnancy and decrease more slowly postpartum in women who develop preeclampsia. Am J Obstet Gynecol. 2005;193:185–191
40. Cudmore M, Ahmad S, Al-Ani B, et al. Negative regulation of soluble Flt-1 and soluble endoglin release by heme oxygenase-1. Circulation. 2007;115:1789–1797
41. Ros HS, Cnattingius S, Lipworth L. Comparison of risk factors for preeclampsia and gestational hypertension in a population-based cohort study. Am J Epidemiol. 1998;147:1062–1070
42. Wolf M, Shah A, Lam C, et al. Circulating levels of the antiangiogenic marker sFLT-1 are increased in first versus second pregnancies. Am J Obstet Gynecol. 2005;193:16–22
43. Bdolah Y, Lam C, Rajakumar A, et al. Twin pregnancy and the risk of preeclampsia: bigger placenta or relative ischemia? Am J Obstet Gynecol. 2008;198:428.e1–428.e6
44. Kahn SR, Almeida ND, McNamara H, et al. Smoking in preeclamptic women is associated with higher birthweight for gestational age and lower soluble fms-like tyrosine kinase-1 levels: a nested case control study. BMC Pregnancy Childbirth. 2011;11:91
45. Jeyabalan A, Powers RW, Durica AR, Harger GF, Roberts JM, Ness RB. Cigarette smoke exposure and angiogenic factors in pregnancy and preeclampsia. Am J Hypertens. 2008;21:943–947
46. Karumanchi SA, Levine RJ. How does smoking reduce the risk of preeclampsia? Hypertension. 2010;55:1100–1101
47. Pickett KE, Rathouz PJ, Kasza K, Wakschlag LS, Wright R. Self-reported smoking, cotinine levels, and patterns of smoking in pregnancy. Paediatr Perinat Epidemiol. 2005;19:368–376