Tobacco Documents Online

DAVID H. IZUUIN PE-rER A. KRASILNIKOFF JOHN M. LEVENI'HAL BIRGITTE WEILE ARNE BERGET Department of Paedratriu and Obstetrics and Gytraecology, University of Copenhagen, Gentofte Hospital, Denmark; Division of Ambulatory Care, Department of Pediatrics, Albert Einstein College of Medicine, Bronx Municipal Hospital Center, New York; and Department of Pe`diatrics, Yale Urtiversity School of Medicine, New Haven, Connecticra, USA Summary 500 consecutive Danish women who had full-tenn babies were interviewed on the third or fourth day post parttun and asked about smoking in all household members. Exposure to smoking by the mother was found to reduce birth-weight, and indirect or passive exposure to smoking by the father had nearly as large (66°„) an effect. On average, birth-weight was reduced by 120 g per pack of cigarettes (or cigar/pipe equivalent) smoked per day by the father. This relation remained statistically significant after controlling for mother's age, parity, alcohol and tobacco consumption during pregnancy, illness during pregnancy, and social class and sex of the baby. The effect of passive smoking was greatest in the lower social classes. Introduction CIGARE'I'1'E smoking during pregnancy is associated with an increased risk of spontaneous abortion and perinatal mortality and reduced matemal weight gain, gestation, and birth-weight.' Birth-weight is highly dependent on the average number of cigarettes smoked per day during prcgnancy,' Attention has latcly been focused on the effects of indirect exposure to tobacco smoke from the mother's proximity to other smokers. Passively inhaled smoke, also called side- stream or second-hand smoke, appears to be hazardous in its own right.3 Studies have found a direct relation between passive smoking and childhood asthma,'S persistent wheezing,° and respiratory illness in the first' and second" years of life. Passive exposure to cigarette smoking has also been shown to reduce pulmonary function in children.9 There are few reports of the effect of passive smoking on birth-weight. In one comparison of pregnant women who were exposed to side-stream smoke for at least 2 h each day with women who were not exposed, the relative risk of having a low-birth-weight baby was found to be raised.10 The aim of the present study was to examine the quantitative effect of smoking in household members on birth-weight. Subjects and Methods This study was conducted at the Gentofte University Hospital, Hcllcrup, Denmark. About 2000 babies are delivered each year at the hospital, which serves the north-east region of greater Copenhagen. As part of a larger, prospective study to investigate the relation between infant feeding and infectious morbidity, consecutive women were interviewed on the third or fourth day after delivery if their babies met the following criteria: birth-weight _> 2000 g; gestational age >_ 36 weeks; and no evidence of serious congenital defects or underlying illness. This population was chosen to exclude any babies with an excess risk of infection from other causes. Each interview was conducted by the principal investigator or one of three research assistants who were trained to use a structured pre-ooded questionnaire in Danish. For the purpose of this study the interview included standard demographic measurements and quantitative assessments of average daily tobacco intake and average weekly alcohol (pttsumption by all household members during the mother's pregnancy. Information about pregnancy and delivery was extracted from medical records. The questionnaire was piloted to ensure that information was obtained and recorded in a uniform manner by the four interviewers. We found no difference between interviewers with respect to questions about demographic variables, tobacco intake, or alcohol eonsumption, Demographic measurements were made with the Hollingshead classification of social class" and standardised Danish measure of social class based on the mother's job and education.' Both classifications use a scale from 1(highest social class) to 5 (lowest). We found no significant difference between the results of the