Factors associated with severe maternal, fetuses and neonates’ outcomes in a university hospital in Rio de Janeiro State

Abstract Objectives: to estimate the burden of parturients, fetuses and neonate’s severe morbidity and mortality and investigate the association between maternal and their conceptus outcomes. Methods: retrospective cohort of 546 parturients and their conceptus in a university hospital, reference for high-risk pregnancy, in the metropolitan region II of Rio de Janeiro State from 2015 to 2017. We classified parturients according to obstetric morbidity (OM) in direct, indirect, or mixed, and their outcomes as: 1) no severity, 2) severe complication (SC), 3) critical intervention/Intensive Care Unit, and 4) greater severity –maternal near-miss (MNM) or death. We evaluated the conceptus as neonatal near-miss (NNM) and fetal and neonatal deaths. We estimated morbimortality indicators and associated factors (multinomial logistic regression). Results: OM was frequent: 29.3% indirect, 22.3% direct, and 15.8% mixed. There were eight cases of NMM, seven with direct MO. Among the conceptus: 7.5% were NNM cases and 4.4%, deaths. The risk of severe maternal outcomes was 16.8 and neonatal, 102.6/1000 live births. Mixed race, inadequate prenatal care, CG and NMM/death, were associated with NNM. Inadequate prenatal care and maternal NM/death were associated with conceptus deaths. Conclusion: even in a reference unit, sociodemographic, and health care inequalities negatively affect mothers and, consequently, their children.


Introduction
Between healthy pregnancy and maternal death, there is a spectrum of morbid conditions, from mild to extremely severe. 1 Maternal death represents the tip of the iceberg of maternal dysfunctions resulting from the pregnancypuerperal cycle: for each maternal death there is a larger number of women who, despite present severe clinical conditions, do not progress to death. 1 The World Health Organization (WHO) has proposed a classification of life-threatening conditions that may result in maternal survival -maternal near miss (MNM) -or death at the extreme end of the severity spectrum.Before constituting a threatening condition, severe maternal complications (of direct obstetric character), critical interventions, or intensive care unit admission should be monitored. 1he most current systematic review on MNM described a worldwide prevalence, between 2012 and 2018, of 18.6 cases of MNM per 1,000 live births (LB), with great variation between continents: 3.1 per 1,000 LB in Europe to 31.8 in Africa.Latin America had an intermediate value of 11.5 per 1,000 LB. 2 A systematic review on MNM in Brazil from 2005 to 2016 identified high frequency and causes similar to those of maternal mortality: hypertensive diseases and hemorrhages. 3The ratio of MNM varied among studies, highlighted those of national basis, with values between 9 and 10 cases per 1000 NL.Inequities, delay in health care, and the protective effect of adequate prenatal care were observed. 4ater studies show that MNM holds steady in the country.6][7] A time series, with national data from Sistema de Internações Hospitalares do Sistema Único de Saúde (SIH-SUS) (Public Health System Hospital Admissions System) showed an increase in MNM in the country between 2010 and 2018, in addition to a higher frequency in the North and Northeast regions. 7][11][12][13][14][15][16][17] In low-and middle-income countries in Africa, Asia, and Latin America, higher rates of fetal, perinatal, and NNM death were observed in MNM cases. 10A metaanalysis in high-income countries reported that women with severe maternal morbidities have a higher chance of perinatal death. 13In both low-and high-income countries, unfavorable perinatal outcomes were more associated with bleeding complications. 11,13][16][17] The present study estimated the burden of severe morbidity and mortality in parturient women and their conceptus and investigated the association between severe maternal outcomes and conceptus outcomes, in a public university hospital in metropolitan region II in Rio de Janeiro State, from 2015 to 2017.

