f165 recurrent wheezing in verypreterm infants · been identified as risk factors for wheezing in...

Archives of Disease in Childhood 1996; 74: F165-F171 F165

Recurrent wheezing in very preterm infants

Dawn E Elder, Ronald Hagan, Sharon F Evans, Helen R Benninger, Noel P French

Departnent ofPaediatrics,Wellington ClinicalSchool ofMedicine,Wellington, NewZealandD E Elder

Department ofNewborn Services,King EdwardMemorial Hospital forWomen, 374 BagotRoad, Subiaco,Western Australia,6008R HaganH R BenningerN P French

Foundation forWomen's and Infants'Health, Subiaco,Western AustraliaS F Evans

Correspondence to:Dr R Hagan.Accepted 20 December 1995

AbstractAims-To document the prevalence of,and identify risk factors for, recurrentwheezing treated with bronchodilators inthe first year of life.Methods-Parental history and neonataldata were coliected prospectively in aregional cohort of very preterm infants(<33 weeks). Data on maternal smoking,siblings at home, breast feeding, respira-tory symptoms, and hospital re-admis-sions were documented at 12 months.Results-Outcome data were available for525/560 (95%) of survivors. The incidenceof recurrent wheeze was 76/525 (14/5%) invery preterm infants and 20/657 (3%) in acohort of term newborns. Significant riskfactors for recurrent wheeze in verypreterm infants were parental history ofasthma, maternal smoking, siblings athome, neonatal oxygen supplementationat 28 days, 36, and 40 weeks of gestation.Conclusions-Wheezing respiratory ill-nesses are common in very preterminfants. The factors involved are similarto those in more mature infants, with theaddition of immaturity and neonatal lunginjury.(Arch Dis Child 1996; 74: F165-F17 1)

Keywords: Very preterm, wheezing in infancy,bronchodilators, maternal smoking, asthma.

Recurrent respiratory illnesses, often requiringadmission to hospital, are common in verypreterm (<33 weeks gestation) or very lowbirthweight (<1501 g) infants in the first fewyears of life. 1-3 Wheezing is a prominent symp-tom in a significant number of these episodes.In the general population several variables havebeen identified as risk factors for wheezing ininfancy.4 These variables are also likely to beassociated with wheezing in very preterm orvery low birthweight infants.These risk factors can be classified as arising

temporally in the prenatal, perinatal, or post-natal stages of development. Prenatal factors,such as race or a family history of bronchialreactivity, have been studied in relation to neo-natal chronic lung disease5-7 but not in relationto later clinical respiratory outcomes in veryimmature infants. Perinatal factors such asmaternal smoking during pregnancy, theperiod of gestation, low birthweight for gesta-tion and acute and chronic neonatal lungdisease have been analysed in relation to subse-quent respiratory symptoms in very immatureinfants, though usually only in small birth-weight delimited cohorts2 38-0 The contri-bution of postnatal factors such as socialclass, siblings in the home, male sex, maternal

smoking and breast feeding has rarely beenanalysed in these cohorts9 10 though they areidentified as being important in more matureinfants.

In a large normal population of 5-11 yearolds Rona et al found that shorter gestationalage was associated with an increased risk ofrespiratory symptoms, especially wheeze,whereas low birthweight for gestation influ-enced only measured lung function in a man-ner that was independent of associatedmaternal smoking."1 This is an exceptionalstudy in that it analysed separately the effectsof gestation and the appropriateness of fetalgrowth. Other, retrospective, studies in child-hood and adolescent populations have docu-mented associations between both pretermbirth,5 low birthweight'2 13 and the develop-ment of bronchial hyperreactivity.The role of neonatal chronic lung disease

has been studied in several preterm or VLBWcohort studies using both prospective andretrospective data238-10 Only two of thesestudies9 10 looked at influences outside theneonatal period that may have modified res-piratory outcome and in all but one, containinga small number of infants,'0 the cohorts werebirthweight defined and likely to contain a vari-able number of low birthweight for gestationinfants. The sensitivity, specificity, and posi-tive predictive value of various periods of neo-natal supplemental oxygen requirement havebeen analysed and the suggestion made thatthe need for supplemental oxygen at 36 weeksof gestational age is the best predictor ofabnormal pulmonary outcome.3 This timeperiod for oxygen supplementation was thussuggested as a better definition for neonatalchronic lung disease than the more usualperiod of 28 days after birth. This study, how-ever, did not measure the influence of knownpostnatal variables contributing to the adversepulmonary outcomes.No study has examined all these temporally

unrelated risk factors in a cohort of verypreterm infants to establish the associationbetween gestational age and the developmentof wheezing in infancy, to examine the effectsof differing patterns of fetal growth, and toinvestigate how these influence the effectsof the variables identified in more matureinfants.The objectives of our study were firstly to

document the respiratory outcome of a largecohort of very preterm infants in the first yearof life, secondly to look for prenatal, perinatal,and postnatal influences on the respiratoryoutcome, and finally to examine the correla-tion between various periods of oxygen supple-mentation in the newborn period and lateroutcome. This study reports our findings withrespect to the outcome of recurrent wheeze

on January 29, 2020 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.74.3.F

165 on 1 May 1996. D

ownloaded from

http://fn.bmj.com/

F166 Elder, Hagan, Evans, Benninger, French

treated with bronchodilators occurring duringthe first year of life in this group.