Hollingshead and Danish social class ratings and have used the Danish classification in this report. Questions about smoking reflected the method and quantity of intake per day by each household member. Method of tobacco intake included any combination of cigarettes, pipe bowls, or cigars. Quantity of intake was the average number of cigarettes andJor pipe bowls and/or cigars smoked per day. Since most of the second smokers in families were fathers (n = 487), all household smokers were defined as such. Questions about alcohol consumption focused on the type and quantity per week of bottles of beer, glasses of wine, and measures of spirits consumed by each household member. A drink was defined as the volume of beverage containing 15 ml of absolute alcohol. This amount is equal to one bottle of 4"„ beer (360 ml), one glass of 12°„ wine (120 ml) or one measure of 80 proof spirit (36 ml). Statistical Methods Multiple regression analyses were used to estimate the effect of paternal smoking on birth-weight, with control of all other independent variablcs. '"" Patemal smoking was tested as a continuous variable (number of ci);arettesJcigars(pipe bowls per day), and as a discrete variable (0, 1-4, 5-14, or > 15 cigarettes/ cigars/pipe bowls per day). In the analyses, the independent variables were: mother's social class; mother's age (17-22, 23-29, 30-36, and 37-41 years); mother's marital status (married or living together [not married]; or single parent); parity (nulliparous or multiparous); complications during pregnancy (dysfunctional placenta, non-specific anaemia, haemorrhage, chronic hypertension, oedema, pre-eclampsia, and normal pregnancy); sex of baby; average quantity of alcohol consumed per week by the mother; and average quantity of tobacco smoked per day by the mother. These variables have been shown to affect birth- \ceight."'" Results Between February and June, 1985, 548 women were invited to participate in the study. 48 women (8 8°„) refused to take part. Characteristics of the study group are given in table 1. 40";, (202/500) of women reported smoking during pregnancy. Of these, 201 reported smoking cigarettes only, T:\KL1i 1-(:I L\K:\C17:KIti IICS UF THIi S"IUUY GKOUI" ilA,• rr n,i vu Y.CI ) r:- \luthcr. 292 ± 46 Fathers 31 9 t 5-6 ,ti.kiel dd.cc 1, 2 3,4,5 Atwil.d ctan,s: Married 5'I"„ Living tugcthcr 38••„ Single mothcr 3•'„ Nirrlrrtrigln: Mean t SD 3339 f 490 (Range) (22(Xl-4(N10) Grrwuo.ml ue: : Mean t Sl7 40I t 1.2 (Range) (36-43) •n = 5(xl.

416 and I woman reported smoking cigarettes and a pipe. The number of cigarettes smoked varied between I and 30 per day (mean and SD, 9-0 ± 5-8). 46% (230/500) of the fathers smoked during their partner's pregnancy. The number of cigarettes smoked varied between 1 and 35 (13-8 ± 6-9); for cigars it varied between I and 9 (3-8t 2-8); and for pipe bowls between I and 8(2-7 t 1•7) per day. The correlation between matemal and paternal tobacco intake was highly significant. The correlation coefficient was 0-25 (p <0-0001, n = 500). 7011% (352/500) of the mothers reported drinking during pregnancy. 209 subjects (59-4%) reported drinking wine, 115 (32-7°l„) beer and wine, 20 (5-7",.;,) beer, 4(1•1'%,) nothing specific, 2(0-6°,0) spirits, and 1(0-3°/„) either spirits and wine or spirits, wine, and beer. Consumption of beer varied between I and 7 bottles (2-0 f 1-5); of wine, betweeri 1 and 29 glasses (2-7 t 2-9); and of spirits, between I and 9 measures (3-3 ± 3-9) per week. Mothers reported that 80°j, of the fathers consumed alcohol during tha period of the pregnancy. Among fathers who drank, the consumption of beer varied between I and 35 bottles (4-5 ± 5-5); of wine, between 1 and 21 glasses (3-4 ± 2-8); and of spirits, between 1 and 14 measures (4-0 ± 3-7) per week. The effect on birth-weight of each variable used in the regression model is shown in table 11. We found a significant relation between birth-weight and (a) matemal and (b) paternal smoking. When smoking was examined as a continuous variable, there was an average loss in birth- weight of 9-2 g per cigarette smoked by the mother l'AIiLE I I-EFM:C. r ON RIRI'H-WEIGMr OF EACH VARIABLE IN