Tobacco Documents Online

_Pm ' Data are given as mean ± SD. SBP indicates sysfolic blood pressure; OBP, diastolic blood pressure- }P<.001e tPS-0L §P<-05 . 1i.~i0?p`. ... ;~2mn-17§.. nn31~i~0.fi:. . . -_.-097._: `Data are given as mean a SD. HDL-C indicates high-densitylipoprofem cholesterol; HOLz C. HDL subfractron 2 cholesterol; LDL-C, low-density lipoprotein chofestemh NS, nonsmoking; and PS, passive smoking. To convert cholesterol from millimoles per liter to milligrams per deciliter, divide millimoles per liter by 0 02586 iP<_.0t. *Ps.007. §P<.05. LIPOPROTEIN, SEX, RACE, AND PASSIVE SMOKING DIFFERENCES The group means for lipoprotein levels by visit, sex, race, and passive smoking status are shown in Table 3. In the oldest children, LDL-C level was higher in blacks than in whites. The difference in LDL-C level between the non- smoking and passive smoking group of children was .,e; ~ant,~v.~w.::: ~- lr~ ca nd 8lood I i~a~ e~athJ~y lac~'e .r t~WW•=i8c n yn . '-11? LLYl ~... ; .19m 1~?4 :-.- :... 2o i'11.237;-~,. ~, 79~i 0 p4 ~.'.:..~ - ~.15i 025f~' ~ ~, ^-03Gi0:1A _ .LI-171 - o.e21aar 0 17 . `. 0 45 x 024Y -/1 x 0_18 ._ 0 16 " greater for whites than for blacks (P<.02). No other group differences were seen for LDL-C. White children had lower HDL-C and HDL subfraction 2 cholesterol (HDL2-C) lev- els than black children at all visits. Levels of HDL cho- lesterol were similar in white and black children in non- smoking families at visit 1(1.24m0.27 vs 1.27s0.20 mmoVL 147.9 z 10.3 vs 49.1 m 7.9 mg/dL] ). However, white children exposed to passive smoke had lower HDL ARCH FEDIATR ADOLESC MEDNOL 153. MAY 1999 449

I t12(55) f-PassFreSnroqnO 7 Nonsmokin9 13~ (53} . ~ Iaz(si) ~ e tYi(49) ~ rsz(n) E .1ae(45) '~ ~ 4.tt(43) ~ i ~ g +.os(41) ( 1.01(39) I 0.BS (3r) .. O.so(3s) Y~sit1 Ns112 Nall9Vlsit4 . flgure 2. High-densltylipoprotein chokstero! (NDL-C) levels by visll and passive smoking status in children with family history ofcardiovzscular disease. Values represent mean xSD. P<.got, repeated-measures analysis of variance between smoking groups for the entire study period. fPassFmSmakinY. .\ Nansmotiny,  PasslveSmokinB 0 Nonsmoklnp. ~ ..Boys Bnys Gles Gies F7gure 3. Nigh-densiryllpopro[ein chdasterol (HOL-Cj ievels by weight sex, and passive smoking stams interactiahs in wfifte children. Va/ues represen7 meanmSD:P<.D7 for model using repeated-measures anarysis of vadance. levels than black children exposed to passive smoke at visit 1(1'.12 x 0.21 vs 1.36 ® 0.22 mmol/L [43.2 z 8.0 vs 52.7m8.4.atgJdL]; P<.02). While HDL-C and HDL2-C levels changed little across visits in the girls, these levels decreased inboys and were significantly lower than the girls' at tlre last visit. Children exposed to passive smoke had lower HDL-C and HDL-C levels than children not exposed.to cigarette smoke at visits T and 3. , HDL-C, RACE, ANTHROPOMETRICS, AND - PASSIVE SMOKING INTERACTIONS We found racial~ differences in the HDL-C levels of chii- dren with a:family history of cardiovascular disease. These d4fferences were exaggerated by exposure to cigarette s moke.-Children from passive smoking families had lower HDL-C levels than those in the nonpassive smoking group (P<:001;~Figure 2): Of those children from passive smoking families, white children had lower HDL-C lev- els than black children (P =.004). . Using a mixed modeling approach with repeated- measures analysis of variance, we explored models for HDLC that included-race, sex, passive smoking status, weight, systolic blood pressure, diastolic blood pres- sure, and all interactions among those variables. Figure 3 illustrates the weight by sex and passive smoking status interaciions (P<.01)..The plot shows fitted HDL C lev- els across weight for white boys and girls in passive smok- ing and nonpassive smoking groups. The data are shown for fixed levels of all other independent variables, ie; mean systolic and diastolic bloodpressures. in white families, as weight increased, HDL-Cleveldecreased.~Boys ex- posed to passive smoking showed the greatest decrease in HDL-C level with increasing weight. -' DIASTOLIC BLOOD PRESSURE;~HDL-C, SEX, AND PASSIVE SMOKING INTERACTIONS We examined the relation of HDL-C level across dias- tolic blood pressure in white children with a family his- tory of cardiovascular disease. Significant differences ex- ist by sex and passive smoking status (P=. .01). Girls not exposed to cigarette smoke showed increasing levelss of - HDL-C with increasing diastolic blood pressure:.ln con- trast, girls who were exposed to cigarette smoke had lower HDL-C levels that remained low with increasing-dias- tolic blood pressure. The HDL-C level did not interact with diastolic blood pressure in boys from nonsmoking families but was lower in boys exposed to passive smok- ing with higher diastolic blood pressure. ' . DIASTOLIC BLOOD PRESSURE, HDL-C,RACE, AND PASSIVE SMOKING INTERACTIONS Among the passive.smoking group, HDL-C