Maternal and Child Health Journal

, Volume 17, Issue 9, pp 1638–1647 | Cite as

Estimation of Preterm Birth Rate, Associated Factors and Maternal Morbidity From a Demographic and Health Survey in Brazil

  • Ricardo P. Tedesco
  • Renato PassiniJr.
  • José G. Cecatti
  • Rodrigo S. Camargo
  • Rodolfo C. Pacagnella
  • Maria H. Sousa
Article

Abstract

To determine the prevalence of preterm birth from self-reports by Brazilian women, to assess complications, interventions and outcomes, to identify factors associated with preterm birth, and to improve the preterm birth rates estimates. This is a secondary analysis of data from a Demographic Health Survey. It interviewed a sample of 4,743 Brazilian women who had 6,113 live births from 2001 to 2007. Estimates of preterm birth rates were obtained per region and per year according to self-reported gestational age. The prevalence rate and 95 % confidence interval (CI) for preterm was determined according to the characteristics of mothers and offspring. Odds ratios and 95 % CI were estimated for complications such as severe maternal morbidity. The preterm birth rate was 9.9 %, with regional variations. Preterm birth was more likely to be associated with neonatal death, low birth weight, and longer hospital stay. Maternal factors associated with preterm birth were: white ethnicity, living in an urban area, history of hypertension or heart disease, twin gestation, non-elective Cesarean section, medical insurance for delivery, low number of antenatal visits, and severe morbidity. A self-report survey has indicated that the preterm birth rate in Brazil is higher than official data suggest, with an increasing trend in more developed areas, and is associated with poor neonatal and maternal outcomes.

Keywords

Preterm birth Risk factors Prematurity Demographic and health survey Severe maternal morbidity 

Abbreviations

BMI

Body mass index

DHS

Demographic health survey

OR

Odds ratio

PR

Prevalence ratio

SINASC

Live births information system

WHO

World Health Organization

Introduction

Neonatal care has markedly improved over a number of years, resulting in decreased infant mortality rates, especially in developing countries [1]. Many countries, however, fall a long way short of the ‘Millennium Development Goals’, proposed by the United Nations [2]. An increase in preterm births in the last two decades has raised alarm among health policy-makers, including those in developed countries.

Neonates born before fetal maturity subsequently require more health care. In addition, many diseases in adulthood are directly associated with the intrauterine living conditions (Barker Theory) [3]. Thus, delivery occurring before the fetus has reached maturity denotes a potential fetal risk, which may possibly result in short or long term consequences.

The most recent and largest study on the topic reported that the incidence of preterm birth rose in the last two decades in the great majority of countries and also that Brazil contributed as the 10th largest number of preterm newborns worldwide [4]. Another study in the USA showed that, in 2006, 12.8 % of births were preterm, an increase of 21 % compared to 1990 [5]. In 2008, at the Surgeon General’s Conference on the Prevention of Preterm Birth in the United States, many experts concluded that preterm represents both social impairment and a serious public health problem, and is becoming more prevalent worldwide [6]. In fact, a recent systematic review estimated 12.9 million preterm births globally for 2005 alone, which would account for 9.6 % of all births worldwide [7].

Preterm, traditionally defined as birth occurring before 37 weeks of gestation [8], is still the leading cause of neonatal morbidity and mortality throughout the world. It is also the major cause of deficiencies acquired after birth, in particular those involving neurological abnormalities [9]. With a multifactorial origin, preterm birth can occur between 22 and 37 weeks, and becomes markedly more severe as the gestational age at birth decreases. Many of the risk factors associated with preterm birth result in inflammation, thus increasing uterus stimulation, such as in local and systemic infections. Some risk factors are associated with maternal sociodemographic characteristics, including African-American ethnicity, low socioeconomic and educational status, young and advanced maternal ages, single marital status, working long hours, and a difficult labor under stress. Other risk factors include maternal psychological stress, tobacco use, alcohol abuse and drug use, pregnancy history (close temporal proximity to a previous delivery and previous preterm birth), actual pregnancy conditions (multiple gestation, vaginal bleeding, polyhydramnios or oligohydramnios), and existing medical conditions, such as low body mass index (BMI), thyroid disease, asthma and diabetes [10].

