Prenatal Inflammation and Perinatal Outcomes

Overview

Preventing poor perinatal outcomes is the goal of all prenatal care, yet just who will go on to develop preeclampsia or have a growth-restricted baby is notoriously difficult to predict. A growing body of evidence suggests inflammatory markers can help predict poor outcomes, even prior to, and beyond, the current pregnancy. Our project will measure the response of one robust inflammatory marker, C-reactive protein (CRP), to a safe immune provocation recommended for all pregnant women (the seasonal influenza vaccine), and to find out whether CRP response is associated with increased risk for gestational hypertension, preeclampsia, preterm delivery, or birth weight. This work will help inform whether inflammatory markers should become part of routine prenatal care.

BACKGROUND Preventing poor perinatal outcomes is the goal of all prenatal care, yet just who will go on to develop preeclampsia or have a growth-restricted baby is notoriously difficult to predict. A growing body of evidence suggests inflammatory markers can help predict poor outcomes, even prior to, and beyond, the current pregnancy. C-reactive protein (CRP) is manufactured by the liver in response to an immune provocation, and helps activate the complement system. Because it is easily assayed via phlebotomy and the less invasive capillary blood spot collection, it is a favoured marker to study inflammation in fields as diverse as immunology, obstetrics and gynecology, primary care, anthropology, and epidemiology. Pregnancy is a pro-inflammatory state, and reproductive physiology studies have shown that CRP is higher in pregnant women than in age-matched non-pregnant women, and higher still in viable pregnancies than in spontaneous abortions or ectopic pregnancies. But a look at the ecology of CRP variation shows trends with clinical relevance. CRP is higher in obese pregnant women than in their leaner counterparts, and is positively correlated with body fat, waist-hip ratio, and HDL. While CRP does not appear to be related to insulin sensitivity in non-diabetic pregnant women, for pregnant women with type I diabetes mellitus, CRP is positively correlated with higher HgBA1C and progression of retinopathy. CRP rises with prenatal iron therapy, during labour, and during the post-partum period. It is higher for women who have had Cesarean sections than for women who delivered vaginally. And physiologic states are not the only variables associated with CRP variation. In a study of 52 women in Denver, Colorado, higher CRP was associated with higher second-trimester life stress scores and lower third-trimester social support scores. In a study of 775 women in North Carolina, higher CRP was associated with self-reported black race, qualifying for food stamps, lack of private insurance, being unmarried, previous preterm birth, and maternal weight. Does an elevated inflammatory marker also correspond to maternal or fetal morbidity? Most research suggests it does. While one study of 32 Finnish women showed no difference in CRP between who had premature prelabour rupture of membranes (PPROM) and gestational age-matched controls, and one study of pregnant teens in New York showed CRP was associated with neither gestational age at birth nor birth weight, most studies of women in the USA, Ecuador, Colombia, the Netherlands, Hungary, Turkey, India, and South Korea have shown that higher third-trimester CRP is associated with increased risk for preeclampsia, intrauterine growth restriction, and preterm delivery. It also appears that elevated inflammatory markers may be associated with long-term maternal and infant morbidity. When CRP was tested in a prospective cohort of post-partum women who underwent diabetes screening at 3 months and 12 months post-partum, those women who developed diabetes had higher CRP at 3 months. And in Spain, high maternal CRP in the first trimester was associated with greater risk of infant wheezing and respiratory tract infections at 14 months (controlling for sex, birth weight, prematurity, maternal smoking and alcohol consumption, parity, maternal pre-pregnancy BMI, maternal total IgE, daycare attendance, and breastfeeding. So while data on the prognostic value of CRP during pregnancy has been accumulating for more than a decade, this variable is not widely investigated in clinical obstetrics practice. Furthermore, most research has focused on static measures of CRP, not on a woman's dynamic response to an immune challenge during pregnancy because of the myriad ethical concerns with clinical trials involving pregnant women: it is possible that a woman's inflammatory response to a challenge is more clinically relevant than her baseline CRP level. RESEARCH OBJECTIVES The investigators propose to answer both of the above research gaps by measuring CRP response to a controlled provocation that is part of the standard of prenatal care in Canada (influenza vaccination), and to follow the results to delivery. Specifically, the investigators wish to answer the following questions: 1. What is the variation in CRP values in a cohort of pregnant women at low risk of obstetrical complications in Calgary, Alberta? 2. How does CRP respond to a standard immune provocation during pregnancy? 3. Does that response vary with age of the mother, ethnicity, gestational age, parity, BMI prior to pregnancy, and weight gain during pregnancy? 4. Does the response vary with risk of gestational hypertension, preeclampsia, preterm delivery, or birth weight? METHODS This will be a quantitative prospective cohort study. Recruitment: At the beginning of 2015 flu season, the investigators will recruit participants through the low-risk prenatal clinic serving the South Health Campus in Calgary, Alberta, where the seasonal influenza vaccine is suggested as part of routine prenatal care, and where Dr. Kidd works. When women check in for their appointments, they will be given a flyer informing them of the study. Women who wish to participate in the study will be asked to present to a dedicated clinic room after their prenatal appointment, Dr. Kidd will obtain informed consent, as well as the patient's name and Alberta Health Care Number for the purpose of data retrieval. Using the standard protocol the physician will collect a capillary whole blood spot sample on Whatman #903 filter paper following a finger stick, and assign an anonymous code, e.g. 001-0. Dr. Kidd or a clinic nurse will then administer the flu vaccine. On day 3, a research assistant will obtain a second capillary blood spot sample, at the participant's home, or other convenient location. The same anonymous code will be used, e.g. 001-3. Both sets of blood spot samples, xxx-0 and xxx-3, will be sent to the Laboratory for Human Biology Research at Northwestern University, for analysis under the direction of Dr. McDade; lab