Evaluation of the fullPIERS model and PLGF as predictors of adverse outcomes in women with hypertensive disorders of pregnancy

• Main Author: Escouto, Daniele Crist?v?o
• Other Authors: Figueiredo, Carlos Eduardo Poli de
• Format: Others
• Language: English
• Published: Pontif?cia Universidade Cat?lica do Rio Grande do Sul 2018
• Subjects: Pregnancy-Induced Hypertension; Prognosis; Biomarker; Angiogenic Proteins; Hipertens?o-Induzida-Pela-Gravidez; Progn?stico; Biomarcador; Indutoresda-Angiog?nese; CIENCIAS DA SAUDE::MEDICINA
• Online Access: http://tede2.pucrs.br/tede2/handle/tede/8270

Submitted by PPG Medicina e Ci?ncias da Sa?de ([e-mail]) on 2018-05-04T19:26:50Z No. of bitstreams: 1 DANIELE_CRISTOV?O_ESCOUTO.pdf: 3899595 bytes, checksum: 60af701fccf65168e901b103c704e2fe (MD5) === Rejected by Sheila Dias ([e-mail]), reason: Devolvido devido ao t?tulo da capa (em ingl?s) estar diferente do t?tulo da folha de rosto e ficha catalogr?fica (em portugu?s). on 2018-05-16T19:01:31Z (GMT) === Submitted by PPG Medicina e Ci?ncias da Sa?de ([e-mail]) on 2018-05-17T11:39:29Z No. of bitstreams: 1 DANIELE_CRISTOV?O_ESCOUTO.pdf: 3899595 bytes, checksum: 60af701fccf65168e901b103c704e2fe (MD5) === Rejected by Caroline Xavier ([e-mail]), reason: Devolvido devido ao t?tulo da capa (em ingl?s) estar diferente do t?tulo da folha de rosto e ficha catalogr?fica (em portugu?s). on 2018-05-28T18:04:10Z (GMT) === Submitted by PPG Medicina e Ci?ncias da Sa?de ([e-mail]) on 2018-09-03T19:01:37Z No. of bitstreams: 1 DANIELE_CRIST?V?O_ESCOUTO.pdf: 2805745 bytes, checksum: c76c0d3c115aae92a7015f9312fd9604 (MD5) === Approved for entry into archive by Sheila Dias ([e-mail]) on 2018-09-05T13:05:03Z (GMT) No. of bitstreams: 1 DANIELE_CRIST?V?O_ESCOUTO.pdf: 2805745 bytes, checksum: c76c0d3c115aae92a7015f9312fd9604 (MD5) === Made available in DSpace on 2018-09-05T13:40:27Z (GMT). No. of bitstreams: 1 DANIELE_CRIST?V?O_ESCOUTO.pdf: 2805745 bytes, checksum: c76c0d3c115aae92a7015f9312fd9604 (MD5) Previous issue date: 2018-03-19 === Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior - CAPES === Introdu??o ? A distin??o adequada dos casos de alto risco para eventos graves nas doen?as hipertensivas gestacionais, n?o apenas pr?-ecl?mpsia, ? um desafio cl?nico. O modelo fullPIERS ? uma ferramenta simples e de baixo custo que utiliza vari?veis cl?nicas para estratificar a probabilidade de eventos adversos em gestantes com pr?-ecl?mpsia. O fator de crescimento placent?rio (PlGF) ? um biomarcador com concentra??es reduzidas no plasma de mulheres com pr?-ecl?mpsia e com crescente emprego na avalia??o de gestantes com suspeita de pr?-ecl?mpsia. Objetivos ? O objetivo deste estudo ? estimar a acur?cia do modelo fullPIERS e do biomarcador PlGF como preditores de desfechos adversos maternos em gestantes com doen?a hipertensiva gestacional. M?todos ? Estudo de coorte prospectiva em um hospital terci?rio em Porto Alegre, Brasil, que incluiu gestantes admitidas com press?o arterial sist?lica ? 140 e/ou press?o arterial diast?lica ? 90 mmHg a partir da 20? semana de gesta??o. Os piores valores de vari?veis cl?nicas e laboratoriais dentro das primeiras 48 horas de admiss?o foram coletados. Desenvolvimento de eventos adversos foi acompanhado por um per?odo de 14 dias. Concentra??es plasm?ticas maternas de PlGF do momento da admiss?o foram mensuradas. Resultados ? 405 gestantes foram inclu?das no estudo. Entre as 351 mulheres inclu?das na an?lise do modelo fullPIERS, 20 (5%) desenvolveram pelo menos um evento adverso materno dentro de 14 dias de interna??o. O modelo fullPIERS teve pouca capacidade discriminativa para prever desfechos em 48 horas [AUC 0,639 (95% CI 0,458-0,819)]. A acur?cia do modelo foi ainda mais baixa dentro de sete semanas da admiss?o [AUC 0,612 (95% CI 0,440-0,783)]; a capacidade discriminativa manteve-se similar dentro de 14 dias da admiss?o [AUC 0,637 (95% CI 0,491-0,783)]. A calibra??o do modelo fullPIERS tamb?m foi ruim: inclina??o 0,35 (95% CI 0,08-0,62) e intercepto 1,13 (95%CI -2,4-0,14). A an?lise do PlGF incluiu 392 gestantes. PlGF <5? percentil esteve associados a eventos adversos maternos