- Email: [email protected]
Tobacco, alcohol, and caffeine use and cessation in early pregnancy
Tobacco, Alcohol, and Caffeine Use and Cessation in Early Pregnancy Phyllis L. Pirie, PhD, Harry Lando, PhD, Susan J. Curry, PhD, Colleen M. McBride, PhD, Louis C. Grothaus, MA Objectives:
Little is known about what happens when individuals attempt to make multiple behavior changes simultaneously. Pregnant women in particular are often in the position of needing to change several behaviors at once, including giving up more than one pleasurable substance. We investigated the success of pregnant women in spontaneously quitting tobacco, alcohol, or caffeine, alone or in combination.
Methods:
Pregnant women (n ⫽ 7489) were identified in the practices of large health maintenance organizations in Seattle and Minneapolis and were interviewed by telephone. Analyses examined the patterns of using and quitting more than one substance, and the extent to which using more than one substance predicts ability to quit other substances.
Results:
Use of the three substances tended to cluster within individuals. Users of multiple substances were less likely to quit each substance than users of single substances. However, in the subgroup of multiple substance users who had quit one substance, having quit a second substance was more, rather than less, common. In multivariate analyses predicting quitting, demographic variables, and not having been pregnant previously were significant predictors of quitting each substance; being a nonsmoker predicted quitting alcohol, and being a nonsmoker and nondrinker predicted quitting caffeine.
Conclusions: The reasons for difficulty in quitting more than one substance are unknown but may include the difficulty of formulating appropriate behavioral strategies or less concern about healthy behavior in pregnancy. Many women in the study successfully quit using two substances, however, and counseling should focus on achieving that outcome. Medical Subject Headings (MeSH): pregnancy, smoking, alcohol, caffeine, risk behavior, life change (Am J Prev Med 2000;18(1):54 – 61) © 2000 American Journal of Preventive Medicine
Introduction
D
espite widespread concern and media attention to the problem of illicit substance use during pregnancy, the use of legal substances such as tobacco, alcohol, and caffeine affect a far greater number of pregnancies each year1,2 and may have more significant health effects.3 Tobacco use during pregnancy has been shown to be causally related to intrauterine growth retardation, prematurity, and low birthweight,4 as well as to late fetal and infant mortality5 and possibly to developmental delays.6,7 Alcohol use during pregnancy is a major cause of preventable mental retardation8,9 and has been linked to a constellation of From the Division of Epidemiology, University of Minnesota (Pirie, Lando), Minneapolis, Minnesota; Group Health Cooperative of Puget Sound, Center for Health Studies (Curry, Grothaus), Seattle, Washington; Duke University Medical Center, Comprehensive Cancer Center (McBride), Durham, North Carolina Address correspondence and reprint requests to: Phyllis L. Pirie, PhD, Division of Epidemiology, University of Minnesota, 1300 South Second Street, Suite 300, Minneapolis, MN 55454-1015. E-mail: [email protected].
54
Am J Prev Med 2000;18(1) © 2000 American Journal of Preventive Medicine
physical and neurologic defects collectively known as fetal alcohol syndrome.10 Lower levels of alcohol use during pregnancy are also known to be linked to neurobehavioral deficits and intrauterine growth retardation.11–13 Caffeine has been demonstrated to produce birth defects and fetal mortality in animal models, but its effects in humans at normal levels of consumption are much less certain. There is some evidence that caffeine may be related to low birthweight, possibly due to intrauterine growth retardation.14 Some evidence suggests an interactive effect on fetal growth when cigarettes, alcohol, and caffeine are all used.15 Pregnant women are often advised to avoid all three substances, which places some women in the unenviable position of being asked to give up two or three pleasurable substances simultaneously. The problem of trying to change several behaviors at once has not been extensively studied. Although it is well known that certain unhealthy behaviors tend to co-occur or “cluster” within individuals, both during pregnancy2 and in the general population,16,17 little attention has been
0749-3797/00/$–see front matter PII S0749-3797(99)00088-4
paid to what happens when individuals attempt to change several behaviors simultaneously. The suggestion has been made, based on theoretical considerations and studies in animal models, that restriction of one reinforcer tends to increase the use of other available reinforcers18; thus, we would expect that reduction in the use of one of these substances would increase the use of others, and quitting both at once would be exceptionally difficult. Observations of attempts to change two behaviors simultaneously outside of laboratory studies are limited, but researchers attempting to modify both smoking behavior and eating (dieting) behavior simultaneously have reported little success.19,20 The problem of simultaneously attempting to change several behaviors is of importance both theoretically and to clinicians who are faced with this situation. The data on clustering of behaviors suggest that encountering a pregnant patient who is using more than one substance is fairly common; however, even basic descriptive data on the magnitude of the problem and its implications for improving health behavior are lacking. The purpose of this paper is to provide descriptive information on the use of three common, legal substances (alcohol, tobacco, and caffeine) prior to pregnancy in a medically well-served, middle-class population of women; the rates of quitting use of these substances in early pregnancy; and patterns of quitting the use of these substances by women using one or more legal substance.
Methods Data were collected by telephone survey conducted with women who had scheduled their first obstetric visit at Group Health Cooperative of Puget Sound (Seattle, Washington) or Park Nicollet Medical Center (Minneapolis, Minnesota). Surveys occurred between April 1993 and February 1995. In each city, interviewing staff used the same survey questionnaire and were trained to use similar interviewing methods. In Minnesota, the clinic sent women an advance letter, alerting them to the survey and requesting passive consent to convey their names to the interviewing organization (Data Collection and Support Services, Division of Epidemiology, University of Minnesota). A return postcard and phone number were provided as options to decline participation. In Seattle, the interviewing organization was part of Group Health Cooperative and could approach the women directly to invite them to participate in the survey. Women were ineligible to participate in the interview if, upon initial contact, they were found to be under age 18, unable to complete the interview in English because of language or hearing problems, or not currently pregnant (due to miscarriage, abortion, or premature delivery). The telephone interview included a variety of questions on health-related issues in
early pregnancy; many of the questions concerned cigarette smoking, because an important purpose of the interview was to identify potential participants for a smoking-cessation trial.21 Caffeine consumption was assessed by items that asked for the number of caffeinated drinks consumed currently and prior to pregnancy; results were categorized as those who had such drinks not at all, less than once a day, or once a day or more. Caffeine quitters were those who reported having consumed caffeinated beverages prior to pregnancy but who reported their current consumption as “not at all.” Alcohol consumption was assessed by items that asked for the number of drinks (defined as: “a can or bottle of beer, a wine cooler, a glass of wine, or one cocktail or shot of liquor”) in a typical week at present, and prior to pregnancy. Results were categorized as not at all, less than once a week, at least once a week, and four or more drinks a week. Alcohol quitters were those who moved from any level of consumption prior to pregnancy to no consumption at all during pregnancy. Cigarette smoking was assessed by items that asked about any cigarette smoking in the past seven days; the number of cigarettes per day in the past seven days; and the number of cigarettes per day on average before becoming pregnant. Results were classified as no smoking, less than one cigarette a day, one to five cigarettes a day, and five or more cigarettes a day. Cigarette quitters were those who moved from any level of consumption prior to pregnancy to no smoking at all in the past seven days during the pregnancy. For all three substances, the phrase “prior to pregnancy” was not specifically defined in the survey. Since interviews were completed either prior to or just after the first prenatal visit, behavior change assessed by these questions would have occurred spontaneously and generally not due to provider intervention. Standardly assessed demographic items included age, education, employment status (employed full time, part time, or not at all), race or ethnic identification, and marital status. Respondents were also asked to supply an estimated gestation, either by stating the projected due date, the date of the last menstrual period, or if neither of those were known, by giving an estimate of the number of weeks they were pregnant. Wanting a pregnancy was assessed by a question asking the woman to self-report whether she “wanted to be pregnant sooner; wanted to be pregnant at this time; wanted to be pregnant but at a later time; or did not want to be pregnant now or in the future.” Those giving the first two responses were categorized as having a wanted pregnancy. Comparisons of the demographic characteristics of users and nonusers of each substance prior to pregnancy were made using chi-square analyses; odds ratios were calculated using logistic regression analysis. Chisquare analyses were used to compare the proportion of the study population using each substance prior to Am J Prev Med 2000;18(1)
55
and during early pregnancy; the proportion of lighter and heavier substance users who quit each substance during early pregnancy; and the proportion of users of single and multiple substances who quit each substance. To compare continuous/ordinal variables such as age, t-tests were used. The expected proportions of users of multiple substances were calculated from the overall population frequency of use of each substance, assuming independence, and the resulting distribution of use was compared to the observed distribution, again using chi-square analysis. For users of multiple substances, odds ratios (OR) and confidence intervals (CI) were computed to see if quitting one substance increased or decreased the odds of quitting the second substance. Multiple logistic regression analysis was used to simultaneously consider the effects of several independent variables on the probability of quitting each substance. All analyses were conducted using Statistical Analysis System.
