Environmental Engineering Reference
In-Depth Information
We collected information for both case and referent mothers concerning their
residence during the 9 months before parturition (or 3 months for women undergoing
induced abortion procedure) and their educational attainment at the Reggio Emilia
Municipality General Registry Offi ce. In case of change of residence during the gesta-
tional period, we considered for exposure assessment in the present study the woman's
residence in the fi rst 3 months after the estimated date of conception: in the three case
showing a change of residence in this early period of the pregnancy, we attributed
exposure status on the basis of the longest period of residence within this time span.
We then geocoded the retrieved maternal addresses, using the database made available
by the Reggio Emilia Province or, in the few cases in which the addresses could not
be found in that database, measuring it on site with a geographical positioning system
device (Garmin GPSmap 60CSx, Garmin Int. Corp., Olathe, KS). A mother was con-
sidered exposed when her address was comprised within the intermediate and high (B
and C) exposure areas, after inputting it in the GIS.
All directly and indirectly nominative data were obtained by the Regional Hospital
Discharge Registry and by the Reggio Emilia General Registry Offi ce and subse-
quently analyzed in accordance with the legal and ethical guidelines for personal data
protection in epidemiological and scientifi c research of the Italian law [15] and with
the ethical guidelines of the IMER Registry [16].
Data Analysis
We calculated the prevalence ratio of having a birth or an aborted foetus with a con-
genital anomaly associated to maternal factors through the OR with its 95% confidence
interval (CI) generated by a conditional logistic regression model, entering as predic-
tive variables the area of residence during gestation, and educational attainment level.
Since no point estimate could be generated by the conditional analysis for the anoma-
lies characterized by the lowest prevalence, we used, in these cases, unconditional
logistic regression adjusting for both maternal age and education. We carried out this
analysis for overall congenital anomalies and for single birth defect categories for the
entire study period and, when numbers of cases made it possible, also for the normal
operation and shut-down periods of operation of the plant. The former period included
all births with congenital anomalies occurring in the periods December 1, 1998-October
31, 2002 and April 1, 2006-December 31, 2006, while that ascribed to the shut-down
period occurred from February 1, 2003 to December 31, 2005 (Figure 2). The remain-
ing time spans, in the first part of the shut-down period and at the beginning of the
reactivation (in 2005) of the plant, were removed from period-specific analysis due
to uncertainties in estimating the actual exposure status of women who delivered in
these periods. For induced abortions linked to an in utero diagnosis of birth defects, the
overall time span considered as “exposed” included the periods January 1, 1998-May 31,
2002 and October 1, 2005-December 31, 2006, while abortions considered as “unex-
posed” occurred from August 1, 2002 until July 31, 2005, and the events occurred in
the remaining periods were removed from analysis. In all these analyses, “period of
exposure” of each control birth was made equal to that of the corresponding matched
case. For all statistical analyses we used the package Stata-10 [17].
Search WWH ::
Custom Search