TABLE 4.-Number of women in the current workforce, classified by occupation (1978) Occupation # of Women # of Women in Thousands Occupation in Thousands White-collar workers 24,594 Professional & Technical 6,083 Biological scientists 22 Chemists 17 Nurses, dietitians, Kz therapists 1,255 Health technologists and technicians 353 Engineering and science technicians 132 Painters and sculptors 83 Photographers 13 Managers and administra- tors, except farm 2,365 Sales workers 2,666 Sales clerks, retail trade 1,672 Clerical workers 13,456 Bookkeepers 1,660 Cashiers 1,222 Secretaries 3,561 Typists 1,009 Blue-collar workers 5,770 Craft and kindred workers 694 Printing craft workers 91 Upholsterers 14 Operatives, except transport 4,317 Assemblers 606 Bottling and canning operatives 25 Checkers, examiners, and inspectors; manufacturing 359 Clothing ironers and pressers 101 Cutting operative, n.e.c. 84 Dressmakers, except factory 113 Drillers, earth 2 Dry wall installers and lathers 1 Filers, polishers, sanders and buffers 38 Furnace tenders, smelters, and pourers, metal 3 Garage workers, and gas station attendants 20 176 Blue-collar workers-cont'd Laundry and dry cleaning operatives, n.e.c. Meat cutters and butchers, except manufacturing Meat cutters and butchers, manufacturing Mine operatives, n.e.c. Mixing operatives Packing and wrappers, excluding meat and produce Painters, manufactured articles 118 13 33 4 3 422 30 Photographic process workers 48 Precision machine operatives 43 Drill press operatives 15 Grinding machine operatives 10 Lathe and milling machine operatives 11 Punch and stamping press operatives 47 Sawyer 14 Sewers and stitchers 772 Shoemaking machine operatives 60 Furnace tenders and stokers, except metal 1 Textile operatives 224 Spinners, twisters, and winders 100 Welders and flame cutters 41 Winding operatives, n.e.c. 37 All other operatives, except transport 1,062 Transport equipment operatives 258 Nonfarm laborers 492 Service workers 8,037 Private households 1,135 Child care workers 477 Cleaners and servants 514 Housekeepers 117 Service workers, except households 6,901 Cleaning workers 858 Table 4 (continued) Occupation # of Women # of Women in Thousands Occupation in Thousands Service workers-cont'd. Health service workers-cont'd. Lodging quarters cleaners Building interior cleaners, n.e.c. Janitors and sextons Food service workers Bartenders Walters' assistants Cooks Dishwashers Food counter and fountain workers Waiters Food service workers, n.e.c. Health service workers Dental assistants Health aides, excluding nursing Nursing aides, orderlies, and attendants 174 462 222 2,951 111 45 678 82 397 1.252 384 1,660 128 238 902 Practical nurses Personal service workers Attendants Barbers Child care workers Hairdresser and cosmetologists Housekeepers, excluding private households Welfare service aides Protective service workers Firefighters Guards Police and detectives Sheriffs and bailiffs Farm workers TOTAL- 390 1,302 175 11 103 483 - 92 115 1 53 28 3 509 38,910 NOTE: n.e.c. is an abbreviation for "not elsewhere classified" and designates broad categories of occupations that cannot be more specifically identified. SOURCE: U.S. Department of Labor (17). impact of occupational exposures (or their interactions with smoking) on the health of women. Because of this cohort effect, any failure to demonstrate an excess risk of a given occupational exposure in women must be interpreted with considerable caution. It may mean only that the women exposed were too young and the exposure too brief for illness to have yet developed. This caution is doubly impor- tant for those attempting to demonstrate an interaction be- tween occupational exposure and smoking on the development of disease in women. Thus, little comfort can be taken from the current low prevalence of occupational disease in women. It is reasonable to expect that any movement of large numbers of women into hazardous occupations will be followed, after an appropriate time lag, by a dramatic increase in the prevalence of occupational illness in women. The Reproductive Role A third reason for examining the effects of occupational expo- sures in women separately from those in men is the difference 177 TABLE 5.