Currently, the maternal medical condition is one of the most important risk factors, and therapeutic preterm birth is an important proportion of the total number of preterm births. Some obstetric conditions, such as hemorrhage and hypertension, are reported as risk factors for preterm birth; for instance, gestational age at birth is usually lower when associated with preeclampsia [11, 12]. Some authors have suggested that maternal morbidity accounts for almost half of preterm births [13]. Another risk factor identified by some authors is related to the mode of delivery. Although Cesarean section is intrinsically associated with complications, there is also an increased risk of unintended iatrogenic prematurity [13, 14] and of neonatal mortality and morbidity [15].

In order to reduce preterm births, it is first important to determine the prevalence, and potential risk factors associated with it. In Brazil, official data show a mean prevalence of roughly 6.5 % [16], below the rates reported by developed countries, and relatively close to the estimate for South America of 7.9 % [7]. This surprisingly low rate may reflect the difficulty of adequately determining gestational age at birth, as well as the likely under-reporting of births, including those preterm, especially in the less developed regions of the country. Indeed, differentiating the proportion of low birth weight newborns that are preterm is extremely difficult where there is no standardized estimation of gestational age. A recent systematic review on the subject found that in low and middle income countries, roughly half of all low birth weight newborns were preterm [17]. Therefore a problem still exists regarding the estimates of preterm birth in settings where the estimation of real gestational age is not standardized nor is the information easily available for great majority of women delivering.

Since 1986, a national population survey has been conducted every 10 years in Brazil to determine the demographic and health profile of the female population of childbearing age and of children under 5 years of age. Such a study can allow an assessment of how maternal and child health has developed in recent years in Brazil. Therefore, the purpose of this study was to assess the self-reported prevalence of preterm birth in a sample of Brazilian women interviewed in the Demographic and Health Survey (DHS) performed in 2006 [18]. The study also aimed at assessing the conditions associated with the health of women who had preterm deliveries, the quality of care received during pregnancy, the occurrence of complications and interventions among pregnant women and their outcomes. Characteristics of preterm and term pregnancies were compared to identify possible factors associated with preterm, and also to compare the data with the official data provided on preterm births in Brazil.

Materials and Methods

This study is a secondary analysis of the DHS data acquired in Brazil in 2006–2007. These data are currently in the public domain [18]. The original study was conducted according to the ethical standards established by the Declaration of Helsinki and approved by the National Review Board. The period of data collection was from 31st October 2006 to 3rd May 2007. The survey was representative of the national population, with assessments in five regions of the country in both rural and urban areas. A total of 15,575 women, aged 15–49 years, were interviewed. In the current analysis, we considered women who had at least one live birth since 2001. The total number of live births during this period in the five regions of the country was 6,160. Thus, data of 4,743 women were extracted from the database. For 47 live births, however, there was no information in the database on the occurrence of preterm birth. All information on twin live births was considered.

The original data of this study was obtained electronically, directly from the website of the Ministry of Health of Brazil, as the database is in the public domain. Additionally, data on the official total number of live births, including those for gestational age groups, for Brazil and its regions were also obtained via the internet. Thus, official preterm birth estimates were compared with the results from the DHS.

For data analysis, initially a descriptive bivariate analysis by region of residence of women was performed. We used the number of pregnancies occurring from 2001 and the live births according to the gestational age as reported by the mother. The prevalence of preterm births was evaluated by region of residence for the period 2001–2007, with information from two sources, the DHS and the live births information system (SINASC) from the Ministry of Health. We stratified the DHS data according to preterm, defined as less than 37 weeks gestation, or term delivery. Associations were then tested using the Pearson Chi-square test. Various factors were assessed, including sociodemographic characteristics, maternal clinical conditions, and features related to pregnancy, prenatal care, and delivery. The analysis determined crude estimates or prevalence rate (PR) with 95 % confidence interval (CI) and also estimates adjusted by other predictors with Poisson multiple regression (PRadj). Finally, the live births were assessed according to maternal complications and/or interventions, or both, with crude odds ratio (OR) and 95 % CI reported as an indirect measure of any association. The characteristics of the complex sampling method (stratum, clustering, weighting) were performed in all analyses. For the analysis, SPSS (version 17.0) and Stata (version 7.0.) softwares were used.