MethodsESTABLISHMENT OF COHORTAll infants <33 weeks of gestation, liveborn orcared for during 1990 and 1991 in KingEdward Memorial Hospital for Women(KEMH), the sole obstetric perinatal tertiaryreferral centre in Perth, Western Australia,were enrolled. 14 5 This centre delivered and/orcared for 93% of the very preterm mother andinfant groups born in Western Australia duringthe period, and for 96% of the survivors. Allmothers were interviewed, their obstetricrecord reviewed, and information verifiedwhile the mother was still an inpatient.Gestational age was determined by the bestobstetric estimation of maturity, with knownmaternal dates and confirmatory obstetricultrasound examination before 20 weeks beingthe gold standard. Less than 5% of mothershad uncertain dates or no early ultrasoundscan. In these cases clinical gestational ageassessment of the newborn using the Ballardscales was used.'6

DATA COLLECTIONDemographic and parental respiratory historyThese were collected prospectively at birth andincluded maternal race, age, smoking duringpregnancy and the presence of other siblings athome. Maternal smoking during pregnancywas scored quantitatively as 0, 1-10, or >10cigarettes per day. A respiratory history ques-tionnaire, based on the American NationalInstitutes of Health respiratory question-naire,17 was used to obtain a history of asthmaor wheezing for both parents. Parents wereassessed as having a positive history if theyreported either doctor diagnosed asthma or ahistory of recurrent wheeze on exertion orexposure to cold at any time in childhood oradult life. Parental history was documented asbeing positive in none, one, or both parents.

Neonatal historyProspectively recorded neonatal data includedgestational age, birthweight, sex, the presenceand severity ofacute respiratory disease, days ofventilation, days on oxygen and method offeeding at discharge. Gestational age wasrecorded as completed weeks of pregnancy andgrouped as: <27, 27-29, and 30-32 weeks.Smallness for gestational age was defined as abirthweight below the 10th centile on age andsex appropriate charts.'8 Acute respiratorydiseases were classified using the criteria ofHjalmarson.19 Hyaline membrane disease(HMD) was defined as respiratory distress withradiographic evidence of air bronchogram anda ground glass appearance of the lung fieldswith increasing oxygen requirements over thefirst 24-36 hours of life persisting to 48 hours inthe absence of infection.'9 Hyaline membranedisease was considered severe if the infantrequired ventilation with a mean airway

pressure -1O cm H20 and FI02 B80% in thefirst 48 hours of life. Infants with no respiratorydistress, mild or transient respiratory distress,infection or immature lungs were classified asnon-HMD. Exogenous surfactant was onlyused in the last six months ofthe cohort recruit-ment period. Chronic lung disease (CIID) wasdefined using the criteria of Bancalari as theneed for supplemental oxygen after 28 days oflife associated with clinical symptoms and anabnormal chest radiograph.20 Persistent oxygenrequirement at 34 through 40 weeks of gesta-tional age was documented.

After dischargeFor the first six months after discharge theinfant's primary carer was asked to keep amonthly diary documenting the method offeeding and any respiratory symptoms. Ahealth questionnaire, completed at 12 monthscorrected age, was used to document and crosscheck the following:

respiratory symptoms using a grid, modi-fied from that of Woodward et al,21 listing13 respiratory symptoms and a frequencytable graded in the following manner:

(a) never (O bouts), rarely (1 bout), sometimes(2-4 bouts), frequently (more than 4bouts) and constantly (most of the time);

(b) need for drug treatment for wheeze;(c) a quantitative assessment of current

parental smoking habits;(d) the number of siblings in the home;(e) total duration of breast feeding after dis-

charge from the neonatal unit (< 1 month,1-2 months, 3-6 months, >6 months); and

(f) the number and nature of admissions tohospital in the first year of life.

Data for those infants admitted were checkedagainst records at Princess Margaret Hospital,the State's only tertiary referral hospital forchildren. Supplementary information was alsoobtained by the neonatal follow up coordinatorwho was in regular contact with local childhealth workers.

MAIN OUTCOME VARIABLESWe have used recurrent wheeze treated withbronchodilators in the first year of life as ourprimary outcome measure. We defined this asthe infant being reported as having two or moreepisodes of respiratory illnesses associated withwheezing for which the general practitioner orpaediatrician prescribed a bronchodilator. Thisoutcome was based on parental report that theyperceived their infant wheezing. This wasconfirmed for those infants admitted to theChildren's Hospital and in our hospital recordsfor those who were seen regularly in our followup clinic. We did not otherwise examine theinfant to confirm the presence of wheezing norassess its severity. No infant was known to havea cardiac or other mechanical cause for wheez-ing. We compared our figures with a localcontrol population from the Western AustralianPregnancy Cohort Study.22 This is a healthypopulation of 2834 newborns being followedup from early pregnancy to late childhood, in


http://fn.bmj.com

/A

rch Dis C

hild Fetal N


165 on 1 May 1996. D

ownloaded from

http://fn.bmj.com/

Recurrent wheezing in very preterm infants F167

Table 1 General clinical and demographic data

Variable No (%) Variable No (%)

