Perinatal factors associated with the respiratory

distress syndrome
Heather Bryan, MD, Peter Hawrylyshyn, MD, Sheilah Hogg-Johnson, MMath,
Susan Inwood, RN, Allen Finley, MD, Mario D'Costa, PhD, and Mary Chipman, MA
Toronto, Ontario, Canada

Perinatal factors related to the incidence of respiratory distress syndrome were analyzed by the multiple
logistic regression statistical method in 263 mothers and their 298 offspring delivered between 24 and 35
weeks' gestation in a 1-year period in a regional referral perinatal center. The risk of respiratory distress
syndrome in white infants rose with decreasing gestational age (p < 0.0001) while prolonged rupture of
membranes of >24 hours in the absence of maternal infection (28% of cases) was highly protective
(p < 0.0001). Compared with vaginal delivery, cesarean delivery without labor increased the risk of
respiratory distress syndrome (p = 0.03). The administration of tocolytic drugs was unrelated to the
incidence of respiratory distress syndrome, but corticosteroid therapy given at least 72 hours before
delivery was protective (p = 0.03). Male and female infants were equally at risk for respiratory distress
syndrome as were black and white infants, but other races had a lower incidence (p = 0.004). Infants with
respiratory distress syndrome were on mechanical ventilators longer than those with other respiratory
illnesses. (AM J OSSTET GVNECOL 1990;162:476-81.)

Key words: Respiratory distress syndrome, delivery mode, prolonged rupture of

membranes, corticosteroids, tocolytics

Respiratory distress syndrome (RDS) in premature ever, about the management of premature rupture of
infants has decreased in incidence over the years since the membranes in preterm pregnancies, which has
the establishment of neonatal intensive care units. I The been reported to occur in about a third of deliveries of
mortality and morbidity, however, still remain signifi- low-birth-weight infants." The obstetrician must decide
cant, particularly in very-low-birth-weight infants. 2 • 3 whether the risks of infection to the mother and infant
Obstetric management of the mother in premature la- outweigh the risks of prematurity with all its associated
bor has become increasingly important in efforts at- problems. 9
tempting to minimize or prevent the onset of RDS in We have analyzed our experience in a tertiary high-
the newborn infant. Stabilization of the at-risk mother risk perinatal unit over a I-year period by examining
with admission to the hospital, bed rest, clinical mon- all pregnancies where infants born alive at <36 weeks'
itoring, and frequent assessment of fetal well-being are gestation were admitted to our neonatal intensive care
important aspects contributing to effective treatment, unit. Our objective was to identify those factors that
as well as controlled labor and atraumatic delivery.4 The might influence our obstetric clinical management
indirect assessment of fetal lung maturity with mea- aimed at reducing the incidence of RDS in premature
surements of amniotic fluid lecithin/sphingomyelin infants.
(LIS) ratios: the administration of steroids to the
mother before delivery to accelerate fetal pulmonary Methods and patients
development,6 and the use of tocolytic therapy to delay Over a continuous 12-month period (December 1984
labor' are also measures adopted to reduce the inci- to November 1985), all infants admitted to the neonatal
dence of neonatal RDS. There is still controversy, how- intensive care unit (levels II and III) at <36 weeks'
gestation were examined prospectively. The relation-
From the Mount Sinai Hospital and the Departments of Pediatrics, ship of prenatal factors to the occurrence of infant RDS
Obstetrics and Gynaecology, Preventive Medicine, and BIOstatIStics was examined statistically by multiple logistic regression
and Clinical BiochemIStry, University of Toronto.
Supported by grant MT5609, Medical Research Counczl of Canada. analysis. \0 The associations investigated included the
Presented at the Forty-Fourth Annual Meeting of The Society influence of the administration of tocolytics and/ or ste-
of Obstetriczans and Gynaecologists of Canada, Vancouver, roids, the mode of delivery, the presence or absence of
British Columbia, june 21-25,1988.
Received for publication june 29, 1989; revised july 31, 1989; prolonged rupture of the membranes (PROM), race,
accepted August 23, 1989. gestational age, and sex of the newborn infant.
Reprint requests: Dr. M. Heather Bryan, Division of Neonatology, The diagnosis of RDS was made from the initial chest
Mount Sinai Hospxtal, 600 University Ave., SUIte 1241, Toronto,
Ontario, Canada M5G 1X5. radiograph obtained within 6 hours of delivery in those
6/1/16425 infants with respiratory symptoms and included the

