Antenatal and Early Life Tobacco Smoke Exposure in an African Birth Cohort Study

INT J TUBERC LUNG DIS 20(6):729–737 Q 2016 The Union http://dx.doi.org/10.5588/ijtld.15.0697 E-published ahead of print 12 April 2016

Antenatal and early life tobacco smoke exposure in an African birth cohort study

A. Vanker,* W. Barnett,* K. Brittain,* R. P. Gie,† N. Koen,‡ B. Myers,§ D. J. Stein,‡ H. J. Zar* *Department of Paediatrics and Child Health, Red Cross War Memorial Children’s Hospital, and Medical Research Council (MRC) Unit on Child & Adolescent Health, University of Cape Town, Cape Town, †Department of Paediatrics and Child Health, Tygerberg Children’s Hospital, Stellenbosch University, Cape Town, ‡Department of Psychiatry and Mental Health and MRC Unit on Anxiety & Stress Disorders, University of Cape Town, Cape Town, §Alcohol Tobacco and Other Drug Research Unit, South African Medical Research Council and Department of Psychiatry and Mental Health, University of Cape Town, Cape Town, South Africa

SUMMARY

BACKGROUND: Exposure to tobacco smoke in African (32%) were active smokers on cotinine testing. At birth infants has not been well studied, despite the high burden and at 6–10 weeks of life, respectively 135/241 (56%) of childhood respiratory disease in these communities. and 154/291 (53%) infants had urine cotinine levels OBJECTIVE: To investigate the prevalence of antenatal indicating tobacco smoke exposure. Household smoking and early life tobacco smoke exposure and associations was prevalent and was associated with positive infant with infant birth outcomes in an African birth cohort, cotinine test results. Antenatal maternal smoking was the Drakenstein Child Health Study. associated with decreased infant birthweight-for-age Z- METHODS: Self-report questionnaires assessing mater- score (0.3, 95%CI 0.1–0.5). nal and household smoking were administered. Mater- CONCLUSION: Antenatal and early life tobacco smoke nal and infant urine cotinine testing was conducted exposure is highly prevalent in this community, and may antenatally, at birth and at 6–10 weeks of life to measure impact on birth outcomes and subsequent child health. tobacco smoke exposure. Multivariate regression mod- Smoking cessation interventions are urgently needed to els explored the associations between exposure to smoke reduce tobacco smoke exposure in African communities. and infant birth outcomes. KEY WORDS: tobacco smoke exposure; birth out- RESULTS: Of 789 pregnant women included, 250 comes; birth cohort; child health; maternal health

EXPOSURE TO TOBACCO SMOKE is an impor- Cotinine, a biomarker of tobacco smoking and tant risk factor for childhood respiratory disease,1–3 exposure, has been used to both confirm and quantify and childhood morbidity and mortality worldwide.4,5 smoking and exposure in pregnant women.9,16,17 Infant Prenatal exposure is associated with an increased risk exposure during the first year of life has been assessed of pneumonia and of wheezing disorders,6–8 and may using blood, urine or hair cotinine measures.16,18–20 also lead to pre-term delivery and decreased birth- However, no studies have used infant cotinine mea- weight, predisposing infants to severe respiratory surements at birth to assess in utero tobacco smoke disease.9,10 Other maternal socio-economic and exposure, and few studies have used infant measures to psychosocial risk factors, including depression and evaluate early life exposure.16,18 Documenting exposure intimate partner violence (IPV), may also impact is especially relevant in low-and middle-income country infant birth outcomes.11–13 (LMIC) settings, which carry the highest burden of Tobacco smoke exposure often begins in utero with childhood respiratory illnesses. active or passive maternal smoking, and may We measured antenatal and early postnatal tobac- continue postnatally. Both prenatal and postnatal co smoke exposure, and investigated the association exposure adversely affect infant health. Nicotine between antenatal exposure and infant birth out- exposure may be directly toxic to the airways or comes in an African birth cohort study. may result in secondary impairment of lung growth due to decreased foetal breathing or cellular dam- METHODS age.14 Furthermore, adult smokers have a higher risk of respiratory infections, with increased risk of A prospective study of smoke exposure was under- pathogen transmission to child contacts.15 taken in pregnant women and infants enrolled in the

Correspondence to: Aneesa Vanker, Department of Paediatrics and Child Health, Red Cross War Memorial Children’s Hospital, Klipfontein Road, Rondebosch, Cape Town, South Africa. e-mail: [email protected] Article submitted 16 August 2015. Final version accepted 10 December 2015. 730 The International Journal of Tuberculosis and Lung Disease

