Tobacco Documents Online

NUTRITION REPORTS INTERNATIONAL 23/5: 881-9oo, Ma~- 1981 7 ~:FFECTS 2F MATERNAL CIGARETTE SMOKING DURING PREGNANCY ON BI~HSIZE, GROW'I'I! OF INFANTS AND INFANT FEED!.~IG PRACTICES/, / A longitudinal nutrition survey of 403 infants was conducted i~ Toronto and Montreal between 1977 and 1979. Infants who were born'to mothers who smoked cigarettes during pregnancy were found to be statistically smaller• than infants whose mothers did not smoke, The former group of infants was also smaller in weight, length and head cir- cumference than the latter group at 6, 12 and 18 months. However, there was no difference in body fatness between the.2 groups of infants" at the specified time of study. The results suggest that cigarette smoking not only has an effect on the outcome of pregnancy, but may also have a long term effect, on the progeny. INTRODUCTION Maternal cigarette smoking during pregnancy exerts an adverse effect on the course of pregnancy. Among cigarette smokers higher incidence of abrupto placentae, oedema, pre- eclampsia, etc. than non-cigarette smokers has been docu- mented ( i, 2). Perinatal morbidity and mortality of new- borns increase with maternal smoking ( 1, 3 , 4 ). There is convincing evidence that infants born to mothers who smoke are smaller in size in both weight an~ length. " The effect of cigarette smoking on subsequent development of the infants "is not as clear cut. While some investigators have found deleterious effect on subsequent physi6al and mental development (5, 6, 7 ], others have not been able to demonstrate any long term effect ( 8, 9 ). Between 1977 and 1979, a. longitudinal survey of. infant nutrition from birth to 18 months was conducted in two met- ropolitan centers in Canada '[i0). in the initial sample of 403 subjects, over 30% of the mothers smoked cigarettes during pregnancy. Because of the large number of cigarette smokers, it permits the examination of the effects, on a :~rospective basis, of cigarette smoking on infant feeding and development up to 18 months of age. MAY 1981 VOL 23 NO. S 887 T104.231136

"NUTRITION REPORTS IN'~ERNATIONAL METHODS Details of the methodology of the survey have been pre- viously described (10). In brief, the survey was conducted longitudinally in Toronto and Montreal at i, 3, 5, 6, 7, 8, 10, 12, 15, and ~8 months of age. The sample was recruited from a list of over 2,000 names of pregnant women. This was compiled from names obtained from hospitals, community health clinics, retail outlets that cater to expectant mothers, etc. The names were stratified basically by mun- icipality and then sampled randomly. In Montreal, it was also'stratified by language spoken (i.e. French and English) The proposed study was fully explained to each subject and written consent was obtained. The survey started with 403 subjects in 1977 and ended with 317 in 1979. Information of the smoking habits, personal and socio- economic backgrounds of the parents was obtained by personal interview shortly before the infants were born. A card was giv@n to each expectant mother at :his time to present to the attending nurse to enter the birthweight and length of the infant at delivery. ~t 3 weeks postpartum, a second questionnaire was administered. In this questionnaire the mothers were asked to provide information on length of gestation, their pre-pregnancy weight and their ~eight gain during pregnancy. Information on the feeding was obtained by home.interviews at subsequent study intervals. Data on the growth of the infants were determined by six anthro- pometric measurements during the home visits from 1 month onwards. Weight was determined on an infant scale (Seca, Vogel and Halkes, Germany). Length was measured on a Frank.- fort Plane (University of Iowa, Iowa), head and mid-left arm circumference were measured with a non-stretchable tape. Tricep and subscapular fatfolds were determined by the Lange caliper (Cambridge Scientific~ Industries, Maryland). All measurements were performed according to procedures des- cribed by Fomon (ii).. For statistical analysis, the ethnic backgrounds of the mothgrs~ have been grouped into five categories, namely: (a) Anglophone Canadian, (b) Francophone Canadian, (c) Euro- pean and American descent, (d) Jewish and (e) others. The ethnic backgrounds of the mothers were personally identifie6 by themsleves and then grouped accordingly. Total family ingome was analyzed according to the fol- lowing groups (a) less than i0,000 (b) i0,000 to 14,999, (c) 15,000 to 19,999, (d) 20,000 to 29,999 and (e] more than 30,000 dollars per annum. The education of'the mothers were presented in 4 groups, namely: (a) did not finish high school (13 grades), (b) graduated from high school, (c) com- ,, / Ti04231137