476
Volume 162 Perinatal factors associated with RDS 477
Number 2

Table I. High-risk pregnancies Table II. Maternal pregnancy factors (white)

Twin pam Newborns Factor n
Race (n) (n)
Tocolytics given 90
White 215 29 244 PROM present 70
Black 24 5 29 Steroids
Other* 24 1 25 o or <24 hr 177
24-72 hr 32
TOTAL 263 35 298 >72 hr 35
Delivery
*Oriental, Asian. Hispanic, Native American, mixed. Vaginal 142
Cesarean section with labor 56
Cesarean section without labor 46
classic findings of air bronchograms with a generalized
reticular granular pattern in the lung parenchyma. To-
colytics given at the discretion of the obstetrician to
mothers in preterm labor were ritodrine and isoxsu- Table III. White premature infants (n 244)
prine. The use of steroids (betamethasone) was ex- n
amined as: (I) none or administered <24 hours before
delivery, (2) administered 24 to 72 hours before birth, Sex
Male 114
or (3) administered at least 72 hours before delivery. Female 130
The mode of delivery included those infants delivered RDS (n = 58)
vaginally and those by cesarean section with labor or Ventilated 58
Died 9
by cesarean section without labor. The absence of labor Non-RDS (n = 186)
was defined as less than an hour of painful contractions Ventilated 47
and no cervical dilatation or effacement. PROM was Died 6
Weeks' gestation
present for longer than 24 hours in hospitalized moth- 24 1
ers who were free of infection as manifested by normal 25 9
temperatures, normal white blood cell counts, and neg- 26 9
27 II
ative cultures. Prophylactic antibiotics were not admin- 28 20
istered to these mothers. Three race groups included 29 11
white infants, black infants, and others who were either 30 23
31 30
Oriental, Asian, Hispanic, Native American, or mixed. 32 20
To examine statistically the associations between the 33 41
covariates and the outcome RDS, a stepwise, backward 34 31
35 38
elimination procedure for variable selection was used
initially with significance to enter at p = 0.05 and to
remove at p = 0.10. The equations for the mathemat-
ical model are given in the Appendix. 1O therapy, including nine who died (13%). Of the 230
Infant survival and the use of mechanical ventilation infants without RDS, 61 (27%) required mechanical
also were examined. Determinations of LIS ratios were ventilation for other respiratory problems: wet lung
determined in many cases either on amniotic fluid ob- (n = 30), apnea (n = 23), pneumonia (n = 3), hydrops
tained by amniocentesis or in vaginal pool fluid" in (n = 3), and pulmonary hypoplasia (n = 2); seven of
cases of premature rupture of the membranes or in these died (3%).
tracheal fluid 12 aspirated from endotracheal tubes A total of III LIS ratios (37%) were obtained from
within 5 hours of delivery in intubated newborn infants. amniotic or newborn tracheal fluid. In 37 infants, all
These data were examined statistically by the Fisher of whom had RDS by clinical and radiographic criteria,
exact test. the LIS ratios were immature (sensitivity 100%). In 74
infants without RDS, 71 LIS ratios were mature while
Results three were immature (specificity 95%). The latter three
During the year of study, 298 live premature infants infants with low LIS ratios did not require mechanical
were delivered to 263 mothers between 24 and 35 ventilation or supplementary oxygen therapy.
weeks' gestation. All infants, including 35 sets of twins, All cases (n = 298) were included in the initial step-
were admitted to the neonatal intensive care unit. The wise logistic regression procedure (Appendix) investi-
population was predominantly white (82%), with 29 gating the association between the covariates (gesta-
black infants and 25 classified as other (Table I). There tional age, sex, race, etc.) and the incidence of RDS.
were 68 infants with RDS, all of whom required assisted Compared with the white population, infants of black
mechanical ventilation and supplementary oxygen mothers had a similar incidence ofRDS (p = 0.9) while
478 Bryan et al. February 1990
Am J Obstet Gynecol

no PROM - C Section no Wxlur

••. ..••.•. .•.•••••••. no PROM· C Section labour

':~:':~::~~:~~""
no PROM· Vaginol 0eI1.-.y
co PROM· Vaginol 0eI1.-.y
ci

(f)
o
a:
,
'" ,
,,
, ,,
,,
,,
,,
,,

"'~", ,.,.
,,
,,
,,
,,
,,
,,
C\I
,,
ci

-.-.
... .............
................ "

