Tobacco Documents Online

Statement by L. G. S. Rao, Ph.D. Bellshill Maternity Hospital Bellshill, Scotland, U.K. Regarding H.R. 4957 S. 1929 My name is Dr. L.G.S. Raor I am Senior Biochemist at Bellshill Maternity Hospital in Glasgow, Scotland. I obtained mv Ph.D. in Biochemistry from the University of Newcastle in 1966. I am the author of numerous scientific publications and have made presentations at scientific meetings in Europe andI the United States. My experience in clinical biochemistry over the past 20 years has been varied and has resulted in the development of interests in several methodological and clinical problems including perinatal medicine. Over the past several years, I have become deeply interested in the investigation of the - causes of the high incidence of low birth weight and perinatal4 mortality which is found among the poorer patients of Bellshil Maternity Fbspital. I have attempted to define in biochemica terms the risk factors associated with the "poor social condi- tions"'which are supposed to be the caus= of the poor reproduc- tive performance of these mothers. I have found, as have otheN researchers in this fiel of the poorer social gr growth ret3rdation is the has been claimed, their s of considerable practical ciencies can be correcte tion. A program of such striking reduction -.in th. nerinatal mortality and m- o' the scientific researct- Introduction - There is a widely h~ -cause of low birthweight (the "causal" hypothesis) cause the statistical ass PNM lacks the specificity significance and because causal hypothesis.

in Glasgow, Scotland. I obtained :)m the University of Newcastle in nerou s scientific publications and scientific meetings in Europe and al biochemistry over the past 20 as resulted in the development of :)dological and clinical problems . Over the past several years, I ted in the investigation of the of low birth weight and perinatal ng the poorer patients of Bellshill attemote d t o define in biochemical ziated with the "poor social condi- be the caus.e of the poor reproduc- )thers. I have found, as have other 607 researchers in this field, that the biological characteristic of the poorer social groups that is of relevance to fetal growth retardation is their poor nutritional status and not, as has been claimed, their smoking habits. This finiing could be of considerable practical iaportance because nutritional defi- ciencies can be corrxted by dietary advice or supplementa- tion. A program of such dietary intervention could lead to a striking reduction in the incidence of low birth weight and perinatal mortality and morbidity. Set forth below is a review of the scientific research supporting the above conclusion: Introduction There is a widely t~eld view that maternal smoking is a -cause of low birthweight (LBW) and perinatal mortality* (PNM) (the "causil" hypothesis). That view is challenged herein be- cause th e statistical associatio n between smoking and LBW and PNM lacks the specificity and the consistency to be of causal significance and because of evidence inconsistent with the causal hypothesis. *The term perinatal mortality includes both stillbirths and infant death within the first few weeks after birth. - 2-

608 Most of the stu9ies which favour the causal hypothesis have not corrected for factors which are already k mwn to have a marked effect on LBW and PNM. The most conspicuous fact that emerges from a scrutiny of all the studies on smoking in preg- nancy is that the so-called "effect" of smoking is seen only in the poorer underprivileged mothers and not in th e mothers who have a good family income. Low family income could lea9 to nutritional deficiencies which can caus e fetal growth retar3a- tion. Therefore, the so-called "effect" of smoking seen in only th e poorer mothers may no t b e du e to smoking itself, but due to deficiencies in maternal nutrition during pregnancy. Evidence for this has come from a recent study on the protein intake in pregnancy in Bellshill Maternity Hospital which ~ showed that among mothers with normal protein intake, there was no difference in the proportion of LBW infants between smokers and non-smokers. Other evidence a gainst the c ausal hypothesis from a review of the available literature is also presented. - Smoking in pregnancy has become such an emotioral issue that it is difficult to be unbiased in the design of the studies on this subject or in the interpretation of the results. As there is a great deal of evidence against the -3- "causal" hypothesis, o the higher incidence perinatal mortality fo, nancy. There is consic growth-retardation is m pregnancy, possibly du denoted as the "nutrit hypothesis is th e view certain type of persor. reproductive performan: -his hypothesis will b thesis. The latter tw common and are compler favour of th e "causal" in the 1979 U.S. Surge discussed in detail, t alternative hypotheses merits discussed. For a valid discuss o-° factors that are a: perinatal mortality nee

the causal hypothesis have already known to have a sost consoicuous fact that :udies on smoking in preg- of smoking is seen only in ad not in the mothers who sily income could lead to ,use fetal growth retarda- =fect" of smoking seen in ue to smoking itself, but itrition during pregnancy. -cent study on the protein Maternity Hospital which protein intake, there was BW infants between smokers 1 hypothesis from a review -esented. : such an emotion3l issue sd in the design of the he interpretation of the . o f ev idenc e aga ins t the 609 "causal" hypothesis, o ther hypotheses are required to explain the higher incidence of low birt'hweight infants and higher perinatal mortality found in sone studies o n smoking and preg- nancy. There is considerable evidence for the view that fetal growth-retardation is mediated through a reduced weight gain in pregnancy, possibly due to undernutrition. This view will be dem ted as the "nutritional" hypothesis. Another alternative hypothesis is the view that smoking is a characteristic of a certain type of person or a group of people who have a poor reproductive performance because of constitutional re3sons. This hypothesis will be denoted as the "constitutional" hypo- thesis. The latter two hypotheses have several features in common and are compleroentarx to each other. The evidence in favour of the "causal" hypothesis is presented in great detail in the 1979 U.S. Surgeon-General's report .(1) and will not be discussed in detail, but the evidence in favour of the two alternative hypotheses will be presente3 and their relative merits discussed. For a valid discussion of smoking and pregnancy, the effect of factors that are already known to affect birthweight and perinatal mortality needs a brief consideration. -4 -

