maternal morbidity in these women (post hoc analysis). 10 The study highlights the need to determine which antihypertensive agent(s) provides optimal control of chronic hypertension in pregnancy to ameliorate these risks.Choice of antihypertensive outside pregnancy depends on ethnicity with those of African/Caribbean family origin receiving calcium channel blockers as first-line agent 11 and is thought to relate to differences in the pathophysiology causing hypertension in those of differing ethnic backgrounds.
12Ethnic disparity in maternal and perinatal outcome in the general pregnant population is well described and likely to be multifactorial. 13 To our knowledge, no randomized controlled trials have investigated the impact of ethnicity on efficacy of antihypertensive treatment in pregnancy. The aims of the PANDA study (Pregnancy and Chronic Hypertension: Nifedipine Versus Labetalol as Antihypertensive Treatment) were 3-fold: to assess feasibility of such a randomized controlled trial, to evaluate mechanistic treatment effects, and to examine the impact of ethnicity on efficacy of nifedipine (a calcium channel blocker with a well-established safety profile in pregnancy) with labetalol (currently recommended as firstline by national UK guidance).
MethodsThe study was an open-label, phase 4, randomized controlled clinical trial (EudraCT Number 2013-003144-23 +6 weeks (to allow for second trimester BP nadir), singleton pregnancies, aged >18 years, and the ability to provide written informed consent. Women were excluded if they had a contraindication (relative or absolute) to either antihypertensive agent, such as labetalol in women with asthma. Details of the randomization process, intervention, and outcome measures are contained in the online-only Data Supplement.
Statistical AnalysisFor the primary analysis, the intention to treat principle was applied; women were analyzed in the groups into which they were randomly allocated regardless of allocation received. The statistical software Stata/SE version 14 for Windows was used for all analyses. The number and percentage were calculated for binary and categorical variables. The mean and SD or the median and interquartile range were calculated for continuous variables. Linear regression with robust SE was used for the primary and other continuous outcomes. Adjustment was made for baseline covariates, including ethnicity (black [determined by self-report of whether the woman had a parent or grandparent who was African or Caribbean] versus non-black [all other ethnicities]), gestational age at randomization, and center. For continuous measures, an adjustment was also made for corresponding baseline measurement (systolic BP at randomization for the primary clinical outcome). For binary outcomes, binary regression with a log link was used to calculate risk ratios (RR). Analysis of the primary clinical outcomes was repeated excluding women delivering their baby before 24 completed weeks of pregnancy because women who deliver before viability did not complete the inten...