--..--- .. -----._---_ _.. _..- ..

o
ci
24 26 28 30 32 34 36
Gestational Age (wks)

Fig. 1. Predicted probability of infant RDS related to gestational age (weeks) in the white population
(n = 244). The risk of RDS was lowest in vaginally delivered infants in the presence of PROM
(n = 49) and highest in those delivered by cesarean section without labor or PROM (n = 42). Those
delivered by cesarean section in labor in the absence of PROM (n = 39) had the same risk of RDS
as vaginal delivery without PROM (n = 93) and cesarean section without labor or PROM. The
probability of RDS rose in all situations as gestational age decreased.

Table IV. Stepwise logistic regression in white patients

Variable Coefficient ({3) SE p Value
Intercept 10.50 1.97 <0.0001
Gestation (wk) -0.38 0.064 <0.0001
PROM -1.80 0.530 0.0002
Cesarean section (no labor)* 0.97 0.445 0.03
Cesarean section (Iabor)* 0.77 0.459 0.09
Steroid (24-72 hr)t 0.12 0.490 0.81
Steroid (>72 hr)t -1.38 0.701 0.03

*Delivery was a categorical variable with three levels where vaginal delivery was the baseline. Indicator variables represented
cesarean section without labor and cesarean section with labor.
tSteroid administration of two doses before delivery was treated as a categorical variable with three levels, where no steroid or
steroid given <24 hours before delivery was the baseline. Indicator variables represented are steroid given 24 to 72 hours or >72
hours before delivery.

infants of other races had significantly less RDS (p = Results from the logistic regression procedure ap-
0.004). Because nonwhite infant numbers were small plied only to the white population are summarized in
(Table I), only white mothers and their infants were Table IV listing the levels of significance. Infant gender
included for further logistic regression analysis. Ma- (male 148, female 150) and the administration of to-
ternal pregnancy characteristics studied are listed in colytic drugs to the mother (n = 109) were not signif-
Table II and infant sex, gestational ages, survival, in- icantly related to RDS, and these two variables were
cidence of RDS, and assisted mechanical ventilation re- therefore removed from the subsequent analyses. The
quirements are shown in Table III. risk of RDS increased with decreasing gestational age
Volume 162 Perinatal factors associated with RDS 479
Number 2

Table V. Delivery mode with or without PROM related to RDS in white patients
PROM No PROM

DelIVery Total No.

I No. wIth RDS Total No.
I No. with RDS

Vaginal 49 5 93 21
Cesarean section
Labor 17 I 39 13
No labor 4 0 42 18
Died 7 3 8 6
TOTAL 70 6 174 52