Drakenstein Child Health Study.21 The site of the Urine cotinine testing study is located 60 km outside Cape Town, South Maternal and infant urine cotinine tests were Africa, in a semi-rural area with a population of low performed using the IMMULITEw 1000 Nicotine socio-economic status (SES).21 More than 90% of the Metabolite Assay Kit (Siemens Medical Solutions population obtain health care in the public sector, Diagnostics, Glyn Rhonwy, UK). This provided a which has a strong primary health care system. quantitative test using a competitive chemilumines- cent immunoassay, which contained solid-phase Study population, participants and procedure beads coated with polyclonal rabbit anti-cotinine Pregnant women at between 20 and 28 weeks’ antibody. The test had a calibration range of 10–500 gestation were consecutively enrolled using conve- ng/ml, with an analytical sensitivity of 2 ng/ml. Urine nience sampling at one of two primary health care cotinine levels were classified as ,10 ng/ml (non- clinics serving different populations: Newman (pre- smoker), 10–499 ng/ml (passive smoker/exposed) or dominantly mixed race) and Mbekweni (predomi- 7500 ng/ml (active smoker), according to the nantly Black African). A second antenatal study visit manufacturer’s directions. was completed at 28–32 weeks’ gestation. All Maternal urine was collected and tested at the deliveries took place at a single central public second antenatal study visit and at birth, with the hospital, Paarl Hospital. Thereafter, mother-infant higher result used to classify smoking levels. Infant dyads attended follow-up visits, including at 6–10 urine was collected at birth and at 6–10 weeks, either weeks’ postpartum.21 via a urine bag or by placing a cotton-wool ball in the diaper from which urine could then be squeezed. Self-reported measures Urine samples were transferred to a clean, preserva- Sociodemographic data were collected at enrolment tive-free, plastic container and transported at tem- using a questionnaire adapted from the South African peratures of between 28 and 88C to the accredited Stress and Health Study.22 A composite SES score was medical laboratory for testing. developed as an internal comparison of SES for this sample, and participants were categorised as lowest, Ethics low–moderate, moderate–high or highest SES.13 The study was approved by the Faculty of Health Maternal tobacco smoking and exposure were Sciences Human Research Ethics Committees of the assessed using self-report questionnaires at enrolment. University of Cape Town and of Stellenbosch Maternal smoking was quantified as pack-years, University (Cape Town, South Africa), and by the whereby one pack-year was defined as 20 cigarettes Western Cape Provincial Health Research Committee smoked daily for one year. Maternal nicotine depen- (Cape Town, South Africa). Written informed con- dence was assessed using the Fagerstrom¨ test, a well- sent was provided by mothers at enrolment. validated questionnaire that scores tobacco dependence as low, low to moderate, moderate or high.23 Statistical analysis The Alcohol, Smoking and Substance Involvement Tobacco smoking and exposure were compared Screening Test (ASSIST) was administered to assess across the recruitment sites using the v2 or the Fisher’s substance use and substance-related risk.24 Partici- exact tests for categorical variables and Wilcoxon pants were categorised as being at low, moderate or rank-sum tests (Mann-Whitney tests) for non-nor- high risk of tobacco-related health problems, and any mally distributed continuous variables. The sensitiv- self-reported antenatal alcohol use was document- ity and specificity of self-reported maternal smoking ed.24 Women who reported antenatal substance use was calculated using maternal urine cotinine as the were counselled about cessation. gold standard. The associations between household Comprehensive psychosocial data were collected and maternal smoking (self-reported and based on antenatally, including an assessment for depression urine cotinine) and infant urine cotinine were using the Beck Depression Inventory (BDI-II) and a assessed using v2 tests, with cotinine levels of 710 questionnaire for any IPV in the past year.11,13 ng/ml used to categorise infants as being exposed to tobacco smoke. Risk ratios (RRs) with 95% confi- Birth outcomes dence intervals (CIs) were calculated to determine the All births were attended by a member of the study strength of these associations. Infant weight-for-age team. Birth outcomes included birth weight, gesta- (WfA) Z-scores were calculated using the revised tional age and presence of respiratory or other Fenton preterm growth charts, and infants with a disease. Gestational age at delivery (in completed WfA Z-score ,10th percentile were classified as small weeks) was calculated from an antenatal ultrasound. for gestational age (SGA).25,26 The associations If no ultrasound was available, then fundal height at between each of maternal and infant cotinine levels, enrolment or maternal recall of last menstrual period potential confounders and infant birth outcomes was used. were explored in regression models. Data were Tobacco smoke exposure in a birth cohort 731

Table 1 Maternal baseline demographic characteristics and infant birth outcomes

Mbekweni T C Newman Total Variable n (%) n (%) n (%) P value Maternal baseline demographic characteristics Number of mothers 412 (52) 377 (48) 789 (100) Age at enrolment, years, median [IQR] 26.7 [22.2 to 31.8] 24.8 [21.3 to 29.1] 25.7 [21.8 to 30.8] ,0.001 Race Black 408 (99) 5 (1) 413 (52) ,0.001 Mixed race 4 (1) 372 (99) 376 (48) Married/cohabiting 145 (35) 159 (42) 304 (39) 0.044 HIV-infected 154 (37) 11 (3) 165 (21) ,0.001 Educational attainment Primary 41 (10) 29 (8) 70 (9) 0.062 Some secondary 219 (53) 189 (50) 408 (52) Completed secondary 124 (30) 143 (38) 267 (34) Any tertiary 28 (7) 16 (4) 44 (6) Unemployed 321 (78) 262 (70) 583 (74) 0.007 Average household income, ZAR/month ,1000 201 (49) 130 (34) 331 (42) ,0.001 1000–5000 163 (40) 176 (47) 339 (43) .5000 48 (12) 71 (19) 119 (15) SES quartile Lowest SES 143 (35) 70 (19) 213 (27) ,0.001 Low-moderate SES 106 (26) 90 (24) 196 (25) Moderate-high SES 96 (23) 97 (26) 193 (25) Highest SES 67 (16) 120 (32) 187 (24) Type of home Informal housing 179 (43) 103 (27) 282 (36) ,0.001 House/flat 233 (57) 274 (73) 507 (64) Household members, n, median [IQR] 4 [3 to 6] 5 [4 to 7] 4 [3 to 6] ,0.001 Antenatal depression: above threshold 87 (21) 76 (20) 163 (21) 0.740 Recent IPV: above threshold 107 (26) 148 (39) 255 (32) ,0.001 Any self-reported antenatal alcohol use 33 (8) 104 (28) 137 (17) ,0.001 Infant birth outcomes Number of infants 415 (52) 377 (48) 792 (100) Sets of twins 3 0 3 Female sex 205 (49) 165 (44) 370 (47) 0.113 Gestation at delivery, weeks, median [IQR] 39 [38 to 40] 39 [37 to 40] 39 [38 to 40] 0.271 Pre-term birth (,37 weeks’ gestation) 65 (16) 58 (15) 123 (16) 0.914 Birth weight, g, median [IQR] 3130 [2800 to 3440] 2980 [2590 to 3350] 3080 [2690 to 3410] ,0.001 WfA Z-score, median [IQR] 0.4 [1.2 to 0.2] 0.7 [1.4 to 0.1] 0.6 [1.3 to 0.1] ,0.001 Low birth weight (,2500 g) 48 (12) 74 (20) 122 (15) 0.002 Small for gestational age 96(23) 110(29) 206(26) 0.053 Respiratory distress at birth 23 (6) 20 (5) 43 (5) 0.883

