Background
There is no tool to accurately predict who is at risk of developing neurologic complications of preeclampsia, and there is no objective method to determine disease severity.
Objective
We assessed whether plasma concentrations of the cerebral biomarkers neurofilament light, tau, and glial fibrillary acidic protein could reflect disease severity in several phenotypes of preeclampsia. Furthermore, we compared the cerebral biomarkers with the angiogenic biomarkers soluble fms-like tyrosine kinase 1, placental growth factor, and soluble endoglin.
Study Design
In this observational study, we included women from the South African Preeclampsia Obstetric Adverse Events biobank. Plasma samples taken at diagnosis (preeclampsia cases) or admission for delivery (normotensive controls) were analyzed for concentrations of neurofilament light, tau, glial fibrillary acidic protein, placental growth factor, soluble fms-like tyrosine kinase 1, and soluble endoglin. The cerebrospinal fluid concentrations of inflammatory markers and albumin were analyzed in a subgroup of 15 women. Analyses were adjusted for gestational age, time from seizures and delivery to sampling, maternal age, and parity.
Results
Compared with 28 women with normotensive pregnancies, 146 women with preeclampsia demonstrated 2.18-fold higher plasma concentrations of neurofilament light (95% confidence interval, 1.64–2.88), 2.17-fold higher tau (95% confidence interval, 1.49–3.16), and 2.77-fold higher glial fibrillary acidic protein (95% confidence interval, 2.06–3.72). Overall, 72 women with neurologic complications (eclampsia, cortical blindness, and stroke) demonstrated increased plasma concentrations of tau (2.99-fold higher; 95% confidence interval, 1.92–4.65) and glial fibrillary acidic protein (3.22-fold higher; 95% confidence interval, 2.06–5.02) compared with women with preeclampsia without pulmonary edema; hemolysis, elevated liver enzymes, and low platelet count; or neurologic complications (n=31). Moreover, angiogenic markers were higher, but to a lesser extent. Women with hemolysis, elevated liver enzymes, and low platelet count (n=20) demonstrated increased plasma concentrations of neurofilament light (1.64-fold higher; 95% confidence interval, 1.06–2.55), tau (4.44-fold higher; 95% confidence interval, 1.85–10.66), and glial fibrillary acidic protein (1.82-fold higher; 95% confidence interval, 1.32–2.50) compared with women with preeclampsia without pulmonary edema; hemolysis, elevated liver enzymes, and low platelet count; or neurologic complications. There was no difference shown in the angiogenic biomarkers. There was no difference between 23 women with preeclampsia complicated by pulmonary edema and women with preeclampsia without pulmonary edema; hemolysis, elevated liver enzymes, and low platelet count; or neurologic complications for any of the biomarkers. Plasma concentrations of tau and glial fibrillary acidic protein were increased in women with several neurologic complications compared with women with eclampsia only.
Conclusion
Plasma neurofilament light, glial fibrillary acidic, and tau were candidate biomarkers for the diagnosis and possibly prediction of cerebral complications of preeclampsia.
Introduction
Eclampsia and other cerebral complications of preeclampsia, including cerebral edema, ischemia, and hemorrhage, are leading causes of maternal morbidity and mortality. Preeclampsia and, in particular, eclampsia are associated with long-term maternal neurologic outcomes, including an increased risk of white matter lesions, stroke, seizure disorders, and vascular dementia later in life. Despite this, there is no available tool to accurately predict who is at risk of developing these complications, and there is no objective method to determine disease severity apart from imaging and symptom assessment. Symptoms thought to predict eclampsia, such as visual disturbances and severe headache, have poor predictive accuracy with a sensitivity of 35% and specificity of 94% for visual disturbances and a sensitivity of 56% and specificity of 83% for headache. Furthermore, there is no objective way of determining disease severity and who is at risk of developing long-term neurologic sequelae.
Why was this study conducted?
Neurologic complications are important contributors to preeclampsia-associated maternal mortality and morbidity. Current diagnostic tools and predictors demonstrate poor sensitivity and specificity. Easily accessible blood biomarkers to identify women at high risk may be useful tools for the prediction of neurologic complications.