dentro de 48 horas em gestantes inclu?das antes de 35 semanas com sensibilidade de 0,80 (0,4-0,96), valor preditivo negativo (VPN) de 0,98 (0,9-0,99) e AUC ROC de 0,672 (IC 95% 0,5-0,9). PlGF <100 pg/mL apresentaram sensibilidade de 0,8 (0,4-0,96), especificidade de 0,6 (0,5-0,7) e VPN de 0,99 (0,94-0,99) em mulheres ap?s 37 semanas de gravidez. PlGF apresentou bom desempenho para prever parto at? 14 dias em gestantes inclu?das antes de 35 semanas. PlGF <5? percentil esteve associado a rec?m-nascido pequeno para idade gestacional (PIG) com sensibilidade de 0,75 (0,6-0,9), especificidade 0,65 (0,5-0,7), NPV de 0,87 (0,79-0,94) e AUC ROC 0,698 (0,6-0,79), em gestantes com <35 semanas, a acur?cia diminuiu com o aumento das idades gestacionais. Conclus?es ? O modelo fullPIERS e a concentra??o de PLGF mostraram baixa acur?cia na predi??o de desfechos adversos maternos em mulheres com doen?a hipertensiva gestacional, incluindo pr?-ecl?mpsia. O modelo fullPIERS teve desempenho inferior na nossa amostra quando comparado com o estudo que validou este teste. O PLGF parece ser um biomarcador para uso como ferramenta adicional na predi??o de parto dentro de 14 dias e rec?m-nascidos PIG, especialmente em gestantes antes da 35? semana gestacional. === Introduction - Singling out high-risk patients from the diverse hypertensive disorders of pregnancy, and not only preeclampsia, is a challenge for clinicians. The fullPIERS model is a simple and low-cost evaluation instrument using clinical variables to stratify the adverse outcomes probability of pregnant women with high-risk preeclampsia. Placental growth factor (PlGF) levels are reduced in preeclampsia and are increasingly being used as a biomarker in the assessment of this disease. Objectives - The aim of the study is to evaluate the performance of the fullPIERS model and PlGF to predict adverse outcomes in women with hypertensive disorders of pregnancy. Methods - A prospective cohort study carried out at a teaching hospital in Porto Alegre, Brazil enrolling pregnant women admitted with a systolic blood pressure ? 140 mmHg and/or a diastolic blood pressure ? 90 mmHg from the 20th week of gestation. First 48 hours of admission worst clinical and laboratory data were recorded and the development of adverse maternal and perinatal outcomes scrutinised up to 14 days. Admission maternal plasma PlGF concentrations were measured. Results ? A total of 405 women were enrolled. From the 351 women included in the fullPIERS model analysis, 20 (5%) developed at least one of the combined maternal adverse outcomes. The fullPIERS model had poor outcomes discrimination at 48h [AUC 0.639 (95% CI 0.458-0.819)]. At the seventh admission day, the model?s accuracy was even lower [AUC 0.612 (95% CI 0.440-0.783)]; the model?s discriminative ability remained similar [AUC 0.637 (95% CI 0.491-0.783)] at 14 days. Calibration of the fullPIERS model was poor: slope - 0.35 (95% CI 0.08-0.62), intercept -1.13 (95%CI -2.4-0.14). PlGF analysis included 392 women. PlGF < 5th percentile predicted maternal adverse outcomes within 48h in women with gestation < 35 weeks with sensitivity of 0.80, NPV of 0.98 and AUC ROC of 0.672 (CI 95%0.5-0.9). The threshold of <100 pg/mL, had best accuracy in women after 37 weeks of pregnancy, sensitivity of 0.8, specificity of 0.6, negative predictive value of 0.99 and PPV of 0.04. PlGF had good performance to predict delivery within 14 days in women presenting before 35 weeks. PlGF <5th percentile predicted delivery of a SGA infant with sensitivity of 0.75, specificity 0.65, PPV of 0.45, NPV of 0.87, and AUC ROC 0.698, in women with gestation < 35 weeks, accuracy decreased at later gestational ages. Conclusion - In conclusion, in our sample the fullPIERS model and PlGF were limited predictors of maternal adverse outcomes in pregnant women with hypertensive disorders of pregnancy, including preeclampsia. The performance of the fullPIERS model in our sample was inferior to that of the original cohort. PlGF as a biomarker appears to be an additional tool to predict delivery within 14 days and SGA newborn in women before 35 weeks gestation.