IQR ¼ interquartile range; HIV ¼ human immunodeficiency virus; ZAR ¼ South African rand; SES ¼ socio-economic status; IPV ¼ intimate partner violence; WfA ¼ weight-for-age. analysed using Stata v. 12 (StataCorp, College mothers screened above the threshold for antenatal Station, TX, USA). depression, and almost a third reported having experienced recent IPV. Self-reported antenatal alcohol use was significantly higher among mixed race (28%) RESULTS than Black African (8%) participants (Table 1). Data from 789 mothers enrolled between March 2012 and October 2014 were included; 792 infants, Maternal smoking including three sets of twins, were born, and data Overall, 24% of mothers reported antenatal smok- from 720 mother-infant dyads at 6–10 weeks’ ing. Most (94%) reported a daily smoking habit of a postpartum were included. median of 1.2 pack-years (IQR 0.4–2.5). Nicotine The median age of the mothers at enrolment was dependence was classified as low or low to moderate 25.7 years (interquartile range [IQR] 21.8–30.8). A in 77% of smokers. However, there was a moderate high prevalence of unemployment (74%) was ob- to high risk of tobacco-related problems in 97% of served, and most participants had not completed the smoking mothers (Table 2). secondary education. Although this is a population Based on urine cotinine levels, 250 (32%) mothers of low overall SES, the mixed race community were classified as active smokers and 366 (46%) as (Newman) was of higher SES than the Black African having passive smoke exposure. The prevalence of community (Mbekweni). Twenty-one per cent of both self-reported and active smoking based on urine 732 The International Journal of Tuberculosis and Lung Disease

Table 2 Antenatal and early postpartum tobacco smoke exposure

Mbekweni T C Newman Total n (%) n (%) n (%) P value Maternal antenatal tobacco smoking 412 (52) 377 (48) 789 (100) Self-reported current smoking 17 (4) 174 (46) 191 (24) ,0.001 Frequency of smoking (n ¼ 191) Daily 12 (71) 163 (94) 175 (92) 0.007 A few times per week/month 5 (29) 11 (6) 16 (8) Cigarettes smoked daily 610 16 (94) 163 (94) 179 (94) 0.620 11–20 1 (6) 5 (3) 6 (3) 21–30 0 2 (1) 2 (1) 731 0 2 (1) 2 (1) Years smoked, median [IQR] (n ¼ 191) 4 [2–6] 6 [4–10] 6 [4–10] 0.010 Pack-years smoked, median [IQR] (n ¼ 191) 0.1 [0.1–0.3] 1.4 [0.5–2.7] 1.2 [0.4–2.5] ,0.001 Nicotine dependence (Fagerstrom¨ test) (n ¼ 191) Low 6 (35) 79 (45) 85 (45) 0.551 Low-moderate 8 (47) 54 (31) 62 (32) Moderate 3 (18) 39 (22) 42 (22) High 0 2 (1) 2 (1) Risk of tobacco-related problems (ASSIST) (n¼183) Lower risk 0 4 (2) 4 (2) 0.799 Moderate risk 10 (83) 128 (75) 138 (75) High risk 2 (17) 39 (23) 41 (22) Urine cotinine (n ¼ 789), ng/ml* ,10 (non-smoker) 134 (33) 39 (10) 173 (22) ,0.001 10–499 (passive/exposed) 221 (54) 145 (38) 366 (46) 7500 (active smoker) 57 (14) 193 (51) 250 (32) Early postpartum tobacco smoke exposure Number of family/household smokers (n ¼ 720) 0 215 (57) 56 (16) 271 (38) ,0.001 1 124 (33) 87 (25) 211 (29) 2 31 (8) 90 (26) 121 (17) 73 6 (2) 111 (32) 117 (16) Infant urine cotinine Urine cotinine at birth, ng/ml (n ¼ 241) ,10 76 (61) 30 (26) 106 (44) ,0.001 10–499 43 (35) 48 (41) 91 (38) 7500 5 (4) 39 (33) 44 (18) Urine cotinine at 6–10 weeks, ng/ml (n ¼ 291) ,10 98 (72) 39 (25) 137 (47) ,0.001 10–499 38 (28) 107 (69) 145 (50) 7500 1 (0.7) 8 (5) 9 (3)

* Antenatal and birth combined, i.e., ‘active smoker’ if cotinine 7500 ng/ml for either timepoint. IQR ¼ interquartile range; ASSIST ¼ Alcohol, Smoking and Substance Involvement Screening Test.