Results Participants In Minnesota, interviews were completed with 3935 of 4513 eligible women (the total number of eligible women includes 315 who declined to have their names forwarded for contact by the survey organization), for a response rate of 87.2%. In Seattle, interviews were completed with 3554 of 3925 eligible women, for a response rate of 90.5%. As shown in Table 1, the resulting population (n ⫽ 7489) was predominantly white, middle class, and married or living with a partner. As compared to nonusers of cigarettes, those who smoked prior to pregnancy were significantly younger, less well educated, and less likely to be employed; they were more likely to be white, less likely to be married or living with a partner, and less likely to be experiencing their first pregnancy. Users of alcohol prior to pregnancy were slightly older than their nondrinking peers, more likely to be white, employed full time, experiencing their first pregnancy, and less likely to be married or living with a partner. Caffeine users differed only slightly from nonusing peers; they were slightly older, not as well educated, more likely to be employed full time, and more likely to be white.
Prevalence of Substance Use Use of the three legal substances prior to pregnancy was common (Table 2), with 65.6% drinking at least some alcohol and 92.6% drinking caffeinated beverages at least occasionally. Only 17.1% reported smoking prior to pregnancy. At the time of the interview in early pregnancy (average 8 weeks gestation in Seattle; 11 weeks gestation in Minneapolis), reported use of all three substances had dropped markedly, with only 9% reporting any cigarette smoking and 5.2% reporting 56
any alcohol consumption. Sixty-seven percent continued to report some caffeine consumption, although the majority of those reported consumption of less than one caffeinated beverage per day. For all three substances, the proportion of both heavy and light users fell, and the proportion of nonusers rose significantly in early pregnancy. As shown in Table 3, for each substance, those who used lower levels of the substance prior to pregnancy were significantly more likely to quit its use than those who used higher levels, but the difference was most pronounced for cigarettes. Among those smoking five or fewer cigarettes per day prior to pregnancy, 88% reported having quit smoking; among those smoking more than five cigarettes per day, 33% reported quitting and 25% reported cutting down to fewer than five per day. A more detailed analysis of quitting cigarettes showed a nearly linear relationship between the number of cigarettes smoked per day prior to pregnancy and the probability of quitting in early pregnancy (data not shown). Self-reported quitting of alcohol was nearly universal, with 84% of those consuming more than three drinks a week reporting quitting, as compared to 94% of those consuming three or fewer drinks a week. For caffeinated beverages, 23% of those drinking at least one beverage a day quit, as compared to 41% of those drinking less than one a day.
Prevalence of Multiple Substance Use Use of these substances tends to cluster within individuals, as indicated in Table 4, which shows the proportion of the population using each substance alone or in combination with other substances, in comparison to the population proportion who would be expected to use each substance alone or in combination if use of the substances were independent. Use of a single substance (smoking alone, alcohol alone, or caffeine alone) occurs less frequently in the population than would be expected by chance, as does the combined use of cigarettes and alcohol without caffeine. Use of combinations of cigarettes plus caffeine and alcohol plus caffeine is seen somewhat more often in the population than would be expected by chance. Use of all three substances is 2.5 times more common than would be expected by chance. As shown in Table 5, users of multiple substances are less likely to quit substance use than users of single substances, and the more substances that are used, the fewer individuals who are able to quit. Thus, for example, 89% of those who used only alcohol reported quitting its use, whereas 86% of those who used alcohol along with either caffeine or tobacco reported quitting alcohol, and 75% of those who used all three substances reported quitting alcohol. The same pattern is seen for caffeine, with over 26% of those who use only caffeine quitting its use, but 16% of those who use one
American Journal of Preventive Medicine, Volume 18, Number 1
Table 1. Characteristics of those usinga and not using cigarettes, alcohol, and tobacco prior to pregnancy; statistical significance of differences between users and nonusers calculated by t-test (age) and chi-square procedures; odds ratios calculated from logistic regression Total population (n ⴝ 7,489)
Cigarettes
Users
Nonusers
Alcohol Odds ratios Users
(n ⴝ 929) (n ⴝ 6558) Age (years)
29.6 27.6
29.9
11
1
35 41 12
62 19 19
61 15 24
29.7
87 13
89 11
3
2
1.32
13 32
11.71 16 5.58 31
16 34
0.94 0.88
17 33
13 34
1.39 1.03
54
(ref)2 51
48
(ref)
48
51
(ref)
62 20 19
87 13
81
95
6
19
5
27 73
p ⬍ 0.001 0.92
94
31 69
29.3
3
32 68
p ⬍ 0.001
NS 1.32 70 (ref) 16 1.71 14
61 20 20
1.46 (ref) 0.89
65 18 17
p ⬍ 0.001 (ref) 96 0.79 4
86 14
0.21 89 (ref)
11
(ref) 0.28
89 11
p ⬍ 0.01
1.32 (ref) 0.73
82 18
(ref) 0.57
p ⬍ 0.001
94
0.50
93
94
0.87
6
(ref)
7
6
(ref)
32 68
0.92 (ref)
p ⬍ 0.001 0.80 35 (ref) 65
55 20 25 p ⬍ 0.001
p ⬍ 0.001
p ⬍ 0.001 Prior pregnancy (percent) No Yes
Odds ratios
2
p ⬍ 0.05 Marital status (percent) Married or living with partner Not married or living with partner
29.5 p ⬍ 0.0001
37.02
Nonusers
30.3
p ⬍ 0.001 Race/ethnicity (percent) Caucasian Other
Odds ratios Users
(n ⴝ 5113) (n ⴝ 2375)
p ⬍ 0.001 Employment status (percent) Employed full time Employed part time Not employed
Nonusers
(n ⴝ 888) (n ⴝ 6598)
p ⬍ 0.0001 Education (percent) Did not graduate 3 from high school High school graduate 16 Some college, 33 vocational, or technical school College graduate 49
Caffeine
31 69 p ⬍ 0.01
NS 1.23 (ref)
31 69
NS
a
Tabled figures are for those at higher levels of use of each substance: 6 or more cigarettes per day; 4 or more drinks of alcohol per week; 1 or more caffeinated beverages per day. ref, reference category for calculation of odds ratios
other substance quitting caffeine, and less than 12% of those who use both other substances quitting. Although there is a similar pattern for cigarettes, with quitting more common among those who used only cigarettes as compared to those who used two or three substances, the differences are small and nonsignificant. Multivariate logistic regression models were devel-
oped to investigate the characteristics of those who quit using each of the three substances (tobacco, alcohol, or caffeine) in early pregnancy. Variables included in each model were pre-pregnancy use of the other two substances, educational level, employment status, ethnicity, whether or not the woman had ever been pregnant before, and whether or not the woman desired to be Am J Prev Med 2000;18(1)
57
Table 2. Proportion of the population using each substance before pregnancy and during early pregnancy, by level of substance use Prior to pregnancy Cigarettes None Up to 5 cigarettes per day More than 5 cigarettes per day Alcohol None Up to 3 drinks per week More than 3 drinks per week Caffeine None Less than daily use Use one or more times daily
During early pregnancy
82.9% 91.1% 4.7% 3.7% 12.4% 5.3% p ⬍ 0.001* 34.5% 94.8% 53.7% 5.1% 11.9% 0.1% p ⬍ 0.001* 7.4% 33.0% 24.3% 39.8% 68.3% 27.2% p ⬍ 0.001*
*p values from 2 df chi-square test comparing distributions for users prior to and during early pregnancy, for each substance.