Gestational age (weeks): HMDt 247 (47-0)<27 51 (9-7) PPVt 298 (56-8)27-29 143 (27 2) CLD** 114 (21-7)30-32 331 (63 0) Oxygen at 36 weeks 63 (12-0)

Birthweight (g): Oxygen at 40 weeks 21 (4-0)<1001 90 (17-1) Breast feeding:1001-1500 202 (38 5) None 154 (29 3)1501-2000 191 (36 4) <1 Month 49 (9 3)

>2000 42 (8 0) 1-2 Months 53 (10-1)Birthweight <10% 167 (31-8) 3-6 Months 126 (24 0)Male sex 286 (54 5) >6 Months 143 (27-2)Caucasian 449 (85 5) Siblings at home 274 (52-2)Parental BHR*: Maternal smoking:

Neither parent 390 (74 3) None 330 (62-9)One parent 116 (22 1) 1-10 per day 75 (14-3)Both parents 16 (3 0) >10 per day 120 (22-9)

*Bronchial hyperreactivity (no information available for three infants), thyaline membranedisease, tpositive pressure ventilation, **chronic lung disease.

whom extensive information on pregnancy,health, development and respiratory outcomesare being collected. A time delimited subset of657 have had their first year data analysed inrelation to the frequency of wheezing episodes.We are unaware of any bias in our medicalcommunity to treat or not treat wheezinginfants with bronchodilators if they have beenvery preterm. Those parents who themselveswere using, or had recently used, broncho-dilators were, however, more likely to discussthe need for bronchodilators in their infantswhen they attended our follow up clinic.

STATISTICSAll data were entered using the StatisticalAnalysis System (SAS) software package.23Relations between variables and the outcomeof bronchodilator treated wheezing were firstassessed by means of x2 analyses and calcula-tion of univariate odds ratios. Logistic regres-sion modelling was then used to evaluate thecorrelation between variables and wheezingwhile controlling for the effects of other con-founding variables by the method of maximumlikelihood. Adjusted odds ratios were calcu-lated for each variable. These analyses wereperformed only in the very preterm cohort.

ResultsThere were 648 infants born <33 weeks gesta-tion and cared for in KEMH during 1990-1.Twenty six died in the labour ward and 62 inthe neonatal unit; 560 were discharged alive.Respiratory follow up was available on 532(95%) of these infants. Eight infants died in

Table 2 Univariate and multivariate odds ratios (OR) and 95% confidence intervals(95% CI) of risk factors for wheeze treated with bronchodilators

Variable Univariate OR (95% CI) Adjusted OR (95% CI)

BHR (any parent vs none) 2-48 (1-48, 4-13) 2-80 (1-61, 4 89)Race (non-caucasian vs caucasian) 0 55 (0-28, 1-07) 0-63 (0 30, 1-35)Male sex 1-10 (0-68, 1-80) 1-31 (0-77, 2 25)Gestation (<30 w vs 330 w) 1-76 (1-08, 2-88) 2-22 (0-81, 6-06)Birthweight <100/ 1-39 (0-84, 2 30) 1-29 (0-72, 2 29)HMD 1-30 (0 80, 2 11) 1 10 (0-62, 1-96)CLD 2-13 (1-26, 3 62) 3-09 (1-45, 6 57)*Oxygen at 36 weeks 3-07 (1-68, 5-62) 4-25 (1-93, 9-35)*Oxygen at 40 weeks 4-89 (1-99, 12-05) 4-35 (1-53, 12-39)*Maternal smoking (any vs none) 1-98 (1-21, 3 23) 1-89 (1-10, 3 24)Siblings at home (any vs none) 2-38 (1-41, 4-01) 2-45 (1-39, 4-31)Breast feeding (21 month vs <1 month) 0-48 (0 30, 0 79) 0 74 (0 43, 1-29)

*Each duration of oxygen requirement entered separately into the logistic regression.

the first year of life. Causes of death were asfollows: bronchiolitis (three cases, one a childwith adrenoleucodystrophy), sudden infantdeath syndrome (two cases), and one case eachof necrotising tracheitis, Leigh's syndrome,and gastroenteritis. One of these infants diednear the end of the first year of life and her datawere included in the analysis. Five hundredand twenty five infants were thus available foranalysis and their clinical and demographicdata are outlined in table 1 In those infantswho died in the neonatal unit 12 (20%) had norespiratory questionnaire data. Ten (20%) ofthe remainder had a positive family history forasthma in one or other parent.

Bronchodilator treatment was prescribedfor recurrent wheeze in 76 (14.5%) infants.Recurrent episodes, defined as two or more, ofwheeze both treated and untreated occurred in125 (23-8%) infants. Admission to any hospitalbecause of a respiratory illness associated withwheezing was required by 78 (14.8%) of infantsand of these 39 never received bronchodilatortreatment. In the reference control group of 657term infants during the first year of life, 20/657(3 0%) infants had wheeze that was treated withbronchodilators and 21/657 (3.20/o) experiencedwheeze requiring admission to any hospital(W Macdonald, personal communication).