Key findings
Cerebral biomarkers have the potential to detect neurologic complications in preeclampsia.
What does this add to what is known?
Plasma concentrations of neurofilament light and tau are increased in preeclampsia but have never been evaluated in women who develop neurologic complications of preeclampsia. Plasma glial fibrillary acidic protein has not been evaluated in preeclampsia. These data contributed to the field by demonstrating the potential usefulness of these biomarkers in neurologic complications in preeclampsia.
The angiogenic biomarkers soluble fms-like tyrosine kinase-1 (sFlt-1), placental growth factor (PlGF), their ratio (sFlt-1–to–PlGF), and soluble endoglin (sEng) can predict the development of preeclampsia. Revealed PlGF testing can reduce the time to diagnosis of preeclampsia in suspected preeclampsia and reduce maternal adverse outcomes. Similarly, the sFlt-1–to–PlGF ratio has been shown to increase the proportion of women diagnosed with preeclampsia within 7 days of sampling. Case-control studies of angiogenic biomarkers concerning cerebral complications demonstrate that PlGF and sFlt-1 are altered in plasma in severe disease, including eclampsia. ,
Cerebral biomarkers are proteins that are highly specific to the central nervous system and can be measured in cerebrospinal fluid (CSF) or blood (plasma or serum). Circulating neurofilament light (NfL), tau, and glial fibrillary acidic protein (GFAP) are useful diagnostic and predictive biomarkers for several neurologic disorders, including Alzheimer disease and traumatic brain injury. NfL and tau increase when there is axonal injury, and GFAP increases when there is glial cell involvement. NfL and tau are increased in women before a diagnosis of preeclampsia and after, but their ability to diagnose or predict cerebral complications in preeclampsia has not been established. GFAP has not been studied in preeclampsia.
First, we assessed whether cerebral biomarkers NfL, GFAP, and tau are increased in maternal plasma in several subtypes of preeclampsia, particularly those with significant neurologic complications, such as eclampsia, stroke, or cortical blindness. Second, we assessed how cerebral biomarkers performed in comparison with the established preeclampsia biomarkers PlGF, sFlt-1, and sEng. Lastly, we correlated concentrations of circulating biomarkers to blood-brain barrier disruption and neuroinflammatory markers in CSF.
Materials and Methods
Study cohort
Women with singleton pregnancies recruited to the Preeclampsia Obstetric Adverse Events (PROVE) biobank at Tygerberg Hospital, Cape Town, South Africa, were included. Tygerberg Hospital is the largest referral hospital in the Western Cape Provence of South Africa and delivers more than 8000 high-risk pregnancies yearly and manages many women with complications of preeclampsia. The exclusion criteria were women with known neurologic or cardiac disease. For normotensive women, additional exclusion criteria included chronic hypertension and diabetes mellitus. Preeclampsia was defined according to the recent American College of Obstetricians and Gynecologists (ACOG) Practice Bulletin, but significant proteinuria was also required to diagnose preeclampsia (protein-to-creatinine ratio of ≥30 mg/mmol [0.3 mg/mg] or ≥0.3 g protein in a 24-hour urine collection or urine dipstick of >1+ in more than 1 occasion). Pulmonary edema was diagnosed when there was worsening dyspnea, fine bibasal inspiratory crackles on auscultation, and features of pulmonary edema on chest x-ray. Hemolysis, elevated liver enzymes, and low platelet count (HELLP syndrome) was defined as a platelet count of <100 × 109/L, aspartate aminotransferase of >70 U/L, and hemolysis as demonstrated by lactate dehydrogenase > 600 U/L or hemolysis on a peripheral blood smear. Eclampsia was diagnosed if generalized tonic-clonic seizures occurred in a woman diagnosed with preeclampsia in the absence of another etiology. Renal impairment was defined as a serum creatinine of >120 umol/L, which is higher than the ACOG definition. Women were followed up from recruitment to discharge. Severe hypertension was defined as a systolic blood pressure of ≥160 mm Hg and/or a diastolic blood pressure of ≥110 mm Hg.