-Most common female job categories, by percentage of the female work force employed Job Percent of Female Work Force Job Percent of Female Work Force Secretary 8.5 Private Household Worker 2.9 Retail Sales Clerk 4.3 Registered Nurse 2.8 Bookkeeper 4.3 Elementary School Teacher 2.8 Waitress 3.2 Typist 2.6 Cleaning Workers 2.2 Cashier 3.1 Sewer & Stitcher 2.0 SOURCE: Rones, P. (14). in their reproductive roles. Toxic occupational exposures in both men and women can reduce fertility and increase fre- quency of teratogenic effects (see Table 6). In addition, however, the g-month duration of gestation provides many opportunities for the fetus to share any adverse toxic exposure of its mother. These risks may interact with the well-established risks of cigarette smoking during pregnancy discussed elsewhere in this report. Table 6 provides a list of hazardous substances in the work environment, some of which are suspected of having ef- fects on reproduction. Another specific concern for women is that of contraception. Substantial numbers of women` in the United States use oral contraceptives (18). These drugs have been shown to interact with cigarette smoking to produce a greatly increased risk of cardiovascular disease, as discussed in this report. In addition, it is possible that oral contraceptives may interact in an adverse manner with physical or chemical agents found in the work place, or that the combination of smoking, occupational expo- sure, and oral contraceptive use may bear special risks. The answers to those questions can be found onIy through the study of populations of working women. One study approached this issue by examining the health status of women involved in the manufacture of oral contracep- tives. Poller, et al. have shown that women working in the man- ufacture of oral contraceptives absorb enough of the drugs to influence the clotting mechanism as well as alter menstrual function (12). Unfortunately, the risk of cardiovascular disease -and the effects of smoking in relation to it-could not be estimated in this population. Because of the established ex- cess risk of cardiovascular disease from concurrent smoking 178 and oral contraceptive use, examination of cardiovascular risk in this group would be of interest. The preceding discussion presents several areas where female-male differences may significantly limit the direct applicability of the results of male smoking studies to the female population. These areas of potential difference present research questions that justify significant, ongoing research activities. !3pedic Interactions Between Occupational Exposure and -lit A review of all the potential risks of occupational exposure for women is beyond the scope of this section. Table 6 lists a number of agents found in the occupational environment and their ob- served organ toxicity. Table 7 presents selected pulmonary ir- ritants and sensitizers in specific occupational settings in rela- tion to the number of women employed in those settings. There is little specific data on the health effects of a given occupational exposure in women. Two clear exceptions exist- exposure to asbestos and to cotton dust. The data from studies of women exposed to these two compounds provide examples of established interactions between smoking and occupational ex- posure in women. ASBESTOS Selikoff, et al. prospectively followed a group of 370 male as- bestos insulation workers. They demonstrated a multiplicative effect of asbestos exposure and cigarette smoking on the risk of development of lung cancer (4,13). Workers who smoked cigar- ettes developed lungcancer at arate 92 times that ofnon-exposed nonsmokers. They observed no deaths from lung cancer among 87 nonsmokers, and 24 deaths from bronchogenic cancer among 283 regular smokers, a number well in excess of the 3 deaths expected. Newhouse, et al. followed a cohort of 900 women first employed between 1936 and 1942 in an asbestos factory making both textiles and insulation materials (2,10,11). They analyzed the group's mortality experience between first employment and 1968, with a minimum of 26 years' follow-up. There was an ex- cess overall mortality partly accounted for by deaths from cancer, observed even among those who worked in jobs with low-to-moderate exposure to asbestos. An excess of cancer of the lung and pleura was found among those who were severely exposed and who had worked less than 2 years. In the group with severe exposure for more than 2 years in the factory, excess deaths from cancer of the lung, pleura, and non- 179 TABLE 6.-Chart of toxins and effects - - mlzlzil - - -- IF- `Cnaa) -- -- I . . . . . . . . ./; Ill il . ii . . CI . I( R l ca R . 1 IW . . I I 4' LII- . . . . . - Al nqq . . . . . Ammal and/or human data Time wmghtsd awra~e 40-hr. week. S-hr day (OSHA). 10.hr. day INIOSHL In pregnancy. blood lead less than 4@ g(jOO gms is suggested 0 Ewdme only male mtertitity no data on females SOURCE: National Institute for Occupational Safety and Health (9). 