levels inter- acted to a greater extent with other risk factors in blacksthan in whites. The interaction obHDL-C level and di- astolic blood pressure in-boys~ from families with a'his- tory of cardiovascular.disease differed by both race and passive smoking status (Ftgure 4);Bothblai 4: and white boys exposed to cigarette smoke and with fauiilv his- tory of cardiovascular disease showed-significant decreases in HDL-Clevel.withincreasing diastolic blood pressure (P<.01)1-However,- in boys not exposed to cigarette smoke, as diastolic blood'pressure increased, HDL-C level decreased more in blackscom- pared with whites.. This-interactive effect of HDL-C level and diastolic blood pressure in boys exposed to passive smoke was greater in those with a family history of cardiovascular disease. SYSTOLIC BLOOD PRESSURE, HDL-C, SEX, AND PASSIVE SMOKING INTERACTIONS We examined the interactions of systolic blood pressure and HDL-C level in children with a family history of car- diovascular disease(Pigure 5). Atlower levels of sys- tolic blood pressure, HDL-C level was lower in boys ex- posed to cigarette smoke than in boys not exposed. As systolic blood pressure increased, the differences attrib- utable to sex and passive smoking diminished. We investigated the relationship of coronary risk factors to changes in HDL-C levels in adolescents by using 2505586025 ^ ARCH PLD7AIR ADOLESC MEDNOL 153. MAY 1999 450

Figure 4. Plot of the diastolic blood pressure (DBP) by race interaction on . hiphdensiry lipoprotein cholesterol (HOL-C) levels in boys with posdive famityhistory of cardlovascular disease. Comparison oiboys not exposed to cigarette smoke (top) with boys exposed to passive smoking (bottom). P<01 for model using repeated-measures analysis of variance. repeated-measures analyses of variance.la•zl Our data show that in children with a family history of cardiovas- cular disease, significant relations exist between HDL-C level and blood pressure that differ by sex and race, and interact with passive cigarette smoking. The mixed modeling approach used allowed for missing informa- tion over visits and allowed modeling of the mean and the variance. Because of the significant interactions in the model, the differences in lipoprotein levels between the races, sexes,and family smoking statuses change. The use of the mixed modeling approach may provide important insights into the mechanisms and interactions of genetic and environmental effects that underlie the childhood antecedents of atherosclerotic heartdisease. - -------- -- The present study confirms our earlier observa- tions and those of others and shows that lower HDL-C levels are found in children exposed to passive smok- ing. We chose to study HDL2 cholesterol in addition to HDL-C for seveml reasons. Although the HDL subfrac- tions are metabolically interrelated, most of the varia- tion in HDL is due to the HDLi-C subfraction." Further- more, Bodurtha et aP' have shown that CAD deaths occur more frequently in families with low levels of HDLz-C. In a previous cross-sectional study of 11-year-old chil- dren, our group showed that preadolescent boys ex- posed to cigarette smoke had lower levels of HDLz-C than boys not exposed..The levels were related to the num- ber of cigarettes smoked daily by the parents of the boys.9 The lowest HDLz-C levels were found in boys exposed to the highest number of cigarettes smoked by their moth- ers. The HDL3-C subfraction was similarly lower in the -PassiveSmokinA. -Nonsmoking, -PassiveSmukinp, -Nonsmokinp, Boys .. Boys Gids Gtns Figure 5. Plot of the systolic blood pressure (SBP) by passive smoking status and sex interactions on high-density lipoprotein cholesterol (HDL-C) levels in white children with positive familyhistory of cardiovascular disease. P<.001 for model using repeated-measures analysis of variance. passive.smoking group of children, with-greater differ- ences seen in the girls. During puberty and early adoles- cence, levels of HDL-C and LDL-C decrease in children and the decrease in HDL-C is more pronounced in boys than in girls.24 The influence of sex hormones and their . changes during puberty are obviously important, with HDL-C levels falling in boys in association with in- creases in testosterone leveLs.'s Passive cigarette smok- ing may further diminish HDL-C and its subfractions that may beassociated with premature atherosclerotic changes. In a recent study, mean HDL-C levels were lower in dyslipidemic children from households with smokers than in those without household smoke expo- sure.16 Passive smoking may worsen the risk profile for early atherosclerosis among such high-risk children. - In this study we found racial differences in LDL-C and HDL-C in pubertal children that were related to pas- sive smoke exposure. Levels of LDL-C were higher in the oldest black children than white children. At all visits, black children displayed higher HDL-C and HDL2-C lev- els than white children. Previous studies have shown that as children reach preschool age, black children begin to have slightly higher levels of HDL-C than white chil- dren and this race-related trend becomes established at about 9 years of age.