Table 3 Associations between tobacco smoke exposure and infant urine cotinine

Infant urine cotinine ,10 ng/ml 710 ng/ml Variable n (%) n (%) RR (95%CI) P value Associations with infant urine cotinine at birth (n ¼ 241) Self-reported maternal antenatal smoking Non-smoker 105 (56) 81 (44) Reference Smoker 1 (2) 54 (98) 2.3 (1.9–2.7) ,0.001 Maternal antenatal urine cotinine* Non-smoker (,10 ng/ml) 42 (93) 3 (7) Reference Passive/exposed (10–499 ng/ml) 63 (51) 61 (49) 7.4 (2.4–22.3) ,0.001 Active smoker (7500 ng/ml) 1 (1) 71 (99) 14.8 (5.0–44.2) ,0.001 Associations with infant urine cotinine at 6–10 weeks (n ¼ 291) Number of family/household smokers 0 74 (76) 23 (24) Reference 1 49 (52) 45 (48) 2.0 (1.3 3.1) ,0.001 2 11 (24) 35 (76) 3.2 (2.2–4.7) ,0.001 73 2 (4) 49 (96) 4.1 (2.8–5.8) ,0.001

* Antenatal and birth combined, i.e., ‘active smoker’ if cotinine 7500 ng/ml for either timepoint. RR ¼ risk ratio; CI ¼ confidence interval. Tobacco smoke exposure in a birth cohort 733

cotinine levels was higher among mixed race than among Black African mothers (Table 2). The sensitivity of self-reported smoking compared to urine (%)

n cotinine was much lower for Black African (26%) than for mixed race (85%) participants. However,

Respiratory distress specificity was high in both groups.

Infant birth outcomes

(%) Only two (0.2%) stillbirths occurred. The median n gestational age at delivery was 39 weeks (IQR 38–40) No distress overall, with 16% (mostly late) pre-term births. Median birthweight was significantly higher in Black

(%) African (3130 g, IQR 2800–3440) than in mixed race SGA n infants (2980 g, IQR 2590–3350). The median WfA Z-score at birth was 0.6 (IQR 1.3 to 0.1) overall, 0.001 0.234

, and was substantially lower in mixed race infants. (%)

n Low birth weight (LBW; i.e., ,2500 g) was observed

AGA/LGA in 15% of infants, and more than a quarter were

large for gestational age. SGA. Only 45 (6%) infants had documented respi- ¼ ratory distress at birth (Table 1). Most mother-infant (%) 2500 g

n pairs (83%) were discharged from hospital within 48 2500 g) Infant SGA at birth Infant respiratory distress at birth ,

, h of delivery. 0.001

, Infant smoke exposure (%)

2500 g Household smokers, including mothers, were report- n Association with: 7 ed in almost two thirds of the homes. The prevalence of household smoking was significantly higher in

0.2] 193 (78) 57 (23) 160 (64) 90 (36) 232 (93)mixed 18 (7) race families, where one or more household appropriate for gestational age; LGA -score members reportedly smoked in 83% of the homes Z ¼ (Table 2). Excluding maternal smokers, the reported 0.001 1.2 to 0.3] 154 (89) 19 (11) 133 (77) 40 (23) 163 (94) 10 (6) 1.1 to 0.2] 323 (88) 46 (12) 293 (79) 76 (21) 354 (96) 15 (4) 1.5 to , number of other family and household smokers was

median [IQR] high at both sites, with 142 (38%) Black African and 0.4 [ 0.5 [ 0.9 [ Infant WfA 264 (77%) mixed race participants reporting other smokers in the home. Urine cotinine measures were obtained from 241 (%) n

Pre-term infants at birth and 291 infants at 6–10 weeks. At

small for gestational age; AGA birth, 56% of infants had urine cotinine indicating ¼ exposure, with 18% having levels comparable with (%) Pre-term birth Infant LBW (

n active smoking; this was significantly higher in mixed Full-term race than in Black African infants (33% vs. 4%)

500 ng/ml for either timepoint. (Table 2). The prevalence of passive exposure was 7 similar across sites.

low birth weight; SGA At 6–10 weeks, half (50%) of the infants had ¼ cotinine levels indicative of passive smoke exposure, with a much higher prevalence in mixed race than in median [IQR] Black African infants (69% vs. 28%). Few infants Gestation at delivery (3%) had levels indicative of active smoking (Table 2). Higher infant urine cotinine levels at birth were

weight-for-age; LBW associated with maternal smoking, both self-reported ¼ and based on cotinine measurements; infants born to mothers classified as active smokers had an almost

500 ng/ml) 39 [37 to 40] 208 (83) 42 (17) 15-fold increased risk of testing positive for smoke 7 10 ng/ml) 39 [38 to 40] 149 (86) 24 (14)

, exposure. At 6–10 weeks, infant urine cotinine levels

Unadjusted associations between tobacco smoke exposure and infant birth outcomes were associated with an increasing number of household smokers, with infants exposed to 73 interquartile range; WfA