180 neoplastic respiratory disease were observed. The authors cal- culated the excess annual mortality due to lung cancer. When workers with low-to-moderate exposure experienced a mean ex- cess lung cancer mortality of 63 deaths (per 100,000 years' expo- sure). Those severely exposed for less than 2 years experienced an excess of 44 deaths, and those severely exposed for 2 years or longer experienced an excess of 241 deaths. Interestingly, an examination of deaths did not reveal any significant association with age at first employment in the asbestos factory. In the sub-sample of workers whose smoking histories were available, those women who had both smoked and were heavily exposed had a risk of developing lung cancer over 30 times that of non- exposed nonsmoking women. The authors concluded that the data suggested that asbestos and cigarette smoking exert mul- tiplicative rather than merely additive effects. In summary, the data on smoking and asbestos exposure in women closely resemble the findings demonstrated for men. COTTON DUST Approximately 250,000 women were employed in the textile industry in 1978; that population included approximately 100,000 women engaged in spinning, twisting, and winding op- erations. Byssinosis is a syndrome characterized by tightness of the chest and shortness of breath in workers exposed to dust of cotton, flax, and hemp. In addition to these acute symptoms, workers have been found to develop chronic bronchitis, and some become severely disabled by their obstructive lung disease (3). Berry, et al. studied the workers in 14 cotton and 2 man- made fiber mills in England (1). They found that men had a greater prevalence of byssinosis than women, and that smokers of both sexes had 1.4 times greater prevalence of byssinosis than nonsmokers. Byssinosis prevalence was also positively as- sociated with length of exposure to cotton dust in both women and men and was positively associated with dust level in the working environment in women. Berry, et al. were unable to determine if the observed difference in prevalence by sex repre- sented a difference in physiologic response or differences in oc- cupational exposure, They also found a higher prevalence of bronchitis in exposed versus nonexposed workers of both sexes. Smoking workers had higher bronchitis rates than nonsmoking workers. Bouhuys, et al. studied 645 active and retired cotton textile workers (including 372 women), aged 45 and older, who had worked an average of 35 years. Their respiratory symptoms and flow-volume curves were compared to those of community resi- 181 10 X of women employed Severe Inorganic in pulmonary irritant sensitizers Occupation thousands Beryllium & Compounds Platinum Salts Electronic Machinery, Phosphorous Trichloride Equipment & Supplies 899 Tellurium (Hexaflouride) -Household Appliances 6'7 Zinc (Chloride fume) -Radio, T.V. & Communication Equipment 216 -Electrical Machinery, Equipment & Supplies 604 Ammonia Phthalic Anhydride Professional & Photo- Chlorine graphic Equipment BE Ozone Watches 238 Sulfuric Acid -Scientific & Controlling Uranium Compounds Instruments 65 Vanadium Compounds -Optical & Health Services (Pentoxide) Supplies 119 -Photographic Equipment & Supplies 36 Acrolein Cobalt, metal Rubber & Misc. Plastic Ammonia fumes & dust Products 267 Cadmium dust Phthalic Anh ydride -Rubber Products 86 Chlorine -Misc. Plastic Products 171 Chromates Dichloroeth yl ether Ethylene Oxide Hydrogen Chloride g TABLE v.--le ofpdmonary irritants and inorganic sensitizers in various occupations where women work Severe pulmonary irritant Inorganic sensitizers Occupation # of women employed in thousands Hydrogen Fluoride Hydrogen Sulfide Phosgene Phosphorous Trichloride Phthalic Anh ydride Sulfuric Acid Tellurium (Hexafluoride) Zinc Compounds Ammonia Phthalic Anhydride Polyvinyl Chloride Chromic Acid & Chromates Chromium, metals Kz insoluble salts Hydrogen Sulfide Phthalic Anh ydride Sulphur Dioxide Ammonia Cadmium dust/fumes Chromic Acid & Chromates Chromium, metal & insoluble salts Fluorine Hydrogen Chloride Nitrogen Dioxide Sulfuric Acid Zinc Chloride fumes Leather & Leather Products -Footwear, except rubber -Leather Products, except footwear Fabricated Metal Products -Cutlery, hand tools, & other hardware -Fabricated structural metal products -Screw machine products -Metal stamping -Misc. fabricated metal products 177 13 40 299 52 78 26 43 101 s TABLE 7.--(Continued) lb Severe pulmonary irritant Inorganic sensitizers Occupation # of women employed in thousands Chlorine Hydrogen Fluoride Chlorine Chlorine Dioxide Chromium, metal & insoluble salts Nitric Acid Sulfur Dioxide Sulfuric Acid Detergents (Enzymatic) Detergents (Enzymatic) Cobalt Personal Services -Laundering, Cleaning, & other Garment Services -Beauticians 231 492 Ammonia Private Households 1,217 Chlorine Hotels & Motels 424 Beryllium & Beryllium compounds Chromic Acid & Chromates Chromium, metal & insoluble salts Iodine Selenium Hexafluoride Zinc Chloride fumes Cobalt, metal fumes & dust Detergentes (Enzymatic) Platinum Salts Professional & Related Services -Hospital Workers -Offices of Physicians -Offices of Dentists -Health Services -Convalescent Institutions 11,931 2,866 506 242 473 # of women Severe pulmonary irritant Inorganic sensitizers Occupation employed in thousands Ammonia Antimony Bromine Cadmium dust/fumes Chlorine Chromates Cotton dust, raw Dichloroeth yl ether Dimeth ylamine Ethylene Chlorohydrin Ethylene Oxide Hydrogen Sulfide Methyl Bromide Nitric Acid Nitrogen Dioxide Sulfur Dioxide Sulfuric Acid Zinc Chloride fumes Cobalt dust Phthalic Anhydride Textile Mill Products -Knitting Mills -Yarn, thread & fabric mills -Misc. Textile mill products Apparel & other fabricated textile products -Apparel & Accessories -Misc. fabricated textile products 409 126 229 23 995 898 97 Hydrogen Sulfide Meat Products SOURCE: National Clearinghouse for Smoking and Health (8), Rones, P. (1~ Stellman, J. (16). 