Methods
All live births in Nova Scotia, Canada, and the United States from 2007
to 2020 were included in the study. Data on live births in Canada were
obtained from the Nova Scotia Atlee Perinatal Database. This
population-based, clinically-focused database, contains information on
maternal characteristics, delivery events and neonatal information for
all births (with a birth weight of at least 500 gram or gestational age
of 20 weeks or more) in the province. Information in the database is
routinely abstracted from antenatal and medical charts by trained
personnel using standardized forms.14 Data for births
in the United States were obtained from the natality files of the
National Center for Health Statistics, which includes information on all
live birth registrations in the United States.15
ACS use in the Natality database of the United States was defined as
“ACS for fetal lung maturation received by the mother before delivery”
and available for all live births. The gestational age at ACS
administration was unknown in both Canada and the United States.
However, in Nova Scotia, information on ACS use in the Nova Scotia Atlee
Perinatal Database included the timing of the first dose administered in
relation to delivery (viz., first dose received <24 hours
prior to delivery, first dose received between 24 hours and 48 hours
prior to delivery, first dose received between 48 hours and 7 days prior
to delivery, and first dose received >7 days prior to
delivery) and this enabled us to distinguish between receipt of a
partial course (one dose) versus a complete course (two doses of
betamethasone) of ACS. Thus, women who received ACS <24 hours
prior to delivery were deemed to have received suboptimal ACS as this
represented insufficient time for receipt of a complete single
course.16-18 Women who received ACS prophylaxis more
than 7 days before preterm delivery at 24 to 34 weeks were also
considered to have received less than optimal therapy since the efficacy
of ACS in reducing respiratory distress syndrome does not extend beyond
7 days.16-18 We, therefore, categorized ACS use as
follows: i) any administration of ACS in the period before delivery; ii)
optimal ACS administration i.e., ACS administration between 24 hours to
7 days before delivery to women who delivered a live birth between 24 to
34 weeks of gestation; and iii) suboptimal ACS administration i.e., ACS
administration <24 hours or >7 days prior to
delivery to women who delivered a live birth between 24 and 34 weeks of
gestation. In Nova Scotia, gestational age was based on the following
hierarchy: the date of early second trimester ultrasound or the date of
the last menstrual period, or a postnatal assessment, and in the United
States it was based on the clinical (obstetric) estimate of gestation.
The time span of the study was divided into two periods, 2007-2016
(i.e., the period before and including the year of publication of the
ALPS trial) vs 2017-2020 (i.e. the period after the publication of the
ALPS trial), with the earlier period used as the reference. Rates of ACS
use were also examined by year. The frequency of ACS administration
within specific categories of gestational age (<24, 24-27,
28-32, 33-34, 35-36, ≥37 weeks) was assessed by calculating rates per
100 live births within each gestational age category in both Canada and
the United States. Odds ratios (OR) were used to quantify temporal
changes in ACS use by gestational age.
In Nova Scotia, Canada, we estimated the frequency of ACS administration
within categories of maternal and clinical characteristics including
mode of delivery. Mode of delivery was categorized as spontaneous
vaginal delivery, instrumental vaginal delivery, cesarean delivery in
labour, and planned cesarean delivery. Temporal trends were assessed by
plotting the frequency of optimal and suboptimal ACS administration
using 2-year moving averages over the study period. The rate
denominators for optimal and suboptimal administration were the number
of live births between 24 and 34 weeks’ gestation. The statistical
significance of a linear pattern in annual rates was assessed using the
Cochran-Armitage chi-square test for linear trend, and also visually to
identify non-linear patterns. The statistical significance of
differences was assessed using two-sided P-values and a P-value
<0.05 was considered statistically significant. Analyses were
performed using SAS software Version 9.2 of the SAS System for Windows
©. The Reproductive Care Program of Nova Scotia and the Research Ethics
Board of the IWK Health Centre provided data access and ethics approval,
respectively.