pregnant at this time or sooner. In the model predicting who would quit smoking, only higher educational level and having never been pregnant before were
Table 3. Percent quitting or reducing use of each substance in early pregnancy, by level of use Percent quit Cigarettes Among those smoking ⱕ5 cigarettes per day (351) Among those smoking ⬎5 cigarettes per day (929) Alcohol Among those drinking ⱕ3 drinks per week (4018) Among those drinking ⬎3 drinks per week (888) Caffeine Among those drinking caffeinated beverages less than daily (1820) Among those drinking caffeinated beverages once a day or more (5113)
Percent reduced to below threshold level*
88 33
Substance
n
Observed Expected Odds percent percent ratio
None 2168 28.9 Cigarettes only 88 1.2 Alcohol only 95 1.3 Caffeine only 3715 49.6 Cigarettes plus alcohol 22 0.3 Cigarettes plus caffeine 627 8.4 Alcohol plus caffeine 579 7.8 Cigarettes, alcohol, 192 2.6 and caffeine
24.5* 3.5* 3.3* 52.7* 0.5† 7.5† 7.1† 1.0*
1.26 0.33 0.38 0.88 0.63 1.13 1.10 2.59
a
Tabled figures are for those at higher levels of use of each substance: ⱖ5 cigarettes per day; ⬎3 drinks of alcohol per week; 1 or more caffeinated beverages per day. *p value for 1 df chi-square test, comparing observed to expected, ⬍0.001. †p value for 1 df chi-square test, comparing observed to expected, ⬍0.05.
significant predictors of quitting. In the model predicting who would quit using alcohol, being a baseline nonsmoker, being nonwhite, and never having been pregnant before predicted quitting alcohol. In the model predicting who would quit using caffeine, being a nonsmoker, being a nondrinker, having a relatively higher education level, being nonwhite, never having been pregnant before, and wanting to be pregnant at this time all predicted quitting caffeine. These results confirm that, even taking other potentially predictive variables into account, simultaneous baseline use of another substance predicts a reduced probability of quitting alcohol or caffeine.
25 Table 5. Percent quitting each substance (cigarettes, alcohol, caffeine) in early pregnancy, among those using that substance alone or in combination with other substances prior to pregnancy
p ⬍ 0.001† 94
Percent quitting 84
15
p ⬍ 0.001 41 23
37
p ⬍ 0.001
*Threshold level is defined as smoking ⱕ5 cigarettes per day, drinking ⱕ3 alcoholic drinks per week, or drinking caffeinated beverages less than daily. †p values are based on 1 df chi-square test comparing quit rates among lighter and heavier users of each substance.
58
Table 4. Proportion of the population using cigarettes, alcohol, or caffeinea prior to pregnancy, alone or in combination; compared to the proportion expected, assuming use of each substance was independent of the other substances
Cigarettes Cigarettes alone Cigarettes plus one other substance Cigarettes plus two other substances Alcohol Alcohol alone Alcohol plus one other substance Alcohol plus two other substances Caffeine Caffeine alone Caffeine plus one other substance Caffeine plus two other substances
34.1 32.5 31.8 NS* 89.5 86.0 75.0 p ⬍ 0.001* 26.4 16.0 12.0 p ⬍ 0.001*
*p value from 2 df chi-square test comparing rates among those using ⬍one, two, or more than two substances.
American Journal of Preventive Medicine, Volume 18, Number 1
Although women using multiple substances were less likely to quit their use than women using single substances, one might ask whether those women using multiple substances who did decide to quit one substance were more or less likely to quit a second substance. Among women using both cigarettes and caffeine prior to pregnancy, women who quit using caffeine during early pregnancy were 2.4 times more likely to quit cigarettes (OR ⫽ 2.4; 95% CI ⫽ 1.7 to 3.8). Among users of both alcohol and caffeine, quitting caffeine was associated with increased odds of quitting alcohol and vice versa (OR ⫽ 6.4; 95% CI ⫽ 2.8 to 14.3). Only for alcohol and cigarettes was quitting one substance not associated with an increased likelihood of quitting the other (OR ⫽ 1.7; 95% CI ⫽ 0.8 to 3.8).
Discussion Smoking rates prior to and during early pregnancy appear to be lower in these middle-class, insured populations than in other populations in the literature, with 17.1% reporting smoking prior to pregnancy and 9.0% reporting smoking during early pregnancy. Recent data reports for national samples show higher smoking rates. Birth certificate data reported by the National Center for Health Statistics (NCHS) indicated that 14.6% of women smoked during pregnancy in 1994.22 The National Health Interview Survey for 1990 reported that 24.6% of women aged 18 to 44 smoked in the year prior to their most recent pregnancy, whereas 15.4% reported smoking during their most recent pregnancy.23 The National Pregnancy and Health Study, which was based on a sample of women interviewed in the hospital after giving birth, reported 20.4% smoking during pregnancy in 1992.24 Whereas the timing of self-reports of smoking is somewhat different in our study than in the national samples, it does appear that smoking in this largely middle-class population is less common than in national samples. Self-report of alcohol use prior to pregnancy was higher in our population than in national samples, whereas self-reported drinking in early pregnancy was lower. We found 65.6% drinking prior to pregnancy, and 5.2% drinking during pregnancy. Behavioral Risk Factor Survey data for 1995 show 50.6% of women of childbearing age drinking prior to pregnancy and 16.5% drinking during pregnancy.25 An analysis of the Pregnancy Risk Assessment Monitoring System data from four states26 reported 47.6% drinking prior to pregnancy, and about 12% drinking during the last three months of pregnancy. Birth certificate data from NCHS, however, show only 1.7% drinking during pregnancy, which the NCHS report describes as “substantially underreported.”22 Because caffeine use during pregnancy has not traditionally been considered a major public health problem, few population-based statistics are available on
caffeine use. Dlugosz and Bracken14 cite unpublished data from Bracken showing that 26% of women giving birth in the Yale–New Haven hospital in 1990 –1991 reported no caffeine intake in the first month of pregnancy, which is similar to our figure of 33% reporting no caffeine intake in early pregnancy. The tendency for the use of alcohol, cigarettes, and caffeine to cluster within the same individuals was not unexpected. That smoking is extremely prevalent in groups being treated for alcohol dependency is well known. Data summarized by Hughes27 show a median rate of 72% heavy smokers in published reports of persons in treatment for alcohol problems. In more general populations, the reported correlations between use of these two substances have been significant but modest. Shiffman and Balabanis28 cite studies showing correlations between any use of alcohol and tobacco on the order of 0.12 and 0.13 and correlations of smoking and drinking rates (amounts) ranging from 0.15 to 0.22. Johnson et al.2 and the National Pregnancy and Health Survey24 also have shown clustering between alcohol and cigarettes, and between these substances and illicit drugs, during pregnancy. Of the three behaviors examined here, women were most likely to quit the use of alcohol during pregnancy, second most likely to quit cigarettes, and least likely to quit caffeine. The low likelihood of quitting caffeine use is most likely attributed to far less attention to this behavior in the mass media and popular sources of advice concerning health in pregnancy. Those who quit caffeine use are likely the most health-conscious subgroup. Both alcohol and tobacco, however, are strongly discouraged during pregnancy; and the dangers of both substances to the developing fetus have been widely publicized. As reported here, of those drinking more than minimal amounts (i.e., more than three drinks a week), 84% reported quitting drinking during early pregnancy. Of those smoking more than minimal amounts (more than five cigarettes per day), 33% quit smoking. Somewhat lower rates of quitting these substances were reported by Kruse et al.,29 who reported 53.2% of drinkers becoming abstinent, whereas 16.7% of smokers did so. They attributed their finding of a greater prevalence of quitting among drinkers than among smokers to an upsurge of publicity concerning the effects of alcohol in pregnancy. Another possible explanation may lie in the probability of physical dependence on the substance in question; it appears that few drinkers in our study were drinking at levels likely to be characteristic of dependent drinkers, whereas many smokers were presumably dependent on tobacco. Few studies exist describing attempts to quit the use of two substances simultaneously, and the data from the existing studies are mixed. Swanson et al.30 suggested that continued use of caffeine by abstinent smokers may increase withdrawal symptoms, because quitting smoking allows blood levels of caffeine to increase and Am J Prev Med 2000;18(1)
59
therefore increases the occurrence of restlessness, impatience, and insomnia. In an experimental study, however, Swanson et al.31 reported no effect of caffeine abstinence on success in smoking cessation at 6 and 12 months. Most withdrawal symptoms in the immediate post-cessation period also did not differ between caffeine abstainers and continued caffeine users, with the exception of fatigue, which was greater among abstainers. Caffeine abstinence occurred only for one week as part of this experiment, so the long-term impact of continued caffeine abstinence is unknown. With regard to simultaneous quitting of smoking and alcohol, on the other hand, in a summary of studies involving alcohol-treatment populations, Bobo et al.32 suggest better drinking-cessation outcomes among smokers who quit smoking than among those who continued to smoke. Data reported here indicate that women using more than one substance may be less likely to quit substance use than women who use only one substance. The reasons for this pattern are unknown, but possibilities include the difficulty of formulating strategies for coping with the effects of quitting multiple substances simultaneously or a generally lower level of concern about the requirements for a healthy pregnancy. On the other hand, among women who quit at least one substance, the odds of having quit another substance were either unaffected or elevated. This pattern suggests a clustering of quitting behavior; women using multiple substances may either quit none of them or several of them, rather than quitting one and continuing others.