(P < 0.0001), while the presence of PROM was asso- (n = 9), compared with 3% (n = 6) in infants without
ciated with a low incidence of RDS (p < 0.0001) (Table RDS (Table III). Prolonged assisted mechanical venti-
IV). Delivery was a categorical variable with three levels lation exceeding 7 days occurred in 47% of the RDS
where vaginal delivery was the baseline. Indicator vari- group compared to 23% of ventilated non-RDS infants
ables represented cesarean section without labor and (p < 0.01, Fisher's exact test) .
cesarean section with labor. Compared with the vaginal
route, delivery by cesarean section in the absence of Comment
labor increased the risk of infant RDS (p = 0.03). In The strikingly reduced incidence of RDS in white
deliveries by cesarean section with labor, the risk of infants delivered vaginally to mothers free of infection
infant RDS fell between but was not significantly dif- after PROM of >24 hours' duration was seen from 24
ferent at the 5% level from the risk associated with to 35 weeks' gestation, rising from 10% at 30 to 35
vaginal delivery (p = 0.09) and delivery by cesarean weeks to a 50% risk at 24 weeks (Fig. 1). This is in
section without labor (p = 0.7) (Table IV). Steroid ad- marked contrast to that incidence occurring in infants
ministration of two doses before delivery was treated where PROM was absent and the delivery was by ce-
as a categorical variable with three levels, where none sarean section without labor. The risk without PROM
or given <24 hours before delivery was the baseline, was approximately twice that seen at all gestational ages
with indicator variables represented as given 24 to 72 for those with PROM , reaching > 80% in the more im-
hours or >72 hours before delivery. Steroid therapy mature infants (Fig. 1).
given to induce fetal lung maturity was associated with The association of PROM with a decreased incidence
a reduction in the incidence of RDS only when admin- of RDS is not a new finding as Bauer et al. 13 first noted
istered at least 72 hours before delivery (p = 0.03) this relationship in 1974. In the National Institutes of
(Table IV). The analysis did not change appreciably Health collaborative trial of antenatal steroid therapy
when twins were removed (n = 58), and therefore both the protective effect of PROM in the placebo group
twins and singletons are included. also was demonstrated. H Garite et aI., 15 however, found
The predicted probability of RDS related to gesta- no difference in the incidence of neonatal RDS in moth-
tional age in the presence or absence of PROM, and ers with PROM and conservative management com-
for the different delivery modes, is illustrated in Fig. 1 pared with a group with PROM delivered 48 hours
from 24 to 36 weeks' gestation. Compared with all de- after steroid administration . The adverse effect of ce-
livery routes without PROM, vaginal delivery with sarean delivery on the incidence of RDS also has been
PROM showed a lower risk for RDS (Fig. 1). Predicted noted in the past,lb and this is most significant when
probabilities in the presence of PROM for other deliv- labor is absent. 14 We have demonstrated again, in this
ery modes were not included because of the small num- study, the link of PROM in vaginal deliveries to a low
ber of such cases (Table V). There were six infants with risk of RDS, in contrast to a higher risk when PROM
RDS in 70 cases of PROM (9%) compared with 52 cases is absent and the delivery is by cesarean section without
in the absence of PROM (30%) (Table V). There were labor. In addition, the risk of RDS changes, rising as
no cases of neonatal sepsis diagnosed by positive cul- the gestational age decreases, the rate being dependent
tures in the infants delivered after PROM where the on the mode of delivery in premature white infants
range of time from membrane rupture to delivery was (Fig. 1).
from > 24 hours to 63 days (median 3.0 days). All in- There was no difference in the effect of race on the
fants from these cases of PROM had septic screening incidence of RDS in this study between white and black
at birth and were placed on a regimen of broad spec- infants, unlike in past reports, 1 , while infants of other
trum antibiotics for 2 to 5 days. races had significantly less RDS as previously noted. II
The death rate in white infants with RDS was 16% Our numbers of infants who were not white were too
480 Bryan et al. February 1990
Am J Obstet Gynecol