Methods
This was a retrospective cohort study of women admitted for childbirth and their conceptus at a maternity in a public university hospital, a reference for the metropolitan region II in Rio de Janeiro State.This is the only hospital in the region qualified at tertiary level for high-risk pregnancy (HRP). 18All chldbirths that occurred from September 2015 to December 2017 were eligible, and multiple pregnancy deliveries were excluded from the analysis because they presented different risks of negative perinatal outcomes, especially for the second twin. 19The follow-up period was from hospitalization for labor (mother) and birth (conceptus) until hospital discharge or death.The sample size was defined based on the study by Abalos et al., 20 which used the following criteria to define the collection time: two months for locations with 6000 or more annual births, three months for locations between 6,000 and 3,000 annual births, and four months for locations with fewer than 3,000 annual births.In the years 2015, 2016, and 2017, 479, 402, and 402 childbirths occurred in HUAP, respectively, reaching a study population equal to 546 childbirths (mother and conceptus pairs).
Data on maternal and conceptus' medical history from the medical record, sociodemographic characteristics, reproductive history, pregnancy, childbirth and newborn history were from the Live Birth Declaration (LBD) or on the medical record (absence of DN for fetal death) were collected by the trained researchers.In cases of fetal, neonatal and maternal death, the causes of death were obtained from the Death Declaration (DD).
The maternal variables analyzed were: sociodemographic characteristics (age range: <20; 20-34 and ≥35 years; race/ color: white, black, mixed, other and schooling: <8 and ≥8 years of schooling, has a partner: yes/no and paid occupation: yes/no), reproductive (no.of children had: 0, 1-3, ≥4), parity: primiparous/multiparous), morbidity pattern (direct and indirect obstetric morbidity) adequate prenatal care (yes/no)and type of childbirth (vaginal/cesarean).The adequacy of prenatal care considered the completion of the month of initiation, and the proportion of prenatal visits performed in relation to those recommended by gestational age 21 (adequate -adequate sum and more than adequate/ inadequate -the others).
The pattern of maternal morbidity was defined according to Chou et al. 22 These authors proposed to analyze maternal morbidity aligned with maternal mortality, classifying the condition as direct, indirect or combined obstetric.
Additionally, categories on maternal severity were created, adapting those proposed by WHO, according to Serruya et al. 23 Thus, maternal outcomes were classified in increasing order of severity: no severity (NS), severe maternal complication (SMC), critical intervention/ICU (IC/ICU), maternal near miss (MNM) and maternal death (MD). 1 When a woman developed more than one severe outcome, she was classified by the worst outcome.
Considering the low frequency of the maternal outcome according to the categories proposed by WHO, 2 the last three categories (IC/ICU, MNM and MD) represented outcomes of greater severity, for analysis, purposes were aggregated into a single category and contrasted with the category of lower risk (NS).
The fetuses and newborns' outcomes were ordered in increasing severity into: no serious outcome (NS), near miss (NM) and death (neonatal -N or fetal -F).
The validated definition of neonatal near miss was adopted (presence of at least one of three pragmatic criteria: 5th minute of Apgar score <7, birth weight <1500g and gestational age <32 weeks). 24he basic cause of death was assessed according to chapter and specific cause of the Classificação Estatística Internacional de Doenças e Problemas Relacionados à Saúde 10 a Revisão -CID 10 (WHO, 1995). 25(International Statistical Classification of Diseases and Health-Related Problems 10th Revision).The presence of congenital anomaly in the MD and DD of neonatal and fetal's deaths was analyzed.Absolute frequencies of specific causes of deaths were calculated.
Absolute and relative frequencies of the study variables were described.Information incompleteness was assessed (excellent, <5; good, 5 to 10; regular, 10 to 20; poor, 20 to 50; and very poor, ≥50%). 26rom the classification of the parturients according to the presence of obstetric morbidity, the absolute and relative percentage frequencies of occurrence of severe maternal and conceptus outcomes were calculated.
The following hospital indicators were calculated: Maternal Mortality Ratio, Maternal Near Miss Ratio, Severe Maternal Outcome Ratio, Ratio between Near Miss and Maternal Death, Maternal Mortality Rate, Neonatal Mortality Rate, Neonatal Near Miss Rate, Severe Neonatal Outcome Rate, Ratio between Near Miss and Neonatal Mortality Rates, Conceptus Mortality Index, Fetal Mortality Rate, and Conceptus Death Rate.The ratios of severe outcomes, were also calculated according to maternal characteristics.
Multinomial logistic regression models were used for severe maternal (main exposure: no morbidity/SMC/ major severity) and conceptus (uncomplicated/ neonatal near miss /fetal and neonatal death) outcomes.For the choice of potentially confounding variables, in addition to literature,excellent completeness of information was considered. 26he selection of variables in the simple and multiple model, incorporating besides each covariate separately, the main exposure (severe maternal outcome, always present) followed the relevance in the outcome of natural history and/or the level of statistical significance less than 20%.In the multiple analysis with the variables selected previously, the level of statistical significance was less than 5%.To define the final multiple model, confounding variables were considered as those that caused changes greater than 10% in the magnitude of the odds ratio (OR) of the main exposure (severe maternal outcome).
Stata SE software (version 12) was used.This study was approved by the Ethics Committee of the Faculdade de Medicina da Universidade Federal Fluminense, opinion no.1826053.