Limitations of the Study The cross-sectional design of this study leaves many important questions unanswered. Specifically, we are unable to address issues of timing: do successful quitters make their behavior changes in sequence, or at the same time? Does quitting more than one substance tend to predict relapse later in the pregnancy? These are important questions that subsequent investigations may need to address. The low self-reported smoking rates and high selfreported quit rates for smoking and alcohol in this study raise concerns about the validity of self-report of these behaviors. Underreporting of proscribed behaviors in health surveys is known to occur and may indeed be the case in this study. However, it does not appear that underreporting of individual behaviors or overreporting of quitting would give rise to the patterns of association shown between quitting different substances that were the focus of this research. The specific question formats concerning substance use are also less precise than we would ideally have liked. In particular, the questions concerning level of alcohol and caffeine use were rather imprecise; more 60
specific information, such as patterns of drinking or binge drinking, would have allowed us to examine possible effects of dependency and more precise analysis of the level of use and quitting.
Conclusions From the point of view of the health care provider confronted with a patient who is using more than once substance, these results indicate that quitting more than one substance at the same time is not impossible, and that having successfully quit one substance does not reduce the probability of having successfully quit a second substance. However, because the data suggest that women using more than one substance show a tendency to continue with the use of multiple substances, providers should be sensitive to the difficulties of making this complex set of behavior changes. These data suggest the need for additional research on the best way to go about quitting more than one substance at a time and the factors that lead to success in doing this. The authors wish to acknowledge the contributions of Ms. Karen Virnig, survey director at the University of Minnesota, and Drs. Robert Junilla and Leslie Pratt of Park Nicollet Medical Center in Minneapolis. This study was supported in part by National Heart, Lung and Blood Institute grant HL48121.
References 1. Vega WA, Kolody B, Hwang J, Noble A. Prevalence and magnitude of perinatal substance exposures in California. N Engl J Med 1993;329:850 – 4. 2. Johnson SF, McCarter RJ, Ferencz C. Changes in alcohol, cigarette and recreational drug use during pregnancy: implications for intervention. Am J Epidemiol 1987;126:695–702. 3. Cotton P. Smoking cigarettes may do developing fetus more harm than ingesting cocaine, some experts say. JAMA 1994;271:576 –7. 4. US Health and Human Services. Reducing the health consequences of smoking: 25 years of progress: a report of the surgeon general. Rockville, MD: US Health and Human Services, Public Health Service, Centers for Disease Control, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1989. 5. Malloy MH, Kleinman JC, Land GH, Schramm WF. The association of maternal smoking with age and cause of infant deaths. Am J Epidemiol 1988;128:46 –55. 6. Sexton M, Fox NL, Hebel JR. Prenatal exposure to tobacco: II. Effects on cognitive functioning at age three. Int J Epidemiol 1990;19:72–7. 7. Olds DL, Henderson CR, Tatelbaum R. Intellectual impairment in children of women who smoke cigarettes during pregnancy. Pediatrics 1994;93: 221–7. 8. Abel EL, Sokol RJ. Fetal alcohol syndrome is now leading cause of mental retardation. Lancet 1986;ii:1222. 9. Spohr H-L, Willms J, Steinhausen H-C. Prenatal alcohol exposure and long-term developmental consequences. Lancet 1993;341:907–10. 10. Jones KL, Smith DW, Ulleland CN, Streissguth AP. Pattern of malformation in offspring of chronic alcoholic mothers. Lancet 1973; i:1267–71. 11. Streissguth AP, Barr HM, Sampson PD, Parrish-Johnson JC, Kirchner GL, Martin DC. Attention, distraction and reaction time at age 7 years and prenatal alcohol exposure. Neurobehav Toxicol Teratol 1986;8:717–25. 12. Fried PA, Makin JE. Neonatal behavioral correlates of prenatal exposure to marihuana, cigarettes and alcohol in a low risk population. Neurotoxicol Teratol 1987;9:1–7.
American Journal of Preventive Medicine, Volume 18, Number 1
13. Shu XO, Hatch MC, Mills J, Clemens J, Susser M. Maternal smoking, alcohol drinking, caffeine consumption, and fetal growth: results from a prospective study. Epidemiology 1995;6:115–20. 14. Dlugosz L, Bracken MB. Reproductive effects of caffeine: a review and theoretical analysis. Epidemiol Rev 1992;14:83–100. 15. Peacock JL, Bland JM, Anderson HR. Effects on birthweight of alcohol and caffeine consumption in smoking women. J Epidemiol Community Health 1991;45:159 – 63. 16. Raitakari OT, Leino M, Raikkonen K, et al. Clustering of risk habits in young adults: the cardiovascular risk in young Finns study. Am J Epidemiol 1995;142:36 – 44. 17. Istvan J, Matarazzo JD. Tobacco, alcohol and caffeine use: a review of their interrelationships. Psychol Bull 1984;95:301–26. 18. Vuchinich RE, Tucker JA. Contributions from behavioral theories of choice to an analysis of alcohol abuse. J Abnormal Psychol 1988;97:181–95. 19. Hall SM, Tunstall CD, Vila KL, Duffy J. Weight gain prevention and smoking cessation: cautionary findings. Am J Public Health 1992;82:799 – 803. 20. Pirie PL, McBride CM, Hellerstedt W, et al. Smoking cessation in women concerned about weight. Am J Public Health 1992;82:1238 – 43. 21. McBride CM, Curry SJ, Lando HA, Pirie PL, Grothaus LC, Nelson JC. Prevention of relapse in women who quit smoking during pregnancy. Am J Public Health 1999;89:706 –11. 22. Ventura SJ, Martin JA, Mathews TJ, Clarke SC. Advance report of final natality statistics, 1994. Monthly Vital Statistics Report 1996;44(suppl 1):l 23. LeClere FB, Wilson JB. Smoking behavior of recent mothers, 18-44 years of age, before and after pregnancy: United States, 1990. In: Advance data from vital and health statistics. Hyattsville, MD: National Center for Health Statistics, 1997.
24. National Institute on Drug Abuse. National Pregnancy and Health Survey. Bethesda, MD: National Institutes of Health, 1996. 25. Durham J, Owen P, Bender B, et al. Alcohol consumption among pregnant and childbearing-aged women United States, 1991 and 1995. MMWR Morb Mortal Wkly Rep 1997;46:346 –50. 26. Bruce FC, Adams MM, Shulman HB, Martin ML. Alcohol use before and during pregnancy. Am J Prev Med 1993;9:267–73. 27. Hughes JR. Clinical implications of the association between smoking and alcoholism. In: Fertig JB, Allen JP, eds. Alcohol and tobacco: from basic science to clinical practice. Bethesda, MD: National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism, 1995. 28. Shiffman S, Balabanis M. Associations between alcohol and tobacco. In: Fertig JB, Allen JP, eds. Alcohol and tobacco: from basic science to clinical practice. Bethesda, MD: National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism, 1995. 29. Kruse J, LeFevre M, Zweig S. Changes in smoking and alcohol consumption during pregnancy: a population-based study in a rural area. Obstet Gynecol 1986;67:627–32. 30. Swanson JA, Lee JW, Hopp JW. Caffeine and nicotine: a review of their joint use and possible interactive effects in tobacco withdrawal. Addict Behav 1994;19:229 –56. 31. Swanson JA, Lee JW, Hopp JW, Berk LS. The impact of caffeine use on tobacco cessation and withdrawal. Addict Behav 1997;22:55– 68. 32. Bobo JK, Walker RD, Lando HA, McIlvain HE. Enhancing alcohol control with counseling on nicotine dependence: pilot study findings and treatment implications. In: Fertig JB, Allen JP, eds. Alcohol and tobacco: from basic science to clinical practice. Bethesda, MD: National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism, 1995.