NUTRITION REPORTS INTERNATIONAL subjects received part-time ~ducation over extended periods of time and thus it was difficult to assess years of edu- cation. . Fatness of the infants was assessed by arm circumfer- ence, tricep and subscapular fatfolds. Mean body mass was represented by the upper arm" muscle circumference, calcu- lated from upper arm circumference and tricep fatfold by the following formula(12): Upper arm muscle circumference (cm) = upper arm circumference - (tricep fatfold) i0 In this report, for brevity, only th~ anthropometric data obtained at birth, 6, 12, and 18 mohths are presente~. Statistical analyses were performed by computer using the Statistical Package for Social Sciences (13). Cigarette smokers or smoking a~e designated by CS and non-cigarette smokers or non-~moking by NCS. CS infants refer to infants whose ~others smoked during pregnancy and NSC infants are those whose mothers did not smoke. Heavy, moderate, light and occasional smokers are subjects who smoked 20 or more, i0 to 19, 1 to 9 and less than I cigarette per day respec- tively. RESULTS i. Maternal Smoking HaD~ts Of the 403 new ~others who started in the survey, 131 or 32.5% smoked cigarettes during pregnancy. Twelve of these were occasional (l~ss than 1 cigarette per day), 46 were light (i to 9 cigarettes per day), 38 were moder&te (10 to 19 cigarettes per day) and 35 were ~eavy smokers (more than 1 pack per day). The effect of CS on birthsize was dose responsive, viz, the birthsi.ze (Table i). multiple regression was performed to search for factors T!04.231138

NUTRITION REPORTS INTERNATIONAL Weight a.nd Length of Newborns Smokin~ l{abit Cigarette smokers :.ton-cigarette smokers t° p** Birthwei@ht Birthlenqth N__~. Kg. N__~. Cm__~. 131 3.25 ~ 0.44 iii 50.73 ~ 2.67 272 3.52 Z 0.46 233 52.09 ~-2.64. 5.36 4.35 ~0.001 ~ 0.001 Heavy' (;. 1 pack/day)35 3.18' ~ 0.36 30 48.89 ~ 2.58 Moderate (10-19/day)38. 3.13 ~ 0.39 34 50.34 ~ 2.48 Light (l-9/day). 46 3.42 ±" 0.48 39 ~ 51~70 ~ 2.67 Occasional (~ /day) 12 3.44 ~ 0.32) 8 52.28 ~ 2.44 F+ 4.62 4.46 P 0.004 0.005 * Mean ~ Sd. ** Probability o Value of t, derived from student's-t test + Value of F, calculated by analysis of variance that might influence birthweight. Initially, continuous and categorical variables were analyzed separately. The continuous variables were maternal pre-pregnancy weight, maternal weight gain during pregnancy, maternal age, pater- nal age and paternal weight. The categorical variables included maternal smoking'habit, length of~gegtation, mat- ernal education, paternal education, total family income, parity and socio-economic status. The variables that ex- T104231139

NUTRITION REPORTS INTERNATIONAL hibited statistical influence on birthweight were cigarette smoking {r=-0.287), length of gestation (r=0~246), maternal prepregnancy weight {r=0.333), maternal weight gain, (r=0.208), and mother's education (0.172). Stepwise mul- tiple regression showed that CS exerted the strongest influ- ence on birthweight of the infants. 3. Post-natal Size of Infants (a) Size at 6, 12, and 18 months: Of the nine para- meters of growth and development shown in Table 2, signifi- cant differences between the two groups of infants were ob- served for body weight, length and head circumference. The weight over height ratio of CS infants was consistently about 2 units less than ~CS. This was significant at 6 months, (P=0.05) but not at later months. No statistical difference was noted between CS and NCS infants for arm circumference, tricep fatfold, ~ubscapular fatfold, sum of tricep and subscapular fatfolds and lean .arm circumference. ~his shows that although weight, length and head circumfer- ence were affected by-CS, body fat and lean body mass were not. (b) Growth in weight and length from birth to 12 months: The gain in weight calculated by difference between 12 m~nths and birth, between CS and NCS was not significant- ly different (t=0.49P=~.627). Whe same was found when gain in length between the two groups wa~ compared (t=0.60P=0.55). This ind.icates that post-natal growth was not influenced by maternal CS habits during pregnancy. Pregnancy Performance Chi Square did not reveal statistical relationship between CS and length of gestation (x2&4.01, df=2, P=0.135). Heither was there significant difference in the pre-pregnan- • cy weight (t=-l.80, P=0.072) and ~eight gain (t=-0.03, P=0.974) during pregnancy between the 2 groups of women. 5. Feedin@ Practices " ' (a) Milk Feeding: Breastfeeding was more popular among NCS and bottle-feeding was more popular among CS ~uring the first week of life (Table ~). Chi Square analy- sis shows that'this discrepancy was statistically signifi- cant. (b) Tim~ of termination of Breastfeeding: Among the mothers who breastfed their infants, CS terminated breast- feeding earlier than NCS. Whe median time of termination of breastfeeding among CS was 2.7 months. Among NCS it was 4.1 months. This difference was sta~istically significant (t = 2.27, P< 0.05). MAY 1981 VOL. 23 NO. 5 891 T104231140