TI IH FULL REGRESSION ANALYSIS Variable Qxfficiertt (SE) pvalues n Constant 3385 (216) <0001 500 \lother: smoking* -9 (4) <0-01 500 Fathcr ti smuking• -6 (3) <0.03 500 Mothrr s alcohol cunsumptiun• -12 (7) NS 5(x) PraFlruu in pryermncv: t Chronic hypcrtti•nsion 94 (305) NS 3 Dysfunctional placenta -644 (234) <0001 7 (kdema 289 (245) NS 6 Hacmorrhagc -306 (209) NS 11 W problc•m -132 (161) NS 451 Anac•mia -453 (200) <0.05 14 Prcarlampsit 0 - 8 lIX: uttrt•,rv: t 17-22yr 11 (116) NS 41 23-29 yr -3 (87) NS 184 30-36 yr -79 (81) NS 239 37-1 t yr 0 - 36 Se.r. t Female ~0 (40) NS 256 \tale 0 - 244 AfJri1,IISlaltLi:t - Married/living together 301 (132) <0-05 487 Single parent 0 - 13 ParirY: t Afultiparuus 59 (45) NS 254 Nuuipataws 0 246 &klJr tlUsS: t 1 129 (83) NS 50 2 194 (79) <005 56 3 9 (59) NS 175 4 9 (65) NS 109 5 0 110 R2 =0-14. NS = rtot significatt. •For these variablesa dtattge in birth-weight (g) can be determined by multiplying the cvelTictirnt by the unit dose (tg, drinks per week or tobacco load per day). Negative values of the ouclficient re•fleex a reduction in birth-v.+eight. tEach coefficient indicates the change in birth-wcight (g) compared with the reference variable (0 coefficient valur.) in each group. Positive values indicate increases in birth-wctight, and negative values indicate reductions in birth-weight. T-tE LANCET, AUGUS-c 23, 1986 ~ 3500 a ~ 3400 t- (9 3300 w 3: 3: 3200 ¢ p<.03' Mother=•-•-• p<-01t Father = 1 1 1 t m . 0 5 10 15 20 AVERAGE TOBACCO LOAD PER DAY (Cigarettes + Cigars + Pipes) Effect of parental smoking on birthweight.• •Corurolled for all variables in regression model. tFather's smoking is controlled for mother's smoking. SMother's smoking is controlled for father's smoking. (p <0-01), and an independent loss in birth-weight of 6-1 g per cigarette (and cigar and pipe bowl) smoked by the father (p <0-03). This effect of the father's smoking was seen while controlling for all independent variables used in the regression, including mother's smoking (see figure). The effect of father's smoking on birth-weight was also significant when smoking was examined as a discrete variable. There was a significant change in the percentage of variance in birth-weight shown by the regression model after the addition of paternal smoking (p <0-05). To determine whether social class affected the relation between father's smoking and birth-weight we examined the differences between social classes 1 and 2 and social classes 3, 4, and 5. The effect of patemal smoking on birth-weight was greatest in the lower three social classes (p<0-03). No relation was found between matemal smoking and social class. We also examined the effect of including the variable gestational age in the original regression model. This increased the percentage variance in birth-weight explained by the model (R2) to 0-25. However, the addition of this variable to the model did not change the size of the effect of maternal or paternal smoking on birth-weight. To investigate potential reporting bias by mothers, we studied the effect ofpatetnal drinking on birth-weight. With the use of the same regression model shown in table 11, we added the variable, father's drinking, which was defined as the average number of drinks per week. The resulting model did not show any effect of father's alcohol consumption on birth-weight (p <0-79). This suggests there was no reporting bias by mothers. Discussion The relation between maternal smoking and fetal development is thought to arise from a direct toxic effect of smoke or from an indirect effect mediated by a reduction in maternal weight gain. t' With either hypothesis, reduction in birth-weight is directly related to the number of cigarettes smoked by the mother.= The relation between passive smoking by fathers and f or other household members and birth-weight is more difficult to explain. One theory is that smoke inhaled passively has the same effect on the fetus as maternal smoking. Several studies have shown that people who are exposed to tobacco smoke excrete high levels of cotinine-the major metabolite of nicotine-in the urine,t"-19 and this may account for the raised morbidity in passive smokets?° N FN+ ~