Significant risk factors for wheeze requiringbronchodilators on univariate analysis in ourvery preterm cohort were lower gestational age(<30 weeks), oxygen requirement at 28 days(CLD), oxygen requirement at 36 weeks andat term, the presence of siblings at home,maternal smoking and a parental history ofbronchial hyperreactivity (table 2) Breast feed-ing was protective.

Logistic regression analysis, controlling forall significant univariate risk factors, showedthat a parental history of bronchial hyperreac-tivity, the presence of siblings in the home,maternal smoking, CLD and oxygen require-ment at 36 weeks and term remained signi-ficant risk factors for wheeze requiringbronchodilators (table 2).

Parental history of bronchial hyperreactivity(BHR) and maternal history of smokingshowed an incremental dose-response relationto wheeze requiring bronchodilators on uni-variate analysis. Eleven per cent (43/390) ofinfants without a family history of BHRwheezed for which bronchodilators wererequired; 17% (20/116) with one parent posi-tive for BHR and 69% (11/16) with bothparents positive also required bronchodilatortreatment. One BHR positive parent increasedthe risk of wheeze to 1-7 (95% CI 1'3, 2 3),while two BHR positive parents positiveincreased the risk to 17 8 (95% CI 10 1, 31*2).

Table 3 shows the relations between mater-nal smoking, duration of breast feeding afterdischarge, and wheeze treated with broncho-dilators in the first year of life. Overall, there isa dose-response to maternal smoking, with theincidence of wheeze increasing as the dailynumber of cigarettes smoked by the motherincreases. In non-smokers any duration ofbreast feeding decreased the incidence oftreated wheeze. In both moderate (1-10 per


http://fn.bmj.com

/A

rch Dis C

hild Fetal N


165 on 1 May 1996. D

ownloaded from

http://fn.bmj.com/


Table 3 Number of infants (%) with wheeze treated with bronchodilators by maternal smoking and breastfeedingexposure

Breastfeeding

Nil <1 Month 1-2 Months 3-6 Months >6 Months Total

No smoking 16/82 (19-5) 3/24 (12-5) 1/32 (3-1) 5/89 (5 6) 12/103 (11-6) 37/330 (11-2)1-lOPerday 3/24(12-5) 3/11 (27 3) 1/4(25 0) 3/15(20-0) 2/21 (9 5) 12/5(16-0)>10 Per day 13/48 (27-1) 3/14 (21-4) 5/17 (29-4) 4/22 (18-2) 2/19 (10-5) 27/120 (22 5)Total 32/154 (20 8) 9/49 (18-4) 7/53 (13-2) 12/126 (9 5) 16/143 (11-2) 76/525 (14-5)

day) and heavy smokers (> 10 per day), it wasonly with prolonged breast feeding (>6months) that a protective effect was seen withno increased incidence of wheeze occurringin this group of infants whose motherssmoked.

Only a small number of infants had motherswho smoked solely either during pregnancy orafter pregnancy, but not in both periods (table4). There was no significant difference inoutcome for their infants between infancyexposure and in utero exposure. In all analysesthe exposure to maternal smoking in infancywas the variable examined.To measure whether any particular period of

prolonged oxygen requirement in the neonatalperiod was a better predictor of our wheezeoutcome, we calculated the sensitivity, speci-ficity, positive predictive values, negativepredictive values and accuracy ofvarying dura-tions of oxygen requirement after birth (table5) The points chosen were oxygen requirementat 28 days (CLD) and weekly incrementalincreases in duration of oxygen supplementa-tion from 34 weeks to 40 weeks gestational age.These were analysed using univariate data sothat comparison with the figures of Shennanet al could be made.3 The sensitivity for eachduration is low and decreases with increasinglength of oxygen requirement. The highestpositive predictive value occurs in those infantsoxygen dependent at term. The positive pre-dictive value for wheeze requiring bronchodila-tors is no better for those infants requiringoxygen treatment at 36 weeks than it is forthose requiring oxygen at 28 days.We calculated the per cent concordance for

the different logistic regression equations foreach weekly increase in length of oxygenrequirement, from 34 to 40 weeks inclusive,and oxygen requirement at 28 days of age.Overall, about 75% of the babies were cor-rectly assigned as developing or not developingwheeze. There was little difference in anyspecific duration of oxygen supplementation,suggesting, therefore, that no particular dura-tion of neonatal oxygen dependency is a betterpredictor of recurrent wheeze requiring bron-chodilators than any other (table 5)

Table 4 Outcome for infants exposured to maternalsmoking soleiy during pregnancy or solely during infancy

Treated RelativeTotal wheeze nsk* 95% CI

No exposure 300 34 (11 3%)Pregnancy alone 30 3 (10-0%) 0-88 0-29, 2-70Infancy alone 32 5 (15-6%) 1-38 0-58, 3-27Pregnancy and

infancy 163 34 (20 9%) 1-84 1 19, 2-84

Relative to no exposure.