few (n = 54) to analyze independently. Male infants fections in this series: this finding is similar to other
were no different from female infants in their risk for reports where the risk of fetal infection was low. 8.9.15.21
RDS, in agreement with previous studies. 17 The exclu- In attempts to reduce the impact of RDS on pre-
sion of twins from our analysis did not alter the risk of mature infant mortality and morbidity our experience
RDS, unlike previous reports that suggest a greater risk in a regional perinatal referral center supports the find-
of RDS for the second-born twin.IH ings that conservative obstetric management of at-risk
The administration of steroids to mothers in pre- mothers with PROM is associated with a decline in the
mature labor has been shown to decre<\s; the incidence incidence of neonatal RDS.8.9. 13 Less clear is the influ-
of RDS in gestations of >27 weeks but not in more ence of the administration of maternal tocolytic drugs
immature infants given dexarriethasone at least 24 to inhibit labor as we and others 7 have not shown a
hours, but not later than 7 days, before delivery. 14. IQ direct decrease in the risk for RDS. Corticosteroid ther-
Our findings with betamethasone concur, even with less apy, appropriate in dose and timing, has a protective
mature infants, although we did not exclude those effect in more mature infants (>27 weeks), 14.19 although
mothers who received steroid more than a week before our experience also included those less mature. Deliv-
delivery. Whether it is the steroid administration alone ery by the vaginal route, where possible and appro-
or the delay in delivery, allowing spontaneous fetal pul- priate, is preferable because cesarean delivery, partic-
monary maturation, or a combination of both that re- ularly if labor is absent, can be associated with an in-
duces the risk of RDS in newborn infants remains to creased incidence of RDS.14· 16
be elucidated.
The administration of maternal tocolytics (isoxsu-
prine or ritodrine) showed no relationship to the risk REFERENCES
of RDS in our infants. A recent review of the efficacy 1. Stahlman MT. Newborn intensive care: success or failure?
of tocolytics in preterm labor summarizing the many J Pediatr 1984;105:162.
publications on this subject concluded that the inci- 2. Wegman ME. Annual summary of vital statistics-1986.
Pediatrics 1987;80:817.
dence of severe respiratory disorders in the newborn 3. Kamper J. Long term prognosis of infants with severe
is not reduced, although an unequivocal beneficial ef- idiopathic respiratory distress syndrome. 1. Neurological
fect was seen in delaying the delivery.7 We did not ex- and mental outcome. Acta Paediatr Scand 1978;67:61.
4. Creasy RK. Preterm labour and delivery. In: Creasy RDD,
amine our population for this latter influence of the Resnik R, eds. Maternal-fetal medicine. Philadelphia: WB
use of tocolytics. Saunders, 1984:418-37.
The diagnosis of RDS depends on clinical and ra- 5. Gluck L, Kulovich MV, Borer RC, Keidel WN. The in-
terpretation and significance of the lecithin / sphingomye-
diologic interpretation. Biochemical estimations of lin ratio in amniotic fluid. AM J OBSTET GYNECOL
LIS ratios did support our early diagnosis. Even in 1974;120:142.
infants of <28 weeks, pulmonary surfactant maturity 6. Liggins GC, Howie RN. A controlled trial of antepartum
glucocorticoid treatment for prevention of the respiratory
is not uncommon.12 distress syndrome in premature infants. Pediatrics
The impact of RDS on mortality in our premature 1972;50:515.
infants was high; 13% of babies with RDS died com- 7. King GF, Grant A, Keirse MJ, Chalmers 1. Betamimetics
in preterm labour: an overview of the randomized con-
pared with 3% of non-RDS infants. Prolonged assisted trolled trials. Br J Obstet GynaecoI1988;95:211.
mechanical ventilation (>7 days) was required in 47% 8. Perkins RP. The neonatal significance of selected perinatal
of the infants with RDS compared with 23% of venti- events among infants oflow birthweight. II. The influence
of ruptured membranes. AM J OBSTET GYNECOL 1982;
lated, non-RDS babies (p < 0.01). There were seven 142:7.
deaths in 70 white infants where PROM occurred 9. Rudd EG. Premature rupture of the membranes. A re-
(Table V), three in infants with RDS, all born at 25 view. J Reprod Med 1985;30:841.
10. SAS Institute Inc. SUGI supplemental library user's
weeks' gestation. Two other neonatal deaths were due guide. Cary, North Carolina: SAS Institute Inc. 1983.
to Potter's syndrome and herpes encephalitis, both con- 11. D'Costa M, Dassin R, Bryan H, Joutsi P. Validation of a
firmed at postmortem examination; one was a late simple rapid high performance liquid chromatographic
method for amniotic fluid lecithin/sphingomyelin ratios.
death at 3 weeks of age in a 26-week infant with nec- Clin Chern 1985;18:27.
rotizing enterocolitis. One other death resulted from 12. D'Costa M, Dassin R, Bryan H. Lecithin/sphingomyelin
severe pulmonary hypoplasia in a 26-week infant of a ratios in tracheal aspirates from newborn infants. Pediatr
Res 1987;22:154.
pregnancy in which membranes had been ruptured for 13. Bauer CR, Stern L, Colle E. Prolonged rupture of mem-
63 days. Pulmonary hypoplasia is a recognized hazard branes associated with a decreased incidence of respira-
to the fetus where membranes are ruptured at or before tory distress syndrome. Pediatrics 1974;53:7.
14. Collaborative Group on Antenatal Steroid Therapy. Ef-
20 to 21 weeks with a subsequent ongoing significant fect of antenatal dexamethasone administration on the
amniotic fluid leak. 20 It is difficult to know whether an prevention of respiratory distress syndrome. AM J OBSTET
earlier delivery would have benefited this infant in the GYNECOL 1981; 141 :276.
15. Garite TJ, Freeman PK, Linzey EM, Braly PS, Dorchester
presence of the marked prematurity «26 weeks). WL. Prospective randomized study of corticosteroids in
None of the infants of gestations with PROM had in- the management of premature rupture of the membranes
Volume 162 Perinatal factors associated with ADS 481
Number 2