Am J Prev Med 2000;18(1)
61
Little is known about what happens when individuals attempt to make multiple behavior changes simultaneously. Pregnant women in particular are often in the position of needing to change several behaviors at once, including giving up more than one pleasurable substance. We investigated the success of pregnant women in spontaneously quitting tobacco, alcohol, or caffeine, alone or in combination.
Methods:
Pregnant women (n ⫽ 7489) were identified in the practices of large health maintenance organizations in Seattle and Minneapolis and were interviewed by telephone. Analyses examined the patterns of using and quitting more than one substance, and the extent to which using more than one substance predicts ability to quit other substances.
Results:
Use of the three substances tended to cluster within individuals. Users of multiple substances were less likely to quit each substance than users of single substances. However, in the subgroup of multiple substance users who had quit one substance, having quit a second substance was more, rather than less, common. In multivariate analyses predicting quitting, demographic variables, and not having been pregnant previously were significant predictors of quitting each substance; being a nonsmoker predicted quitting alcohol, and being a nonsmoker and nondrinker predicted quitting caffeine.
Conclusions: The reasons for difficulty in quitting more than one substance are unknown but may include the difficulty of formulating appropriate behavioral strategies or less concern about healthy behavior in pregnancy. Many women in the study successfully quit using two substances, however, and counseling should focus on achieving that outcome. Medical Subject Headings (MeSH): pregnancy, smoking, alcohol, caffeine, risk behavior, life change (Am J Prev Med 2000;18(1):54 – 61) © 2000 American Journal of Preventive Medicine
Introduction
D
espite widespread concern and media attention to the problem of illicit substance use during pregnancy, the use of legal substances such as tobacco, alcohol, and caffeine affect a far greater number of pregnancies each year1,2 and may have more significant health effects.3 Tobacco use during pregnancy has been shown to be causally related to intrauterine growth retardation, prematurity, and low birthweight,4 as well as to late fetal and infant mortality5 and possibly to developmental delays.6,7 Alcohol use during pregnancy is a major cause of preventable mental retardation8,9 and has been linked to a constellation of From the Division of Epidemiology, University of Minnesota (Pirie, Lando), Minneapolis, Minnesota; Group Health Cooperative of Puget Sound, Center for Health Studies (Curry, Grothaus), Seattle, Washington; Duke University Medical Center, Comprehensive Cancer Center (McBride), Durham, North Carolina Address correspondence and reprint requests to: Phyllis L. Pirie, PhD, Division of Epidemiology, University of Minnesota, 1300 South Second Street, Suite 300, Minneapolis, MN 55454-1015. E-mail: [email protected].
54
Am J Prev Med 2000;18(1) © 2000 American Journal of Preventive Medicine
physical and neurologic defects collectively known as fetal alcohol syndrome.10 Lower levels of alcohol use during pregnancy are also known to be linked to neurobehavioral deficits and intrauterine growth retardation.11–13 Caffeine has been demonstrated to produce birth defects and fetal mortality in animal models, but its effects in humans at normal levels of consumption are much less certain. There is some evidence that caffeine may be related to low birthweight, possibly due to intrauterine growth retardation.14 Some evidence suggests an interactive effect on fetal growth when cigarettes, alcohol, and caffeine are all used.15 Pregnant women are often advised to avoid all three substances, which places some women in the unenviable position of being asked to give up two or three pleasurable substances simultaneously. The problem of trying to change several behaviors at once has not been extensively studied. Although it is well known that certain unhealthy behaviors tend to co-occur or “cluster” within individuals, both during pregnancy2 and in the general population,16,17 little attention has been
0749-3797/00/$–see front matter PII S0749-3797(99)00088-4
paid to what happens when individuals attempt to change several behaviors simultaneously. The suggestion has been made, based on theoretical considerations and studies in animal models, that restriction of one reinforcer tends to increase the use of other available reinforcers18; thus, we would expect that reduction in the use of one of these substances would increase the use of others, and quitting both at once would be exceptionally difficult. Observations of attempts to change two behaviors simultaneously outside of laboratory studies are limited, but researchers attempting to modify both smoking behavior and eating (dieting) behavior simultaneously have reported little success.19,20 The problem of simultaneously attempting to change several behaviors is of importance both theoretically and to clinicians who are faced with this situation. The data on clustering of behaviors suggest that encountering a pregnant patient who is using more than one substance is fairly common; however, even basic descriptive data on the magnitude of the problem and its implications for improving health behavior are lacking. The purpose of this paper is to provide descriptive information on the use of three common, legal substances (alcohol, tobacco, and caffeine) prior to pregnancy in a medically well-served, middle-class population of women; the rates of quitting use of these substances in early pregnancy; and patterns of quitting the use of these substances by women using one or more legal substance.
Methods Data were collected by telephone survey conducted with women who had scheduled their first obstetric visit at Group Health Cooperative of Puget Sound (Seattle, Washington) or Park Nicollet Medical Center (Minneapolis, Minnesota). Surveys occurred between April 1993 and February 1995. In each city, interviewing staff used the same survey questionnaire and were trained to use similar interviewing methods. In Minnesota, the clinic sent women an advance letter, alerting them to the survey and requesting passive consent to convey their names to the interviewing organization (Data Collection and Support Services, Division of Epidemiology, University of Minnesota). A return postcard and phone number were provided as options to decline participation. In Seattle, the interviewing organization was part of Group Health Cooperative and could approach the women directly to invite them to participate in the survey. Women were ineligible to participate in the interview if, upon initial contact, they were found to be under age 18, unable to complete the interview in English because of language or hearing problems, or not currently pregnant (due to miscarriage, abortion, or premature delivery). The telephone interview included a variety of questions on health-related issues in
early pregnancy; many of the questions concerned cigarette smoking, because an important purpose of the interview was to identify potential participants for a smoking-cessation trial.21 Caffeine consumption was assessed by items that asked for the number of caffeinated drinks consumed currently and prior to pregnancy; results were categorized as those who had such drinks not at all, less than once a day, or once a day or more. Caffeine quitters were those who reported having consumed caffeinated beverages prior to pregnancy but who reported their current consumption as “not at all.” Alcohol consumption was assessed by items that asked for the number of drinks (defined as: “a can or bottle of beer, a wine cooler, a glass of wine, or one cocktail or shot of liquor”) in a typical week at present, and prior to pregnancy. Results were categorized as not at all, less than once a week, at least once a week, and four or more drinks a week. Alcohol quitters were those who moved from any level of consumption prior to pregnancy to no consumption at all during pregnancy. Cigarette smoking was assessed by items that asked about any cigarette smoking in the past seven days; the number of cigarettes per day in the past seven days; and the number of cigarettes per day on average before becoming pregnant. Results were classified as no smoking, less than one cigarette a day, one to five cigarettes a day, and five or more cigarettes a day. Cigarette quitters were those who moved from any level of consumption prior to pregnancy to no smoking at all in the past seven days during the pregnancy. For all three substances, the phrase “prior to pregnancy” was not specifically defined in the survey. Since interviews were completed either prior to or just after the first prenatal visit, behavior change assessed by these questions would have occurred spontaneously and generally not due to provider intervention. Standardly assessed demographic items included age, education, employment status (employed full time, part time, or not at all), race or ethnic identification, and marital status. Respondents were also asked to supply an estimated gestation, either by stating the projected due date, the date of the last menstrual period, or if neither of those were known, by giving an estimate of the number of weeks they were pregnant. Wanting a pregnancy was assessed by a question asking the woman to self-report whether she “wanted to be pregnant sooner; wanted to be pregnant at this time; wanted to be pregnant but at a later time; or did not want to be pregnant now or in the future.” Those giving the first two responses were categorized as having a wanted pregnancy. Comparisons of the demographic characteristics of users and nonusers of each substance prior to pregnancy were made using chi-square analyses; odds ratios were calculated using logistic regression analysis. Chisquare analyses were used to compare the proportion of the study population using each substance prior to Am J Prev Med 2000;18(1)
55
and during early pregnancy; the proportion of lighter and heavier substance users who quit each substance during early pregnancy; and the proportion of users of single and multiple substances who quit each substance. To compare continuous/ordinal variables such as age, t-tests were used. The expected proportions of users of multiple substances were calculated from the overall population frequency of use of each substance, assuming independence, and the resulting distribution of use was compared to the observed distribution, again using chi-square analysis. For users of multiple substances, odds ratios (OR) and confidence intervals (CI) were computed to see if quitting one substance increased or decreased the odds of quitting the second substance. Multiple logistic regression analysis was used to simultaneously consider the effects of several independent variables on the probability of quitting each substance. All analyses were conducted using Statistical Analysis System.