When infants with any history of parentalBHR, exposure to passive smoking, CLD inthe neonatal period or cohabitation withsiblings were excluded, the overall incidence ofwheeze treated with bronchodilators in theremaining very preterm infants was 5/95(5-3%). This is still higher (difference 2-2%(95% CI -2-5 to 6.9%)) than in the terminfants.

DiscussionWheezing is a common symptom in the firstyear of life in this population of infants born atless than 33 weeks of gestation. We restrictedour risk analysis to those infants requiringbronchodilator treatment for wheeze to avoidincluding those infants wheezing only once aspart of an identifiable viral illness such as bron-chiolitis. The role of bronchodilators in thetreatment of wheeze in infants in the first yearof life is controversial, hence the lack of treat-ment in this age group may reflect varyingmedical practice rather than severity of thedisease. This could have led us to both under-and over-reporting of clinically importantwheeze. The incidence of wheeze treated withbronchodilator in our very preterm cohort wasalmost five times higher than in our term con-trol group. We did not examine these infantsand are thus dependent on local practitionersfor accuracy of diagnosis. However, bothpreterm and term control populations hadtheir diagnosis ascertained in an identical man-ner.We selected this as our measure of an

adverse respiratory outcome, because it is themost commonly reported adverse respiratoryoutcome.2 8 10 There is also evidence thatsuggests that VLBW and very preterm infantshave increased asthma24 and troublesomecough25 later in life. A number of infants whowheeze during infancy will develop asthmalater in childhood.26 We are following up thiscohort through their early school years.

Significant risk factors for wheeze treatedwith bronchodilators identified in our verypreterm infants are similar to those identifiedin infants in the general population.4

Comparability of reported figures forwheeze outcome is compromised because ofvariations in definitions of wheeze and in theages at which follow up has been documented.An increased incidence ofwheeze symptoms inpreterm compared with term infants has beenfound in birthweight defined cohort studies2 8but no group has yet reported this outcome ina gestational age based cohort. Ford et al foundthat 409% ofVLBW infants had experiencedone or more wheezing episodes by the age of 2


http://fn.bmj.com

/A

rch Dis C

hild Fetal N


165 on 1 May 1996. D

ownloaded from

http://fn.bmj.com/


Table S Assessment ofpredictive value of various durations of oxygen dependence for theoutcome ofwheeze treated with bronchodilators

Sensitivity Specificity PPV* NPV* Accuracy Concordance

CLDa 25-0* 90-2 30-2 87-7 80-9 74-4Oxygen (at weeks):34 30 3 80-4 20-7 87-2 73-1 73-935 27-6 87-3 22-6 87-3 75-8 74-136 25-0 90-2 30-2 87-7 80-8 74-137 19-7 89-8 24-6 86-9 79-6 73-838 15-8 91-8 24-5 86-6 80-8 73-439 14-5 92-9 25-6 86-5 81-5 73-440 11-8 97-3 42-9 86-7 85-0 73-6

O'Chronic lung disease, *positive predictive value, tnegative predictive value, tall figures arepercentages.

years compared with 4%/o of a term controlgroup.2 No comment about treatment ofwheeze was made. Kitchen et al, reporting onthe same cohort at 5 years, diagnosed asthma ifwheezing episodes had been treated with bron-chodilators in the previous year, and found theincidence to be 14-7% compared with 10/6%in a normal birth weight population.27 Lucaset al reported that 22-5 % of infants weighing<1850 g in their cohort had wheezed by theage of 18 months.9 The incidence for thosewith wheeze or cough treated with drugs was5-9%. Our prevalence of 14 5% was five timeshigher than in our reference population and isconsistent with these other studies in showinga much higher prevalence of wheezing in thevery preterm infant.

Increased bronchial responsiveness has beendocumented previously in preterm infants bothin cohort studies28 and in groups selected forhistory oflung disease in the neonatal period.29These findings are consistent with an effecton small airways. The prevalence of wheezesymptoms seems to decrease over the firstdecade of life, and while bronchial hyper-responsiveness persisting into child and adult-hood has been documented in those with ahistory of severe chronic lung disease in theneonatal period, in many of those testedclinical symptoms have been mild or evenabsent.27 29 This group may well represent thelow birthweight for gestation infant who aremore at risk of severe neonatal chronic lungdisease.6 The differential effects of pretermdelivery and low birthweight for gestation onclinical symptomatology and pulmonary func-tion have been clearly shown by Rona et al,'1though only studies defined on gestational agerather than birthweight will be capable ofdemonstrating these differential effects.

In our study, when only those infants whowere very preterm but had no significant riskfactors for wheeze were considered, the inci-dence of wheeze treated with bronchodilatorswas 5/95 (5 3%). This is not significantly dif-ferent from the 3% incidence found overall inour normal reference population (P=0 35).This reference control population does, how-ever, include infants with some of those samerisk factors, such as maternal smoking,parental bronchial hyperreactivity or siblings athome, suggesting, therefore, that very pretermbirth itself must increase the risk of wheeze inthe first year of life. There was no furtherincrease in risk at the lowest extreme of gesta-tional age in our cohort.