Results
Among the 546 parturients, 65% were from Niterói, 21.8% from São Gonçalo and 13.2% from other cities of the Metropolitan Region I and II.Women were predominantly 20-34 years old, mixed color, with 8 or more years of schooling, unmarried, unpaid and multiparous (Table 1).About 50% had an adequate prenatal care, and 60.4% had a cesarean delivery.Except for the variables, maternal schooling, paid occupation, and having a partner, the incompleteness of the information was excellent (<5%) (Table 1).
Figure 1 presents the classification of parturients regarding to obstetric morbidities and the conditional probabilities of serious maternal and perinatal outcomes.
Indirect obstetric morbidities were most frequent followed by direct (22.3%) and the combination of direct and indirect (15.8%).Among the direct ones, hypertensive syndromes of pregnancy predominated, followed by gestational diabetes (GD) and urinary tract infection (UTI).There were few cases of hemorrhage.In the presence of two or more direct morbidities, often the gestational hypertensive syndrome was one of the morbidities.As for the indirect ones, hypertension was the most frequent, and in the case of direct and indirect association, the superposition of gestational hypertensive syndrome to pre-existing hypertension was common.All 178 women without morbidity evolved well.The occurrence of severe maternal complications was higher among women with direct and indirect obstetric morbidity (44.2%) and of more severe maternal outcomes among women with direct obstetric morbidity (9.8%).There were no cases of severe maternal complications among parturients with indirect obstetric morbidity.Four women required ICU admission, three for gestational hypertensive syndrome and one without obstetric morbidity, had Hodgkin's lymphoma, and passed away.Eleven women required hemotransfusion, including one case of NMM.
Eight cases of MNM were identified, seven with direct obstetric morbidity (five cases of gestational hypertensive syndrome) and one with chronic kidney disease.There was a predominance of clinical criteria (three cases of coagulopathy, one case of jaundice associated with gestational hypertensive syndrome and one of respiratory alterations).One maternal death (indirect cause) occurred during the study period, she had lymphoma.
Among the conceptus, 481 (88.1%) progressed without complications, 41 (7.5%) were classified as neonatal near miss, and there were 14 neonatal deaths (2.6%) and 10 fetal deaths (1.8%).Most conceptus of parturients without maternal morbidity and maternal outcome were without complications (87.6%).The occurrences of neonatal near miss and death were equal among the conceptus of mothers with higher severity outcome and indirect obstetric morbidity, and higher than the other categories (except NMN mixed obstetric morbidity and higher severity).
For every thousand live births there are about 15 cases of maternal near miss and 17 severe maternal outcomes (Table 2).Mortality among severe maternal outcomes was 11.1%.The rates of severe neonatal outcomes and conceptus mortality were 119 and 44 per 1,000 births, respectively.About three cases of neonatal near miss occurred for every neonatal death.
There was a higher risk of severe maternal complication (SMC) in women aged 20 to 34 years, white, with higher education, and primiparous, and for the outcome of greater severity (OGS), in those aged 35 years and older, black, with low education, and multiparous.Absence of paid Probabilities of maternal and newborn outcomes according to parturients' morbidities, reference university maternity hospital for high-risk pregnancy in metropolitan region II of RJ, 2015-2017.occupation represented a higher risk for both maternal outcomes (Table 3).Women with inadequate prenatal care had higher rates of SMC, but not of OGS.Women who underwent cesarean section had higher rates of both outcomes (Table 3).
The highest rates of neonatal near miss occurred among conceptus of women 35 years or older, mixed color, unpaid, inadequate prenatal care, and cesarean section.The highest rates of death were among children of women 20 to 34 years old, white, primiparous, unpaid, inadequate prenatal care, and vaginal delivery (Table 3).
The parturient that evolved in maternal death, had unspecified Hodgkin's disease (code C81.9-ICD 10R) and other diseases and complicated disorders due to the pregnancy, childbirth and puerperium (code O99.8-ICD 10R), and this last cause, after coding and applying selection rules, was constant as underlying cause in the SIM.
As for the conceptus' deaths (n=28), three neonatal deaths were not found in the SIM, making it impossible to analyze them.Congenital malformations (Chapter XVII) predominated among neonatal deaths and did not occur as a basic cause among fetal deaths.All causes of fetal deaths were classified as Some Conditions Originating in the Perinatal Period (chapter XVI).As for specific causes, the most frequent underlying cause among conceptus was related to hypertensive syndromes of pregnancy (n=3), followed by underlying causes related to infectious and parasitic diseases of the mother and unspecified sepsis of the newborn.
In the simple regression model, only age group was not associated with the fetal outcomes.In the presence of severe maternal outcomes (main exposure), all covariates showed association with the outcome of the conceptus (p<0.2):mixed color and inadequate prenatal care for both conceptus outcomes and age ≥35 years, only death and more severe maternal complications, only for neonatal near miss (Table 4).In the multiple model, only age group did not remain associated with conceptus outcome (p≥0.05).Finally, mixed color mother, inadequate prenatal care, severe maternal complications, and severe maternal outcomes were associated with neonatal near miss; inadequate prenatal care and severe maternal outcomes were associated with conceptus death (Table 4).