Results Participants In Minnesota, interviews were completed with 3935 of 4513 eligible women (the total number of eligible women includes 315 who declined to have their names forwarded for contact by the survey organization), for a response rate of 87.2%. In Seattle, interviews were completed with 3554 of 3925 eligible women, for a response rate of 90.5%. As shown in Table 1, the resulting population (n ⫽ 7489) was predominantly white, middle class, and married or living with a partner. As compared to nonusers of cigarettes, those who smoked prior to pregnancy were significantly younger, less well educated, and less likely to be employed; they were more likely to be white, less likely to be married or living with a partner, and less likely to be experiencing their first pregnancy. Users of alcohol prior to pregnancy were slightly older than their nondrinking peers, more likely to be white, employed full time, experiencing their first pregnancy, and less likely to be married or living with a partner. Caffeine users differed only slightly from nonusing peers; they were slightly older, not as well educated, more likely to be employed full time, and more likely to be white.
Prevalence of Substance Use Use of the three legal substances prior to pregnancy was common (Table 2), with 65.6% drinking at least some alcohol and 92.6% drinking caffeinated beverages at least occasionally. Only 17.1% reported smoking prior to pregnancy. At the time of the interview in early pregnancy (average 8 weeks gestation in Seattle; 11 weeks gestation in Minneapolis), reported use of all three substances had dropped markedly, with only 9% reporting any cigarette smoking and 5.2% reporting 56
any alcohol consumption. Sixty-seven percent continued to report some caffeine consumption, although the majority of those reported consumption of less than one caffeinated beverage per day. For all three substances, the proportion of both heavy and light users fell, and the proportion of nonusers rose significantly in early pregnancy. As shown in Table 3, for each substance, those who used lower levels of the substance prior to pregnancy were significantly more likely to quit its use than those who used higher levels, but the difference was most pronounced for cigarettes. Among those smoking five or fewer cigarettes per day prior to pregnancy, 88% reported having quit smoking; among those smoking more than five cigarettes per day, 33% reported quitting and 25% reported cutting down to fewer than five per day. A more detailed analysis of quitting cigarettes showed a nearly linear relationship between the number of cigarettes smoked per day prior to pregnancy and the probability of quitting in early pregnancy (data not shown). Self-reported quitting of alcohol was nearly universal, with 84% of those consuming more than three drinks a week reporting quitting, as compared to 94% of those consuming three or fewer drinks a week. For caffeinated beverages, 23% of those drinking at least one beverage a day quit, as compared to 41% of those drinking less than one a day.
Prevalence of Multiple Substance Use Use of these substances tends to cluster within individuals, as indicated in Table 4, which shows the proportion of the population using each substance alone or in combination with other substances, in comparison to the population proportion who would be expected to use each substance alone or in combination if use of the substances were independent. Use of a single substance (smoking alone, alcohol alone, or caffeine alone) occurs less frequently in the population than would be expected by chance, as does the combined use of cigarettes and alcohol without caffeine. Use of combinations of cigarettes plus caffeine and alcohol plus caffeine is seen somewhat more often in the population than would be expected by chance. Use of all three substances is 2.5 times more common than would be expected by chance. As shown in Table 5, users of multiple substances are less likely to quit substance use than users of single substances, and the more substances that are used, the fewer individuals who are able to quit. Thus, for example, 89% of those who used only alcohol reported quitting its use, whereas 86% of those who used alcohol along with either caffeine or tobacco reported quitting alcohol, and 75% of those who used all three substances reported quitting alcohol. The same pattern is seen for caffeine, with over 26% of those who use only caffeine quitting its use, but 16% of those who use one
American Journal of Preventive Medicine, Volume 18, Number 1
Table 1. Characteristics of those usinga and not using cigarettes, alcohol, and tobacco prior to pregnancy; statistical significance of differences between users and nonusers calculated by t-test (age) and chi-square procedures; odds ratios calculated from logistic regression Total population (n ⴝ 7,489)
Cigarettes
Users
Nonusers
Alcohol Odds ratios Users
(n ⴝ 929) (n ⴝ 6558) Age (years)
29.6 27.6
29.9
11
1
35 41 12
62 19 19
61 15 24
29.7
87 13
89 11
3
2
1.32
13 32
11.71 16 5.58 31
16 34
0.94 0.88
17 33
13 34
1.39 1.03
54
(ref)2 51
48
(ref)
48
51
(ref)
62 20 19
87 13
81
95
6
19
5
27 73
p ⬍ 0.001 0.92
94
31 69
29.3
3
32 68
p ⬍ 0.001
NS 1.32 70 (ref) 16 1.71 14
61 20 20
1.46 (ref) 0.89
65 18 17
p ⬍ 0.001 (ref) 96 0.79 4
86 14
0.21 89 (ref)
11
(ref) 0.28
89 11
p ⬍ 0.01
1.32 (ref) 0.73
82 18
(ref) 0.57
p ⬍ 0.001
94
0.50
93
94
0.87
6
(ref)
7
6
(ref)
32 68
0.92 (ref)
p ⬍ 0.001 0.80 35 (ref) 65
55 20 25 p ⬍ 0.001
p ⬍ 0.001
p ⬍ 0.001 Prior pregnancy (percent) No Yes
Odds ratios
2
p ⬍ 0.05 Marital status (percent) Married or living with partner Not married or living with partner
29.5 p ⬍ 0.0001
37.02
Nonusers
30.3
p ⬍ 0.001 Race/ethnicity (percent) Caucasian Other
Odds ratios Users
(n ⴝ 5113) (n ⴝ 2375)
p ⬍ 0.001 Employment status (percent) Employed full time Employed part time Not employed
Nonusers
(n ⴝ 888) (n ⴝ 6598)
p ⬍ 0.0001 Education (percent) Did not graduate 3 from high school High school graduate 16 Some college, 33 vocational, or technical school College graduate 49
Caffeine
31 69 p ⬍ 0.01
NS 1.23 (ref)
31 69
NS
a
Tabled figures are for those at higher levels of use of each substance: 6 or more cigarettes per day; 4 or more drinks of alcohol per week; 1 or more caffeinated beverages per day. ref, reference category for calculation of odds ratios
other substance quitting caffeine, and less than 12% of those who use both other substances quitting. Although there is a similar pattern for cigarettes, with quitting more common among those who used only cigarettes as compared to those who used two or three substances, the differences are small and nonsignificant. Multivariate logistic regression models were devel-
oped to investigate the characteristics of those who quit using each of the three substances (tobacco, alcohol, or caffeine) in early pregnancy. Variables included in each model were pre-pregnancy use of the other two substances, educational level, employment status, ethnicity, whether or not the woman had ever been pregnant before, and whether or not the woman desired to be Am J Prev Med 2000;18(1)
57
Table 2. Proportion of the population using each substance before pregnancy and during early pregnancy, by level of substance use Prior to pregnancy Cigarettes None Up to 5 cigarettes per day More than 5 cigarettes per day Alcohol None Up to 3 drinks per week More than 3 drinks per week Caffeine None Less than daily use Use one or more times daily
During early pregnancy
82.9% 91.1% 4.7% 3.7% 12.4% 5.3% p ⬍ 0.001* 34.5% 94.8% 53.7% 5.1% 11.9% 0.1% p ⬍ 0.001* 7.4% 33.0% 24.3% 39.8% 68.3% 27.2% p ⬍ 0.001*
*p values from 2 df chi-square test comparing distributions for users prior to and during early pregnancy, for each substance.