=','a The black-white difference in HDL-C has-been attributed to differences in lipoprotein particle number, with blacks having an inherently more efficient lipid-clearing mechanism. While these studies adjusted for active cigarette smoking in older children, the effects of passive smoke exposure and physical ac- tivity were not considered."•23 We believe that racial differences in lipoprotein lev- els are related, at least in part, to long-term passive ciga- rette smoke exposure. At age 11 years, black and white children not exposed to cigarette smoke showed similar HDL-C levels whereas white children had lower HDL-C levels than black children among the passive smoking group. Among the passive smoking group, our sample size of black children was insufficient to test for differ- ences from white children at subsequent visits. How- ever, passive smoke exposure was greater at all visits in white than in black families. That fathers smoked more than mothers is consistent with current information ob- AACH PEDtATa ADOLESC MEDNOL 153, MAY 1999 451

SUBJECTS AND METHODS POPULATION As part of an ongoing genetic.longitudinal study of devel- opmental changes in cardiovascular risk factors during ado- lescence, we recruited families with twins from nearby schoolsystems. Using school rosters, we identified twin pairs living in the Commonwealth of Virginia. Descriptions of = our study were sent by the school; affirmatively replying families were invited to participate. A total of 408 twih pairs (89% of affirmatively replying families) participated in the study. All twins were examined as close as possible to their 11th birthday. None of the twins were active cigarette smok- ers at entrance to the study and none initiated active ciga- rette smoking during the study period. Families participated in a protocol that included the collectiorrof data on family and health histories,.smoking and alcohol consumption (questionnaire and personal in- terview), self-reported weekly exercise level, blood pres- sure, and collection of blood samples for biochemieal as- says. The protocol was repeated at 18-month intervals. Parents were asked about the family history, including the incidence of heart disease. A family history of cardiovas- _ cular disease was defined as hypertension or early cardio- vascular death (before age 55 years) in a parent or in a first- degree relative of the parent or heart disease in either parent The number of cigarettes smoked each day by the parents was recorded- Serum cotinine level was used both as the measure of smoke exposure and to verify nonsmoking sm- tus. No attempt was made to prescreen enrollees for the presence or absence of cardiovascular risk factors. In- formed written consent, which had been approved by the Committee on the Conduct of Human Research, was ob- tained frorn each family before it entered the study. PROCEDURES Anthropometrics . . Height and weight of each subject in stocking feet were mea- sured with a calibrated stadiometer and digital scale, re- Spectively. Anthropometric data were obtained in dupli- cate and averaged. Sexual.staging was performed using a 5-scale score based on Tanner criteria.'s ~.~Blood Pressure Casual blood pressure was measured using a mercury sphyg- momanometer with the appropriate compression cuff avith .. the subject in the sitting position. Chosen cuff size was big enoughtoencircleatleasthalftheupperarmwithoutover- .- over- lapping. The rubber bladder rested over the artery being compressed and had sufficient width to cover at least two ~ tlvrdsoftheupperami.Pressurewithinthecomptrssioncufl . uidiratedby the level of the mercury column at the murmur for family clustering.° Nonetheless, environmental ef fects account for a significant portion of the variance of coronary risk factors in pubertal children. Cigarette~smoking is an important environmental determinant of theearly stages of atherosclerosis in ado- lescents andyoung adults.' The pathogeriesis may re- of the first- and fourth-phase Icorotkoff sound was re- corded. Blood pressure was recorded 3 times and avemged. The interobserver intmclass correlation coefficients for blood pressure were 0.89 and 0.84 for systolic and diastolic blood pressures,respectively. Blood Samples A venipuncture sample of whole blood was obtained, stored on ice, and processed within 1 hour for quantitative lipo- protein cholesterol measurements using [he vertical spin ultracentrifugation technique.1B The time elapsed from the last meal to the time of blood drawing was recorded and was found not to correlate with any of the lipoprotein lev- els. Serum cotinine concentration was quantitated by ra- dioimmunoassay methods." STATISTICAL