by Sofia Reynoso, Co-Managing Editor
My introduction to preeclampsia came from binge watching Downton Abbey in middle school, when (spoiler alert) Lady Sybil Crawley died during childbirth. As an avid fan of Sybil, I was shocked that the writers of the show would shatter the sacred image of pregnancy and kill off my favorite character. Unfortunately, this condition is not uncommon offscreen – preeclampsia is a serious, life-threatening condition killing approximately 50,000 people a year and is diagnosed in up to 1 in 12 pregnancies (1,2).
Preeclampsia is a pregnancy complication that develops usually after 20 weeks, which includes high blood pressure accompanied by proteinuria (protein in the urine) and/or end-organ dysfunction (i.e., damage to the brain, liver, kidneys, etc.). Once the condition develops into eclampsia, patients often experience seizures (3).
It is thought that the condition originates from the placenta, an organ responsible for supplying the fetus with oxygen and nutrients. The placenta may have a defect in its formation and/or it could abnormally attach to the uterine wall, increasing the risk for developing preeclampsia. Often this flaw is associated with the formation of blood vessels, hence the high blood pressure (1).
Surprisingly, doctors still use a similar treatment today as they did back in Sybil’s era: delivery (4). A review published in 2019 on the condition even stated that delivering the baby is still the best treatment method, although aspirin is another prevention measure used (1). Magnesium sulfate and drugs prescribed for hypertension are also used (1). Even with these treatments and others in development (5), there is only so much to do when you don’t see the diagnosis coming.
Along with a lack of universally effective treatment, physicians are unable to predict whether a woman will develop preeclampsia later in pregnancy. Invasive tests to sample DNA or RNA from the fetus have been used to begin understanding each stage in development, including abnormalities indicative of preeclampsia, but these tests run serious risks and are ethically questionable when repeated (6).
A simple blood test on the mother may do the trick (2). The key lies in a nucleic acid known as cell-free RNA. Cell-free DNA/RNA is released by cells during necrosis and apoptosis (both are forms of cell death) and retains properties of their original cell/tissue. Cell-free DNA from the fetus found in maternal blood is currently being used as a method of noninvasive genetic testing on fetuses for chromosomal abnormalities along with other disorders (7).
In a recent study, researchers honed in on cell-free RNA (cfRNA), which provides better understanding of which genes are being expressed at different stages of development (7). To achieve this goal, the study gathered the largest and most diverse dataset of maternal gene expression, also known as the transcriptome (2).
Traditionally, gestational age, or the age of the fetus, has been predicted via ultrasound. However, the researchers developed a machine learning algorithm using 1,908 cfRNA profiles as the training set that used cfRNA levels to represent gene expression levels in samples. This algorithm accurately predicted the gestational age of 474 test samples within an error of around 2 weeks. Compared to the traditional ultrasound, their technique was as good as ultrasound dating in the second trimester and better in the third trimester relative to ultrasound dating (2).
Since they established that there was enough variation in cfRNA levels throughout pregnancy to allow for gestational dating, they next tackled the separation of cfRNA into their tissues of origin and understand the current status of their original tissues. In particular, they focused on maternal, fetal, and placental cell types to understand how gene expression changes in these tissues over time. The genes they studied exhibited changes over the pregnancy, consistent with previously understood physiological changes (2).
Having established the power of cfRNA to understand the normal progression of pregnancy, they moved to preeclampsia. Researchers used blood draws from 72 preeclamptic and 452 normal patients taken on average 14.5 weeks before the delivery of the baby, unlike previous studies that gathered blood from patients once diagnosed with preeclampsia past 14.5 weeks before delivery. Therefore, many blood draws were done on asymptomatic individuals to trace back the expression patterns predictive of the condition (2).