¼ household smokers having a four-fold increased risk value 0.135 0.715 IQR Table 4 Maternal antenatal urine cotinine* * Antenatal and birth combined, i.e., ‘active smoker’ if cotinine Non-smoker ( Passive/exposed (10–499 ng/ml) 39 [38 to 40] 312 (85) 57 (15) Active smoker ( P of testing positive for smoke exposure (Table 3). 734 h nentoa ora fTbruoi n ugDisease Lung and Tuberculosis of Journal International The

Table 5 Adjusted associations between tobacco smoke exposure and infant birth outcomes

A) Association with WfA Z-score B) Association with LBW (,2500 g) C) Association with SGA Variable Regression coefficient (95%CI) Adjusted regression coefficient (95%CI) RR (95%CI) aRR (95%CI) RR (95%CI) aRR (95%CI) Maternal urine cotinine* Non-smoker (,10 ng/ml) Reference Reference Reference Reference Reference Reference Passive/exposed (10–499 ng/ml) 0.1 (0.2 to 0.1) 0.004 (0.2 to 0.2) 1.1 (0.7 to 1.9) 1.0 (0.6 to 1.6) 1.1 (0.6 to 2.0) 0.8 (0.6 to 1.2) Active smoker (7500 ng/ml) 0.5 (0.7 to 0.3) 0.3 (0.5 to 0.1) 2.1 (1.3 to 3.4) 1.4 (0.8 to 2.4) 2.3 (1.3 to 4.1) 1.3 (0.9 to 1.9) Recruitment site T C Newman Reference Reference Reference Reference Reference Reference Mbekweni 0.3 (0.1 to 0.4) 0.1 (0.05 to 0.3) 0.6 (0.4 to 0.8) 0.6 (0.4 to 0.9) 0.8 (0.6 to 1.0) 0.9 (0.7 to 1.2) Maternal age at enrolment 0.02 (0.01 to 0.03) 0.01 (0.001 to 0.03) 1.0 (1.0 to 1.02) 1.0 (1.0 to 1.0) Gravidity Multigravida Reference Reference Reference Primigravida 0.4 (0.5 to 0.2) 1.2 (0.9 to 1.7) 1.6 (1.3 to 2.0) SES quartile Highest SES Reference Reference Reference Reference Reference Reference Moderate-high SES 0.3 (0.5 to 0.1) 0.3 (0.5 to 0.1) 1.6 (0.9 to 2.7) 1.6 (0.9 to 2.8) 1.7 (1.2 to 2.5) 1.7 (1.1 to 2.4) Low-moderate SES 0.2 (0.4 to 0.001) 0.2 (0.4 to 0.02) 1.6 (1.0 to 2.8) 1.7 (1.0 to 2.9) 1.6 (1.1 to 2.3) 1.5 (1.0 to 2.2) Lowest SES 0.2 (0.4 to 0.02) 0.2 (0.4 to 0.003) 2.2 (1.3 to 3.6) 2.4 (1.4 to 4.1) 1.6 (1.1 to 2.4) 1.7 (1.1 to 2.5) Maternal HIV infection Non-HIV-infected Reference Reference Reference HIV-infected 0.1 (0.03 to 0.3) 0.8 (0.5 to 1.2) 0.9 (0.6 to 1.2) Antenatal depression Below threshold Reference Reference Reference Above threshold 0.2 (0.4 to 0.004) 1.0 (0.7 to 1.5) 1.3 (1.0 to 1.7) Recent IPV Below threshold Reference Reference Reference Above threshold 0.2 (0.3 to 0.0001) 1.5 (1.1 to 2.0) 1.2 (1.0 to 1.6) Antenatal alcohol use No alcohol use Reference Reference Reference Reference Reference Any alcohol use 0.4 (0.6 to 0.2) 0.3 (0.4 to 0.1) 1.4 (0.9 to 2.0) 1.5 (1.1 to 1.9) 1.3 (1.0 to 1.7)

* Antenatal and birth combined, i.e., ‘active smoker’ if cotinine 7500 ng/ml for either timepoint. WfA ¼ weight-for-age; LBW ¼ low birthweight; SGA ¼ small for gestational age; CI ¼ confidence interval; RR ¼ risk ratio; aRR ¼ adjusted RR; SES ¼ socio-economic status; HIV ¼ human immunodeficiency virus; IPV ¼ intimate partner violence. Tobacco smoke exposure in a birth cohort 735

