Prediktivni model za nastanak bronhopulmonalne displazije kod novorođenčadi porođajne mase ispod 1500 grama

Abstract

Uvod: Bronhopulmonalna displazija (BPD) je najčešća i najteža respiratorna posledica prematuriteta. Utvrđivanje najznačajnijih faktora rizika za nastanak BPD kod novorođenčadi porođajne mase (PM) ispod 1500g može omogućiti procenu rizika za nastanak bolesti i identifikaciju novorođenčadi u visokom riziku, što je važno za pružanje informacija roditeljima o prognozi, planiranje preventivnih i terapijskih mera i stratifikovanje novorođenčadi koja su u riziku za sprovođenje budućih istraživanja. Cilj: Utvrđivanje incidencije, stepena težine BPD, smrtnosti, identifikacija najznačajnijih prenatalnih i postnatalnih faktora rizika za nastanak BPD, konstrukcije modela predikcije za nastanak BPD. Materijal i metode: Istraživanje je sprovedeno na 504 prevremeno rođene novorođenčadi PM<1500g koja su rođena u porodilištima u AP Vojvodini i lečena u tercijarnom Centru za neonatologiju i intenzivnu negu i terapiju, na Institutu za zdravstvenu zaštitu dece i omladine Vojvodine u periodu od 2006.-2011. godine. Retrospektivno je analizirano prisustvo BPD, prema stepenima težine, smrtnost. Podaci su izdvojeni iz istorija bolesti za svako novorođenče, 30 potencijalnih prenatalnih i postnatalnih faktora je opisano deskriptivnom i univarijantnom statistikom. Statstički najznačajniji faktori su uneti u multifaktorsku logističku regresionu analizu u cilju konstrukcije prediktivnih modela za nastanak BPD u 1.,14. i 21. danu neonatalnog života. Podaci su obrađeni u StatSoft-ovom programskom paketu Statistica 10.0. Validacija modela predikcije je sprovedena u prospektivnom delu istraživanja, na 100 prevremeno rođene novorođenčadi<1500g, u periodu od 2012-2013. godine. Rezultati: U retrospektivnom delu istraživanja, od 504 novorođenčeta PM<1500 grama, umrlo je 17.65%, BPD je imalo 45.43% (blagu BPD 19.44%, srednje tešku 19.84%, tešku 6.15%), srednje tešku i tešku 25.99%.Antenatalna primena kortikosteroida je zastupljena u 47.02%, surfaktant je primenjen kod 69.78% novorođenčadi. Najznačajniji prenatalni prediktivni faktor rizika za nastanak BPD/smrtnog ishoda je horioamnionitis (OR 5.72; 95% CI 3.42-9.62), dok su protektivni faktori: prenatalna primene kortikosteroida (OR 0.41; 95%CI 0.29-0.60), porođaj carskim rezom (OR 0.24; 95% CI 0.16-0.36). Najznačajniji postnatalni prediktivni faktori rizika su: GS (p≈0.00), PM (p≈0.00), reanimacija u porođajnoj sali (OR 7.01; 95% CI 4.12-12.01), rana neonatalna sepsa (OR 7.35; 95%CI 3.79-14.58), RDS (p≈0.00), primena surfaktanta (OR13,3;95%CI 8,2 - 21,67), DAP (OR 4.12; 95%CI 2.47-6.89), dok je ženski pol protektivan (OR 0.61; 95% CI 0.42-0.89). FiO2 i IPPV su u svim posmatranim danima značajni faktori rizika. Primena IPPV u 1. danu (OR 10.71; 95% CI 6.67-17.26); u ostalim danima rizik od BPD raste prema rastućoj invazivnosti respiratorne potpore. Konstruisani su modeli predikcije za 1, 14 i 21. dan života, modeli imaju visoku prediktivnu vrednost: ukupan procenat uspešnosti modela je 84.26%-90.80%, modeli sa nešto većim uspehom predviđaju prisustvo (85.36%-94.12%), nego odusustvo BPD (81.72-86.56%). OR modela je 28.07-103.04. Modeli su uspešno validirani na 102 pacijenta sa ukupnim procentom uspešnosti (82-90%), PPV (0.86-0.94) i NPV (0.76-0.87). Zaključak: Korišćenjem prenatalnih i postnatalnih kliničkih podataka moguće je predvideti nastanak BPD ili smrtnog ishoda.