THE LANCET, AUGUST 23, 1986 Few studies have focused on the risk to the fetus from passive smoke. One study found that perinatal mortality was higher in the children of fathers who were heavy smokers than in the children of non-smoking fathers?t In a study of the relation between passive smoking and birth-weightt" passive smoking was defined as exposure to another person's cigarette smoke for at least 2 h each day during pregnancy either in the home or at work. About a quarter (23 •6°;,) of the women in the study had not smoked during pregnancy but were exposeld to passive smoke. Among these, exposure was significantly related to having a low-birth-weight baby (<2500 g). The relation was seen only in full-term (>_ 37 weeks) babies. The relative risk of having a low-birth- weight child for exposed women compared with unexposed women was 2• 17 (95°,u, confidence limits = 1•05, 4•50) after adjustment for confounding factors. Babies delivered to mothers exposed to passive smoke were an average of 24 g lighter than those delivered to unexposed mothers. Since the only information on smoking and drinking in other household members came from the mothers' reports, we considered the possibility that our findings were distorted by biased reporting. For example, smoking mothers might have had a tendency to exaggerate their reports of tobacco use by other family members. In this case, low birth-weight due to maternal smoking would appear to be the result of tobacco use by other family members. We do not believe that such a distortion is present in our data since the effect is not seen in relation to alcohol use. Mothers might also under-report their own tobacco use but report fully that of other family members. In this case under-reporting of alcohol consumption would also be expected; yet a large proportion (70"„) of the mothers in our study reported alcohol consumption during pregnancy. Another study of pregnant Danish women reported a similarly high percentage (77".„) of drinkers.' That the women in our study generally gave accurate reports of tobacco use also is suggested by the close agreement of our findings with those of other studies.2'15 Although we were able, with our model, to explain only 14"/, of the variance in birth-weight, our findings are not very different from the 17";, variance found by Dougherty and Jones.15 They used similar variables in their model but did not include father's or household smoking_ The differences in the models may also be partly due to the omission of maternal and paternal height in our model. We examined paternal smoking as both a continuous and a discrete variable. Both showed a significant effect of paternal smoking on birth-weight while controlling for the effect of maternal smoking. Although the paternal effect was less than the maternal effect these results suggest that, in addition to direct exposure to smoke from the mother, indirect or passive exposure to smoke from the father may result in an independent effect on birth-weight. Dr Rubin was the recipient of a Fulbright Fcllouship. "Dlis study was supported in part by the Dr Louiscs Borne hospitals Forskningsfond. We thank Mrs Jytte Purtoft and Mrs Liz Krogh for technical help, Prof Henrik Spiild, Prof Bjame Nielsen, and Prof Thorkild Nielsen of the Institute for Mathematical Statistics and Operations Research at the Danish "1'echnical University, and Dr Marsdat Wagner and Mrs Elizabeth Hclsing of the World Health Organisation, Regional Office for Europe, for their help with the projcct. We also thank Mr Jonathon Logan, Dr Ruth Stein, Dr Polly Bijur, and Mr Mathew Kurz"n for their advice, and Miss Catherine Carroll for typing the manuscript. Correspondence should be addressed to D. H. R., Division of Ambulatory Care, Departsnent of Pediatrics, Bronx Municipal Hospital CcrttcrJB South 19, Pelham Parkway South and Eastchester Road, Bronx, New York 10461, USA. Rtfareates ar foot of naxt coGmm 417 PROTECTIVE EFFECT OF NATURALLY ACQUII2ID HOMOTYPIC AND IIETF.ROTYPIC ROTAVIRUS ANTIBODIES SHUNZO CHIBA' TAKASHI YOKOYAMA' SHUJI NAKATAI YASUYUKI MORITA' TOMOKO URASAWA2 KOKI