We could find no effect of low birthweightfor gestation, as measured by birthweight<10%. This is similar to the results of Ronaet al, 1 who found that clinical symptoms wererelated to immaturity but that pulmonaryfunction changes related to low birthweight forgestation.A family history of asthma increases the like-

lihood of bronchial hyperresponsiveness inearly infancy.30 While this association is alsotrue for low birth weight infants <2000 g,9 28the relation between a parental history ofbronchial hyperreactivity and the outcome ofwheeze has not been documented in a gesta-tional age based cohort. We also had clearevidence of an incremental increase in riskwhen both parents reported symptoms ofbronchial hyperreactivity and this effect per-sisted after controlling for all confounding vari-ables in our logistic regression analysis. Someauthors have postulated that familial bronchialhyperresponsiveness might be a risk factor foridiopathic preterm labour and chronic lungdisease postnatally.31 32 This correlation withneonatal chronic lung disease may relate to anincrease in severity in the disease in thoseinfants with a family history of bronchialhyperresponsiveness5 7 rather than an increasein the prevalence.Our rate of matemal smoking is comparable

with that of other studies, although rates dovary considerably, partly depending on mater-nal race.33 34 We used parental report to gaugesmoking exposures as this has been shown byothers to be reliable.35 While the effect ofmaternal smoking on infant lung function maybe most profound when the infant has beenexposed in utero,36 exposure to environmentaltobacco smoke in the first years of life seemsto have a separate and significant effect onrisk for respiratory illness and hospitalisation,especially in low birthweight infants.30 35 37This effect seems to be more severe in childrenwith a predisposition to asthma or wheeze.38 Inour cohort maternal smoking was a major con-tributing factor, almost doubling the risk of theinfant developing wheeze requiring treatment.In infants of mothers who smoke duringpregnancy a variety of pulmonary abnormali-ties have been reported and they may have arole in the development of bronchial hyper-reactivity in preterm infants with chronic lungdisease.34 39 4"2We were unable to differenti-ate the effects of in utero and infancy exposure,because we had so few mothers who smokedsolely either during pregnancy or after delivery.The presence of siblings in the home or

sharing a room has been reported to havevariable effects on the risk of wheezing ill-nesses in infancy. 12 33 43 44 We did not quanti-tate the number of siblings in the home, butthere was an increased risk of wheeze in thoseinfants with any sibling at home. This mayrelate to increased transmission of mild viralinfections that are associated with wheezing ininfancy whether or not there is bronchialhyperreactivity.

Neonatal chronic lung disease is known tobe associated long term with bronchial hyper-reactivity.29 45 Our study confirms this. We did


http://fn.bmj.com

/A

rch Dis C

hild Fetal N


165 on 1 May 1996. D

ownloaded from

http://fn.bmj.com/


not demonstrate a significant incrementalincrease in risk if oxygen requirement persistedbeyond 34 weeks gestation. Definitions of neo-natal chronic lung disease should have somelong term predictive value, as suggested byShennan et al,3 yet there is no accepted conceptof adverse pulmonary outcome. We have takenthe single outcome measure of recurrentwheezing treated with bronchodilators ratherthan the definition of adverse pulmonary out-come used in Shennan's study which was verybroad. One of their adverse outcomes wasrequirement for oxygen at 40 weeks gestationalage - part of the spectrum of duration of sup-plemental oxygen for chronic lung disease. Wefound that no particular duration of oxygensupplementation was a better predictor thanany other for our adverse outcome. This is incontrast to Sheenan et al, who reported thatrequirement for oxygen at 36 weeks gestationwas the best predictor of abnormal pulmonaryoutcome.3 Shennan's study, which was birth-weight based, did not control for the influenceof other factors such as parental bronchialhyperreactivity and passive smoking. Sensi-tivity and specificity values calculated onunivariate analyses are thus likely to present anincomplete and inaccurate picture.The overall protective effect of breast feed-

ing against the development of wheeze did notpersist after controlling for other factors. Lucaset al showed in their group of preterm infantsweighing less than 1850 g that early exclusivefeeding of breast milk did not protect againstthe development of wheezing.46 In term infantsa protective effect has been reported only in thefirst 4 months of life, though breast feeding hasbeen reported as protective against the nega-tive effect of maternal smoking on risk ofrespiratory illness'247 Our data suggest thatthis is true as we found a protective effect fromprolonged breast feeding in infants of heavysmokers.Wheezing illnesses are common in very

preterm infants during their first year of lifeand the factors associated with an increasedrisk are similar to those reported in the generalpopulation. Immaturity, as measured by gesta-tional age, and the acute lung injury that com-monly accompanies immaturity are associatedwith a higher risk. Low birthweight for gesta-tion is not associated with an increase in risk.This study has again highlighted at least onemodifiable factor that may decrease the inci-dence ofwheeze in this group. This is maternalsmoking which in this cohort occurred in37-1%. We found that many women who stopsmoking in pregnancy often restart in the post-natal period. This critical time must be usedfor support and education to help mothersprolong their period of non-smoking. Furtheradvances in the prevention and treatmentof neonatal chronic lung disease may alsoimprove respiratory outcome in this group andwe have yet to see the long term effects of earlysurfactant use. Little can be done to amelioratethe effects of the presence of siblings inthe house or a parental history of bronchialhyperreactivity, but awareness of these riskfactors may encourage earlier diagnosis and

appropriate treatment of preterm infants pre-senting with wheeze in the first year of life.We are grateful to Chris Campbell, Debra Chiffings, DoreenCoakes and Judith Onslow, who assisted with data collectionand entry, and to Colleen Ryan who assisted with data entry.