Results

In each of the five regions the majority (70–80 %) of women interviewed, who had at least one live birth, reported that they had only one pregnancy in the period 2001–2006. The northern and north-eastern regions exhibited the highest proportion of women (4.1/6.1 %) with three to five pregnancies in that period (Table 1).
Table 1

Percentage distribution of women with at least one live birth child from 2001 to 2007, according to the number of pregnancies by region (weighted analysis)

Number of pregnancies from 2001 to 2007

Region

Total (4,743)a

North (953)

Northeast (896)

Southeast (935)

South (968)

Midwest (991)

1

70.4

76.0

83.6

86.7

81.6

80.5

2

23.5

19.9

15.3

12.1

15.9

17.0

3–5

6.1

4.1

1.1

1.2

2.5

2.5

Total expanded

1,359,528

3,974,414

5,815,517

1,975,857

1,128,229

14,253,545

aNumber of women interviewed with at least one live birth child (without expansion)

The prevalence of preterm live births according to self-reports ranged from 6.4 % in the north-eastern region to 15.2 % in the southern region, and the difference between these two regions was statistically significant. The overall prevalence of preterm birth (with weighting) reported by women in the study period was 9.9 %, as shown in Table 2. In the figures from the different regions, there were significantly more preterm births in south Brazil in 2003 and 2004 and in southeast Brazil in 2005 compared with other regions (Table 3). In north-eastern Brazil there was a significantly lower prevalence in 2003 and 2005 compared with other regions of Brazil. Although the prevalence of preterm births tended to increase over the years overall, there were no significant differences in the self-reported prevalence of live preterm births during the years included in the survey. Table 3 presents the official data on the prevalence of preterm live births in the same period, showing virtually no change throughout the study period for each region of the country, with values ranging from 5 to 7.5 %, lower than that estimated from the self-reported data. However, no comparative analysis was performed among the data, because theoretically the DHS sample was also contained in the SINASC database, and also because data were obtained in different ways.
Table 2

Percentage distribution of live birth children from 2001 to 2007, according to self-reported gestational age at birth by region (weighted analysis)

Self-reported gestational age (months)

Region

Total (6,113)a

North (1,351)

Northeast (1,186)

Southeast (1,156)

South (1,163)

Midwest (1,257)

9

91.8

93.6

89.0

84.9

89.6

90.1

8

5.8

4.2

9.0

12.1

7.8

7.5

7

2.1

2.0

1.4

2.4

1.7

1.8

6

0.3

0.2

0.6

0.6

0.8

0.5

5

0.0

0.0

0.0

0.1

0.1

0.0

Prevalence of preterm with weighting (%)

8.2

6.4

11.0

15.1

10.4

9.9

Total expanded

1,849,146

5,094,567

6,875,189

2,248,132

1,366,957

17,433,991

All information on twin live births was included (50 twins)

aTotal number of live birth children; information missing for 47 live births

p < 0.001 (encompassing all the study variables)

Table 3

Prevalence of preterm live births according to the year of birth by region from 2001 to 2007, using two information sources

Years

Region

Total

p*

North

Northeast

Southeast

South

Midwest

Prevalence of preterm with weighting (DHS)

 2001

3.6

6.7

9.6

10.4

8.0

7.9

0.515

 2002

5.7

5.3

7.3

12.0

8.2

7.3

0.412

 2003

9.1

4.3#

8.1

24.0

13.7

9.6

<0.001

 2004

8.8

9.9

7.4

18.1#

14.1

10.2

0.132

 2005

10.1

1.8

18.8

11.0

10.6

11.7

<0.001

 2006–2007

11.3

9.7

14.7

14.9

8.8

12.4

0.498

 p**

0.239

0.074

0.098

0.143

0.592

0.205

 

(Total live births)a

(1,351)

(1,186)

(1,156)

(1,163)

(1,257)

(6,113)

 

Official data (SINASC)

 2001

6.0

5.7

7.0

6.8

6.7

6.4

 

 2002

5.4

5.5

7.0

6.9

7.2

6.4

 

 2003

5.2

5.3

7.2

7.2

6.8

6.4

 

 2004

5.1

5.6

7.4

7.3

6.5

6.5

 

 2005

5.2

5.7

7.5

7.3

6.7

6.6

 

 2006

5.1

5.7

7.7

7.6

6.3

6.7

 

 2007

4.5

5.4

7.9

7.7

6.5

6.7

 

Values in bold mean that the differences are statistically significant (p < 0.05)