ANALYSIS Because twins share genes and environments and represent nonindependent observations, data from a single twin ran- domly ascertained from each family was used to determine group means for statistical testing. A random number table was generated and sequential twin pairs were assigned num- bers.If an odd numberwas assigned to a pair, twin 1 was cho- sen from that pair. Changes in passive smoking stams from one visit to the next are associated with alterations in lipo- protein leveis.'a Therefore, data from families that changed smoking status (from smoking to nonsmoking or vice versa) anytime after the firstvistt were dropped from further analy- ses. Statistical differences between group means were as- sessed by 2-sided [ tests, taking into account whether group variances were equal. Pooled t tests were used to test for dif- ferences in variables between boys and girls. Spearman non- parametric correlation coefficients were used when it was ap- parentthatagivenvariablewasnotnonnallydistributed,such as cigarettes smoked each day and HDL-C levels. Regression analysis was used to test for effects of confounding vari- ables. Data are presented as mean z SD. Repeated-measures analyses of variance using a mixed modeling approach was performed for HDL-C levels."All analyses were performed in the SAS system using PROC MIXED.}9 The mixed modeling approach allows for miss- ing information over visits and allows for modeling of the variance and the mean. The missing data in our data set were missing completely at random and not covariate de- pendent missing at random. PROC MIXED will delete any observations with missing values for any variable in the model statement, ie, fixed effects. The analysis and results are valid with respect to the missing data provided that the cause of the missing data are independent of the outcome variable. For example, if twin pairs are treated as a re- peated measure, unpaired twins will contribute to the re- gression, but not to the esti¢iate of the covariance. Models for HDL-C included race, sex, passive smoking status, weight, systolic blood pressure, diastolic blood pressure, and all interactions among those variables e late to the inverse dose-dependent relationship between cigarette smoking and HDL-C levels a In addition to ac- tive cigarette smoking, environmental tobacco smoke or passive smoking affects both plasma lipids and endothe- lium-dependent vasodilation- Adolescents exposed to their parents' smoke have depressed levels of HDL-C and its ARCH PEDLATR ADOLESC MEDNOL 153, MAY 1999 447 4

ARTICLL• Childhood Passive Smoking, Race, and Coroiiaiy Artery Disease Risk 1 he 21nCV Twit1 Study - - A ePliani 13.~ Jvloshowitr, MD; P¢ntela F. Schwartz, PhD; RiThard M.:Schiehen, MD Background: Fhildren with long-term exposure topas-sivc'cigarette smoke may beat elevated risk for the de- velopment'of pteulature coronary artery disease (CAD). C+6jective:To examine how CAD risk factors, expo-e sure in pa_ sii-e smoking, sex, and race are related inpu- be.rral childrern ar.d to determine if there is an identifi- able:childhood riskprofrle.(ie, does passive smoking lnteract.wtth other coronary:risk factors to increase the risk'ofdevelopingprematuteCAD): Design: Cohort analytic study. Setting: The MedicaICollege of Virginia (MCV) Twin 'Sttudy, Richutond, Va. . Su6jeetsc Randomly selected twins from.40811-year-. : old twin pairs recruited from nearby schools. Methodn Data collection occurred at 18-month intervals on family and health histories, smoking and alcohol con- somption; blood pressure, anthro pometrics, and biochemi- . calassays: Datafromcohortsof 11-year-oldssmdied thmngh age 15yearswere analyzedby repeated-measures analyses of variance using a mixed modeling approach. Models forltigh-densitylipopi•otein cholesterol (HDL-C) included race, sex, pa'ssive smoking status, weight, systolic and di- astolic blood pressures, and all interactions. -. Resalts: Passive smoke exposure was greater in white fa,milietliasin black families. Levels of HDL-C and . HDL-C (HDL subfraction 2 cholesterol) were lower in white children-than in black children (visit 1: HDL-C. - mean t SD, 1.21 x 0.26vs 1.31 x 0.26 mmol/L{47.0 i 10.1 vs 50.6 x 10.1 mg/dLj, P5.01; HDLi, ttlean m SD, 031m0.18vs0.41m0.I9mmol/1.