Data sources

The Scottish Health Survey (SHS) is undertaken periodically to monitor the health and health-related risk factors of the Scottish population and has cardiovascular disease as its principal focus. The survey uses multistage, stratified random sampling to provide a representative sample of the Scottish population, and recruits different participants to each survey, enabling the results of individual surveys to be combined. In addition to face-to-face interviews and physical measurements, participants provide blood samples for biochemical assays. Serum CRP concentrations are measured using the N Latex monoimmunoassay on the Behring Nephelometer II Analyzer (CSL Behring UK Ltd, West Sussex, UK).

The Scottish Morbidity Record 2 (SMR2) is a Scotland-wide, administrative database that collects information on all pregnancies, including maternal, obstetric, and infant characteristics. Data collection began in 1961, and the most recent data quality assessment was undertaken in 2008 through 2009. It compared the database information with medical records and found that for the mandatory variables used in the current study the data matched in ≥90% of records. Specifically, induction of labor matched in 93% of records, estimated gestation in 92%, and mode of delivery in 87%.

This study is exempt from institutional review board approval, because it is a retrospective analysis of routinely collected, anonymized data.

Inclusion and exclusion criteria

We linked women who participated in the 2 most recent SHSs (1998 and 2003) to their pregnancy records. The study was restricted to women who were aged between 15-71 years when they participated in the SHS and whose first pregnancy predated the survey and resulted in a liveborn, singleton infant delivered between 24-43 weeks’ gestation. Stillbirths were excluded, as were women who were pregnant or acutely ill at the time that CRP was measured.

Definitions

The SHS provided information on a number of potential confounding factors, including smoking status, age, height, body mass index, and essential hypertension. Postcode of residence was used to derive socioeconomic quintiles using the Carstairs index. This applies information collected in the 2001 census on 4 parameters (social class, car ownership, unemployment, and overcrowding) to postcode sectors of residence (mean population 5698). The SMR2 provided information on maternal age, gestation and birthweight at delivery, preeclampsia, and previous spontaneous abortions. It also provided information on total parity prior to CRP measurement. Preterm delivery was defined as <37 weeks’ gestation, and term delivery as ≥37 weeks’. Indicated delivery was defined as induced vaginal birth or cesarean section without onset of labor. Spontaneous delivery was defined as either cesarean section following onset of labor or vaginal delivery in the absence of induction of labor. Cases of preterm delivery following premature rupture of fetal membranes were defined as spontaneous preterm delivery. Birthweight, infant sex, and gestation were used to append sex- and gestation-specific deciles derived from all deliveries in Scotland over the study period.

Statistical analysis

Categorical variables were summarized using percentages and groups compared using the χ ² test. Continuous variables were summarized using the median and interquartile range and groups compared using the Mann-Whitney U (MWU) test. The association between preterm delivery and CRP was examined using univariate and multivariate linear regression analyses. In all analyses, CRP concentration was transformed by log to normalize the distribution of residuals. The covariates included in the multivariate analyses were body mass index, smoking status, age, parity, height, and socioeconomic deprivation quintile at the time of CRP measurement, and birthweight decile, previous spontaneous abortions, essential hypertension, and preeclampsia during the index pregnancy.