Print version ISSN 0021-7557
J. Pediatr. (Rio J.) vol.80 no.4 Porto Alegre July/Aug. 2004
Multicentric networks and the quality of neonatal care
Fernando C. BarrosI; José Luis Diaz-RosselloII
Centro Latino Americano de Perinatologia, OPAS/OMS, Montevideu, Uruguay
IIProfessor, Department of Neonatology, Universidade de la República Oriental del Uruguay. Researcher, Centro Latino Americano de Perinatologia, OPAS/OMS, Montevideu, Uruguay
One of the major public health problems today concerns the increase in preterm births on a worldwide basis and their growing importance as a cause of infant mortality. The studies that demonstrate an increase in prevalence of preterms in the United States and Canada1,2 show that their most frequent causes are increase in obstetric interventions, increase in the number of multiple births, and improvements in the quality of gestational age determination, due to the replacement of date of last menstrual period by estimations made by fetal ultrasound early in pregnancy.
In Brazil, there seems to be an increase in preterm births too, as shown by several studies, including one conducted in Ribeirão Preto, state of São Paulo, and another one in Pelotas, state of Rio Grande do Sul. In Ribeirão Preto3 there was a significant rise in the prevalence of preterm births from 7.6% in 1978-1979 to 13.6% in 1994. The authors suggest that the larger number of cesarean sections may have contributed to this increase, although it is quite hard to rule out problems with reverse causality in this case.
In Pelotas, the prevalence of preterm births increased from 5.6% to 7.5% between 1982 and 1993.4 Currently, we are conducting a new perinatal study in this city, and the results for the first four months of 2004 indicate an important increase in preterm births, to around 18%. This increment apparently occurs in large newborns with 35 and 36 weeks of gestation and is observed both in vaginal and cesarean deliveries. This finding suggests that we should regard all forms of interventions (not only C-sections, but also induced labor) as possible causes for this problem (Barros et al.; unpublished data).
Since preterm newborns are responsible for a significant proportion of neonatal and infant morbidity and mortality in any population, the topic discussed by the Brazilian Neonatal Research Network (BNRN) in the current issue of Jornal de Pediatria the antenatal use of corticosteroids in preterm labor is of extreme importance today.5 Corticosteroid therapy is considered to be highly effective as a preventive measure, but is often underused. Recent meta-analyses have shown that the use of corticosteroids in preterm labor or prior to the elective termination of preterm pregnancy, may substantially reduce neonatal mortality and severe morbidities such as hyaline membrane disease and intraventricular hemorrhage. Therefore, it is worrying that the recent and acclaimed series on Infant Survival, published by The Lancet, has estimated the use of this intervention to be only 5% on a worldwide basis.6
In Latin America, the prevalence of antenatal corticosteroid therapy in preterm labor is not well known, since there is a paucity of population-based studies that allow its determination. A study conducted in Montevideo, Uruguay, and published in Jornal de Pediatria, showed that the use of corticosteroids in very low birth weight newborns amounted to 59.7%, 65.6% in public hospitals and 53.5% in private ones.7
In Brazil, a methodologically well-designed study carried out in public hospitals of Rio de Janeiro, including data from the early 1990s, showed that corticosteroids were used in only 4.3% of preterm labors between 28 and 34 incomplete weeks of gestation.8 More recently, between 1998 and 1999, the BNRN has published information on antenatal corticosteroid therapy for very low birth weight newborns, revealing a mean prevalence of 29%, with a range from 10 to 39%.9
In the current issue of Jornal de Pediatria, the authors of the BNRN describe the characteristics and outcome of a group of preterm newborns and their mothers, who were split into two categories: those who received antenatal corticosteroid therapy and those who did not.5 The BNRN consists of eight neonatal care units located in teaching hospitals in southeastern and southern Brazil, and the current study is a multicenter observational investigation.
The first positive finding of the BNRN study was that the use of antenatal corticosteroid therapy in preterm labor at less than 34 weeks reached 61%. As far as we know, this figure is much higher than in any other Latin American study. The authors describe maternal and newborn characteristics in detail, including morbidity and mortality. This serves a dual purpose: to identify possible maternal risk factors for nonuse of corticosteroids and to evaluate the possible effect of corticosteroids on the health of preterm newborns. With regard to the latter purpose, it should be noted that the group who received corticosteroids did not have a decrease in the incidence of hyaline membrane disease, but had an increase in necrotizing enterocolitis, which is inconsistent with the results reported in the literature. The inclusion criteria may have contributed to this fact. Interestingly enough, the prevalence of maternal infection and/or prolonged rupture of membranes, diabetes, and hypertension was higher in the group that received corticosteroids.
The comparison of preterm newborns who received corticosteroids with those who did not may yield results that are difficult to interpret, since these groups are different in other aspects than the use or not of corticosteroids. Mothers who did not receive corticosteroids might have been those who were already in labor at admission and therefore too late to receive preventive management (use of tocolytics in less than 2%) or those who did not meet the inclusion criteria.