Tobacco smoke exposure and birth outcomes affecting participants were considerable, with a high Associations were observed between maternal urine prevalence of antenatal depression and recent IPV. cotinine and a number of adverse birth outcomes in The high prevalence of self-reported smoking was crude analysis. Maternal smoking was significantly confirmed by maternal urine cotinine testing, with associated with a reduction in infant WfA Z-score, surprisingly high sensitivity of self-report, especially and with an increased risk of LBWand SGA (Table 4). in the mixed race community (85%). The lower Although associations between maternal urine sensitivity and lower prevalence of smoking in the cotinine levels and LBW and SGA were no longer black African population may reflect cultural differ- significant in adjusted analyses, active smoking ences in acceptability of cigarette smoking, perhaps 30 remained significantly associated with decreased contributing to under-reporting. WfA. Infants born to mothers who were active Although nicotine dependence was low or low-to- smokers during pregnancy had, on average, a 0.3 moderate in most mothers, this may be explained by (IQR 0.1–0.5) Z-score decrease in WfA at birth patterns of cigarette use in these impoverished compared to infants born to non-smoking mothers, communities, where the cost of cigarettes often independent of recruitment site, maternal age, SES results in fewer cigarettes smoked. In contrast, and antenatal alcohol use. In addition, antenatal ASSIST scores indicated that the majority were at moderate-to-high risk of tobacco-related health alcohol use remained an independent predictor of problems and could benefit from tobacco cessation decreased WfA, and was associated with a 0.3 (IQR programmes. While the most effective smoking 0.1–0.4) Z-score decrease in WfA (Table 5). cessation programmes rely on a combination of Similar results were observed when comparing counselling and medication interventions, given the birth outcomes across categories of infant urine potentially low rates of physical dependency (based cotinine measured at birth. No associations were on Fagerstrom¨ test findings), brief behavioural observed between infant cotinine and any of the interventions that include building readiness and following: gestational age at delivery, pre-term birth, motivation to change may be useful for facilitating LBWor respiratory distress at birth. In crude analysis, reductions in tobacco-related health problems; how- however, infant cotinine was significantly associated ever, medication-assisted therapy may be necessary with both a reduction in WfZ Z-score and an for those showing higher rates of dependency.31,32 increased likelihood of SGA. It should be noted that Household smokers were another major source of the association between infant cotinine levels and ongoing smoke exposure, with household smoking decreased WfA persisted in adjusted analyses. Infants again more prevalent in the mixed race (84%) than in in the active smoker category had, on average, a 0.6 the Black African community (43%). An increasing (IQR 0.2–1.1) Z-score decrease in WfA at birth number of household smokers was associated with compared to infants in the non-smoker category, positive urine cotinine measurements in infants, regardless of recruitment site, SES, maternal age and confirming more exposure. maternal antenatal alcohol use (results not shown). Such intense, early exposure to tobacco smoke may have profound effects on child health. In this study, DISCUSSION smoke exposure was associated with decreased birth weight, a well-described association in both middle- This study shows an alarmingly high prevalence of and high-income countries.9,10,33 This association objectively assessed antenatal and early life tobacco persisted even when adjusted for potential confound- smoke exposure in infants in this low SES population. ers. Smoke-exposed children have a higher risk of Almost two thirds of infants had evidence of smoke developing pneumonia, wheezing disorders and exposure at birth, while more than 50% were chronic respiratory diseases.6,7,15,34 The cumulative exposed at 6–10 weeks of age. Furthermore, the and long-term effects of early and ongoing exposure degree of smoke exposure is worrying, with 18% of on child health are currently being studied longitudi- infants having birth urine cotinine values in the nally in this cohort. ‘active smoker’ range. This prevalence of smoke Several limitations of these findings must be noted, exposure is much higher than previously reported including the fairly small sample size. In addition, globally.2 Notably, smoke exposure was associated although maternal urine cotinine was obtained from with lower birth weight, a risk factor for respiratory all mothers, urine collection was somewhat less disease and other long-term health effects.9,10,27 successful in infants. As the half-life of cotinine is Mothers were a major source of smoke exposure, 17 h, the prevalence of active smoking or smoke especially in the mixed race population, with 46% of exposure may have been under-estimated if a urine mixed race women reporting antenatal smoking. This sample was obtained after this time period. These is as much as 10 times the global estimate for data therefore provide a minimum estimate of pregnant women in LMICs, particularly in Afri- smoking and exposure. This analysis was limited to ca.2,28,29 Furthermore, the psychosocial stressors tobacco smoke exposure in the antenatal and early 736 The International Journal of Tuberculosis and Lung Disease postnatal period; further study of ongoing smoke weight gain, and infant birthweight. Addict Behav 2003; 28: exposure is underway. 