12,986 dents who acted as controls (3). Textile workers of both sexes had significantly increased prevalence of chronic cough, wheez- ing, and dyspnea. Work in the textile mills was the major vari- able associated with symptom prevalence, with smoking as an additional significant variable. The lung function data con- firmed the association of both smoking and working in the mills with decreased lung function. Nonsmoking female workers were slightly more likely to report chronic cough than nonsmoking men, but smoking male workers were almost twice as likely to report this symptom as smoking women. A similar pattern was seen for wheezing and chest tightness, but not for dyspnea. Kilburn, et al. studied the prevalence of byssinosis and bron- chitis in 1,046 women textile workers and showed an interaction of smoking and work exposure in producing a higher prevalence rate of both byssinosis and bronchitis at a given dust level (5). In summary, women have clearly been shown to have a higher risk of developing byssinosis, chronic bronchitis, and chronic obstructive lung disease because of exposure to cotton dust in the workplace. Cigarette smoking has been shown to interact with some work exposures to increase this risk, al- though it is not established whether this interaction is additive or multiplicative. Men employed in occupations where they are exposed to cotton dust have a greater prevalence of bronchitis and respiratory disability than women. Clarification is neces- sary to determine whether this is a sex difference or a dif- ference in exposure (either occupational or smoking). 1. The 1979 Surgeon General's Report identified the ways in which smoking cigarettes may interact with the occupational environment. They include: a) Facilitation of absorption of physical contamination of cigarettes, b) Transformation of workplace chemicals into more toxic substances, c) Addition of the exposure to a toxic constituent of to- bacco smoke to a concurrent exposure to the same con- stituent present in the workplace, d) Addition of a health effect due to environmental expo- sure to a similar health effect due to smoking, e) Synergy of exposures, and f) Causation of accidents. 2. Women are entering occupational environments with greater frequency, and thus may be experiencing greater expo- sures to physical and chemical agents. 186 3. Cohorts of women with a greater prevalence of smoking are currently reaching the ages of maximal disease occurrence, re- placing earlier cohorts with lower cigarette exposures. 4. Physiologic differences in hormonal status between males and females constitute a potential source of differing responses. 5. In the workplace women who are pregnant present a g-month exposure opportunity, including potential teratogenic and perinatal mortality effects. 6. Concurrent exposure of women to smoking and asbestos resulted in a clear excess of cancer of the lung. 7. Women smokers exposed to cotton dust run a higher risk of developing byssinosis, bronchitic syndromes, and abnormal pulmonary function tests than nonsmoking women. References (1) BERRY, G., MOLYNEUX, M.K.B., TOMBLESON, J.B.L. Relationships between dust level and byssinosis and bronchitis in Lancashire cotton mills. British Journal of Industrial Medicine 31: 18-27, 1974. (2) BERRY, G., NEWHOUSE, M.L., TUROK, M.E. Combined effect of as- bestos exposure and smoking on mortality from lung cancer in factory workers. Lancet 2(7775): 476-479, September 2, 1972. (8) BOUHUYS, A., SCHOENBERG, J.B., BECK, G.J., SCHILLING, R.S.F. Epidemiology of chronic lung disease in a cotton mill community. Lung 154: 167-186, 1977. (4) HAMMOND, E.C., SELIKOFF, I.J. Relation of cigarette smoking to risk of death or absestos-associated disease among insulation workers in the United States. In: Boguvski, P., Gilson, J.C., Timbrell, V., Wagner, J.C., Davis, W. (Editors). Biological Effects of Asbestos. International Agency for Research on Cancer, Scientific Publication No. 8, Lyon, France, International Agency for Research on Cancer, 1973, pp. 312- 317. (5) KILBURN, K.H., KILBURN, G.G., MERCHANT, J.A. Byssinosis: Mat- ter from lint to lungs. American Journal of Nursing 73(11): 1962-1956, November 1973. (6) NATIONAL CENTER FOR HEALTH STATISTICS. Health Interview Survey, 1976. Department of Health, Education, and Welfare, Public Health Service, National Center for Health Statistics. (Unpublished data) (7) NATIONAL