SD-OCT Measurement of the Human Retina in Pregnancy With Pre-existing or De Novo Hypertension and Without Hypertension

This is a prospective observational cohort study that was undertaken at Foothills Medical Centre, University of Calgary, Calgary, Canada. Recruitment began in October 2013 and clinical follow-ups completed in May 2017. Data was collected at enrolment (\< 20 weeks gestation), at every follow-up clinical encounter with the Foothills Hospital High-risk Obstetrical clinic between 20 weeks gestation and delivery, at delivery itself, and at least 1 additional encounter in the non-pregnant state, usually postpartum. Demographic information including participant age, gestational age, medical co-morbidities, medications, clinical symptoms, weight, height, automated office blood pressure and all investigations ordered by managing physicians were recorded. In addition, macular thickness was measured at each encounter using spectral domain optical coherence tomography (SD-OCT). Obstetrical and neonatal outcomes were collected from standardized birth and delivery records. SD-OCT images were performed by a trained physician or research assistant according to the following standardized technique. The same Zeiss Cirrus 4000 SD-OCT instrument was used for all retinal images. Two SD-OCT images were obtained from each eye at each clinical encounter without dilation of the pupil at a scanning resolution of 512 A-scans x 128 B-scans over a 6 mm square grid focused on the fovea. Between sequential sessional measurements (i.e., right eye first measurement, right eye second measurement, etc.) participants were instructed to remove their face from the examining platform and the instrument was reset to default parameters in order to match the effect of an independent scan. Scans were examined systematically for signal strength, definition of the vessel architecture, centrality and motion artifact to determine image quality, and the better of the 2 images was prospectively selected for analysis. Macular thickness was measured from the internal limiting membrane to the retinal pigment epithelium, a standard method completed by the instrument on every retinal scan. All scans were referenced to the image obtained from the same individual in the non-pregnant state to determine interval change over the pregnancy. Given the exploratory nature of this study and its small subject numbers, the research team considered both standard summary statistics and measures of individual response. In respect to the latter, a decision support tool was derived before starting the study whereby 'clinically meaningful change' was arbitrarily defined as a directionally identical difference ≥ ±4 µm (the test re-test coefficient of repeatability of the instrument) in 3 or more contiguous segments on the Early Treatment of Diabetic Retinopathy Study (ETDRS) grid \[1\] in a single eye. This summary measure was subsequently tested against the frequency of appearance of all possible combinations of ETDRS segments possessing that and other levels of differential change in a null difference distribution comprised of 1370 paired images collected on all participants enrolled in SD-OCT studies managed by the PI as of May 2017. This tool identifies a statistically significant interval change compared to the expected null difference at a P-value = 0.046. Assessment of this parameter is ongoing and will be published separately. Participant characteristics were summarized by descriptive statistics using means and standard deviation (SD) for continuous variables and frequencies for categorical variables. A 95% confidence interval (95%CI) was computed for each outcome parameter. Statistical significance was defined as a P-value \<0.05. A linear mixed-effects model was used to estimate macular thickness at the 3 gestational intervals with random effects considered at three hierarchical levels: patient level, eye side (right or left) and position on the ETDRS grid. A continuous autocorrelation structure was used to adjust for correlation of variables repeatedly measured at differing clinical encounters.