[12.3x7.Ovs15.9m7.4 mg/dL], P:~.001). Children with a family history of car- diovascular disease had differences in HDL-C levels re- lated to race that were worsened by exposure sociga- rette smoke. In these children, HDL-C level was lower in those exposed to passive smoking(visit 1:~ 1.18 t 0.23 vs 1.25 x 0.23 mmol/L (45.6 x 9 vs 48:2 t 9 mgldL]: and- visit 4: 0.98 x 0.10 vs 1.19 m 0.18 mmol/L [37.8 t 4 vs - 46.0s7 mg/dL]; P<.001), with white children having lower HDL-C levels than black children (visir1: 1.12 s 0.21 vs 1.36 m 0.23 mmollL [43.2 i 8 vs 52.7 m 9mg/ dL] and visit 4: 0.97x0.31 vs 1.01t0.31 inmol/L [37.6 m 12 vs 39.0 m 12 mg/dLl; P=.004). In whitefami- lies, as weight increased, boys exposed to passive smok- ing showed the greatest decrease in HDL-C level (P<.01 for weight by sex and passive smoking interaction). Risk factors for CAD, such as blood pressure, interacted with HDL-C and these relationships varied by race and;sex. Conclusions: Pubertal children with long-term passive cigarette smoke exposure havelower HDi. (- h-vels; Ra= cial differences in-HDL-C levels are related to_ passive _smokeexposure:Inchildrenwitha#amilyhistoryofcar- . diovaseular disease, interactions exist between passive -. smoking, HDL-C level,and.blood prrssurr that diffrr by sex and race. White males exposed to passive smoking ... who have a family historybf c(rdiovascular dise:+se and higher weights and diastolic blood pressurec may bc at . special risk for premature CAD. I Arch PediatrAdolescMed- I999;753:446--153 --Ediror•s Note: The results of this intriguing study should be dis---cussed with all parents who smoke. If they ie going to light up near their chil_dren,-they might as well feel guilty. - . CatherineD:DeAngelis,MT) Frorn tlw Division of Pediatric Cardiology, Department of Pediatrics, Cliiidren's Medical Center (Drs Moskowitz and Schteken), and Department of Biosmttstia (Dr Schwartz), Medical College of Virginia,. Virginia Commonwealth University, Richmond A 1HEROSCLEROirc;Cltartges- found-iti~middle age be- gin in childhood. The tnechanisms-tuay relate to abnonaallevels of risk fac- tors. Certainriskn factors; such as serum lipid and lipoprotein levels, hyperten- sion, and smoking, aretkought to be re- lated to-the eacliest stags oi'atberoscle- i rotic coronary artery disease (CAD), We know that(in adults, high levels ofIow- - density lipoprotein cholesterol- (LDL-C) and low levels of high-density lipopro-: tein cholesterol'(HDL-C) and'its'niajor.--subfractions are associated with myocar- : dialinfarction.' Familiesvaith special risk of CAD can be identified by the aggrega-- ' tion of lipoprotein levels and low levelsof-` HDL-C and its-subfractions in the chil- dren.'-' Because serum lipid and jipopro- tein levels and other risk factors in early pubertal children are under strong ge- netic influence, they may be responsible ARC11 Pm1AiRAnO1FSCMFD/VOL 153, MAY 1999 2505586021 L1i6

tained from a recent population-based study.39 There- fore, long-term passive.cigarette smoke exposure must be considered when examining lipoprotein levels in pe- diatric populations. ' We found an inverse relation between HDL-C and weight in white pubertal boys and girls. Despite smok- ing status, boys showed a greater reduction in HDL-C with increasing weight than girls. More important, the greatest reduction in HDL-C with increasing weight was in children exposed to passive cigarette smoke long- term. Other authors have found an inverse relationship between HDL-C and weight in teenage boys30 and girls.'. Glueck et aP' found a significant inverse relationship be- tween HDL-C level and Quetelet index in children rang- ing in age from 12 to 16 years. None of these studies ex- amined the effects of cigarette smoke exposure. The consistent inverse associations between passive smok- ing, weight, and HDL-C level show that the effects of these variables on HDL-C levels and on CAD risk in adults are already present in early adolescence. - We have previously shown that the varianceof HDL-C and HDL2-C levels are influenced both by genetic and environmental influences e0 When we com- pared the self-reported weekly exercise level in our twins, the number of exercise episodes each week were similar for boys exposed to passive smoking and those who were not (4.7 ± 2.0 vs 4.2 t 2.3 times per week) and for girls exposed to passive smoking and those not (4.3 m 2.3 vs 4.7 ± 21 times per week). Analysis using X2 tests showed no association between smoking status and exercise in either the boys (X'= 1.7, P<.2) or the girls(xr=0.5, P<.5). Exercise therefore was not exam- ined as a variable in this study but was our focus in a recent study.31 Dietary changes, exercise, and weight reduction have resulted in sustained increases in HDL-C levels in obese children.33 Eliminating cigarette smoke exposure from a child's home environment may be part of an effective interventional strategy to increase HDL-C levels. Offspring with a family history of cardiovascular dis- ease often show adverse risk factor clusters bypu- berry." The data from our study show that in children with a family history of cardiovascular disease signifi- cant associations exist between HDL-C level and blood pressure that differ by sex and race. These variables may interact proatherogenically with passive cigarette smok- ing. Racial and sex diffeiences in lipoprotein levels and other CAD risk factors in adults with cardiovascular fam- ily history have been previously reported.'s•36 Data from the CARDIA Study showed that blacks had higher sys- tolic blood pressure and