* Comparison between regions (considering all the design variables)

** Comparison between years (considering all the design variables)

#p < 0.02,  p < 0.001 (comparisons of each region vs. others)

Comparison between years (total): 2001–2002 versus other years (p = 0.023); 2001–2003 versus other years (p = 0.033)

DHS demographic health survey, SINASC national information system on live births

Source: DATASUS/MS

aValid number, without expansion of live birth children; information missing on 47 live births

Women who had at least one preterm birth in the period were more frequently white (PRadj = 1.49; 95 % CI: 1.12/1.97), with urban residence (PRadj = 2.28; 95 % CI: 1.54/3.38) and with a history of hypertension or heart disease (PRadj = 1.89; 95 % CI: 1.35/2.66) (Table 4). Preterm birth was also significantly associated with twin pregnancies (PRadj = 4.85; 95 % CI: 3.32/7.08), Cesarean section (PRadj = 1.74; 95 % CI: 1.31/2.29), delivery funded by private health insurance (PRadj = 1.51; 95 % CI: 1.03/2.23) and a low number of prenatal care visits (PRadj = 1.69; 95 % CI: 1.24/2.31) (Table 5).
Table 4

Percentage distribution of women with a live birth child from 2001 to 2007, according to characteristics, grouped by preterm and term births (weighted analysis)

Characteristic

Group

PR (95 % CI)

PRadj. (95 % CI)c

Only pretermb

Only term

Age (years)

 15–19

12.8

9.3

1.46 (0.82–2.60)

 

 20–29

49.2

54.4

1.00 (ref.)

 

 30–39

30.3

29.8

1.04 (0.90–1.20)

 

 40–49

7.6

6.5

1.06 (0.92–1.23)

 

 (n)a

(388)

(4,177)

  

Race/ethnicity

 Caucasian

46.3

34.7

1.55 (1.15–2.08)

1.49 (1.12–1.97)

 Other

53.7

65.3

1.00 (ref.)

1.00 (ref.)

 (n)a

(385)

(4,133)

  

Household situation

 Urban

90.7

79.2

2.41 (1.61–3.62)

2.28 (1.54–3.38)

 Rural

9.3

20.8

1.00 (ref.)

1.00 (ref.)

 (n)a

(388)

(4,177)

  

Education level

 Until primary school

52.2

54.2

1.00 (ref.)

 

 > primary school

47.8

45.8

1.08 (0.74–1.56)

 

 (n)a

(385)

(4,094)

  

Marital status

 With partner

83.5

85.1

0.90 (0.55–1.46)

 

 Without partner

16.5

14.9

1.00 (ref.)

 

 (n)a

(385)

(4,176)

  

Parity

 1

47.2

45.5

1.00 (ref.)

 

 ≥2

52.8

54.5

0.94 (0.66–1.34)

 

 (n)a

(388)

(4,177)

  

Paid work

 Yes

43.4

45.9

1.00 (ref.)

 

 No

56.6

54.1

1.10 (0.78–1.55)

 

 (n)a

(388)

(4,175)

  

Total household income

 Until R$ 500.00

40.9

47.3

1.00 (ref.)

 

 >R$ 500.00

59.1

52.7

1.27 (0.89–1.82)

 

 (n)a

(341)

(3,650)

  

Health insurance

    

 Yes

26.1

21.8

1.24 (0.84–1.83)

 

 No

73.9

78.2

1.00 (ref.)

 

 (n)a

(388)

(4,173)

  

Smoking

 Yes

13.0

15.5

1.00 (ref.)

 

 No

87.0

84.5

1.21 (0.72–2.02)

 

 (n)a

(388)

(4,177)

  

Hypertension or heart disease

 Yes

24.5

13.2

1.96 (1.37–2.80)

1.89 (1.35–2.66)

 No

75.5

86.8

1.00 (ref.)

1.00 (ref.)

 (n)a

(387)

(4,170)

  

Diabetes

 Yes

2.1

1.7

1.18 (0.56–2.49)

 

 No

97.9

98.3

1.00 (ref.)

 

 (n)a

(386)

(4,164)

  

Anemia

 Yes

35.3

37.6

0.91 (0.63–1.31)

 

 No

64.7

62.4

1.00 (ref.)