At inclusion, women with preeclampsia were divided into 4 groups. These groups were preeclampsia with neurologic complications, preeclampsia with HELLP syndrome, preeclampsia complicated by pulmonary edema, and preeclampsia without pulmonary edema, HELLP syndrome, or neurologic complications. All women could have severe hypertension. We used a hierarchical system where women with neurologic complications could also suffer from HELLP syndrome and pulmonary edema at inclusion, women with HELLP syndrome could suffer from pulmonary edema but no neurologic complication, and women with pulmonary edema could not have HELLP syndrome or neurologic complications. After inclusion, any later complications were recorded, but the women remained in their initial groups. If women with preeclampsia without pulmonary edema, HELLP syndrome, or neurologic complications subsequently developed any complications, they were excluded from the study. No pregnant woman with a normotensive pregnancy developed hypertension. We further subdivided preeclampsia with neurologic complications into eclampsia only (1 eclamptic seizure with no other neurologic symptom) or several neurologic complications (women who had multiple seizures, stroke, a Glasgow Coma Scale [GCS] of >13, or eclampsia together with other organ complications).
Baseline data were obtained by interview and extraction from medical records. All data were entered and stored in a Research Electronic Data Capture database and double-checked for accuracy.
Sample collection
Plasma samples were collected in ethylenediaminetetraacetic acid tubes at inclusion after a diagnosis of preeclampsia or after admission for delivery (normotensive pregnancies). Women could be included before or shortly after delivery. CSF was collected in a subset at the time of spinal anesthesia at delivery. Samples were centrifuged, aliquoted, and frozen at −80°C. Samples were shipped to Melbourne, Australia, for analysis of angiogenic biomarkers and to a neurochemistry laboratory in Gothenburg, Sweden, for analyses of cerebral biomarkers and neuroinflammatory markers.
Biomarker assays
Plasma concentrations of tau, NfL, and GFAP were measured using the single-molecule array (Simoa) Neuro 4-Plex kit on an HD-X Analyzer, as described by the kit manufacturer (Quanterix, Billerica, MA). Calibrators were run in duplicates, whereas samples were run in singlicates with a 4-fold dilution. Of note, 2 quality control (QC) samples were run in duplicates at the beginning and end of each run. For GFAP, a QC sample with concentrations of 49.6 pg/mL resulted in a repeatability of 11.9% and an intermediate precision of 11.9. Intra- and interassay coefficients of variation were 4.3% and 15%.
The concentrations of sFlt-1 and PlGF were measured with a commercial electrochemiluminescence immunoassay platform (Roche Diagnostics, Basel, Switzerland). sEng was measured using Endoglin CD/105 DuoSet enzyme-linked immunosorbent assay kits (R&D Systems, Minneapolis, MN) according to the manufacturer’s instructions. The samples were run in singlicates, with a 200-fold dilution, and 2 QC samples were run with each plate.
CSF and plasma albumin concentrations were measured by immunonephelometry on a Beckman IMMAGE Immunohistochemistry System (Beckman Instruments, Beckman Coulter Inc, Brea, CA). The CSF-to-plasma albumin ratio was calculated as CSF albumin (mg/L)/serum albumin (g/L) and was used as a measure of the blood-brain barrier integrity. CSF concentrations of interleukin (IL)-1β, IL-6, IL-8, and tumor necrosis factor α (TNF-α) were measured using the Meso Scale Discovery 4-Plex Proinflammatory Panel II according to instructions from the manufacturer (Meso Scale Discovery, Rockville, MD) and have been published previously. Laboratory technicians were blinded to the groups.
Statistical methods
Demographic and clinical characteristics were presented as mean (standard deviation), median (interquartile range [IQR]), or number (percentage).