diastolic blood pressure than whites with racial differences being greater in women than in men.37 The present longi,tudinal study demonstrates the complicated relations of passive smoking and CAD risk factors within a biracial population of sexually ma- turing adolescents. Other lines of investigation, such as gene by environment interaction (ie, genetic back- ground influence on the response to a given environ- mental stimulus), may have the potential to explain the genetic (racial, sex, developmental, and family history) and environmental (passive smoking) interrelation- ships of CAD risk variables. We agree with the importance of establishing, early in life, a healthy lifestyle that includes regular exercise, avoidance of weight gain, and abstinence from tobacco use. Given the evidence that irreversible atherosclerotic changes may be caused by passive smoke exposure and that passive smoke exposure may act synergistically with other risk factors such as hypertension,'a.avoidance of long-term passive smokeexposure during childhood is important, especially for children with known prema- ture cardiovascular disease in their family. Our results . suggest that white males exposed to passive cigarette smoke who.have a family.history of cardiovascular dis- ease, weigh more than average; and have higher than av- erage diastolic blood pressure may be individuals at spe- -.cial risk for developing premature atherosclerotic heart disease. These risk profiles may be helpful in develop- ing preventive cardiovascular strategies for children. Accepted for publication September 22, 1998. 2This study was sponsored in part by the National In- s stitutes of Health, National Heart, Lung, and Blood Insti- tute, Bethesda, Md (R29-HL-38878 and RO1-HL-31010). Presented in abstract form at the Annual Meeting of the Society for Pediatric Research, San Diego, Ca1if, May 10_, 1995. Reprints: William B. Moskowitz, MD, Medical Col- lege of Virginia, PO Box 980543, Richmond, VA 23298. 1. 6uring JE, 0'Connor GT, Goldhaber 5Z, et at. Decreased HDLi and HDI3 cho- Iesterol, apoA-I and apo A-Il,and increased risk of myocardial infarction. Cir- culation. 1992;85:22-29. - 2. MoIIPP,SingCF,WeidmanWH,etal.Totalcholesterolandlipoprojeinsinschool children:predictionofcoronaryheartdiseaseinaduHrelatives.Circufation.1983: 67:127-134. , 3. Lee J, lauer RM, Clarke WR. Lipoproteins in the progeny of young men with coronary arterydlsease: children with Increased risk. Pediatrics. 1986;78:330- 337. 4. Bodurtha JN, Schieken R, SegrestJ, Nance WE. High-density lipoprotein cho- lesterol subtractions In adolescent twins. Pediabics. 1987;79.181-189. 5. Ksnnel WB, Felnleib M, McNamara PM, Garrison RJ, Castelll WP. Investi0allon of coronary heart disease In families: the Framingham Offspring Study. AmJFpi- demiot 1979;110:281-290. 6. Bodurtha JN, Chen CW, Mosteller M, Nance WE. Schieken RM, SegrestJ. Ge- nefic and environmental contributions to cholesterol antl iLs subtractions In 11- yearbld twins. ArtedoscferThromb.1991;11:844-850. . - 7. PathobiologialDetenninanlsofAtheroscle'rosisInY°ulh(PDAY) ResearchGroup. Relationship of atherosclerosis in young men to serum lipoproteln.cholrsterol concentrations and smoking: a preliminary regort from the Pathobiological De- terminants of Atherosclerosis in Youth (PDAY) Research Group. JAMA.199g: 264:3018-3024. B. Shepherd FJ, Cox M, West C. $anre factnrs influencing serum lipid levels in a working population. Atherosclerosis.198P,35:287-300. 9. MoSkowRzWB,MostelterM,SchlekenRM,etal:Lipoproteinandoxygentrraos- portaherations in passive smoking preatlolescentchildren: the MCVTwin Study. Circufation.1990:81:586-592-. . 10. Celernajer D5, Adams MR. Clarkson P, et al. Passive smoking and Impaired en- dothelium-dependentanerialdBZgoninhea0hyyoungaduhs.NErp/JMe0.1996; 334:150-154. 11. Kteinbaum DG, Kupper LL; Cassel JC; Tyroler HA. Multivariate analysis of coro- nary heart disease in Evans County, Georgia. Arch interri Med. 1971;128.943- 948. 12. Tyroler HA, Glueck CJ, Christensen g, Kwlerovich P0. Plasmahigh-densityli- poprotein cholesterol comparisons In blackand white populations: the Llpltl Re- search Clintcs Program Prevalence Study. CUcu/a6ad.1980;62(suppl IV):IV-99- IV-107. 13. Watkins L0, Neaton JD, Kutler LH: Hsdal differences In high-densly lipoprotein ARCH PEniATR ADOLESC MEDNOL 153, MAY 1999 452 2505586027