55–66. To our knowledge, this is the first study to quantify 11 Koen N, Wyatt G E, Williams J K, et al. Intimate partner violence: associations with low infant birthweight in a South exposure to tobacco smoke at birth using a combi- African birth cohort. Metab Brain Dis 2014; 29: 281–299. nation of maternal and infant measures as well as 12 Quispel C, Lambregtse-van den Berg M P, Steegers E A, subjective and objective assessments, and highlights Hoogendijk W J, Bonsel G J. Contribution of psychopathology, the urgent need to develop and implement effective psychosocial problems and substance use to urban and rural interventions for smoking cessation in pregnant differences in birth outcomes. Eur J Public Health 2014; 24: 15,31,32 917–923. women. In addition, public health interven- 13 Stein D J, Koen N, Donald K A, et al. Investigating the tions and educational initiatives that highlight the psychosocial determinants of child health in Africa: the risk of household tobacco smoking to adult and child Drakenstein Child Health Study. J Neurosci Methods 2015; health are urgently needed among vulnerable groups 252: 27–35. in poor communities. 14 Sly P D. The early origins of asthma: who is really at risk? Curr Opin Allergy Clin Immunol 2011; 11: 24–28. Acknowledgements 15 van Zyl-Smit R N, Pai M, Yew W W, et al. Global lung health: the colliding epidemics of tuberculosis, tobacco smoking, HIV The authors thank the study and clinical staff at Paarl Hospital and COPD. Eur Respir J 2010; 35: 27–33. (Paarl) and Mbekweni and Newman Clinics, and the laboratory 16 Jordanov J S. Cotinine concentrations in amniotic fluid and staff at the Red Cross War Memorial Children’s Hospital (Cape urine of smoking, passive smoking and non-smoking pregnant Town) and Metropolis Laboratory (Cape Town, South Africa). women at term and in the urine of their neonates on 1st day of They also thank the participants and their families. life. Eur J Pediatr 1990; 149: 734–737. The study was funded by Bill & Melinda Gates Foundation 17 Searles Nielsen S, Dills R L, Glass M, Mueller B A. Accuracy of (Seattle, WA, USA; OPP1017641), the Discovery Foundation prenatal smoking data from Washington State birth certificates (Johannesburg, South Africa) (1997/1), the South African Thoracic in a population-based sample with cotinine measurements. Ann Society (Cape Town, South Africa) AstraZeneca Respiratory Epidemiol 2014; 24: 236–269. Fellowship, National Research Fund, South Africa (Pretoria, South 18 Spector L G, Murphy S E, Wickham K M, Lindgren B, Joseph A Africa; 87760), the Clinical Infectious Diseases Research Initiative M. Prenatal tobacco exposure and cotinine in newborn dried Wellcome Trust Clinical Fellowship (London, UK; 084323), and blood spots. Pediatrics 2014; 133: e1632–e1638. the South Africa Medical Research Council Clinician Scholarship 19 Eliopoulos C, Klein J, Phan M K, et al. Hair concentrations of (Tygerberg, South Africa). nicotine and cotinine in women and their newborn infants. Conflicts of interest: none declared. JAMA 1994; 271: 621–623. 20 Ashford K B, Hahn E, Hall L, Rayens M K, Noland M, Collins References R. Measuring prenatal secondhand smoke exposure in mother- baby couplets. Nicotine Tob Res 2010; 12: 127–135. 1 Tobacco Free Initiative, World Health Organization. Protection 21 Zar H J, Barnett W, Myer L, Stein D J, Nicol M P. Investigating from exposure to second-hand tobacco smoke. Policy the early-life determinants of illness in Africa: the Drakenstein recommendations. Geneva, Switzerland: WHO, 2007. Child Health Study. Thorax 2015; 70: 592–594. 2 Oberg¨ M, Jaakkola M S, Woodward A, Peruga A, Pruss-Ust¨ un¨ 22 Stein D J, Seedat S, Herman A A, et al. Findings from the first A. Worldwide burden of disease from exposure to second-hand South African Stress and Health Study: policy brief. Cape smoke: a retrospective analysis of data from 192 countries. Town, South Africa: Medical Research Council, 2007. Lancet 2011; 377: 139–146. 23 Heatherton T F, Kozlowski L T, Frecker R C, Fagerstrom K O. 3 Ferkol T, Schraufnagel D. The global burden of respiratory ¨ disease. Ann Am Thorac Soc 2014; 11: 404–406. The Fagerstrom¨ Test for nicotine dependence: a revision of the 4 Zar H J, Ferkol T W. The global burden of respiratory disease— Fagerstrom¨ Tolerance Questionnaire. Br J Addict 1991; 86: impact on child health. Pediatr Pulmonol 2014; 49: 430–434. 1119–1127. 5 Walker C L, Rudan I, Liu L, et al. Global burden of childhood 24 Humeniuk R E, Henry-Edwards S, Ali R L, Poznyak V, pneumonia and diarrhoea. Lancet 2013; 381: 1405–1416. Monteiro M. The Alcohol, Smoking and Substance 6 Stern D A, Morgan W J, Halonen M, Wright A L, Martinez F D. Involvement Screening Test (ASSIST): manual for use in Wheezing and bronchial hyper-responsiveness in early primary care. Geneva, Switzerland: World Health childhood as predictors of newly diagnosed asthma in early Organization, 2010. adulthood: a longitudinal birth-cohort study. Lancet 2008; 372: 25 Fenton T R, Nasser R, Eliasziw M, Kim J H, Bilan D, Sauve R. 1058–1064. Validating the weight gain of preterm infants between the 7 Jones L L, Hashim A, McKeever T, Cook D G, Britton J, reference growth curve of the fetus and the term infant. BMC Leonardi-Bee J. Parental and household smoking and the Pediatr 2013; 13: 92. increased risk of bronchitis, bronchiolitis and other lower 26 Fenton T R, Kim J H. A systematic review and meta-analysis to respiratory infections in infancy: systematic review and meta- revise the Fenton growth chart for preterm infants. BMC analysis. Respir Res 2011; 12: 5. Pediatr 2013; 13: 59. 8 Ferrante G, Antona R, Malizia V, Montalbano L, Corsello G, 27 Borja J B. The impact of early nutrition on health: key findings La Grutta S. Smoke exposure as a risk factor for asthma in from the Cebu Longitudinal Health and Nutrition Survey childhood: a review of current evidence. Allergy Asthma Proc (CLHNS). Malays J Nutri 2013; 19: 1–8. 2014; 35: 454–461. 28 Mackay J, Amos A. Women and tobacco. Respirology 2003; 8: 9 Wang X, Tager I B, Van Vunakis H, Speizer F E, Hanrahan J P. 123–130. Maternal smoking during pregnancy, urine cotinine 29 Caleyachetty R, Tait C A, Kengne A P, Corvalan C, Uauy R, concentrations, and birth outcomes. A prospective cohort Echouffo-Tcheugui J B. Tobacco use in pregnant women: study. Int J Epidemiol 1997; 26: 978–988. analysis of data from demographic and health surveys from 54 10 Secker-Walker R H, Vacek P M. Relationships between low-income and middle-income countries. Lancet Global cigarette smoking during pregnancy, gestational age, maternal Health 2014; 2: e513–e520. Tobacco smoke exposure in a birth cohort 737