pregnant at this time or sooner. In the model predicting who would quit smoking, only higher educational level and having never been pregnant before were
Table 3. Percent quitting or reducing use of each substance in early pregnancy, by level of use Percent quit Cigarettes Among those smoking ⱕ5 cigarettes per day (351) Among those smoking ⬎5 cigarettes per day (929) Alcohol Among those drinking ⱕ3 drinks per week (4018) Among those drinking ⬎3 drinks per week (888) Caffeine Among those drinking caffeinated beverages less than daily (1820) Among those drinking caffeinated beverages once a day or more (5113)
Percent reduced to below threshold level*
88 33
Substance
n
Observed Expected Odds percent percent ratio
None 2168 28.9 Cigarettes only 88 1.2 Alcohol only 95 1.3 Caffeine only 3715 49.6 Cigarettes plus alcohol 22 0.3 Cigarettes plus caffeine 627 8.4 Alcohol plus caffeine 579 7.8 Cigarettes, alcohol, 192 2.6 and caffeine
24.5* 3.5* 3.3* 52.7* 0.5† 7.5† 7.1† 1.0*
1.26 0.33 0.38 0.88 0.63 1.13 1.10 2.59
a
Tabled figures are for those at higher levels of use of each substance: ⱖ5 cigarettes per day; ⬎3 drinks of alcohol per week; 1 or more caffeinated beverages per day. *p value for 1 df chi-square test, comparing observed to expected, ⬍0.001. †p value for 1 df chi-square test, comparing observed to expected, ⬍0.05.
significant predictors of quitting. In the model predicting who would quit using alcohol, being a baseline nonsmoker, being nonwhite, and never having been pregnant before predicted quitting alcohol. In the model predicting who would quit using caffeine, being a nonsmoker, being a nondrinker, having a relatively higher education level, being nonwhite, never having been pregnant before, and wanting to be pregnant at this time all predicted quitting caffeine. These results confirm that, even taking other potentially predictive variables into account, simultaneous baseline use of another substance predicts a reduced probability of quitting alcohol or caffeine.
25 Table 5. Percent quitting each substance (cigarettes, alcohol, caffeine) in early pregnancy, among those using that substance alone or in combination with other substances prior to pregnancy
p ⬍ 0.001† 94
Percent quitting 84
15
p ⬍ 0.001 41 23
37
p ⬍ 0.001
*Threshold level is defined as smoking ⱕ5 cigarettes per day, drinking ⱕ3 alcoholic drinks per week, or drinking caffeinated beverages less than daily. †p values are based on 1 df chi-square test comparing quit rates among lighter and heavier users of each substance.
58
Table 4. Proportion of the population using cigarettes, alcohol, or caffeinea prior to pregnancy, alone or in combination; compared to the proportion expected, assuming use of each substance was independent of the other substances
Cigarettes Cigarettes alone Cigarettes plus one other substance Cigarettes plus two other substances Alcohol Alcohol alone Alcohol plus one other substance Alcohol plus two other substances Caffeine Caffeine alone Caffeine plus one other substance Caffeine plus two other substances
34.1 32.5 31.8 NS* 89.5 86.0 75.0 p ⬍ 0.001* 26.4 16.0 12.0 p ⬍ 0.001*
*p value from 2 df chi-square test comparing rates among those using ⬍one, two, or more than two substances.
American Journal of Preventive Medicine, Volume 18, Number 1
Although women using multiple substances were less likely to quit their use than women using single substances, one might ask whether those women using multiple substances who did decide to quit one substance were more or less likely to quit a second substance. Among women using both cigarettes and caffeine prior to pregnancy, women who quit using caffeine during early pregnancy were 2.4 times more likely to quit cigarettes (OR ⫽ 2.4; 95% CI ⫽ 1.7 to 3.8). Among users of both alcohol and caffeine, quitting caffeine was associated with increased odds of quitting alcohol and vice versa (OR ⫽ 6.4; 95% CI ⫽ 2.8 to 14.3). Only for alcohol and cigarettes was quitting one substance not associated with an increased likelihood of quitting the other (OR ⫽ 1.7; 95% CI ⫽ 0.8 to 3.8).
Discussion Smoking rates prior to and during early pregnancy appear to be lower in these middle-class, insured populations than in other populations in the literature, with 17.1% reporting smoking prior to pregnancy and 9.0% reporting smoking during early pregnancy. Recent data reports for national samples show higher smoking rates. Birth certificate data reported by the National Center for Health Statistics (NCHS) indicated that 14.6% of women smoked during pregnancy in 1994.22 The National Health Interview Survey for 1990 reported that 24.6% of women aged 18 to 44 smoked in the year prior to their most recent pregnancy, whereas 15.4% reported smoking during their most recent pregnancy.23 The National Pregnancy and Health Study, which was based on a sample of women interviewed in the hospital after giving birth, reported 20.4% smoking during pregnancy in 1992.24 Whereas the timing of self-reports of smoking is somewhat different in our study than in the national samples, it does appear that smoking in this largely middle-class population is less common than in national samples. Self-report of alcohol use prior to pregnancy was higher in our population than in national samples, whereas self-reported drinking in early pregnancy was lower. We found 65.6% drinking prior to pregnancy, and 5.2% drinking during pregnancy. Behavioral Risk Factor Survey data for 1995 show 50.6% of women of childbearing age drinking prior to pregnancy and 16.5% drinking during pregnancy.25 An analysis of the Pregnancy Risk Assessment Monitoring System data from four states26 reported 47.6% drinking prior to pregnancy, and about 12% drinking during the last three months of pregnancy. Birth certificate data from NCHS, however, show only 1.7% drinking during pregnancy, which the NCHS report describes as “substantially underreported.”22 Because caffeine use during pregnancy has not traditionally been considered a major public health problem, few population-based statistics are available on
caffeine use. Dlugosz and Bracken14 cite unpublished data from Bracken showing that 26% of women giving birth in the Yale–New Haven hospital in 1990 –1991 reported no caffeine intake in the first month of pregnancy, which is similar to our figure of 33% reporting no caffeine intake in early pregnancy. The tendency for the use of alcohol, cigarettes, and caffeine to cluster within the same individuals was not unexpected. That smoking is extremely prevalent in groups being treated for alcohol dependency is well known. Data summarized by Hughes27 show a median rate of 72% heavy smokers in published reports of persons in treatment for alcohol problems. In more general populations, the reported correlations between use of these two substances have been significant but modest. Shiffman and Balabanis28 cite studies showing correlations between any use of alcohol and tobacco on the order of 0.12 and 0.13 and correlations of smoking and drinking rates (amounts) ranging from 0.15 to 0.22. Johnson et al.2 and the National Pregnancy and Health Survey24 also have shown clustering between alcohol and cigarettes, and between these substances and illicit drugs, during pregnancy. Of the three behaviors examined here, women were most likely to quit the use of alcohol during pregnancy, second most likely to quit cigarettes, and least likely to quit caffeine. The low likelihood of quitting caffeine use is most likely attributed to far less attention to this behavior in the mass media and popular sources of advice concerning health in pregnancy. Those who quit caffeine use are likely the most health-conscious subgroup. Both alcohol and tobacco, however, are strongly discouraged during pregnancy; and the dangers of both substances to the developing fetus have been widely publicized. As reported here, of those drinking more than minimal amounts (i.e., more than three drinks a week), 84% reported quitting drinking during early pregnancy. Of those smoking more than minimal amounts (more than five cigarettes per day), 33% quit smoking. Somewhat lower rates of quitting these substances were reported by Kruse et al.,29 who reported 53.2% of drinkers becoming abstinent, whereas 16.7% of smokers did so. They attributed their finding of a greater prevalence of quitting among drinkers than among smokers to an upsurge of publicity concerning the effects of alcohol in pregnancy. Another possible explanation may lie in the probability of physical dependence on the substance in question; it appears that few drinkers in our study were drinking at levels likely to be characteristic of dependent drinkers, whereas many smokers were presumably dependent on tobacco. Few studies exist describing attempts to quit the use of two substances simultaneously, and the data from the existing studies are mixed. Swanson et al.30 suggested that continued use of caffeine by abstinent smokers may increase withdrawal symptoms, because quitting smoking allows blood levels of caffeine to increase and Am J Prev Med 2000;18(1)
59
therefore increases the occurrence of restlessness, impatience, and insomnia. In an experimental study, however, Swanson et al.31 reported no effect of caffeine abstinence on success in smoking cessation at 6 and 12 months. Most withdrawal symptoms in the immediate post-cessation period also did not differ between caffeine abstainers and continued caffeine users, with the exception of fatigue, which was greater among abstainers. Caffeine abstinence occurred only for one week as part of this experiment, so the long-term impact of continued caffeine abstinence is unknown. With regard to simultaneous quitting of smoking and alcohol, on the other hand, in a summary of studies involving alcohol-treatment populations, Bobo et al.32 suggest better drinking-cessation outcomes among smokers who quit smoking than among those who continued to smoke. Data reported here indicate that women using more than one substance may be less likely to quit substance use than women who use only one substance. The reasons for this pattern are unknown, but possibilities include the difficulty of formulating strategies for coping with the effects of quitting multiple substances simultaneously or a generally lower level of concern about the requirements for a healthy pregnancy. On the other hand, among women who quit at least one substance, the odds of having quit another substance were either unaffected or elevated. This pattern suggests a clustering of quitting behavior; women using multiple substances may either quit none of them or several of them, rather than quitting one and continuing others.