Discussion
In the cohort of parturients and their conceptus at the public university hospital in metropolitan region II in Rio de Janeiro State from 2015 to 2017, women with obstetric morbidities evolved more frequently to severe maternal outcomes that negatively impacted their conceptus' health.
Indirect obstetric morbidity was more frequent, an expected result, considering the referenced care of the university hospital, with high-risk prenatal care and several clinical specialties.Arterial hypertension stood out, as already shown for Latin America. 23he worst maternal outcomes (IC/ICU, MNM and MD) were more frequent among women with direct or combined obstetric morbidity.Pregnancy hypertensive syndromes were the most frequent morbidities in all severity categories.This result is in agreement with the national literature, [3][4][5]7 and is in line with that found for maternal mortality in Brazil. 27In other low-and middle-income countries, bleeding disorders are more frequent causes of maternal death, with the exception of Latin America, where hemorrhage and hypertension are equivalent.28 The profile of causes of maternal death has been changing, reducing deaths from direct causes and increasing deaths from indirect causes, although this change is slow in the country.27 Our assumption is that women were accompanied in the university hospital to have their pre-existing diseases well controlled, avoiding more unfavorable outcomes, while those with complications arose during the pregnancy may have been identified and/or referred late from the usual risk prenatal care to high-risk prenatal care.
The burden of severe maternal morbidity, expressed by the near miss rate of the analyzed hospital (14.9 p/1000 LB) was lower than another university hospital in Maceió, AL (54.8 p/1000 LB), 3 and higher than other places in the country where hospitals of different natures were included. 5,16,17][5] It should be emphasized that the only death that took place was a patient with cancer of unfavorable evolution, and this death cannot be attributed only to the poor quality of obstetric care.Therefore, these two indicators, which depend on the number of deaths, were impaired.
The high neonatal near miss rate found in the study, different from that found in six maternity hospitals in the states of Rio de Janeiro and São Paulo (TNNM between 6.2 and 35.9/1000LB), 29 can be explained by the low frequency of births/year, as well as by being a reference unit for HRP.The severe neonatal outcomes rate was also high: 2.3 times the maximum rate found in the same study (50.3/1000LB). 29onsidering the relationship between mothers and their conceptus, severe maternal outcomes were associated with both neonatal near miss and infant deaths, while maternal complications were associated only with MNM.These results are similar to those found in the literature, but with more balance between fetal and neonatal deaths. 10,11,17In these studies, the association with hemorrhagic complications and a greater impact on fetal deaths was emphasized, a fact not observed in the present study.
It is also noteworthy that there were no fetal deaths due to congenital malformation; however, deaths due to maternal conditions occurred among women with morbidity or who developed HRP.Among women without complications, deaths were from placental/ umbilical cord disorders.Although such conditions may be triggered by specific maternal conditions, such as hypertension, no maternal conditions that would classify them differently from "uncomplicated" were observed in the medical records.Therefore, it can be inferred that the death classification was adequate.Regarding neonatal death, the highest frequency was among women without complications, and all causes are related to the specific conditions of the newborn or to the malformation.Only among the women with morbidity, there were neonatal deaths related to maternal disorders, in addition to deaths due to malformation, which also occurred in this group of women.These results explain why the study observed a higher risk of death among conceptus in women without morbidity.