'NS indicates nonsmoking: PS, passrve smoking; and fH%, percentage of families with a positive family history of cardiovasculardisease, subfractiotLs.9 Passive smoke exposure also causes, in a dose-dependent manner, endothelial dysfunction.!o Child- hood passive cigarette sinoking may therefore acceler- ate atherosclerotic changes and elevate the risk of devel- oping CAD. . - Despite-a high prevalence of hypertension in black males, early reports suggest that the incidence of prema- mre CAD is lower in black males than in white males." The lower CAD prevalence in black males despite the higher prevalence of hypertension may be.engendered by higher levels of HDL-C in blacks.12 In a 1986 study,. HDL-C distributioit did not appear to fully account for the racial difference.l' While the risk of CAD may vary between blacks and whites, the standard major risk fac- tors for CAD appear predictive for both blacks and whites.14 Racial differences in CAD risk factors ob- served in middle-age are incompletely documented in childhood and adolescence in part because of thee con- founding of pubertal changes. Because of the known dif- . ferences in HDLGlevels in black-compared with white adults,.we pursued a study in adolescence to ttyto de- tect different effects of passive-smoking:onchanges in HDL-C levels in a biracial adolescent population. The pres- ent study examines thetollowing specific questions: (1) How are CAD risk factors, passivesmoking; sex, and race related in pubertal children? (2) Are.the effects.of pas- sive smoking different by race in familieswitha history of cardiovascular disease? (3) Is there an identffiablechild- hood risk profile, ie, does passive~smokeexposure in- teract with other coronary risk factors to increase the risk of developing premature CAD? . ' STUDY POPULATION AND,SMOKE EXPOSURE Population sample size and distribution for the entire study are shown in Table l. A significant proportion of the subjects were lost to follow-up. These losses were due to deletion of families that changed their smoking status, and losses from families moving away or drop- ping out of the study. Selection bias was not operative as the sample population subjects completing the study were similar to those subsequently norincluded by sex (males, 43% vs 44%), race (black, 22% vs 25%), and all  elaskMOrher  WhfteMnerer  BladcFZt6er 0 WhileFaher za .... 1s f u 1e t ~ t 5 0 . Nsfi1 V6vR2 Ys83 Yrstt4 Figure 1. Average number uf cigarettes smoked daily by the parents by visit Within each visit, a greater dairynumber of cigarettes were smoked by white parents than by black parents. Asteriskindicates Pc.05 and dagger, P<.07 lorstztistical comparisons otblack vs white parentz anthropometrics measures. The percentage of both black families and families with a history of cardiovascular dis- e ease was stable across visits at approximately 21%. The prevalence of smoking families decreased during the study from 35% to 27%. In smoking families, fathers began smoking at 18.2 m 6.2 years of age andpresently smoked 14.3 x 14.7 cigarettes per day. Mothers began smoking at 18.4 t 4.3 years of age and presently smoked 8.2 m 10.4 cigarettes per day. Thetotal daily number-of.cigarettes smoked by the parents at visit 1 ranged from1 to 10 in 17%, 11 to 20 in 32%, and was more.than 20 in 51%. The total daily number of cigarettes smoked by the par- ents did not vary significantly from visit to visit. How- ` ever, within each visit,.agreater daily number.ofciga-rettes were smokedby-whiteparents than by black parents (Figvre 1). Cotinine was not detected in chil- dren who were not exposed to cigarette smoke but was present in children exposedto°passive smoke (8.52 x 17.6 nmol/L). ANTHROPOMETRIC AND BLOOD PRESSURE DATA. Anthropometric and.blood pressure data are.shown in ~Tabte 2, No significant differences were seen between the sexes for weight and height at earlier ages, although boys were significantly heavier and taller than girls at later visits. Black children were significantly heavier and taller than white children at the earlier visits. Significant dif- ferences existed for systolic blood pressure between older boys and girls. In younger children, systolic blood pres- sure was higher in black children than in white chil- dren. Diastolic blood pressure was similar in black and white children. Girls and black children overall were more sexually mature by Tanner staging areach of the first 33 visits. Passive smoking and nonpassive smoking groups by sex were similar in all anthropometric measure- ments. No difference was noted in weekly self-reported physical activity (the number of times each week vigor- ous exercise was performed) in any age, race, or sex group. Alcohol consumption was absent by self-report during the first 3 visits and only 1 twin at visit 4 reported any alcohol consumption. ARCH PEOIAIIt ADOLESC MEDNOL 153. MAY 1999 2505586023 448