Limitations of the Study The cross-sectional design of this study leaves many important questions unanswered. Specifically, we are unable to address issues of timing: do successful quitters make their behavior changes in sequence, or at the same time? Does quitting more than one substance tend to predict relapse later in the pregnancy? These are important questions that subsequent investigations may need to address. The low self-reported smoking rates and high selfreported quit rates for smoking and alcohol in this study raise concerns about the validity of self-report of these behaviors. Underreporting of proscribed behaviors in health surveys is known to occur and may indeed be the case in this study. However, it does not appear that underreporting of individual behaviors or overreporting of quitting would give rise to the patterns of association shown between quitting different substances that were the focus of this research. The specific question formats concerning substance use are also less precise than we would ideally have liked. In particular, the questions concerning level of alcohol and caffeine use were rather imprecise; more 60
specific information, such as patterns of drinking or binge drinking, would have allowed us to examine possible effects of dependency and more precise analysis of the level of use and quitting.
Conclusions From the point of view of the health care provider confronted with a patient who is using more than once substance, these results indicate that quitting more than one substance at the same time is not impossible, and that having successfully quit one substance does not reduce the probability of having successfully quit a second substance. However, because the data suggest that women using more than one substance show a tendency to continue with the use of multiple substances, providers should be sensitive to the difficulties of making this complex set of behavior changes. These data suggest the need for additional research on the best way to go about quitting more than one substance at a time and the factors that lead to success in doing this. The authors wish to acknowledge the contributions of Ms. Karen Virnig, survey director at the University of Minnesota, and Drs. Robert Junilla and Leslie Pratt of Park Nicollet Medical Center in Minneapolis. This study was supported in part by National Heart, Lung and Blood Institute grant HL48121.
References 1. Vega WA, Kolody B, Hwang J, Noble A. Prevalence and magnitude of perinatal substance exposures in California. N Engl J Med 1993;329:850 – 4. 2. Johnson SF, McCarter RJ, Ferencz C. Changes in alcohol, cigarette and recreational drug use during pregnancy: implications for intervention. Am J Epidemiol 1987;126:695–702. 3. Cotton P. Smoking cigarettes may do developing fetus more harm than ingesting cocaine, some experts say. JAMA 1994;271:576 –7. 4. US Health and Human Services. Reducing the health consequences of smoking: 25 years of progress: a report of the surgeon general. Rockville, MD: US Health and Human Services, Public Health Service, Centers for Disease Control, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1989. 5. Malloy MH, Kleinman JC, Land GH, Schramm WF. The association of maternal smoking with age and cause of infant deaths. Am J Epidemiol 1988;128:46 –55. 6. Sexton M, Fox NL, Hebel JR. Prenatal exposure to tobacco: II. Effects on cognitive functioning at age three. Int J Epidemiol 1990;19:72–7. 7. Olds DL, Henderson CR, Tatelbaum R. Intellectual impairment in children of women who smoke cigarettes during pregnancy. Pediatrics 1994;93: 221–7. 8. Abel EL, Sokol RJ. Fetal alcohol syndrome is now leading cause of mental retardation. Lancet 1986;ii:1222. 9. Spohr H-L, Willms J, Steinhausen H-C. Prenatal alcohol exposure and long-term developmental consequences. Lancet 1993;341:907–10. 10. Jones KL, Smith DW, Ulleland CN, Streissguth AP. Pattern of malformation in offspring of chronic alcoholic mothers. Lancet 1973; i:1267–71. 11. Streissguth AP, Barr HM, Sampson PD, Parrish-Johnson JC, Kirchner GL, Martin DC. Attention, distraction and reaction time at age 7 years and prenatal alcohol exposure. Neurobehav Toxicol Teratol 1986;8:717–25. 12. Fried PA, Makin JE. Neonatal behavioral correlates of prenatal exposure to marihuana, cigarettes and alcohol in a low risk population. Neurotoxicol Teratol 1987;9:1–7.
American Journal of Preventive Medicine, Volume 18, Number 1
13. Shu XO, Hatch MC, Mills J, Clemens J, Susser M. Maternal smoking, alcohol drinking, caffeine consumption, and fetal growth: results from a prospective study. Epidemiology 1995;6:115–20. 14. Dlugosz L, Bracken MB. Reproductive effects of caffeine: a review and theoretical analysis. Epidemiol Rev 1992;14:83–100. 15. Peacock JL, Bland JM, Anderson HR. Effects on birthweight of alcohol and caffeine consumption in smoking women. J Epidemiol Community Health 1991;45:159 – 63. 16. Raitakari OT, Leino M, Raikkonen K, et al. Clustering of risk habits in young adults: the cardiovascular risk in young Finns study. Am J Epidemiol 1995;142:36 – 44. 17. Istvan J, Matarazzo JD. Tobacco, alcohol and caffeine use: a review of their interrelationships. Psychol Bull 1984;95:301–26. 18. Vuchinich RE, Tucker JA. Contributions from behavioral theories of choice to an analysis of alcohol abuse. J Abnormal Psychol 1988;97:181–95. 19. Hall SM, Tunstall CD, Vila KL, Duffy J. Weight gain prevention and smoking cessation: cautionary findings. Am J Public Health 1992;82:799 – 803. 20. Pirie PL, McBride CM, Hellerstedt W, et al. Smoking cessation in women concerned about weight. Am J Public Health 1992;82:1238 – 43. 21. McBride CM, Curry SJ, Lando HA, Pirie PL, Grothaus LC, Nelson JC. Prevention of relapse in women who quit smoking during pregnancy. Am J Public Health 1999;89:706 –11. 22. Ventura SJ, Martin JA, Mathews TJ, Clarke SC. Advance report of final natality statistics, 1994. Monthly Vital Statistics Report 1996;44(suppl 1):l 23. LeClere FB, Wilson JB. Smoking behavior of recent mothers, 18-44 years of age, before and after pregnancy: United States, 1990. In: Advance data from vital and health statistics. Hyattsville, MD: National Center for Health Statistics, 1997.
24. National Institute on Drug Abuse. National Pregnancy and Health Survey. Bethesda, MD: National Institutes of Health, 1996. 25. Durham J, Owen P, Bender B, et al. Alcohol consumption among pregnant and childbearing-aged women United States, 1991 and 1995. MMWR Morb Mortal Wkly Rep 1997;46:346 –50. 26. Bruce FC, Adams MM, Shulman HB, Martin ML. Alcohol use before and during pregnancy. Am J Prev Med 1993;9:267–73. 27. Hughes JR. Clinical implications of the association between smoking and alcoholism. In: Fertig JB, Allen JP, eds. Alcohol and tobacco: from basic science to clinical practice. Bethesda, MD: National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism, 1995. 28. Shiffman S, Balabanis M. Associations between alcohol and tobacco. In: Fertig JB, Allen JP, eds. Alcohol and tobacco: from basic science to clinical practice. Bethesda, MD: National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism, 1995. 29. Kruse J, LeFevre M, Zweig S. Changes in smoking and alcohol consumption during pregnancy: a population-based study in a rural area. Obstet Gynecol 1986;67:627–32. 30. Swanson JA, Lee JW, Hopp JW. Caffeine and nicotine: a review of their joint use and possible interactive effects in tobacco withdrawal. Addict Behav 1994;19:229 –56. 31. Swanson JA, Lee JW, Hopp JW, Berk LS. The impact of caffeine use on tobacco cessation and withdrawal. Addict Behav 1997;22:55– 68. 32. Bobo JK, Walker RD, Lando HA, McIlvain HE. Enhancing alcohol control with counseling on nicotine dependence: pilot study findings and treatment implications. In: Fertig JB, Allen JP, eds. Alcohol and tobacco: from basic science to clinical practice. Bethesda, MD: National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism, 1995.
Am J Prev Med 2000;18(1)
61