It is also worth noting the adequacy of prenatal care as a confounder of the association investigated.The relationship between prenatal care and negative outcomes for both women and fetuses is well documented. 3,4,6,10,14n addition, maternal skin color was a confounder only for neonatal near miss: the chance of MNM among mixed color mothers was about four times higher than among white and black mothers, the latter without statistical significance. 29The relationship between non-white color and occurrence of maternal near miss was described in a systematic review in national studies. 4he study of maternal and perinatal near miss has become increasingly necessary for public management, since its monitoring allows the timely to organize the health system. 1 Brazil has not incorporated the use of these markers in policies aiming at maternal and child health, as it occurred in Cuba. 30However, there was the first recommendation for the incorporation of monitoring of maternal near miss criteria by health services in CIT Resolution Nº 1, March 30, 2021.
Although many public policies have been developed for the promotion of maternal and child health after the implementation of the Sistema Único de Saúde (SUS) (Public Health System), there have been advances and setbacks throughout this period.Investing in women's health during the reproductive years is relevant not only for the women themselves, but for the impact on the health and development of the next generation.Furthermore, assertive policies will impact perinatal mortality rates and birth conditions in the country. 21his study has its limitations, we highlighted first, the low number of hospitalizations that occurred during data collection, caused by partial closure of the unit and that impaired the sample size.Additionally, regarding the quality of the information, two variables showed regular completeness: woman's schooling and occupation.These are variables that could broaden the understanding of socioeconomic conditions related to maternal death.The predominance of women with eight years or more of schooling was noteworthy, although many had no paid occupation.This result, different from that observed in other studies, 4 may be the result of bias information.Or perhaps the access to high-risk treatment, even mediated by vacancy regulation systems, reveals inequalities and women with higher educational levels find it easier to reach units such as the university hospital analyzed.
As a strength of the study, we point out the use of more expanded proposals for the classification of maternal morbidity, aligned to the classification of direct and indirect maternal deaths and the concept of severity gradient, 22,23 allowing deepening the knowledge of women's health and the impacts on child's health.
The results found corroborate socioeconomic, racial, and health care inequalities that negatively affect women's lives.The incorporation of the evaluation and monitoring, at least, of the near miss criteria by the health services, according to Resolution CIT No. 1, of 03/30/2021, will help future research.It is necessary that future studies consider the severity gradient and the classification of direct and indirect maternal morbidity.In conclusion, we reinforce the need for investments in women's health beyond reproductive health policies, ensuring healthy living conditions, redistribution of income, education, female empowerment, as well as equity in the health services.

Figure 1
Figure 1 Probabilities of maternal and newborn outcomes according to parturients' morbidities, reference university maternity hospital for high-risk pregnancy in metropolitan region II of RJ, 2015-2017.

Table 1
Absolute and proportional distribution of parturients according to maternal characteristics and degree of incompleteness of information.Cohort of parturients' conceptus, university maternity hospital, Niterói, 2015-2017; (N=546).

Table 2
Indicators of burden of severe morbidity and mortality in the cohort of parturients and conceptus of a university maternity referral hospital for high-risk pregnancy, metropolitan region II of RJ, 2015-2017.

Table 3
Maternal and newborn outcome rates according to maternal sociodemographic and care characteristics, in a university maternity hospital, reference for high-risk pregnancy, metropolitan region II of RJ, 2015-2017 (N=546).

Characteristics Multiple, incorporating maternal outcome (main exposure) and each covariate separately Final multiple
per thousand live births; b per thousand births; c excluding fetal deaths.Associations between severe conceptus and maternal outcomes -cohort of parturients and conceptus, reference university maternity hospital for high risk, metropolitan region II of RJ, 2015-2017. a