Biomarkers were presented as median (IQR) and compared among groups as fold changes with 95% confidence intervals. Unadjusted analyses were performed using the Welch analysis of variance on log-transformed outcomes, and adjusted analyses were performed using the Welch analysis of covariance on log-transformed outcomes, adjusting for age, parity, gestational age (GA) at blood sampling, and time from eclampsia to plasma sample collection. Angiogenic biomarkers were further adjusted for time from delivery to plasma sample. Furthermore, we performed a separate analysis of angiogenic biomarkers on the subgroup of women with blood samples obtained before delivery.
In statistical evaluations with <6 observations per group, analyses were performed using nonparametric permutation tests. Corresponding confidence intervals were calculated by test inversion. Comparisons were performed on both unadjusted and adjusted for GA at blood sampling.
Correlations between biomarkers in plasma and neuroinflammatory markers in CSF were analyzed using Pearson correlations between log-transformed variables. Because of small sample sizes, P values were calculated nonparametrically using exact permutation tests.
In all hypothesis tests, a 2-sided P value of <.05 was considered statistically significant. Data and statistical analyses were performed using SPSS (version 26.0; SPSS; PASW Statistics, Chicago, IL), Stata/MP (version 16.0; StataCorp, College Station, TX) for Mac software package, and SAS software (version 9.4; SAS Institute, Cary, NC).
Sample size
Previous studies comparing women with normotensive pregnancies with women with preeclampsia used a sample size of 10 in each group to detect a difference in plasma concentrations of NfL with an alpha error of 0.05 and a power of 0.8. To study the subgroups and to study preeclampsia with neurologic complications, the sample size was set to at least 20 in each group and at least 60 in the group with preeclampsia and neurologic complications.
Ethics approval and registration details
Ethics approval was obtained (protocol number N18/03/034; Federal Wide assurance number 00001372; Institutional Review Board approval number IRB0005239). All participants or their guardians signed informed consent. The biobank is registered (ISRCTN10623443) and the protocol is published.
Data availability
Anonymized data will be made available on request from any qualified investigator after approval.
Results
We included women from April 2018 to March 2020. A total of 177 plasma samples were available for analysis: 28 were normotensive pregnancies and 146 had preeclampsia. Of note, 31 women had preeclampsia without pulmonary edema, HELLP syndrome, or neurologic complications, 23 women had pulmonary edema, 20 women had HELLP syndrome, and 72 women had severe neurologic complications. We had CSF samples for 8 women with preeclampsia or eclampsia and for 7 women with normotensive pregnancies ( Figure 1 ).
Background characteristics
Maternal characteristics and pregnancy outcomes are presented in Table S1 . Women with preeclampsia were younger, more often nulliparous, more often used alcohol and/or methamphetamine, and had a higher body mass index. They delivered at an earlier gestation, had more low birthweight infants, and experienced more stillbirth than women with normotensive pregnancies ( Table S1 ). Women with preeclampsia were divided into subgroups as presented in Table 1 .