610 Factors that are known to affect fetal arowth: 8iological Factors The biological factor which shows the best correlation with the weight of the infant is the functional capacity of the placeita, which is determined mainly by the size of the pla- centa and the quantity and quality of th_ blood supply to the fetus. The size of the pla centa is largely determined by the height and weight of the mother. The factors that influence the quantity of blood flow are not well understood except in pathological states such as pre-eclamosia and extensive infarc- tions in the placenta. There is some evidence that uterine blood flow is under hormonal control (26). The quility of the blood supply is mainly determined by the nutritional status of the mother. It is well-known that in poorer countries with nutritional d eficiencies, the weight of the mother and the infant are both lower than that of those in the more prosperous countries (2) . Thus, maternal stature and the nutritional status during the pregnancy appear to be important factors on theoretical grounds and are, in fact, found to be so in pra c tice. The other biological factors which are known to affect birthweight are:- 1. The length of gestation, the shorter the gestational period, the smaller the baby, 2. Sex of the infant, males being slightly heavier than females (about 200g) after about 36 weeks of gesta- tion, and -5- 3. The pregnancy n lighter than lat Socio-e conon ic f actors In addition to these that there are important to influ ence birthweight status could have an in tries, probably b y affec even in some prosperous infant is related to the husband. 1hus, social c have a higher birthweig -perinatal mortality and 1 in social classes 4 ar. These socio-economic fact in their effect on perir Thus, the perinatal mort 1,000 births and was 32.0 - ~: The striking effect c appears to be mainly duc rity, low birthweight bab fact, social class has st

ws the best correlation with functional capacity of the ly by the size of the pla- of tha blcod supply to the s largely d etermined by the The factors that influence t well u nderstood except in amnsia and extensive infarc- some evidence that uterine 1 (26). The quality of the ~y the nutritional status of t in poorer countries with 3ht of the mother and the thos e i n th e mor e prosperou s tature and the nutritional to be important factors on .ct, found to be so in prac- s which are known to affect 611 3. The pregnancy number, the first chi13 being slightly lighter than later children (about 200g). Socio-econanic factors In addition to these biological factors, it is well known that there are imnortant socio-°conomic factors which are known to influ ence birthweight. It is not surprising that economic status could have an influence on birthweight in poor coun- tries, probably by affecting the nutrition of the mother, but even in some prosperous western countries the weight of the infant is related to the social a nd educatioaal status of the husband. Thus, social classes I and 2 (professional workers) have a higher birthweight„ longer gestational periods, lower -perinatal mortalit y and lower congenital deformities than those in social classes 4 and 5 (manual and unskilled workers). 'hiese socio-economic factors appear to be particularly striking in their effect o n perinatal mortality in the Unitei Kingdom. 'Ihus, the perinatal mortality in social class 1 was 9.5 per 1,000 birtl-s and was 32.0 per 1,000 in social class 5(3). The striking effect of social class on perinatal mortality appears to b e mainly due to the higher incidence of prematu- rity, low birthweight babies, and congenital abnormalities. In fact, s ocial c lass h as such a striking effect on the health of -6- Y -

612 the pooulation in certain countries (4) that there is a higher incidence of shorter mothers (449! ) in social classes 4 an9 5 comoarefl to social classes I an3 2 (?0.5$). The nunber of mothers delivered before 36 weeks gestation was twice as high for c lasses 4 and 5( 8.48 ) a s that o f I an3 2 (4.2$ ). ( See Table 1). This higher incidence of prematurity alone can ac- count for the higher perinatal problems of the social classes 4 and 5. The socio-economic differences persist even when mater- nal stature has been accouzted for (5) . It has been shown that the incidence of prematurity is associated with th e social class of the mother's father, for any given social class of the mother's husband (6). This indicated that the socio-economic status (prcbably nutritional status) of the mother when she was a child had a striking effect on her growth in chill3hoo3 and -her reproductive performance in later life. Thus, in Britain, social class appears to affect the maternal size and nutrition, and could b e a very important f actor which affects birthweight and perinatal mortality. In the United States also, similar socio-econonic factors appear to affect perinatal mortality. Thus, black populations in general and the less educate3 among the whites have a higher perinatal mortality than the better educated whites (29). In a large study on perinatal mortality conducted in Canada, the hospital status of the mother, whether private or ward patient, which is probably determine d by the educational status of the father, has been shown to be an -7- imoortant risk factor, the perinatal mortalitv than tha Thus, there are sever 'actors which affect birth :hese must be taken into ac tion between smoking and per Sone Methodoloaical consider Most studies on the sut not taken the above factor, clusions, therefore, are op Seneral's report has pointe of adjustment for differenc in the distribution of su economic status and race smoking to perinatal mortali ?he usua 1 definition of '"eighe3 less than 2500 grams zient definition, it cou1d 9estational age, the parity taken into consideratio n t o `-etal growth retardation. I

