Volume 199, Issue 5 , Pages 443-444, November 2008
Preeclampsia: new approaches but the same old problems
Article Outline
Elsewhere in the Journal, Dr Buhimschi1 and her colleagues use the new “discovery science” to help understand preeclampsia. They are to be commended. The proteomics approach used and the analysis are elegant and clearly presented. Discovery science relies on the power to identify many analytes and analyze large amounts of data to search out differences present in disease without preconceived ideas. In contrast to more traditional research approaches, discovery science is not hypothesis driven but rather hypothesis generating, and this is exactly how it was used in this study.
See related article, page 551
The investigators reported that SERPINA1 and albumin fragments were present in excess in the urine of preeclampsia cases. They then demonstrated that despite increased SERPINA1 excretion, there was an excess of circulating SERPINA1 in cases. They point out that in other settings, increases in SERPINA1 and more specifically the misfolded SERPINA1 they identified are associated with physiological dysfunctions relevant to preeclampsia including liver and vascular disease. Finally and very importantly, they present preliminary data that women destined to develop preeclampsia have an increase in the markers prior to evident disease. Not only does this suggest SERPINA1 may be a predictor, but it is reassuring that SERPINA1 is not simply a consequence of preeclampsia.
Rather than using the diagnostic features of preeclampsia, gestational hypertension, and proteinuria as an endpoint, they determine cases by outcome, women with preeclampsia with indicated deliveries. This selection approach is commendable. The markers identified are linked to cases of preeclampsia that are important because they relate to maternal and neonatal outcome. The authors describe the power of the proteomic technique to resolve difficult diagnoses.
This may be a bit overstated. For example, perhaps the most difficult diagnosis is differentiating chronic hypertension that is worsening from hypertension with the superimposition of preeclampsia. I am not convinced this approach is a practical or precise way of making this differential. However, in the midst of the discussion of the diagnostic potential of the technique is a very insightful comment. The authors rightly point out that markers such as the ones they identify should be advantageous over purely clinical diagnoses to classify preeclampsia.
This concept is based on the observation that whereas most of the individuals with severe outcomes had the markers, the markers were also present in some subjects with mild preeclampsia. Rather than considering this a failure of the markers, they suggest that this may be evidence of different forms of preeclampsia. This concept is worth exploring.
The advent of powerful analytical strategies such as the proteomic analysis used in this study has the capacity to vastly increase our understanding of preeclampsia. Unfortunately, we are victims of the availability of million-dollar analysis (both cost and value) and a “fifty-cent diagnosis.” The choice of gestational hypertension and proteinuria to identify a pregnancy disorder that could be rapidly progressive and affect both mother and baby was the result of historical circumstance and not driven by receiver operator characteristic curves.2
In the late 19th century, hypertension and proteinuria were the first 2 clinically recognizable abnormalities identified during and prior to what had been considered a pregnancy-specific seizure disorder, eclampsia. Once identified as predicting a risk for seizures, it was soon recognized that even without seizures, these findings identified an important adverse condition of pregnancy. Nonetheless, this combination is neither sensitive nor specific at predicting adverse outcome.3
It is likely that these nonspecific endpoints are defining several different disorders. These disorders might share only the diagnostic findings and few if any other pathophysiological features. As we begin to use sophisticated analyses to understand preeclampsia, we must seek to identify a more homogeneous group of subjects than can be identified by gestational hypertension and proteinuria.
The importance of this is not merely theoretical. Being able to identify subsets of women with the preeclampsia diagnosis would enormously increase our ability to understand and perhaps even manage the disorder (think type 1 vs type 2 diabetes).
One approach to accomplish this goal is to do as Dr Buhimschi and her coauthors and define disease by important outcomes. This outcome-based definition of cases does not guarantee that only 1 disease is identified, but it does indicate a form of preeclampsia that “matters.” However, with this approach the presence of the markers in individuals with mild diseases is a “failure.” As the authors of the proteomics study indicate, it is also possible that the marker is identifying a subtype of preeclampsia that is likely more morbid than other forms but that also exists as a spectrum from less to more severe.
Is there another, better approach to identify a homogeneous group and subtypes of preeclampsia? There may be. Certainly it would be reasonable to be guided by the differences we already know about potential subsets based on epidemiological findings. It is evident based on this knowledge that preeclampsia that occurs only in first pregnancies has very different implications for later life cardiovascular disease, compared with preeclampsia that occurs in more than 1 pregnancy in which these outcomes are far more common.4, 5 A similar argument could be made for preeclampsia occurring before and after 37 weeks' gestation. Growth restriction is not present in excess in preeclamptic pregnancies after 37 weeks, and the earlier in pregnancy preeclampsia occurs, the more likely the disorder is to be recurrent and predictive of cardiovascular disease later in life.6, 7
We should also be able to exploit laboratory differences. Years ago Leon Chesley8 encouraged the inclusion of elevated uric acid in the diagnosis of preeclampsia. We find that preeclampsia with elevated uric acid is associated with more indicated preterm births and growth-restricted babies.9 Perhaps high and low SERPINA1 identifies different forms of preeclampsia. It would be useful to examine these subsets pathophysiologically, looking for differences in angiogenic factors, inflammation, oxidative stress, and endothelial dysfunction.
This elegant presentation represents our entry into a new era for understanding preeclampsia. A pragmatic strategy of relating new markers to important adverse outcomes is appropriate. However, we should also use strategies including discovery science to identify potential subsets of the disorder to guide future rational therapy.
References
- Proteomic profiling of urine identifies speciic fragments of SERPINA1 and albumin as biomarkers of preeclampsia. Am J Obstet Gynecol. 2008;199:551;e1-e16
- . Hypertension in pregnancy. J Reprod Med. 1979;23:97–151
- . Prediction of adverse outcomes by common definitions of hypertension in pregnancy. Obstet Gynecol. 2001;97:261–267
- . The remote prognosis of eclamptic women: sixth periodic report. Am J Obstet Gynecol. 1976;124:446–459
- Long-term mortality after preeclampsia. Epidemiology. 2005;16:206–215
- . Long term mortality of mothers and fathers after pre-eclampsia: population based cohort study. Br Med J. 2001;323:1213–1217
- . Association of preeclampsia with high birth weight for gestational age (Reply). Am J Obstet Gynecol. 2002;186:1106–1107
- . Diagnosis of preeclampsia. Obstet Gynecol. 1985;65:423–425
- Uric acid is as important as proteinuria in identifying fetal risk in women with gestational hypertension. [see comment] Hypertension. 2005;46:1263–1269
PII: S0002-9378(08)00789-8
doi:10.1016/j.ajog.2008.07.005
© 2008 Mosby, Inc. All rights reserved.
Refers to article:
-
Proteomic profiling of urine identifies specific fragments of SERPINA1 and albumin as biomarkers of preeclampsia
Volume 199, Issue 5 , Pages 443-444, November 2008