This study was funded by the Foundation for Women's andInfant's Health, King Edward Memorial Hospital, WestAustralian Health Promotion Foundation (Healthway).

1 Cunningham CK, McMillan JA, Gross SJ.Rehospitalization for respiratory illness in infants of lessthan 32 weeks gestation. Pediatrics 1991; 88: 527-32.

2 Ford GW, Rickards AL, Kitchen WH, Lissenden JV, RyanMM, Keith CG. Very low birthweight and normal birth-weight infants: a comparison of continuing morbidity.Med JAust 1986; 145: 125-8.

3 Shennan AT, Dunn MS, Ohlsson A, Lennox K, HoskinsEM. Abnormal pulmonary outcomes in prematureinfants: prediction from oxygen requirement in the neo-natal period. Pediatrics 1988; 82: 527-32.

4 Morgan WJ, Martinez FD. Risk factors for developingwheezing and asthma in childhood. Ped Clin North Am1992; 39: 1185-203.

5 von Mutius E, Nicolai T, Martinez FD. Prematurity as arisk factor for asthma in preadolescent children. J Pediatr1993; 123: 223-9.

6 Palta M, Gabbert D, Weinstein MR, Peters ME.Multivariate assessment of traditional risk factors forchronic lung disease in very low birth weight neonates.PPediatr 1991; 119: 285-92.

7 Hagan R, Minutillo C, French N, Reese A, Landau L, LeSouef P. Neonatal chronic lung disease, oxygen depen-dency and a family history of asthma. Pediatr Pulmonol1995; 20: 277-83.

8 Greenough A, Maconochie I, Yuksel B. Recurrent respira-tory symptoms in the first year of life following pretermdelivery. J Perinat Med 1990; 18: 489-94.

9 Lucas A, Brooke OG, Cole TJ, Morley R, Bamford MF.Food and drug reactions, wheezing, and eczema inpreterm infants. Arch Dis Child 1990; 65: 411-15.

10 Myers MG, McGuinness GA, Lachenbruch PA, KoontzFP, Hollingshead R, Olson DB. Respiratory illnesses insurvivors of infant respiratory distress syndrome. Am RevRespirDis 1986; 133: 1011-18.

11 Rona RJ, Gulliford MC, Chinn S. Effects of prematurityand intrauterine growth on respiratory health and lungfunction in childhood. BMJ 1993; 306: 817-20.

12 Weitzman M, Gortmaker S, Sobol A. Racial, social andenvironmental risks for childhood asthma. Am J Dis Child1990; 144: 1189-94.

13 Schwartz J, Gold D, Dockery DW, Weiss ST, Speizer FE.Predictors of asthma and persistent wheeze in a nationalsample of children in the United States. Am Rev Respir Dis1990; 142: 555-62.

14 Hagan R, Benninger H, Chiffings D, Evans S, French N.Very preterm birth: A regional study. 1. Maternal andobstetric factors. Br J Obstet Gynaecol 1996; 103: 230-8.

15 Hagan R, Benninger H, Chiffings D, Evans S, French N.Very preterm birth: A regional study. 2. The very preterminfant. BrJ Obstet Gynaecol 1996; 103: 239-45.

16 Ballard JL, Novak KK, Driver MA. A simplified score forassessment of fetal maturation of newly born infants.JPediatr 1979; 95: 769-74.

17 Ferris BG. Epidemiology standardization project. Am RevRespirDis 1978; 118: 1-120.

18 Blair E, Stanley FJ. Intra-uterine growth charts. AustralianGovernment Publishing Service, Canberra: 1985.

19 Hjalmarson 0. Epidemiology and classification of acute,neonatal respiratory disorders. Acta Pcediatr Scand 1981;70: 773-83.

20 Bancalari E, Abdenour GE, Feller R, Gannon J.Bronchopulmonary dysplasia: Clinical presentation.JPediatr 1979; 95: 819-23.

21 Woodward A, Douglas RM, Graham NMH, Miles H.Acute respiratory illness in Adelaide children: breast feed-ing modifies the effect of passive smoking. J EpidemiolComm Health 1990; 44: 224-30.

22 Newnham JP, Evans SF, Michael CA, Stanley FJ, LandauLI. Effects of frequent ultrasound during pregnancy:a randomised controlled trial. Lancet 1993; 342: 887-91.

23 SAS Institute. SAS/STAT user's guide. Release 6.03. Cary,North Carolina: SAS institute, 1988.

24 Seidman DS, Laor A, Gale R, Stevenson DK, Danon YL. Islow birthweight a risk factor for asthma during adoles-cence? Arch Dis Child 1991; 66: 584-7.

25 Chan KN, Elliman A, Bryan E, Silverman M. Respiratorysymptoms in children of low birth weight. Arch Dis Child1989; 64: 1294-304.