30 Steyn K, Bradshaw D, Norman R, Laubscher R, Saloojee Y. 32 van Zyl-Smit R N, Allwood B, Stickells D, et al. South African Tobacco use in South Africans during 1998: the first tobacco smoking cessation clinical practice guideline. S Afr demographic and health survey. J Cardiovasc Risk 2002; 9: Med J 2013; 103: 869–876. 161–170. 33 Levy D, Jiang M, Szklo A, de Almeida L M, Autran M, Bloch 31 Everett-Murphy K, Steyn K, Mathews C, et al. The effectiveness M. Smoking and adverse maternal and child health outcomes in of adapted, best practice guidelines for smoking cessation Brazil. Nicotine Tob Res 2013; 15: 1797–1804. counseling with disadvantaged, pregnant smokers attending 34 Martinez F D. The origins of asthma and chronic obstructive public sector antenatal clinics in Cape Town, South Africa. Acta pulmonary disease in early life. Proc Am Thorac Soc 2009; 6: Obstet Gynecol Scand 2010; 89: 478–489. 272–277. Tobacco smoke exposure in a birth cohort i

RESUME

CONTEXTE : L’expositiona ` la fumée de tabac n’a pas test de nicotine. A la naissance eta6` ` a 10 semaines de et´´ e bienetudi´ ´ ee chez les nourrissons africains en dépit du vie, respectivement 135/241 (56%) et 154/291 (5%) des lourd fardeau constitué par les maladies respiratoires des nourrissons avaient des niveaux de nicotine urinaire enfants dans ces communautés. témoignant d’une expositiona ` la fumée de tabac. Le OBJECTIF : Rechercher la prévalence de l’expositiona ` tabagismea ` domicileetait ´ prévalent etetait ´ associéà une la fumée de tabac anténatale et au début de la vie et les recherche urinaire de nicotine positive chez le bébé. Le associations avec l’état du bébéà la naissance dans une tabagisme anténatal maternel aet´ ´ e associéàune cohorte africaine de nouveau-nés, dans le cadre de diminution du score Z du poids de naissance du bébé l’étude Drakenstein sur la santé de l’enfant. (0,3 ; IC95% 0,1–0,5). ME´ THODE : Des questionnaires ontet´ ´ e remplis afin CONCLUSION : L’expositionalafum´ ` ee de tabac d’évaluer le tabagisme maternel et domiciliaire. La anténatal et au début de la vie est très prévalente au recherche de nicotine dans l’urine de la mère et du sein de cette communauté et a un impact sur l’état du bébéaétéréalisée avant la naissance, lors de la naissance bébéà la naissance et l’état de santéultérieur de l’enfant. eta6` ` a 10 semaines de vie afin de mesurer l’expositiona ` Des interventions visantal’arrˆ ` et du tabac sont la fumée de tabac. Des modèles de régression multivariée nécessaires de fa¸con urgente afin de réduire l’exposition ont exploré les associations entre l’expositiona ` la fumée des communautés africainesa ` la fumée de tabac. et l’état du bébéà la naissance. RE´ SULTATS : Sur 789 femmes enceintes incluses, 250 (32%) ontet´ ´ e classées comme fumeuses actives selon le

RESUMEN

MARCO DE REFERENCIA: La exposicion ´ de los RESULTADOS: De las 789 embarazadas que lactantes africanos al humo de tabaco no se ha participaron en el estudio, 250 eran fumadoras activas estudiado plenamente, pese a la alta carga de segun ´ la prueba de la cotinina (32%). Los recién nacidos morbilidad por enfermedades respiratorias de la presentaron concentraciones urinarias de cotinina infancia en estas comunidades. indicativas de exposicion ´ al humo de tabaco al OBJETIVO: Investigar la prevalencia de exposicion ´ nacimiento (135/241; 56%) y entre la sexta y la prenatal y en los primeros meses de vida al humo de décima semana de vida (154/291; 53%). El tabaco y su asociacion ´ con los resultados obstétricos en tabaquismo en el hogar fue predominante y se asocio ´ una cohorte de nacimiento africana del estudio con pruebas de cotinina positivas en los lactantes. El Drakenstein sobre la salud de los ninos.˜ tabaquismo prenatal se asocio ´ con un bajo peso al nacer ME´ TODOS: Se utilizaron cuestionarios (puntuacion ´ de la desviacion ´ estańdar 0,3; IC95% 0,1– autoadministrados con el fin de evaluar el tabaquismo 0,5). materno y el tabaquismo en los hogares. Se practicola ´ CONCLUSIO´ N: En la comunidad estudiada es muy prueba de cotinina en orina de la madre y del hijo antes frecuente la exposicion ´ prenatal y temprana de los del parto, en el momento del nacimiento y entre la sexta lactantes al humo de tabaco y puede tener consecuencias yladécima semana de vida, con el proposito ´ de medir la en los resultados obstétricos y en la salud futura del nino.˜ exposicion ´ al humo de tabaco. Mediante modelos de Se precisan con urgencia intervenciones de apoyo al regresion ´ multivariante se examinaron las asociaciones abandono del tabaquismo con el fin de disminuir esta entre el humo de tabaco y los resultados obstétricos. exposicion ´ en las comunidades africanas.