Acid-Base and Point of Care Ultrasound in Severe Preeclampsia

I. Introduction / Specific Aims Preeclampsia is a major cause of maternal mortality (15-20 % in developed countries) and acute and long-term morbidity for the mother and newborn. It is a multi-system disorder characterised by abnormal vascular response to placentation, associated with increased systemic vascular resistance, enhanced platelet aggregation, activation of the coagulation system, and endothelial dysfunction. Preeclampsia is defined by a number of signs and symptoms, but is characterized by multi-organ changes that encompass the central nervous system (CNS), lungs, liver, kidneys, systemic vasculature, coagulation, and the heart .

Derived from other clinical arenas of intensive care medicine, new potential markers of disease severity in women diagnosed with preeclampsia have recently been identified. Using noninvasive point-of-care (POC) ultrasound, pulmonary interstitial edema and raised intracranial pressure were identified in 25% and 20% respectively of women showing features of severe preeclampsia. In addition, echocardiographic and laboratory studies in women diagnosed with preeclampsia indicate an incidence of diastolic dysfunction and increased left ventricular end-diastolic pressure (LVEDP) in up to 30%. These findings could have a significant impact on maternal outcome and the detailed methodology of anesthesia care for the parturient with severe preeclampsia (PreE), including the mode of anesthesia, positioning of the patient, fluid management, and hemodynamic goals.

In a recent investigation performing a comprehensive physicochemical acid/base analysis in pre-eclamptic women (AB-PreE-Trial), the presence of simultaneous hypoalbuminaemic alkalosis and hyperchloraemic acidosis was identified. In addition, investigators found indications that hypoalbuminaemic alkalosis is more marked in early versus late onset disease and is associated with mode of delivery. In other critical care arenas it has been shown that reduced albumin and acid-base abnormalities are associated with pulmonary interstitial edema, elevated intracranial pressure (ICP), and poor clinical outcomes. Venous acid-base analysis may therefore be a widely available and inexpensive marker of pulmonary interstitial edema and elevated ICP in preeclampsia, with potentially important impact on anesthesia care. Therefore, one aim is to evaluate the association between elevated base excess to changes in albumin concentration (BE(Alb)) with pulmonary interstitial edema and increased ICP (both evaluated with non-invasive ultrasound) in patients with late onset severe preeclampsia. Investigators will further evaluate for an association with urgent delivery (fetal or maternal indication).

Long-term goal will be to use findings to guide clinical decision-making by anaesthetists to identify individual patients with acute organ system failures, helping better plan anaesthetic management during Cesarean delivery. Also, the findings from a larger future study could predict maternal and fetal complications, and show differences between early- and late onset disease. This should be of significant future benefit in the South African context, where preeclampsia is a leading cause of maternal morbidity and mortality.

II. Aims

Aim 1: To explore the association between ultrasound findings and/or venous acid-base abnormalities, with urgent delivery.

It is hypothesised that patients with acid-base changes secondary to albumin, interstitial edema and/or elevated ICP will experience a greater proportion of urgent deliveries, compared to patients without these abnormalities.

Aim 2: To determine the association between pulmonary interstitial edema and increased ICP with changes in BE(Alb).

Primary Hypothesis: Based on prior studies showing the association of low albumin with increased extravascular lung water and worse cerebral edema it is hypothesised that the mean BE(Alb) will be higher in patients with pulmonary interstitial edema and increased ICP (assessed with POC ultrasound), compared to patients without interstitial edema and normal ICP. Point-of-care ultrasound is a validated method of identifying both pulmonary interstitial edema and elevated ICP.

Sample size calculations:

Aim 1: Among patients with late onset disease, prior data suggests that 30% of patients will have a BE(Alb) > 4 meq/L, with a rate of fetal distress >70% in this population, as compared to a fetal distress rate of 25% in women with a BE(Alb) < 4 meq/L. Based on the observation that 70% of women in late onset disease will have a BE(Alb) < 4 meq/L at diagnosis, and considering an alpha level of 0.05 and power of 0.8, 70 patients will be needed to reject null hypothesis of predicting delivery status with BE(alb) > 4mEq/L.

Aim 2: In addition, available data suggests that the incidence of elevated intracranial pressure and interstitial pulmonary edema diagnosed by ultrasound in severe preeclampsia is approximately 25%, and an elevated BE (Alb) is found in approximately 30% of patients with severe preeclampsia. Prior data in critically ill patients indicates a mean albumin of 2.8 g/dL (SD 0.7) in patients with normal lung water, compared a mean albumin of 2.2 g/dL (SD 0.7) in patients with increased lung water. When converted to the BE (Alb) scale (assuming a mean pH of 7.4) and with the biologically plausible assumption of a similar relationship with ICP, in order to show a statistically meaningful difference in the mean BE (Alb) level among patients with severe preeclampsia with and without interstitial edema and with and without elevated ICP, we will need to recruit a total of 80 patients, with an alpha level of 0.05 and a power of 0.9. Furthermore, on the BE (Alb) scale, the mean BE (Alb) of the low albumin group (mean albumin of 2.2 g/dL) is approximately equal to a BE (Alb) of 4 meq/L and may represent an important clinical endpoint. Thus, assuming a normal distribution, 50% of this group is expected to have a BE (Alb) > 4 meq/L. The increased albumin group (mean albumin of 2.8 g/dL) lies approximately 1 standard deviation from the decreased albumin group, and assuming a normal distribution, we would expect 18% of this group to exceed the BE (Alb) > 4 meq/L cut point. Using these two proportions, a chi-squared test with a sample ratio of 3:1, power of 0.8, and alpha level of 0.05 will require a total sample size of 84 patients (21 with the ultrasound findings and 63 without findings).