NUTRITION REPORTS INTERNATIONAL TI04231141

NUTRITION REPORTS'INTERNATIONAL (c) Time of Introduction of Solids: ~he frequency of early.and late solid introduction between CS and NCS is shown in Table 3. Two months was used as the cut-off point for early and late introduction of solids in the diets of the infants, since the.median time of solid introduction by the entire sample was two months. Alt~ough slightly more CS than ~CS provided solids to their infants during the early period, Chi Square analysis showed that the difference was not statistically significant. 6. Inter-relationship between Maternal Education, Ethnic Background, Socio-economic Status and CS on Size of 12-month old infants. Comparison of weight, length and head circum- ference between CS and NCS twelve-month-old infants in the different groups of ethnic background, education and socio- Table 3 Cigarette Smoking and Infant Feeding Practices Type of milk feeding during first week postpartum NCS CS Row total Breastfeeding 218 69 287 76.0% 24.0% 71.1% 76.8% 58.0% Bottlefeedin~ .66 50 .116 56.9% 43.1% 28.9% 23.2% 42.0% Total 284 119 403 Chi Square = 13.52 D.e~ree of Freedom = 1 Probabilit7 = 0.002 Time of Introduction of Solid Foods NCS CS Row Total Before 2 months 122 68 190 64.2% 35.8% 49.7%" 47.5% 54.4% After 2 months 135 ~7 192 70.3~ 29.7% 50.3% 52.5~ 45.6.% Total 257 125 382 Chi Square = 1.35 De~ree of Freedom = I Probability = 0.245 T!0423 1 "142

Z SR 95 ~'0- 50"0 • 9~ S~ 86 ~£'0 SR S6 SN 08 ~£'0 SN 08 S~ 9£ E0"0- SN B~ S~ 6T gg"0 S~ 61 OZ I~'O EE 6L'O OOI. LI'O E8 ~'0 L6 SR 001 ~00"0> 18 mopee~ ~o ee~Sea o S~ ~u~ S~ uee~eq u~e~ aR~ ~o e~ueze~a + LL'I SN 9Z T~'0 s~pn~s e~npmx6-~sod pe~Idmoo "p 9~'0 SN i0[ E0"0 ~SXSA~Un mox~ pe~np~x6 TI'I SN 8~T Z~'0 TooqDs q6~q pe~eIdmom "q S~'T **SN ~ 05"0 ~ooqos qSTq ~sTu~ ~,,uPTO 96"0 SN 96 ~0"0 666'6Z - 000'0~ "P "0L'0 SN 08 TZ'0 666'6T - 000'~T "~ B'/'0 SN 8~ I£'0" 666'~ - 00O~01 "q 60"0 SN. 6T 9T'0- 000'0T> "~ • (~) e~oouI ~Ty~~ EB'T- SN OE LL'O- sxeqwo 'e ZL'E SN ~ TS"0 qs~aD "p LB'0 SH OOI 60"0 ~ueosap u~o~xeun~ ~ u~do/n~ "o £9"T TO'O) £8" Tg"0 u~peu~D euoqdoaue~ "q 0S'0 SS 96 T~'0 u~p~u~D auoqdoTSu~ .~unog5~o~ ~ "~P SD-SDN ~ "~P SD-SDN .~ "gP +SD-SDN ~uexe~unox~O q~ueq ~qB~eM ~ eTq=~