CLEARINGHOUSE FOR SMOKING AND HEALTH. Adult Use of Tobacco, 1975. Department of Health, Education, and Welfare, Public Health Service, Center for Disease Control, Bureau of Health Education, National Clearinghouse for Smoking and Health, June 1976. (8) NATIONAL CLEARINGHOUSE FOR SMOKING AND HEALTH. Sur- vey of Health Professionals: Smoking and Health, 1975. Department of Health, Education, and Welfare, Public Health Service, Center for Dis- ease Control, Bureau of Health Education, National Clearinghouse for Smoking and Health, June 1976, 42 pp. (9) NATIONAL INSTITUTE FOR OCCUPATIONAL SAFETY AND HEALTH. Occupational Diseases (Revised Edition). Department of Health, Education, and Welfare, Public Health Service, Center for Dis- 187 ease Control, National Institute for Occupational Safety and Health, June 1977,608 pp. (10) NEWHOUSE, M.L. Cancer amongworkers in the asbestos textile indus- try. In: Buguvski, P., Gilson, T.C., Timbrell, V., Wagner, J.C., Davis, W. (Editors). Biological Effects of Asbestos. International Agency for Re- search on Cancer, Scientific Publication No. 8, Lyon, France, Interna- tional Agency for Research on Cancer, 1973, pp. 203-208. (11) NEWHOUSE, M.L., BERRY, G., WAGNER, J.C., TUROK, M.E. A study of the mortality of female asbestos workers. British Journal of Indus- trial Medicine 29: 134-141, 1972. (12) POLLER, L., THOMSON, J.M., OTRIDGE, B.W., YEE, K.F., LOGAN, S.H.M. Effects of manufacturing oral contraceptives on blood clotting. British Medical Journal 1: 1761-1762, June 30, 1979. (13) PROCTOR, N.H., HUGHES, J.P. Chemical Hazards of the Workplace. Philadelphia, J.B. Lippincott Company, 1978, 533 pp. (14) RONES, P., LEON, C. Employment and unemployment during 1978: an analysis. Special Labor Force Report 218. Department of Labor, Bureau of Labor Statistics, 1979. (15) SELIKOFF, I.J., HAMMOND, E.C., CHURG, J. Asbestos exposure, smoking, and neoplasia. Journal of the American Medical Association 204(2): 106-112, April 8, 1968. (16) STELLMAN, J., DAUM, S.M. Work is Dangerous to Health. New York, Pantheon Books, 1973, 448 pp. (17) U.S. DEPARTMENT OF LABOR. Employment and unemployment dur- ing 1978: An analysis. Department of Labor, Bureau of Labor Statis- tics, Special Labor Force Report 218, 1979. (18) U.S. PUBLIC HEALTH SERVICE. Smoking and Health. A Report of the Surgeon General. Department of Health, Education, and Welfare, Public Health Service, Office of the Assistant Secretary for Health, Office on Smoking and Health, DHEW Publication No. (PHS) 79-50066, 1979, 1251 pp. 188 PREGNANCY AND INFANT HEALTH. PREGNANCY AND INFANT HEALTH Introduction A woman who smokes during pregnancy not only risks her own health, but also changes the conditions under which her baby develops. Studies have identified specific areas in which the effects of maternal smoking during pregnancy may occur. These include fetal growth, most often determined by compar- ing birth weights of smokers' babies with those of otherwise similar nonsmokers' babies; spontaneous abortions, fetal deaths, and neonatal deaths; pregnancy complications, includ- ing those that predispose to preterm delivery; possible effects on lactation; and long term effects on surviving children. The relationships between maternal smoking and these outcomes have been established by clinical, pathological, and especially epidemiological studies. Understanding of mechanisms by which smoking may produce the observed effects has been gained by physiological studies in humans and experimental studies in animals, The Chapter on Pregnancy and Infant Health in the 1979 Surgeon General's Report is a detailed review of past studies of the effects of smoking in pregnancy, with a comprehensive bib- liography. This section summarizes current knowledge in major areas of study, describes important new studies, and points out areas requiring further research (146). Smoking, Birth Weight, and Fetal Growth Babies born to women who smoke during pregnancy are, on the average, 200 grams lighter than babies born to comparable women who do not smoke. Since 1957, when Simpson reported this finding from her original study (1381, it has been confirmed in more than 45 studies of more than half a million births (146). Results of these studies are expressed as mean birth weights of smokers' and nonsmokers' babies or, alternatively, as the per- centage-of babies who weigh less than a specified amount, usu- ally 2,500 grams. To illustrate the association between maternal smoking and an increased proportion of low-birth-weight infants, the results of five studies with an aggregated total of almost 113,000 births in Wales, the United States, and Canada are summarized in Table 1. In these populations, 34 to 54 percent of the mothers smoked during pregnancy and on the average the smokers had twice as many low-birth-weight babies as the nonsmokers. Also in these populations, from 21 to 39 percent of the incidence of 191 TABLE l.