Introduction: Bronchopulmonary dysplasia (BPD) is the most common serious pulmonary morbidity in very low birth weight (VLBW) infants. It is of clinical importance to determine clinical variables that are associated with BPD in order to identify infants who are at risk of developing BPD; it contributes to BPD prevention, may enable prognostic information for parents and future studies design. Objective: The aim of this study was to determine the incidence and severity of BPD, mortality rate in VLBW infants, to identify prenatal and postnatal predictive risk factors for bronchopulmonary dysplasia and competing outcome of death and to develop predictive models. Materials and Methods: Study was conducted in 504 VLBW infants born in the maternity hospitals in Vojvodina and admitted to tertiary Center for newborn and neonatal intensive care at the Institute for Child and Youth Health Care of Vojvodina, from January 2006. to December 2011. Data were retrospectively collected from clinical records for outcomes BPD or death; prenatal and postnatal factors associated with BPD were collected at three postnatal ages and examined by descriptive and univariate statistics; factors that were significantly associated with BPD and/or death were entered into a multivariate logistic regression analysis for develop predictive models. Data were analyzed using StatSoft's software package Statistica 10.0. Validation of the models were conducted in a prospective study in 102 VLBW infants born from January 2012. to December 2013. Results: There were 504 very low birth weight infants who were eligible for this study, 17.65% died, 45.43% developed BPD (mild BPD 19.44%, moderate 19.84%, severe 6.15%), moderate and severe 25.99%. The mean birth weight for the cohort was 1125.6±280.9g, the mean gestation age was GS 28,78±3,01, 49.21% were male. Surfactant received 69.78%, antenatal steroids 47.02% newborns. Key risk factors for BPD and/or death were: chorioamnionitis and maternal infections at delivery (OR 5.72; 95% CI 3.42-9.62); protective prenatal factors were: antenatal corticosteroid therapy (OR 0.41; 95%CI 0.29-0.60), cesarean delivery (OR 0.24; 95% CI 0.16-0.36). Postnatal rick factors were: GS (p≈0.00), birth weight (p≈0.00), delivery room resuscitation (OR 7.01; 95% CI 4.12-12.01), early neonatal sepsis (OR 7.35; 95%CI 3.79-14.58), RDS (p≈0.00), surfactant (OR13,3;95%CI 8,2 - 21,67), DAP (OR4.12; 95% CI 2.47-6.89), while female gender was protective (OR 0.61; 95% CI 0.42-0.89). At each time point studied, FiO2 was significantly higher in BPD/death, as well as respiratory support; on the first day invasive respiratory support was significantly associated with BPD/death (IPPV and HFOV) (OR 10.71; 95% CI 6.67-17.26), in other days BPD was associated with increasing invasiveness of respiratory support. In multifactorial logistic regression analysis separately predictive models were developed at three postnatal ages, at 1st, 14th and 21st day. Models had high predictive performance: total success of the models were 84.26% - 90.80%, models successfully predicted the presence of BPD in 85.36% -94.12%, absence of the BPD in 81.72 - 86.56% cases. OR of models were 28.07-103.04. The models were successfully validated on 102 patients with a total percentage of success 82 - 90%, with PPV 0.86-0.94 and NPV 0.76-0.87. Conclusion: Using prenatal and postnatal clinical data it is possible to predict the development of BPD and/or death in very low birth weight infants. It is very important to identify risk factors for BPD development in order to decrease the incidence of BPD and mortality rate.