TANIGUCHI2 SHOZO URASAWA2 TOORU NAKAO' Departnrents of Paediatricst and Hygiene and Epidenriology,z Sapporo Medical College, Sapporo, Hokkaido,Japan Summar'y To assess serotype specificity of immune resistance to rotavirus gastroenteritis, the relation between pre-existing neutralising antibodies to homotypic and heterotypic rotaviruses and protection against infection or clinical illness was investigated. The subjects were 44 orphans exposed once or twice to consecutive outbreaks of gastroenteritis due to type 3 rotavirus in an orphanage in Sapporo. Sera were collected throughout these outbreaks and the serum levels of neutralising antibodies against four different serotypes of group A human rotavirus were measured before and after the outbreaks. Protection against rotavirus gastroenteritis seemed to be serotype specific and to be related to levels of antibody against homotypic virus. A neutralising antibody level of 1/128 or greater seemed to be protective. The protective effect was of short duration, which was probably the explanation for recurrent attacks of gastroenteritis due to D. H. RUItIN AND Ol1iERS: (tL'FEKENCES 1. Johnsron C. Ggarcttc srru*ing and thc ourc,mtc of hunun prcgn:m i 5: a sutus repon on the mnxquenas. Clin Tnaic"I 1981; 18: 1N4-?09 2. Hotxlau• DS, Topham AL. 'Ihe etlir> of sm,dcing orr ( Yal, nc.xural and dtildhond development. I'edinrr A~w 1978; 7: 105- 36. 3. Cole Ii. Studying rcproduaive risks, sm,~t:ing. JA.S1A 1986; 255: 22-23. 4. Gorvnaker SL, KI tin-Walker D, JaaAh. 1, 11, Rudt-R- 11. 1'aratratsrnoking and th< risk of childhood a<thma. Am J lSrAlic NcaLh 195?; 72: 574 -79. 5. Horwood LJ, Fergusson DM, Shann.m Fl'. S.cial and funilial f.aon in the dcw<lopmrnt of arly childhood a,thma. l'cdilr- 1985; 7 5: 854-68. 6. WeissST,TagerIS,Speiu F1,RounYli.l'crsi,ccnt..hrcu•.ArnRn, Ncsp1)ir 1980; 122: 697-707. 7. 1'edrein FA, Guandolo VL, Feroli EJ, Mella G%T', Wd„ I1'. Involunury smul:ing and in idence "f respirau~ illmss during the frnt ycar of life. Prdiarrics 1985; 75: 594-97. . 8. Fngus,um DM, Horu•,a,d LJ, Shanmm 1'1-,'fa.i,v l3. I'arenral srtx,king and Iknvcr rrspiratury• illnr.cc in the fnt three years uf life. J F. piJ,ml, J G+rnu Heulrh 19H 1; 35: 1811-84. 9.lhger 113, Wciss ST, Munoz A, R-ner It, Spcizer Fti. Lungirodinal study of the efrcasofrraterrul smokingon pulmona~ funai,m in dtildnT. n' 1:,,grJM,d 1983; 309: 699-703. 10. ManinTK,F3radcmMlS.A>wcvrxmuflowbinhw~ight.vithpa..vvcsmokccxpusurc in pregnancy. Arn J l:pidrna:.! (in pnxc). 1 L Hollingshead AR, Redlich FC. So ial da,s and mutul i0ncs. Ne.v Yark: John Wile'y, 1958. 12. Hanscn GJ. Ungd.xn "g Uddan<Isc tI. 1'ublic-aiun no 47, Sl vl ! r Skningsinai- tuner. Kntenham: Tekni,k Forlag, 1971. 13. Kldnbaum DG, Kupper LL. Applied regres,"em amlysis arnd uthcr muhivariable nu~rhods. Iksrun: Duxburv 1'rnv, 1978. 14. Cuhcn J, Cohm 1'. Applied muhiplc egnssi- aorrctni.m aruly,rs f,r rlu txhavioral scirna . Ncw Jersey: La- Erltuum A,s,wvrts, 1983. 15. U,u,gheny CKS, Jon es AD. The daemtinant, of birth.veighr. A.u J(3hsrer fi.+tecd 19H2; 144: 190-200. 16. Showstadc JA, Kudcni 1'1', Minkkr 1). Faaun a.v,cia~ d.vith birthweighc an axplontitn uf the mks uf prenatal eare and lengrh r+f gsrat6n. An, J 1'ablic Health 19ft4; 74: 1003-1R1. 17. Davi<s DI', Gray 01', GIlw.xd PC, Abemcthy A. Cigarctte srm.king in pregnanc-y: as.vtciati"rts widr mat<mal weight gain and i tal gn,..th. I.+acer 1976; i: 3N5-87. 18. Greenberg KA, Haley NJ,!•:rzel RA, Loda FA. Mnwring the aq+"surc of infanrs to wbacco snwke, niattine and axinine in urine an.1 saliva. N t:arf J McJ 1984; 310: 1075-7N. 19. Musukun S, Taminato T, Kiran" N, n al. l:ffuas ofemtir.nrnrntal tubaan snwkeon urinary cotinine acretion in rwnvnnken: evideur fix p-ive snxtking. N Lrytf J Med 1984; 311: 828- 32. 20. Fielding J C. Smuking: halrh etfcm and antrul. A't:ntl JAfeJ 1985; 313:491 -98 arxl 555-61. 21. Mau G, Nctter 1'. Th<e/rccts of patenul cigarette snwtking on pmnatal mortaliry and the incidcncr of malfonnations. Druh Mrd R''arMruhr 1974; 99: 1113-18. 22. olsrn J, Radntwtin P, Schibdt AV. AIa,M>f usc corKCptixr timc ard binhwci6fit. J EpiMr" Canan Healrh 1983; 37: 63- 75.