-Birth weight under 2,500 grams by maternal smoking habit, relative and attributable risks derived from published studies Study Cardiff US Collaborative White Black California, Kaiser Permanente White Black Montreal Ontario Nonsmokers No. 7,176 8,466 11,252 3,189 934 3,954 27,316 Smokers Propor- No. tion 6,238 .465 9,781 .536 7,777 .409 2,145 .402 479 .338 3,004 .432 21,062 .435 Births < 2,500 grams Relative Attribut- Non- risk able smoker Smoker smoker: risk* (76) (o/o) nonsmoker (%) 4.1 8.1 1.98 31 4.3 9.5 2.21 39 10.7 17.5 1.64 21 3.5 6.4 1.83 25 6.4 13.4 2.09 27 5.2 11.4 2.19 34 4.5 9.1 2.02 31 *Percentage of total birth weights < 2,500 gm attributable to maternal smoking. Attributable risk in population = b(r- 1) divided by b(r-1) +l where b = proportion of mothers who smoke and r = relative risk of low weight = smoker rate/nonsmoker rate. SOURCE: Meyer, M.B. (86). 10 - Nonsmokers o =-==** Smokers Birth Weight (Scale in Pounds: Intervals of 4 oz.) FIGURE l.-Percentage distribution by birth weight of infants of mothers who did not smoke during pregnancy and of those who smoked one pack or more of cigarettes per day SOURCE: MacMahon, B. (77). low birth weight could be attributed to maternal smoking (3,15,38,86,102,106,107). One study in which rates of low birth weight were simulta- neously adjusted for multiple factors showed that maternal smoking had a more significant relationship to birth weight than did previous pregnancy history, hospital pay status, mother's prepregnant weight, height, age-parity, or sex of child. Adjusted rates of birth weights under 2,500 grams were 49 per thousand for nonsmokers, 76 per thousand for smokers of less than a pack per day, and 114 per thousand for smokers of a pack per day or more. The risk of having a low-birth-weight baby therefore increased 53 percent and 130 percent for light and heavy smokers, respectively, compared with nonsmokers (86). Population studies that illustrate whole distributions of birth weights by maternal smoking levels show a downward shift of all birth weights in proportion to the amount smoked (74, 77,83,114,136,160) (see Figure 1). 193 These studies show that the relationship between smoking and reduced birth weight is independent of all other factors that influence birth weight, such as race, parity, maternal size, socioeconomic status, sex of child, and other factors that have been studied. It is also independent of gestational age. There is a dose-response relationship: that is, the more the woman smokes during pregnancy, the greater the reduction in birth weight. If a woman gives up smoking by her fourth month of gestation her risk of delivering a low-birth-weight baby is simi- lar to that of a nonsmoker. PLACENTAL RATIOS Analyses of placental weights by maternal smoking habits have noted that these weights were either not affected or were less affected by maternal smoking than were birth weights (57,61,91,104,155). The placental ratio, the ratio of placental weight to birth weight, tended to be larger for smokers than for nonsmokers, mainly because of the dose-related reduction in birth weights with increasing number of cigarettes smoked. Wingerd and colleagues have studied placental ratios based on data from 7,000 pregnancies among members of the Kaiser Foundation Health Plan in Oakland, California (156). Smoking information was obtained early in pregnancy, and placentas were handled according to Benirschke's standardized protocol. Figure 2 shows placental ratios by smoking level and gestation for single live births. At each gestational age, from 37 through 43 weeks, the more the mother smoked during pregnancy, the higher was the placental ratio. These ratios were higher for black than for white women and tended to increase as maternal hemoglobin level decreased (156). Christianson's recent report, based on standardized examina- tions of these placentas, has shown that the increase in placen- tal ratio with maternal smoking level was due to considerable decreases in mean birth weight, accompanied by slight in- creases in mean placental weight. In addition, smokers' placen- tas were significantly thinner than those of nonsmokers, and their minimum diameters were larger (19). Maternal smoking leads to significant increases in car- boxyhemoglobin in maternal and fetal blood, with a consequent reduction in the oxygen carrying capacity of both, and a reduc- tion of the pressure at which oxygen is delivered to the fetal tissues (70,72,146). Christianson .discusses the similarity be- tween studies of placental ratios by smoking level, altitude, maternal anemia, and maternal cyanotic heart disease. She suggests that the changes in placental ratio represent an adap- 194 8 14.5 T- x 14.0 13.5 13.0 / 0 37 38 39 40 41 42 43 `5e5i 15.0 Week of gestation FIGURE 2.