26 Martinez FD, Wright AL, Taussig LM, Holberg CJ,Halonen M, Morgan WJ, and the Group Health MedicalAssociates. Asthma and wheezing in the first six years oflife. N EnglJ Med 1995; 332: 133-8.

27 Kitchen WH, Ford GW, Doyle LW, Rickards AL, KellyEA. Health and hospital readmissions of very-low-birth-weight and normal-birth-weight children. Am J Dis Child1990; 144: 2213-18.

28 Chan KN, Noble-Jamieson CM, Elliman A, Bryan EM,Aber VR. Airway responsiveness in low birthweightchildren and their mothers. Arch Dis Child 1988; 63:905-10.

29 Northway WH, Moss RB, Carlisle KB, Parker BR, PoppR L, Pitlick PT, et al. Late pulmonary sequelae of


http://fn.bmj.com

/A

rch Dis C

hild Fetal N


165 on 1 May 1996. D

ownloaded from

http://fn.bmj.com/


bronchopulmonary dysplasia. N Engl Jf Med 1990; 323:1793-9.

30 Young S, Le Souef PN, Geelhoed GC, Stick SM, TurnerKJ, Landau LI. The influence of a family history ofasthma and parental smoking on airway responsiveness inearly infancy. N Engl Jf Med 199 1; 324: 1168-73.

31 Nickerson BG, Taussig LM. Family history of asthma ininfants with bronchopulmonary dysplasia. Pediatnics 1980;65: 1140-4.

32 Bertrand J-M, Riley SP, Popkin J, Coates AL. The long-term pulmonary sequelae of prematurity: the role of famil-ial hyperreactivity and the respiratory distress syndrome.N EnglJ Med 1985; 312: 742-5.

33 Wright AL, Holberg C, Martinez FD, Taussig L and GroupHealth Medical Associates. Relationship of parentalsmoking to wheezing and nonwheezing lower respiratorytract illnesses in infancy. J Pediatr 1991; 118: 207-14.

34 Hanrahan JP, Tager IB, Segal MR, Tosteson TD, CastilleRG, Van Vunakis H, et al. The effect of maternal smokingduring pregnancy on early infant lung function. Am RevRespir Dis 1992; 145: 1129-35.

35 Chilmonczyk BA, Salmun LM, Megathlin KN, Neveux LM,Palomaki GE, Knight GJ, et al. Association between expo-sure to environmental tobacco smoke and exacerbations ofasthma in children. NEnglJMed 1993; 328: 1665-9.

36 Taylor B, Wadsworth J. Maternal smoking during preg-nancy and lower respiratory tract illness in early life. ArchDis Child 1987; 62: 786-91.

37 Chen Y. Environmental tobacco smoke, low birth weight,and hospitalization for respiratory disease. Am J RespirCrit Care Med 1994; 150: 54-8.

38 Sherrill DL, Martinez FD, Lebowitz MD, Holdaway MD,Flannery EM, Herbison GP, et al. Longitudinal effects of

passive smoking on pulmonary function in New Zealandchildren. Am Rev Respir Dis 1992; 145: 1136-41.

39 Collins MH, Moessinger AC, Kleinerman J, Bassi J, RossoP, Collins AM, et al. Fetal lung hypoplasia associated withmaternal smoking: a morphometric analysis. Ped Res1985; 19: 408-12.

40 Chen MF, Kimizuka G, Wang NS. Human fetal lungchanges associated with maternal smoking during preg-nancy. PediatrPulmonol 1987; 3: 51-8.

41 Johnson DE, Lock JE, Elde RP, Thompson TR.Pulmonary neuroendocrine cells in hyaline membranedisease and bronchopulmonary dysplasia. Ped Res 1982;16: 446-54.

42 Aguayo SM. Determinants of susceptibility to cigarettesmoke. Am _J Respir Cit Care Med 1994; 149: 1692-8.

43 Gardner G, Frank AL, Taber LH. Effects of social andfamily factors on viral respiratory infection and illness inthe first year of life. 7 Epidemiol Comm Health 1984; 38:42-8.

44 Woodward A, Douglas RM, Graham NMH, Miles H.Acute respiratory illness in Adelaide children - the influ-ence of child care. MedJAust 1991; 154: 805-8.

45 Silva DT, Hagan R, Sly PD. Home oxygen therapy for neo-natal chronic lung disease in Western Australia. J PaediatrChild Health 1995; 31: 185-8.

46 Lucas A, Brooke OG, Morley R, Cole TJ, Bamford MF.Early diet of preterm infants and development of allergicor atopic disease: randomised prospective study. BMJ1990; 300: 837-40.

47 Wright AL, Holberg CJ, Martinez FD, Morgan WJ, TaussigLM and Group Health Medical Associates. Breast feedingand lower respiratory tract illness in the first year of life.BMJ 1989; 299: 946-9.


http://fn.bmj.com

/A

rch Dis C

hild Fetal N


165 on 1 May 1996. D

ownloaded from

http://fn.bmj.com/

f165 recurrent wheezing in verypreterm infants · been identified as risk factors for wheezing in...

Documents