personnel will not have access to participant information. Eight months later (when all participants' pregnancies will have completed), prenatal records with completed delivery information will be obtained from the electronic medical record in the clinic. Prenatal records will provide independent variables (ethnicity, gestational age, parity, BMI prior to pregnancy, and weight gain during pregnancy) and dependent variables (systolic blood pressure, diastolic blood pressure, gestational age at delivery, method of delivery, and birth weight). Any proteinuria results will be obtained through the provincial EMR, Netcare. Once datasets are complete, participants' information will be de-identified. CRP results will be added to the dataset as an additional dependent variable. The investigators will am for a convenience sample of 100 women with pre- and post-vaccination blood spots. Laboratory analysis: The investigators will analyze blood samples for CRP (mg/L) using an improved version of a high sensitivity ELISA protocol developed specifically for dried blood spot (DBS) samples. Prior validation of assay performance indicates that the method has excellent sensitivity, precision, and reliability, and a high correlation between matched plasma and DBS samples across the entire assay range (Pearson R=0.96). For all analyses the investigators will apply standard laboratory quality control procedures, including running samples in duplicate, inspecting calibration curves for fit and day-to-day variation, and measuring quality control samples with each assay. DATA MANAGEMENT AND ANALYSIS All data will be kept confidential and stored in a secure location in the Department of Family Medicine at the University of Calgary. Electronic data will be password-protected. Only Dr. Kidd will have access to participant medical records, and only the research team will have access to the data. The distribution of CRP values is likely to be skewed at baseline and day 3, and values will therefore be log transformed (base 10) to normalize the distribution. However, descriptive analyses and statistical tests will be implemented using untransformed as well as transformed results to confirm similar patterns of results. Non-parametric tests of differences in medians will be implemented using Wilcoxon matched-pairs signed-ranks and rank sum tests, and Spearman rank-order correlation will be used to assess strength of association. Log-transformed results will be subjected to paired t tests, Pearson correlation, and ordinary least squares regression analyses. Descriptive statistics will be use to characterize baseline concentrations of CRP, and the CRP response to vaccination, in relation to prior research. Bivariate analyses will investigate the association between CRP/CRP response and the following independent variables: maternal age, ethnicity, BMI prior to pregnancy, weight gain in pregnancy, gestational age, and parity. Bivariate analyses will also investigate the association between CRP/CRP response and pregnancy complications and birth outcomes, including: systolic blood pressure, diastolic blood pressure, proteinuria, gestational age at delivery, and birth weight. Multivariate regression analyses will investigate the associations between maternal factors and CRP/CRP response, independent of potential confounding influences. Similarly, regression analyses will investigate CRP/CRP response as independent predictors of birth outcomes. Limitations of the study include the relatively small sample size, although prior research in this area has drawn conclusions from smaller samples. An additional limitation is the potential for loss to follow-up at two junctures: women may not be available for a research assistant to obtain a DBS on day 3, and some women's prenatal care will be transferred to obstetricians when complications arise and therefore their complete records will not be available to the research team. If it seems that too much attrition is happening before day 3, the investigators will begin asking women to collect their own DBS with disposable lancets the investigators will provide. The investigators will not be able to control for the women who are transferred to obstetrics, though Dr. Kidd's experience at the clinic is that this happens for less than 20 per cent of women initially considered to be at low-risk of complication. ETHICAL CONSIDERATIONS The researchers have no conflicts of interest to disclose. Health risks to participants are limited to minimal risk associated with routine vaccinations, for example pain, redness or swelling at the injection site or low-grade fever. More serious risks include allergic reaction to the vaccine (dyspnea, hives, hypotension), and Guillain-Barré syndrome, a rare ascending paralysis that affects approximately 1 in 1 million flu vaccine recipients. Risks of maternal and fetal complication from influenza are thought to far exceed risks from potential flu vaccine reaction. Risks of blood spot collection are limited to discomfort associated with a finger prick. Participants' information will be de-identified and participants will not be identified by name in any presentation or publication of study findings. RELEVANCE This work has important public health implications as it may help us better predict and prevent poor perinatal outcomes. For example, women with high CRP values, or high CRP responses, might benefit from closer fetal surveillance, or might need to go on medical leave earlier in their pregnancy - few guidelines are currently available to help with this clinical decision. Going forward, as new research continues to show that baseline pregnancy levels of CRP are correlated with pre-pregnancy levels, awareness of inflammatory markers in prenatal care may help us identify pre-pregnancy conditions that may lead to poor perinatal outcomes, such as social inequity and adversity. Perhaps CRP might one day be used to monitor the "readiness" of a woman contemplating pregnancy, much in the way that pre-pregnancy BMI can help predict risk of complications. DISSEMINATION OF FINDINGS Selected findings will be presented at appropriate venues including conferences, academic seminars and publications in the biomedical sciences and anthropology. Our study will take place over 9 months, beginning in October, 2015. TEAM Dr. Monica Kidd is a family physician and low-risk obstetrics provider in Calgary, Alberta. She has a background in evolutionary biology and her research interests include child and maternal health. Dr. Thomas McDade is a biological anthropologist specializing in human biology. His research focuses on the causes and consequences of variation in inflammation in community-based settings. He is the director of the Laboratory for Human Biology Research, which has promoted the use of dried blood spot sampling as a minimally-invasive alternative to venipuncture.