- Ratio of placental weight to birth weight by length of gestation and maternal smoking category SOURCE: Wingerd, J. (156). tation to relative fetal hypoxia (19). An adaptive advantage for survival might occur because a larger placenta with an in- creased area of attachment would deliver more oxygen, and a smaller fetus would have a decreased oxygen demand. If so, it is extremely important to know whether this reduction in size is accompanied by any long-term costs in later growth and devel- opment. GESTATION AND FETAL GROWTH In early studies the consistent finding that mean birth weights were lower and the frequency of births under 2,500 grams higher for women who smoked during pregnancy than for similar nonsmokers raised the obvious question of whether this might be due to a smoking-related reduction in gestation. This is not the case. Studies consistently show that mean gesta- tion is minimally reduced by maternal smoking (less than 2 days) (3,13,146,159) and that birth weight is lower for infants of smokers than for infants of nonsmokers at each gestational age (3,15,83,146). 195 The finding that maternal smoking does not cause an overall downward shift in the distribution of gestational ages, as was shown for birth weights of smokers' infants, leads to the conclu- sion that the lower weight must be due to direct retardation of fetal growth. In other words, these infants are small-for-dates rather than preterm. The type of fetal growth retardation as- sociated with maternal smoking is characterized by an abnor- mally short crown-heel length for gestational age (89,90). Smokers' babies are smaller than corresponding nonsmokers' babies in all dimensions measured, including length, head cir- cumference, chest circumference, and shoulder circumference (10,30,31,52,57,61,102,104,146,157). Previous studies of these measurements at birth have in- ferred that birth size reflects the rate of fetal growth; this has been confirmed by a definitive study in which fetal biparietal diameters were measured serially during gestation. Persson and coworkers studied 5,715 pregnancies prospectively, making ultrasonic measurements of biparietal diameters (BPD) from 18 to 20 weeks through term. Separate growth curves of BPD were constructed for fetuses of smokers and nonsmokers who were delivered between 266 and 294 days after the last menstrual period. The BPD increased faster in the nonsmoking group; the difference from the smoking group was significantly apparent from the 28th week and was positively correlated with the aver- age number of cigarettes smoked (Figure 3). Measurements taken at birth showed that the distributions of birth weights and lengths shifted downwards in proportion to the level of smoking. Figure 4 illustrates this shift (114). These findings cor- roborate Miller's characterization of smokers' babies as nor- mally proportioned but short as well as light for dates, and smaller in all dimensions than babies of nonsmokers (90). The data are also consistent with the speculation that relative fetal hypoxia results in a slower mitotic rate, a baby with fewer cells, and a reduced oxygen demand. LONG-TERM GROWTH AND DEVELOPMENT Possible long-term consequences of maternal smoking during pregnancy are also of concern. Several long-term studies pro- vide evidence that children of smoking mothers have slight but measurable deficiencies in physical growth, intellectual and emotional development, and behavior (95). Because these complex outcomes are affected by many known and unknown factors, it is important to take these other factors into account in any attempt to measure long-term effects of maternal smoking. Several well-controlled studies have shown 196 mm 8 7( 6( BPD = - = = 0 nonsmokers 0 heavy smokers I 1 I 25 1 30 35 40 weeks FIGURE 3.-Fetal Biparietal Diameters (BPD) values [means and standard error of means @EM)] of nonsmokers and heavy smokers (10 cigarettes/day) plotted in relation to postmenstrual age against the normal range (shaded area depicts 95% confidence interval) SOURCE: Persson, P.H. (114). 197 that the physical growth of smokers' babies remains behind that of nonsmokers' babies as measured at 7 to 14 days (31); 1 year, 4 years, and 7 years (pairs of births matched for race, date of delivery, maternal age and education, and sex of child) (52); 5 years (adjusted for other factors) (157); up to 6% years (prospec- tive study) (35); and at ages 7 and 11 (follow-up studies of the 17,000 children from the British Perinatal Mortality Study, with the adjustment for other social and biological factors) (16,30,33). Associations have also been noted between maternal smoking and deficiencies in neurological and intellectual development of the child. Hardy and Mellits analyzed findings for 88 pairs of children of smokers and nonsmokers, matched for race, date of delivery, maternal age and education, and sex of the child. Al- though they reported no significant differences in intellectual function between children born to