Participants
The study was based on interview data and blood samples from the Danish National Birth Cohort (DNBC), which is a nationwide cohort of 100,418 pregnant women and their offspring.
Enrolment in the DNBC took place between 1996 and 2002. The women were recruited in connection with the first antenatal visit to the general practitioner. Information on variables reflecting exposures before and during the early part of pregnancy was collected by means of a computer assisted telephone interview scheduled around gestational week 12. A second interview was scheduled in week 30 (interview forms are available at the website for the cohort).
During pregnancy, two blood samples were collected; one between gestational weeks 6 to 12, the second in gestational week 24. A sample was also drawn from the umbilical cord.
The interviews were performed if the women were reached within four phone calls, and if they agreed to participate.
The interviews covered reproductive history, age, smoking status, domestic contact to animals as well as very detailed questions regarding occupational contact to different animals.
A detailed description of the cohort can be found elsewhere [14].
In women who participated in the first interview and who also provided a blood sample (n = 95000) the study population was defined as follows: Occupational contact with livestock (n = 195), domestic contact with cattle or sheep (n = 202) and a randomly selected sample with no contact to livestock (n = 459). Blood samples from these 856 women were analyzed for antibodies against C. burnetii.
Pregnancy outcome was defined as:
Spontaneous abortion: fetal loss before 154 days (22 weeks) after the first day of the last menstrual period with gestational age estimated from the participants’ self- reported last menstrual period.
Preterm delivery: delivery (live births and stillbirths) between gestational weeks 22 + 0 days and 36 weeks + 6 days.
Small for gestational Age (SGA): for children born from week 37 + 0 and onwards, SGA was defined as a birth weight corresponding to the 10th percentile in gram and below. Children with a birth weight above the 10th percentile were used as reference group.
The relationship between serological status, birth weight and gestational age, respectively, was also evaluated.
We also evaluated late induced abortions and stillbirth.
Detection of antibodies against C. burnetii
C. burnetii expresses two antigens, phase I and phase II. When infected, phase II IgG and IgM antibodies are elevated, and they may remain positive for months to years. A large study from Australia and England found that phase II IgG antibodies persisted after four and 12 years, respectively [15].
In acute Q fever, primarily antibodies against phase II are raised, and titers are higher than antibodies against phase I. As with most other infections, IgM antibodies appear first.
In chronic forms of the disease, antibodies against phase I are elevated.
In order to determine antibodies against C. burnetii, we chose a two-step approach. First, all samples were screened in a commercial enzyme-linked immunosorbent assay (ELISA). Positive samples from the ELISA were confirmed with an immunofluorescence antibody test (IFA) which is considered to be gold standard when diagnosing Q fever.
The commercial ELISA kit were purchased from Panbio (Queensland, Australia) (cat. no. E-QFB01G and E-QFB01M) and used according to the manufacturer’s instructions with minor modifications; due to low sample volume the samples were diluted differently from what was prescribed in the instructions but the same dilution factors were used.
Samples which were positive for either IgG or IgM antibodies in the ELISA were confirmed with an IFA test from Focus Diagnostics (ca. no. IF0200G and IF0200M). The test was performed according to the instructions provided by the manufacturer, with the following minor modifications: due to a low amount of sample material, the diluted samples 1:10 from the ELISA were used to further dilute the samples as described by the manufacturer. The effect of the dilution in the Panbio buffer was tested prior to the use on patient samples and did not show any influence on the results (results not shown).
A local cutoff adjusted to the Danish population has been defined [16], including negative, equivocal and positive titers. When the ELISA positive samples in our study were reanalyzed using IFA, a modified version of the Danish cutoff was used. A sample was considered IFA positive when any of the phases were 1:128 or above.
For women without animal exposure, only the blood sample from the first trimester was analyzed. In women with contact to livestock, blood samples from the umbilical cord or mid-pregnancy were analyzed initially (n = 361 women) and therefore seroconversion during pregnancy could be monitored.
In order to detect a possible seroconversion throughout pregnancy, our strategy was to initially analyze the last existing blood sample (for 79 women this was the mid pregnancy sample and for 282 it was the umbilical cord sample). If this sample was tested positive in ELISA, the first blood sample from pregnancy week 12–16 was analyzed using ELISA.
In order to select which of the ELISA positive samples from the beginning of pregnancy were to be reanalyzed in IFA, the following criteria had to be met: a change in ELISA from negative in the beginning of pregnancy to positive in the mid-pregnancy or umbilical cord sample or a doubling in the adjusted ELISA OD-value throughout pregnancy.
In analyses of pregnancy outcome, women with seroconversion as well as women who were seronegative in the midpregnancy or in the umbilical cord sample were classified as seronegative.
All serological analyses were performed in a certified laboratory at Statens Serum Institut, Denmark. Laboratory personnel were blinded for exposure status and samples were always analyzed in the same batch of commercial kits.
Statistical analysis
Associations between positive serology (IFA), spontaneous abortion, preterm birth and Small for Gestational Age (SGA) were analyzed by logistic regression. The association between gestational age at birth (which does not follow a normal distribution) and positive IFA serology was tested using a non-parametric (Wilcoxon) test. We examined the association between positive serology (IFA), birth weight and gestational age for children born at term, respectively, by fitting multiple linear regression models.
Maternal age (<25 years, 25–34 years, 35+ years), number of previous pregnancies (0, 1+) and smoking during pregnancy (0, 1–10, 11+ cigarettes per day) were a priori selected as potential confounders.
All analyses were carried out using STATA statistical software, version 11.
Women enrolled in the Danish National Birth Cohort gave both verbal and written consent to participate. The women gave permission to include interview information, blood samples, and health information from other registers in the Danish National Birth Cohort. This study was approved by the Danish National Birth Cohort, the Danish Data Protection Board, and the Danish Regional Scientific Ethical Committee.