NUTRITION REPORTS INTERNATIONAL economic classes show that, in most qases, but not all, CS infants were smaller than NCS (Table 4). ||oweger, many of the diffemences were not statistically significant, probably uecause of the small sample size in some of the subgroups. |~evertheless, the general tendency of CS to be smaller than tiCS infants does exist. DISCUSSION Available statistics show that in the decade between 1965 and 1975 there was a gradual decline in the number of CS in the Canadian population 15 years of age and over &4 ). ~his was largely attributable to the decline in the percent- age of smoking males. The percentage of female smokers in- creased slightly between 1965 and 1974, but stabilized between 1974 and 1975. }lowever, it remained lower than males. In 1975, 32%'of the female population smoked cigar- ettes regularly. The highest percentage of female smokers were in the age groups 20 to 24 (38.3%) and 25 to 44 (37%) years. In the present survey, the percent of pregnant women who smoked regularly was 32.5%. This was surprisingly close to the national average for women in 1975. However, since 98% of the survey population recruited in 1977 was over 20 years of age, the percent of CS was 5% to 6% below the values provided by Health and Welfare Canada for females 20 or more years of age &4). This difference could be due to decline in the ihcidence of CS among females between 1975 and 1977 or to pregnancy which m~ght have d~ssuaded some women from smoking. UnfQrtunately, there are no national statistics on the incidence of CS among pregnant women for cdmparison with the present observations. On the average the birthweight and length of .CS. infants were 230 g lighter and 1.0 cm shorter than NCS infants. "These differences were Within range of studies reviewed by .Iolsclaw and Topham {3 )" and confirmed that. indeed CS during ~regnancy reduced the s~ze of the newborn. The observati6n ~hat the size of the newborn decreased with increasing num- .~er of cigarettes s~oked also agreed with earlier findings (15,16). Furthermore, cf the various factors that were tested and shown to affect birthsize of the infant.s, CS ex- hibited the strongest influence. The ~ame was reported by Ailler et al (17). Thus, it is clear that maternal smoking habit during pregnancy is an important determinant of birth- size. CS has been reported to.reduce the length o~ ~estation and maternal weight gain during pregnancy(18~ 19, 20). The lack of difference in the length of ge6tation and weight gain of CS and NCS mothers in the present survey was prob- ably the resul£ of the method of recruitment of the sample which screened-out overt premature and "unhealthy" newborns. MAY 1981 VOL. 23 NO. $ " .. 895 Ti04231144

NUTRITION REPORTS INTERNATIONAL However, length of gestation and weight gain of the mothers did independently affect birthsize. Therefore, the results suggest that CS can affect birthsize without the mediatory effect of length of gestation and pregnancy weight gain. CS not only affects the size of infants at birth, b~t also the size of the infants in later life. CS infants were significantly smaller in weight, length and head circumfer- ence taken prospectively from 1 to 18 months. There was no difference in weight over longth, arm cirucmference, fatfold thicknessess and lean arm circumference between the two groups of infants. This suggests that there was no differ- ence ~n fatness between NCS and CS infants and that the cb- served differences are accountable only by stature. Further more, since there was no difference in postnatal rate of weight and length gain between the two groups of infants, it implies that there was no catch-~p growth among CS infants, at least not during the survey period. The long term effect of maternal smoking during preg nancy on height of children has been well established by ~ number of investigators (5, 6, 9). The effect of CS on long-term weight development is not as clear cut. Russell et al (9) were able to show a weight advantage of ~CS over CS infants at.6 months but not at I year. Similarly Hardy and Mellitus (8) did not detect weight differences between the 2 groups of infants at 1 year. These two studies sug- gest that there is catc11~up growth by CS infants. On the contrary Dunn et al (6) showed a continued advantage of" weicht, height, and rate of wciaht ~ain and .length-over- weight gain of NCS infants was still detectable at 6 years. The present study supports Dunn et al's (6) observation of an effect of CS on weight and length up to 18 months, but not on rate of weight or length gain. It is interesting that the tw~ studies showing prolonged effects of CS on growth were both conducted in Canada. Whether there are other environmental effects that might have resulted in the different observations in Canada and elsewhere remain to be determined. Of greater importance than physical growth is the poss- ible effect CS may have on mental development. Although the CS infants in the present survey were foundto have head circumference significantly smaller than NCS infants, this cannot be interpreted to mean that cS infants |]ave slower mental development than NCS infants sinc4 psychological tests were not performed. However, Butler and Goldstein ~5) and Dunn et al (7) have provided evidence of behavioural and academic stunting among school age children whose mothers smoked during pregnancy. Additionally Denson et al (21)observed a positive relationship between CS and hyper- ]:inesis among a small group of infants. The results did not show a cause and effectrelationship between CS and mental development since other environmental and social factors may T!04231145