smoking and nonsmoking mothers, the direction of difference on almost all tests was in favor of the nonsmokers' babies. Fewer smokers' than nonsmokers' children had normal neurological status at age 1 year, both in the original 88 matched pairs and in the additional set of 55 pairs of children of smokers and nonsmokers, matched for birth weight as well as for the other cited factors. In both sets, smokers' children had lower scores on the majority of tests of intelligence and intellectual function at ages 4 and 7 (52,146). Similarly, Dunn evaluated neurological, intellectual, and be- havioral status in a prospective study of low-birth-weight in- fants, including 76 who were "small-for-dates" (term and pre- term), 92 "truly premature" (preterm with birth weight be- tween 11 and 89 percentile) and 151 full-birth-weight control infants. Neurological abnormalities, including minimal cerebral dysfunction and abnormal or borderline electroencephalo- grams, were slightly more common among children born to women who smoked (Table 2). In a battery of psychological tests, the mean scores of children of nonsmoking mothers were better than those of smokers' chil- dren in 45 out of 48 correlations, and the difference was signifi- cant in 14 of these. Some significant differences in favor of nonsmokers' children were also demonstrated with respect to behavior ratings and school placement (35). These results are very similar to those of Hardy and Mellits in that the direction of the differences was almost always in favor of the nonsmoker's child. Small numbers and population selection factors were not a problem in the longitudinal follow-up of the population origi- nally included in the British Perinatal Mortality Study, com- prising approximately 17,000 births, an estimated 98 percent of all births in England, Scotland, and Wales during the week of 198 TABLE 2.-Incidence of neurological abnormalities at about 6% years, by maternal smoking habits Percent of Children with Diagnosis Maternal Smoking Habits Diagnosis Smoker Nonsmoker P Minimal cerebral dysfunction Total neurological abnormalities EEG borderline or abnormal Low-birth-weight children Full-birth-weight children NS = not significant. SOURCE: Dunn, H.G. (35). 20.0 11.0 < .05 29.4 19.5 < .05 46.3 32.4 NS 28.2 21.6 NS March 3 to 9,1958. These children have been traced and studied again at ages 7 and 11, to describe their behavior, their health, their physical development, their educational standards, and their home environment. At ages 7 and 11 years, physical and mental problems due to maternal smoking during pregnancy were found, and these increased with the number of cigarettes smoked. Children whose mothers smoked 10 or more cigarettes a day during pregnancy were on average 1.0 centimeter shorter and 3 to 5 months retarded in reading, mathematics, and general abil- ity, as compared with the offspring of nonsmokers. After allow- ing for associated social and biological factors, all of these dif- ferences were highly significant, as illustrated in Figure 5 (p C 0.001) (16,30). Denson's case-control study of hyperkinesis reported a highly significant association of hyperkinesis with heavy maternal smoking, which at a mean level of 23.3 cigarettes per day was more than three times the average for two control groups. The authors concluded that their findings were "consistent with the hypothesis that smoking during pregnancy is an important cause of the hyperkinetic syndrome" (31). A recent comparison by Saxton of behavioral patterns of in- fants of mothers who smoked during pregnancy with infants of mothers who did not smoke found that these patterns can be influenced by smoking in pregnancy, and that the auditory senses are particularly affected. Fifteen smokers of more than 15 cigarettes per day and 17 nonsmokers were selected for study, matched for maternal age, social class, and parity. All infants were spontaneous term deliveries of normal birth weight. Sex distribution, length of labor, analgesia, and obstet- rical factors were similar for the two groups. Examiners who did 199 - Nonsmokers - - - All smokers 44 46 48 50 52 54 56 FIGURE 4.- Distribution of birth lengths Length, cm SOURCE: Person, P.H. (114). not know the smoking status of the mother evaluated the in- fants at 4 to 6 days of age, using the Brazelton Neonatal Behav- ioral Assessment Scale. The scale includes a total of 20 tests and maneuvers. While many of these showed no statistically signifi- cant differences, auditory tests or tests with auditory compo- nents were significantly different. Recorded "overall im- pressions " of the infants at the end of the test showed that the smokers' infants tended towards "irritability, decreased ability for self-control, and a general lack of interest, whereas the nonsmokers, infants tended to be less irritable and better oriented." The author concluded that some effect on the normal 200