The Norwegian Mother and Child Cohort Study (MoBa)
MoBa is a prospective population-based pregnancy cohort study conducted by the Norwegian Institute of Public Health [
27]. Participants were recruited from all over Norway from 1999 to 2008 and 40.6% of all women that were invited consented to participate. The cohort now includes 114,500 children, 95,200 mothers, and 75,200 fathers. Three questionnaires were answered during pregnancy, with follow-up questionnaires answered post-partum at regular intervals. Pregnancy and birth records from the Medical Birth Registry of Norway (MBRN) were linked to the MoBa cohort [
28]. The quality-assured data files entitled ‘version 8’ were used. Response rates for the MoBa version 8 questionnaires can be found elsewhere [
29,
30].
Questionnaires and variables
The data were primarily collected from four self-administered questionnaires (Q). The first one was answered in gestational week 15 (Q1), and the subsequent questionnaires were administered around week 20 (Q2), week 30 (Q3), and 6 months post-partum (Q4).
Q1 probed into details regarding maternal health and lifestyle, demographics, previous pregnancies, and early first trimester reports of NP, NVP, and PGP.
From Q1 we used maternal height (cm) and weight (kg) at the start of pregnancy to calculate body mass index (BMI, kg/m2), education (seven categories collapsed into: ≤12 years, 13–16 years, or ≥ 17 years), smoking during pregnancy (no, sometimes, or daily), age at menarche (years), incidences of irritability before menses (five categories collapsed into ‘yes/no’), number of previous pregnancies, previous experiences of NVP (yes or no), and experiences of NP, NVP, and PGP in the first 15 weeks of pregnancy (separated into 4 week intervals; yes or no). The incidences of irritability before menses were defined as having reported an alleviation of such symptoms after menses. Years from menarche to pregnancy were calculated by subtracting age at menarche (from Q1) from maternal age at time of birth (from MBRN). Due to obvious erroneous reporting, only values ≥5 years and < 35 years from menarche to current pregnancy were used. Likewise, ages at menarche < 7 years and > 25 years of age were considered to be erroneously reported.
Q2 included a food frequency questionnaire and was answered in mid-pregnancy. Version two of Q2 included detailed questions regarding nausea and vomiting, therefore only women answering this version were included in the sample. While the questions regarding NVP and PGP in Q1 were structured simply as ‘yes/no’ for the corresponding 4 week intervals in early first trimester, the NVP questions in Q2 probed experiences of nausea or vomiting during pregnancy (yes or no), the gestational week of onset and cessation, and whether women were still experiencing nausea and/or vomiting at the time of answering the questionnaire (around gestational week 20). These NP and NVP answers were cross-checked with answers provided in Q1 to avoid inconsistent responses, as detailed in the ‘Study Sample’ section below. Prior to calculating the duration of either nausea or vomiting from variables stating the gestational week of onset and the week of cessation of these conditions, recoding for some variables was required, as described in detail previously [
5].
Data from Q3 (gestational week 30) enabled us to differentiate NVP women from women diagnosed with hyperemesis gravidarum (HG), defined as prolonged nausea and vomiting during pregnancy requiring hospitalisation before week 25 of gestation [
31]. Q3 also provided comprehensive information regarding pain distribution and pain intensity (none/mild/severe) in the pelvic girdle during pregnancy. We defined PGP as self-reported pain in both the pubic symphysis in the anterior pelvis and bilaterally in the sacroiliac joints in the posterior pelvis (yes or no). Severe PGP (sPGP) was defined as having reported severe pain in all the above pain locations [
21]. These conditions are referred to as PGP and sPGP, respectively. We also obtained responses regarding experiences of NP, NVP, and PGP from week 15 to week 29 or more of pregnancy (separated into 4 week intervals; yes or no). All ‘yes/no’ responses for NP, NVP, and PGP from Q1 and Q3 were cross-checked with identified cases based upon the more comprehensive definitions of NP, NVP, and PGP defined above.
We based the reporting of PGP or sPGP during the last part of pregnancy and/or 4-6 months post-partum upon questions in Q4 (six months post-partum) regarding pain in the anterior and bilateral posterior pelvis [
32]. Women reporting PGP or sPGP during the last part of pregnancy were added to those women reporting PGP or sPGP in Q3, producing the total number of women reporting PGP or sPGP during pregnancy.
Data concerning maternal age at time of birth (year), parity (primiparous, multiparous), and gestational length (weeks) were obtained from the MBRN. Questions allowing only for an answer of ‘yes’ had non-answers recoded to ‘no’ or ‘none’ from MoBa and MBRN [
24].
Statistical analyses
The study sample was divided into three groups: symptom-free (SF), nausea alone (NP), and both nausea and vomiting (NVP). Furthermore, we categorised the women according to whether they were with or without all forms of PGP during pregnancy. Two variables were dichotomised according to the median value: nausea duration (< 8 or ≥ 8 weeks) and vomiting duration (< 7 or ≥ 7 weeks). Only women reporting either NP, NVP, or PGP both in the simple ‘yes/no’ questions as well as the more detailed definitions used for these conditions were included in the analyses regarding the reported week of onset.
Descriptive results are presented as means (standard deviations; SDs) or frequencies (%). Comparisons between women with or without PGP, stratified by condition (SF, NP and NVP), were performed by two sample t-test and chi-squared test. PGP and sPGP at 4-6 months post-partum were explored using logistic regression analysis. We present crude and adjusted odds ratios (cOR and aOR, respectively), with 95% confidence intervals (CI). We adjusted for maternal age, BMI, smoking during pregnancy, parity, education, age at menarche, and incidences of irritability before menses after reviewing previous literature regarding NP, NVP, and PGP. Additional adjustment for length of gestation, years from menarche to pregnancy, and use of hormonal contraceptives prior to pregnancy did not change the results, therefore these results are not reported.
A significance level of 0.05 was used except when we studied statistical interaction effects, where 0.01 was used to allow for multiple testing. All analyses were performed using SPSS 22.0 (IBM Corp, Armonk, NY, USA). We followed the guidelines in the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement (
http://www.strobe-statement.org).