;4) that there is a higher n social classes 4 and 5 (?0.58). The number of station was twice as high of 1 an3 2 (4.2$). (See prematurity alone can ac- :ms of the social classes 4 r life. Thus, in Britain, ternal siz e an d nutrition, - which affects birthweight .ited States also, similar 'fect perinatal mortality. .: .. ~. t -- ~ . • . . id the less educated among m ortality than the better ad y o n perinata 1 mortality itus of the mother, whether ,robably determined by the has been shown to be an 613 imoortant risk factor, the private patients having a much lower perinatal mortality than that of the ward patients (21) . Thus, there are several import3-it biological ani social factors which affect birthweight and perinatal mortality and these must be taken into account when cor.sidering the associa- tion between smoking and perinatal mortality. Some Methodological considerations Most studies on the subject of smoking and pregnancy have not taken the above factors into consideration and their con- clusions, therefore, are ope n to question. The 1979 Surg eon- , General's report has pointed out that "problems arise from lack of adjustment for differences between smokers and non-smokers in the distribution of such factors as age, parity, socio- economic status and race when th e relationship of maternal smoking to perinatal mortality is under study." (1) The usua 1definition of a low-birthweight baby i s one that weighed less than 2500 grams at birth. Although it is a conve- nient definition, it co.ild be subject to serious errors. The gestational age, the parity, and the sex of the infant must be taken into consideration to obtain some degre e of validity of fetal groath retardation. It is obvious that for a small woman -8-

614 weighing 100 pounds, a 2500 gram baby born at 34 weeks is not a small baby but it is for a tall woaan weighing 160 poilds and delivering at 40 weeks of her pregnancy. In fact, the 2500 gram limit appears to be absurd in some situations as it has been shown that the perinatal morta?ity for the under 2500 gram babies increases with the increase in maternal height (7). Although it would become too cumbersome to take the mother's height and weight into consideration in defining low birth- weight, at least the gestational age, parity and sex of the in fan t shoul d b e take n int o account . An important c oncept in d efining fetal growth retardation was described by GYuenwal3 (8) and adopted in a study by Miller and co-workers (9) . According to this concept, fetal growth . -retardation is divided into two types. One is the 1ong, thin baby," which is the result of wasting that occurs during a period of days prior to birth, resulting in a low ponderal index. The second type has a general decrease in growth pro- bably extending over a period of weeks before birth, with the result that the deficits in body length and weight at birth are proportional resulting in a normal ponderal index. 'Ihe second type is described as short for dates (SHFD). The infants of the first type with the low ponderal index, have large appe- tites and will catch up within a few months with the weight of 'infants with a normal birthweight. On the other hand, the SHFD -9- infants a nd their sibling disease, suggestive off a be seen later in this pape fetal growth retardation i question of smoking ?ind pr The "Oausal" T.•lypothesis - Several studies have birthweight babies and a h be statistically a ssociate rancy. FYbwever. „there a ssoci ation i s not _-;eppr association has bee n inter -in spite of considerable e worthwhile to assess _crit this hypothesis. If maternal smoking is tion and an increase in smoking should be found }n the population. 5b~.$ve.r, effect is seen only in p upper socio-economic,group

~y born at 34 weeks i s not a man weighinq 150 Do1n3s and gnancy. In fact, the 2500 some situations as it has !ity for the under 2500 gram e in maternal height (7) : arsome to take the mother's .on in defining low birth- ige, parity and sex of the zg fetal growth retardation idopted in a study by Miller this concept, fetal growth es. One is the "long, thin sting that occurs during a ssulting in a lo•+ ponderal ~ral decrease in growth pro- eeks before birth, with the ngth and weight at birth are oonderal index. ltie second tes (SHFD). The infants of :al index,-have large appe- !w months with the weight of On the other hand , the SHFD 615 infants a nd their siblings have a notable incidence of organic disease, suggestive of a genetic or familial pattern. It can be seen later in this paper that this tvpe of classification of fetal growth retardation is useful in the understanding of the question of smoking and pregnancy. 'lh e "Causal " !iypothesi s Several studies have reported a higher incidence of low birthweight babies and a higher rate of perinatal mortality to be statistically associated with maternaL smoking during preg- n3ncy. Fbwever, there are some studies in which. this association is not reported. Nevertheless, this statistical association has been interpreted as having causal significance, I -in soite of considerable evieience against this view. : It may be worthwhile t o asses s_critically the evidence for and against this hypothesis. If maternal s moking is,the cause of fetal growth retarda- tion and an increase in perinatal mortaLity, this effect of smoking should be found in all countries and in a1l sections of the population. [iowever, this is not found to be true. 'ihis effect is seen only in poorer social groups, but not in the upper socio-economic groups (See Table 2). Many studies have -10-