cholesterol and coronary heart disease Incidence in the usua( care group of the Multiple Risk Factor Interventional Trial. Am J Cardiof.1986;57:538-545. 14. Keil JE, Sutherland SE, Knapp RG, Lackland OT, Gazes PC, Tyrofer HA. Mortaliry rafesanddskfacforsforcoronarydlseaselnblackasonmparedwKfhwhhemen and women. NEnBI J Med.1993;329:73-78. - 15. TannerJM. 1962GrowthatAdolescence.2nded.OxfoM.England:BlackwellSci- entific:1962:40. . 16. Cone JT, Segrest JP, Chung BH, Ragland JB, Sabesin SM, Glasacack A Com- putedzed mptdhlgh resolution quantilaiive analyses of plasmatipoproteln based on stngle'ver8ral spin uifracentrituBatidn. J Upid Aes.1982;23:923-935. 17. PaNishall EN, Sirope GL, EnefRA, Helms RW. Haley NJ. Denny FW, Serum co- ~ linine as a measure of tobacco smoke exposure in children. AJDC.1985;139: 1101-1104.' - 18. Moskowitx WB, MostellerM, Schieken RM. Change in passive smoking status ahem Iiptd tracking in childhood: the MCV Twin Study. PediatrReS. 1993;33(4, pt2):24A. . 19. JennrichBl,SchluchterMD.Unbatancedrepeatedmeasuresmodetswithstruc- turetl covariance matrices. @iometrics. 1986;42:805-820. 20. SASInstitutelnc.SASTechniralRepartP229,SAS6TATSottwareChanpasand Enhancements, Release 6.07. Cary, NC: SAS Instilute tnc;1992. 21. McLean RA,Sanders WL, Stroup W W. A unlFred approach to mixed linear mod- ets. Am Statlstician.1991;45:54-64. 22. Patsch W, Schohteld G, Gotto AM, Patsch J. Charadenation at human high- Aensity Itpoproteins by zonal uRracenbituga8on. JAinlChem. 1980;255:3178- 3185.-- - ' 23. BodurthaJN. Schleken RM. Segrest J. Nance WE HDL cholesterol subtractions in adotescent twins. Pedietdrs.1987;79:181-189. - 24. BerensonGS.SdnrvasanSR.CresanmJL.FosterTA,WebberLS.Dynamicchanges of serum lipoproteins inchlldren during adolescence and sexual maturation. Am - JEpidemPoL 1981;113:157-170. 25. Kiddand RT, Keenan 65, Probstfleid JL, atal. Decrease in plasma high-densiry Gpoprotein cholesterol levels at puberty in boys with delayed adolescence: cor- relation with piasma testosierone level. JAMA 1987;257:502-507. 26. NerrteldEJ,MietusSnyderM,BeiserAS-BakerAL,NewburgerJW.Passiveciga- rette smoking and reduced HDLcholesterol levels in childrenwfth high-risk lipid profiles. CirculaNon.1997196:f403-1407. . - 27. Srinivasan SR, Frerichs RR, Webber LS. Berenson GS Serum lipoproteln pro- file in children irom a biracial community: the Bogalusa Heart SWdy. Circula- fioo- 1976:54t309-318. 28. Berenson G5, Faster TA, Frank GC, et al. Cardipvascular disease risk factor vari- ables at the preschool age: the Bogalusa Heart Study. (]mulation.1978:57:603- 612. 29. Hahn LP, Folsom AR, Sprafka JM, Norsted SW. Cigarette smoking and cessa- tion behaviors among urban blacks and whites. Public Heatth Aep. 1990;105: 290-295. ' " - 30. Frerichs RR, Webber LS, Srinivasan SR, et al. Relation of semm lipidsand tipo- protelns to obesity and sexual maturity In white and black children. Am J Ep'i- demioL1978;508:486-496. 31. Glueck CJ, Taylor HL, Jacobs 0, etal. Plasma high-density lipoproteln choles- teml: association with measurements of body mass-the Lipid Research Clin- ics Program Prevalence Study. Circulation. 1980;62(suppl4):IVE2-IV-69. 32. Van den Bree MBM, Schieken RM, Moskowitz W8, Eaves Ld. Genetic regulation oi hemodynzmicvadables during exercise: the MCV Twin Study. Crrculatlnn.1996; 94:1864-1889. 33. EpsteinLH,KuIlerLH,WingRR,ValdskiA,McCurleyJ.Theeffactofweightcnn- trol on lipid changes in obese children. AJDC.1989:143:454-457. 34. Bao W, Srinivasan SR, Valdez R, Greenlund KJ, Watligney WA, Berenson GS. Longitudinal changes In cardiovascular risk from childhood to young adulthood in offspring a/ parents with coronary artery disease: the Bogalusa Heart Study. JAMA.1997;278:1749-1754. 35. BaoW,SrinivasanSR,WattigneyWA.BerensonGS.Therelationofparentaicar- diovasculardisease to nsktactars in children and young adutts- Circufafion.1995 91:365-371. ' 36. Srinivasan SR, Dahlen GH, Jarpa RA, Webber L5. Berenson GS. Racial (black- whee) differences in serum lipoprutein(a) distribution and Its relation to paren- tal myocardial infarc9on in children. Circuladan.1991:84:160-167. 37. Llu K, Ruth KJ, Flack JM, et al. Blood pressure in young blacks and whites: rel- evance at obesity and lifestyle factors in determining ddferences: the CAROIA Study. Circulation. 1996;93:60-86. 38. Howard G, Wagenknechl LE, Burke GL, et al. Cigarette smoking and progres- sion of atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) Study. JAMA.1998;279:119-124. . Conrctiou Errors in Text. In the article titled"Head Tmuma in Children Younger Than 2 Years: Are There Predictors for Conrpiica- tions?," published in thejanuary issue of the ARCt1tvE5 (1999;153:15-20), several editorial errors were made in the abstract and text. In the "Results" section of the abstract, the third sentence should have read, °Seven percent of complications from SF/ICA resulted from falls 3 ft (0.9 m) or less." The last sentence in that paragraph should have read, "Among children who fel13 ft (0.9 m) or less and had no loss of consciousness, emesis, seizure, behavioral change, or scalp abnormality, none of 31 (95%confidence interval (CII, 0-0.10) children younger than 24 months and none of 20 (95%Cl, 0-0.15) children younger than 12 months had SFFICA." Similarly, in the first iull paragraph on page 17, the next to last sentence should have read, "OE the 138 children who fell 3 ft (0.9 m) or less, 7% had an SF andlor ICA." The ARCHtvps regrets these errors. ARCH PEDIATB ADOLESC MEDNOL 153, MAY 1999 453