They found seven genes of interest that were expressed differently in those that developed preeclampsia: CLDN7, PAPPA2, SNORD14A, PLEKHH1, MAGEA10, TLE6 and FABP1. Four of these have previously known functions in the condition. Using another machine learning algorithm, the researchers developed a mathematical model based on the seven genes of interest to predict whether someone would develop preeclampsia. The model had a sensitivity of 75%, meaning it could predict ¾ of true cases; it also had a positive predictive value of 32.3%, indicating around ⅓ of patients that were predicted to have preeclampsia eventually developed the disease (2).
The powerful model was also shown to predict preeclampsia regardless of maternal BMI, age or race. Given the previously demonstrated worse outcomes for women of color (8), for example, in terms of maternal mortality due to pregnancy-related complications, this algorithm could be a vital, universally actionable tool for improving the health of all pregnant individuals (2).
Since aspirin is most effective before 16 weeks of pregnancy, hopefully this model can provide accurate predictions to get pregnant people treated earlier, possibly even in their first trimester (6). It could also help us understand how and why the disease manifests and allow us to develop better targeted therapeutics.
Although this is an impressive find, the model has a long way to go. One potential limitation is that 13.7% of the individuals sampled developed preeclampsia, while the generally accepted global incidence is about 2-8% of pregnant women. Therefore, the positive predictive value would likely be lower (more false positives) in reality than in this sample. In addition, the inability to predict preeclampsia has proven dangerous, but false positives, which could become an issue with this model, can be dangerous as well through overtreatment and the stress of the diagnosis (6).
Even if there is a long way to go, this study provides reassurance for prospective parents. Preeclampsia has been affecting pregnant individuals for centuries, including women like Sybil Crawley, and to finally have a predictive understanding could prove vital to improving maternal and infant mortality.
Edited by Autumn Pereira
References:
- Phipps, E.A., Thadhani, R., Benzing, T., & Karumanchi, S.A. (2019). Pre-eclampsia: pathogenesis, novel diagnostics and therapies. Nat Rev Nephrol, 15(5), 275–289. https://doi-org/10.1038/s41581-019-0119-6
- Rasmussen, M., Reddy, M., Nolan, R., Camunas-Soler, J., Khodursky, A., Scheller, N. M., Cantonwine, D. E., Engelbrechtsen, L., Mi, J. D., Dutta, A., Brundage, T., Siddiqui, F., Thao, M., Gee, E. P. S., La, J., Baruch-Gravett, C., Santillan, M. K., Deb, S., Ame, S. M., … McElrath, T. F. (2022). RNA profiles reveal signatures of future health and disease in pregnancy. Nature, 601(7893), 422–427. https://doi.org/10.1038/s41586-021-04249-w
- Magley M, Hinson MR. Eclampsia. [Updated 2021 Jun 12]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2022 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK554392/
- Bell M. J. (2010). A historical overview of preeclampsia-eclampsia. Journal of obstetric, gynecologic, and neonatal nursing : JOGNN, 39(5), 510–518. https://doi.org/10.1111/j.1552-6909.2010.01172.x
- Eddy, A. C., Bidwell, G. L., 3rd, & George, E. M. (2018). Pro-angiogenic therapeutics for preeclampsia. Biology of sex differences, 9(1), 36. https://doi.org/10.1186/s13293-018-0195-5
- Shook, L. L., & Edlow, A. G. (2022). A blood test to predict pregnancy complications. Nature News & Views. 601. https://www.nature.com/articles/d41586-021-03801-y
- Drag, M. H., & Kilpeläinen, T. O. (2021). Cell-free DNA and RNA—measurement and applications in clinical diagnostics with focus on metabolic disorders. Physiological Genomics, 53(1), 33–46. https://doi.org/10.1152/physiolgenomics.00086.2020
- CDC (2019, September 5). Racial and Ethnic Disparities Continue in Pregnancy-Related Deaths. CDC. https://www.cdc.gov/media/releases/2019/p0905-racial-ethnic-disparities-pregnancy-deaths.html
The Abstract 