Multivariate logistic regression analysis was used to determine which of the variables were associated with outcomes such as mechanical ventilation, positive blood culture and survival. The aim was to check whether the protective effect of the use of corticosteroids against mortality and the use of mechanical ventilation, and its possible effect of facilitating infections were still the same after adjusting for differences between mothers and infants who received and did not receive corticosteroids, including important variables such as birth weight, gestational age and perinatal conditions in the delivery room. In this study, the analysis confirmed that the effect is maintained, which is in agreement with the literature.
However, these results should be taken with caution, due to the methodological problems inherent to this process. Some difficulties lay in some of the variables included in the model, which clearly were intermediate variables of the analyzed outcomes. This group includes the use of oxygen at 36 weeks, SNAPPE II and mechanical ventilation (for the other two outcomes). Moreover, other variables included in the model have a very particular relationship with the use or not of corticosteroids, not being fully investigated as they should have been. These variables encompass the use of tocolytics, maternal hypertension and type of delivery. On top of that, in multivariate analyses there is always the possibility of not including in the model important variables that could have changed the association between exposure and outcome. This residual confounding is a problem in observational studies and may be minimized by repeated observations and careful measurements. Therefore, the results of the multivariate analyses should take into account these limitations and should be regarded as interesting findings to be explored in the future using a more appropriate methodology.
Multicenter studies have been very much in vogue and can be found in the most renowned medical publications all around the world. They often are randomized clinical trials (RCTs) recruiting a large number of patients in a short period of time for studies that require large samples. Also, because these studies are conducted in several places, they can increase the external validity, allowing the results to be generalized to a larger group of patients.
But another great advantage of multicenter studies is the use of standardized variables and case-notes, which allow the effectiveness of preventive or therapeutic measures previously tested in RCTs to be assessed through observational studies, providing results on the local populations. This is the design of the current study and there are many other examples of neonatal networks, as the US-based Vermont-Oxford Network and the Neonatal Network of the National Institute of Child Health and Human Development (NICHD). It is our hope that the BNRN can be expanded by including new centers from other Brazilian regions, and actively contribute to the promotion of more appropriate practices and to the improvement of the quality of neonatal care in Brazil.
1. Joseph KS, Kramer MS, Marcoux S, Ohlsson A, Wen SW, Alexander A, et al. Determinants of preterm birth rates in Canada from 1981 through 1983 and from 1992 through 1994. N Eng J Med. 1998;339:1434-9.
2. Demissie K, Rhoads GG, Ananth CV, Alexander GR, Kramer MS, Kogan MD, et al. Trends in preterm birth and neonatal mortality among blacks and whites in the United States from 1989 to 1997. Am J Epidemiol. 2001;154:307-15.
3. Bettiol H, Rona RJ, Chinn S, Goldani M, Barbieri MA. Factors associated with preterm births in Southeast Brazil: a comparison of two birth cohorts born 15 years apart. Paediat Perinat Epidemiol. 2000;14:30-8.
4. Horta BL, Barros FC, Halpern R, Victora CG. Baixo peso ao nascer em duas coortes de base populacional no sul do Brasil. Cad Saúde Pública. 1996;12 Supl 1:27-31.
5. Rede Brasileira de Pesquisas Neonatais. Uso antenatal de corticosteróide e evolução clínica de recém-nascidos pré-termos. J Pediatr (Rio J). 2004;80:277-84.
6. Jones G, Steketee RW, Black RE, Bhutta ZA, Morris SS. How many child deaths con we prevent this year? Lancet. 2003;362:65-71.
7. Matijasevich A, Barros FC, Forteza C, Diaz-Rossello JL. Atenção à saúde de crianças de muito baixo peso ao nascer de Montevidéu, Uruguai: comparação entre os setores públicos e privado. J Pediatr (Rio J). 2001;77:313-20.
8. Krauss Silva L, Pinheiro da Costa T, Reis AF, Iamada NO, Azevedo AP, Albuquerque CP. Avaliação da qualidade da assistência hospitalar obstétrica: uso de corticóides no trabalho de parto prematuro. Cad Saúde Publica. 1999;15:817-29.
9. Leone CR, Sadeck LSR, Vaz FC, Almeida MFB, Draque CM, Guinsburg R, et al. Brazilian Neonatal Research Network (BNRN): Very Low Birth Weight (VLBW) infant morbidity and mortality. Pediatr Res. 2001;49:405A.
Mas as redes multicêntricas apresentam uma outra grande vantagem, que é a utilização de registros padronizados de variáveis, que permitem a realização de estudos de observação que avaliam a efetividade de técnicas preventivas ou terapêuticas já testadas em RCTs e apresentam resultados para as populações locais. É deste tipo o presente trabalho, e são muitos os exemplos de redes neonatais internacionais, como as norte-americanas Vermont-Oxford e a Rede Neonatal do Instituto de Saúde Infantil do Instituto Nacional de Saúde (NICHD). Esperamos que a RBPN possa ser ampliada, incorporar novos centros de outras regiões do país e colaborar ativamente na promoção das práticas mais adequadas e nas melhorias da qualidade da atenção neonatal no Brasil.