Assessment of blood-brain barrier integrity and neuroinflammation in preeclampsia

Background: Although blood-brain barrier integrity is intact under normal pregnancy conditions, animal studies suggest that blood-brain barrier impairment occurs in preeclampsia. Yet, human data are limited, and the integrity of the blood-brain barrier has not been assessed in women with preeclampsia. Objective: We sought to test the hypothesis that the integrity of the blood-brain barrier is impaired and that neuroinflammation is increased in women with preeclampsia. Study Design: We performed an observational case-control study in pregnant women >24 weeks gestation who underwent spinal anesthesia for elective cesarean delivery or combined spinal epidural analgesia for labor. Cases were women with preeclampsia, and control subjects were women with either healthy pregnancy, chronic hypertension, or gestational hypertension. Paired samples of blood, urine, and cerebrospinal fluid were collected from each subject before delivery. We measured albumin, C5a, C5b-9, tumor necrosis factor-α, and interleukin-6 concentrations in plasma and cerebrospinal fluid, and albumin, C5a, and C5b-9 concentrations in urine, using colorimetric or enzyme-linked immunosorbent assays. The ratio of albumin in cerebrospinal fluid to plasma (Q_alb) was used as a surrogate for maternal blood-brain barrier integrity. Cerebrospinal fluid concentrations of C5a, C5b-9, tumor necrosis factor-α, and interleukin-6 were used as surrogate markers of neuroinflammation. Differences in Q_alb and cerebrospinal fluid protein concentrations between groups were assessed by nonparametric test of medians. Results: Forty-eight subjects were enrolled, which included 16 cases with preeclampsia, 16 control subjects with healthy pregnancy, and 16 control subjects with either chronic or gestational hypertension. Q_alb values were not increased in preeclampsia cases compared with healthy or hypertensive control subjects (Q_alb median, 3.5 [interquartile range, 2.9–5.1] vs 3.9 [interquartile range, 3.0–4.8] vs 3.9 [interquartile range, 3.0–4.8]; P=.78]. Moreover, Q_alb values were not increased in the subset of women with preeclampsia with severe features (n=8) compared with those without severe features (n=8; Q_alb median, 3.5 [interquartile range, 3.3–4.9] vs 3.7 [interquartile range, 2.3–5.5]; P=.62]. Cerebrospinal fluid concentrations of C5a, C5b-9, tumor necrosis factor-α and interleukin-6 were not increased in cases of preeclampsia, compared with control subjects with either healthy pregnancy, chronic hypertension, or gestational hypertension (P>.05, all comparisons). In contrast to the negative findings in cerebrospinal fluid, plasma concentrations of both C5b-9 and interleukin-6 and urine concentrations of C5a and C5b-9 were increased in cases of preeclampsia. Conclusion: Through measurements of albumin, complement proteins, and cytokines in paired samples of blood and cerebrospinal fluid at the time of delivery, we found no evidence of blood-brain barrier impairment or neuroinflammation in preeclampsia. Larger studies that will investigate a wider range of proteins are suggested to validate our findings.