| Characteristic | Preeclampsia a | Pulmonary edema | HELLP | Neurology |
|---|---|---|---|---|
| n | 31 | 23 | 20 | 72 |
| At baseline | ||||
| Maternal age (y) | 24.9 (5.2) | 30.4 (8.0) | 28.6 (7.4) | 22.8 (6.2) |
| Nulliparous | 17 (55) | 11 (48) | 7 (35) | 50 (69) |
| HIV | 5 (16) | 5 (22) | 4 (20) | 7 (10) |
| Smoking | 2 (7) | 1 (4) | 2 (11) | 12 (17) |
| Alcohol use | 1 (3) | 0 (0) | 1 (5) | 7 (10) |
| Methamphetamine use | 0 (0) | 0 (0) | 1 (5) | 2 (3) |
| Diabetes mellitus | ||||
| Pregestational | 1 (3) | 0 (0) | 0 (0) | 0 (0) |
| Pregnancy induced | 0 (0) | 1 (5) | 0 (0) | 1 (1) |
| Chronic hypertension | 6 (19) | 1 (5) | 3 (15) | 6 (9) |
| BMI (kg/m 2 ) | 27.9 (7.8) | 32.7 (8.7) | 31.2 (5.7) | 25.7 (5.0) |
| Missing | 3 (10) | 3 (13) | 6 (30) | 14 (19) |
| After inclusion | ||||
| GA at delivery (wk) | 33.5 (4.2) | 31.6 (4.5) | 30.3 (5.4) | 33.4 (4.4) |
| Sample taken before delivery | 11 (36) | 3 (13) | 5 (25) | 33 (46) |
| Sampling in relation to delivery (d) | ||||
| Antepartum samples | −0.5 (0.7) | −0.7 (0.6) | −0.2 (0.4) | −0.2 (0.6) |
| Postpartum samples | 2.3 (1.8) | 1.7 (1.2) | 2.1 (1.3) | 2.8 (5.3) |
| Mode of delivery | ||||
| Vaginal delivery | 8 (26) | 6 (26) | 7 (35) | 23 (32) |
| Elective CD | 1 (3) | 0 (0) | 0 (0) | 1 (1) |
| Emergency CD | 22 (71) | 17 (74) | 13 (65) | 48 (67) |
| Live-born infant | 26 (84) | 19 (83) | 13(65) | 60 (83) |
| Birthweight (g) | 2014.1 (947.9) | 1747.8 (997.3) | 1366.3 (689.3) | 2090.3 (901.5) |
| Angiogenic biomarkers | ||||
| PlGF (pg/mL) | 17.5 (9.7–44.4) | 18.6 (10.6–50.2) | 9.5 (7.4–28.5) | 18.0 (10.6–28.7) |
| sFlt-1 (pg/mL) | 4617.0 (981.0–8099.0) | 1929.0 (840.8–4297.0) | 4663.5 (1170.5–12779.0) | 4415.5 (2090.8–11197.8) |
| sFlt-1–to–PlGF ratio | 117.5 (60.7–328.2) | 70.2 (46.1–119.6) | 162.6 (58.5–1636.0) | 171.0 (100.3–561.0) |
| sEng (pg/mL) | 183,516.3 (130,565.2–236,011.3) | 152,344.6 (121,787.1–31,273.7) | 287,394.2 (168,641.7–376,916.7) | 228,487.8 (146,314.4–301,254.5) |
| Maternal complications | ||||
| Maternal death | 0 (0) | 0 (0) | 0 (0) | 2 (3) |
| ICU admission | 0 (0) | 1 (4) | 2 (10) | 9 (13) |
| OCCU admission | 3 (10) | 22 (96) | 13 (65) | 52 (72) |
| Eclampsia | 0 (0) | 0 (0) | 0 (0) | 68 (94) |
| Recurrent eclampsia | 0 (0) | 0 (0) | 0 (0) | 23 (32) |
| Stroke b | 0 (0) | 0 (0) | 0 (0) | 3 (4) |
| GCS<13 | 0 (0) | 0 (0) | 0 (0) | 17 (24) |
| Cortical blindness | 0 (0) | 1 (4) | 0 (0) | 4 (6) |
| Pulmonary edema | 0 (0) | 23 (100) | 6 (30) | 2 (3) |
| Inotropic support | 0 (0) | 0 (0) | 1 (6) | 1 (1) |
| Renal impairment | 0 (0) | 3 (13) | 9 (45) | 14 (19) |
| Dialysis | 0 (0) | 1 (4) | 1 (6) | 0 (0) |
| HELLP syndrome | 0 (0) | 1 (4) | 20 (100) | 17 (24) |
| Increased INR (>1.2) | 0 (0) | 3 (13) | 3 (15) | 9 (13) |
| Severe hypertension | 8 (26) | 16 (70) | 16 (80) | 29 (40) |
| Sepsis | 0 (0) | 4 (15) | 2 (13) | 8 (11) |
| Venous thromboembolism | 0 (0) | 0 (0) | 0 (0) | 2 (3) |
| Placental abruption | 1 (3) | 1 (4) | 2 (10) | 4 (6) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