and the premature gestation. AM J OBSTET GYNECOL mentary library. If p represents the probability of RDS,
1981;141:508. then the logistic regression model is of the form:
16. Usher RH, Allen AC, McLean FH. Risk of respiratory
distress related to gestational age, route of delivery, and (p)
maternal diabetes. AM J OBSTET GYNECOL 1971; III :826. Log., - - = ~o + ~,x, + . . . +~"xp
1 - P
17. Farrell PM, Avery ME. Hyaline membrane disease. Am
Rev Respir Dis 1975;111:657. where X" ... , Xp are the explanatory variables (which
18. Driscoll SG, Yen SB. Neonatal pulmonary hyaline mem- may include both continuous variables and indicator
brane disease: Some pathologic and epidemiologic as-
pects. In: Villee CA, Villee DB , Zuberman J, eds. Respi- variables for categories) and 130, 13" . .. , I3 p are the co-
ratory distress syndrome. New York: Academic Press, efficients to be estimated; 0 is the intercept. This equa-
1973:161. tion could be reexpressed as:
19. Avery ME, Aylward G, Creasy R, Little AB, Stripp B.
Update on prenatal steroid for prevention of respiratory p = exp (~o + ~,x, + . . . ~.xp)
+
distress. Report of a conference-September 26-28,1985. 1 + exp (~o + ~,X, + . . . ~"xp)
+
AM J OBSTET GYNECOL 1986; 155:2.
20. Nimrod C, Varela-Gittings F, Mackin G, Campbell D, Wes- A backward stepwise regression procedure starts with
senbert R. The effect of very prolonged membrane rup-
ture on fetal development. AM J OBSTET GYNEGOL
all explanatory variables of potential interest in the
1984; 148:540. model. At each step one variable may be deleted; all
21. Druzin ML, Toth M, Ledger WJ. Nonintervention in pre- variables included in the model are examined, and the
mature rupture of the amniotic membranes. Surg Gynecol one with the smallest maximum test statistic, provided
Obstet 1986; 163:5. it has a significance level greater than a specified level
for removal, is withdrawn. After one variable is re-
Appendix moved, if any are eligible, each variable not in the
A backward stepwise logistic regression procedure model is examined, and if any has a significance level
was used to analyze the association between various smaller than some specified level for entry, it is added.
explanatory or predictor variables and the probability The procedure continues until no further variables are
of RDS. This was accomplished with the LOG 1ST pro- eligible to be added or deleted. to
cedure from the Statistical Anaiy~;s System's supple-

Bound volumes available to subscribers

Bound volumes of the AMERICAN JOURNAL OF OBSTETRICS AND GYNECOLOGY are avail-
able to subscribers (only) for the 1990 issues from the Publisher, at a cost of $56.00 ($79.00
international) for Vol. 162 (january-June) and Vol. 163 (july-December). Shipping charges
are included . Each bound volume contains a subject and author index and all advertising
is removed. Copies are shipped within 60 days after publication of the last issue in the
volume. The binding is durable buckram with the JOURNAL name, volume number, and
year stamped in gold on the spine. Payment must accompany all orders. Contact The C.V.
Mosby Company, Circulation Department, 11830 Westline Industrial Drive, St. Louis,
Missouri 63146-3318, USA; phone (800) 325-4177, ext. 7351.
Subscriptions must be in force to qualify. Bound volumes are not available in place
of a regular JOURNAL subscription.