B3. Reproducibility and assessment of longitudinal confounders in the measurement of maternal haemodynamics in women with chronic hypertension in pregnancy

Abstract

Introduction: To assess the reproducibility of Tensiomed Arteriograph measurements taken as part of a longitudinal cohort study of women with chronic hypertension (CHT).

Methods/Results: chartPulse wave velocity (PWV) and augmentation index (AIX) were measured three times (lying and sitting) per visit at 4–8 weekly intervals from 10 to 34 weeks gestation. Three hundred and ninety-one observations were available from 40 women. PWV and AIX were obtained in 95.4% and 94.6% of attempts. Intraclass coefficients (ICC) were calculated across six gestation windows. The median PWV ICC was 0.71 (range 0.22–0.92), improving to 0.84 (0.78–0.88) for readings with standard deviation <1.5m/s (79.6% of readings). Corresponding AIX ICCs were 0.88 (0.54–0.93) and 0.94 (0.92–0.97). For lying and sitting position measurements, PWV ICC was 0.82 (0.71–0.89) and AIX 0.81 (0.57–0.87). A 10 beat increase in HR was associated with 4.1–5.2% decrease in AIX, adjustment for HR reduced the gestational effect (Figure 1). The effects of HR and gestation on PWV were smaller, but statistically significant.

Figure 1. Mixed level regression models (with 95% CI) assessing the interaction between HR, gestation and haemodynamic measurements (fixed effects), adjusted for individual and visit number (random effects).

Discussion: PWV and AIX have acceptable reproducibility for use in clinical practice. Changes AIX attributable to HR are consistent with non-pregnant studies and interact with the gestational effect. Reproducibility was unaffected by maternal position, but unacceptable in 20% of readings and additional measures of haemodynamic status may therefore be required.