 

 (n)a

(386)

(4,159)

  

Bold values indicate statistically significant

All variables were considered in the analysis

Real R$—Brazilian currency—(1US$ = 1,57R$)

aValid number, without expansion of women interviewed with at least one live birth

bPreterm delivery, but includes women with losses

cMultiple Poisson regression analysis; selection method ‘backward’

Table 5

Percentage of live birth children from 2001 to 2007, according to medical factors, grouped by preterm and term births (weighted analysis)

Characteristic

Group

PR (95 % CI)

RPadj (95 % CI)c

Preterm

Term

Twin

 Yes

6.6

0.5

6.34 (4.16–9.65)

4.85 (3.32–7.08)

 No

93.4

99.5

1.00 (ref.)

1.00 (ref.)

 (n)a

(587)

(5,527)

  

Prenatal care

 Yes

98.5

98.6

1.00 (ref.)

 

 No

1.5

1.4

1.10 (0.57–2.12)

 

 (n)a

(587)

(5,515)

  

Cesarean section

 Yes

51.2

42.0

1.39 (1.04–1.87)

1.74 (1.31–2.29)

 No

48.8

58.0

1.00 (ref.)

1.00 (ref.)

 (n)a

(586)

(5,510)

  

Scheduled cesarean

 Yes

17.4

19.8

0.87 (0.63–1.20)

0.57 (0.40–0.81)

 Not scheduled or not Cesarean

82.6

80.2

1.00 (ref.)

1.00 (ref.)

 (n)a

(585)

(5,492)

  

Place of prenatal careb

 NHS

72.8

78.2

0.77 (0.54–1.09)

 

 Health insurance

22.6

15.4

1.51 (1.03–2.23)

 

 Private

7.4

8.4

0.88 (0.55–1.41)

 

 Others

0.2

 <0.1

2.02 (0.48–8.59)

 

 (n)a

(574)

(5,327)

  

Number of prenatal care visits

 Up to 5

25.5

16.6

1.61 (1.15–2.27)

1.69 (1.24–2.31)

 >5

74.5

83.4

1.00 (ref.)

1.00 (ref.)

 (n)a

(544)

(5,065)

  

Bold values indicate statistically significant

All design variables were considered in the analysis

aValid number of live birth children without expansion

bNon-exclusive categories (each site vs. others)

cMultiple Poisson regression analysis; selection method ‘backward’; the model was also controlled for the variables: mother’s age, race or color, residence situation, education, marital status, parity, paid work, health insurance/health plan, smoking, hypertension, diabetes, anemia

Table 6 presents a comparative analysis between preterm and term newborns, according to the neonatal clinical conditions as reported by the mothers. Live newborns with low birth weight, kept in hospital after maternal discharge, and those who died by the date of the interview, were eight, seven and five times, respectively, more often reported by the women with at least one live preterm newborn, compared with the women who reported only term live births. These differences were significant. The table also presents the reasons for the longer hospital stay in the newborns.
Table 6

Characteristics of live birth children from 2001 to 2007, grouped by preterm and term delivery (weighted analysis)

Characteristic

Group

p*

Premature

Term

Child dead at the survey date

5.7

1.0

<0.001

(n)a

(587)

(5,527)

 

Low birth weight

31.0

3.8

<0.001

(n)a

(556)

(5,240)

 

Children remained hospitalized after maternal discharge

20.0

2.9

<0.001

(n)a

(580)

(5,288)

 

Reason for hospitalization

  

<0.001

 Gaining weight

44.6

11.7

 

 Neonatal jaundice

5.7

29.9

 

 Had infection

9.7

12.6

 

 Born prematurely

35.0

0.2

 

 Other

4.9

45.6

 

 (n)a

(128)

(112)

 

Hospital stay (≥ 8 d) of children who were hospitalized after maternal discharge

81.7

47.7

<0.005

(n)a

(125)

(115)

 

* All design variables were considered in the analysis

aValid number, without expansion of live birth children

Table 7 presents data on the association of severe maternal morbidities with preterm birth, including eclampsia, hemorrhage, admission to intensive care, interhospital transfer, need for mechanical ventilation, and hospital stay exceeding 7 days.
Table 7

Percentage of live birth children from 2001 to 2007, according to complications and/or interventions in mothers, grouped by preterm and term delivery (weighted analysis)

Complications and interventions in mothers

Group

Total

OR (95 % CI)

Preterm

Term

Complications

 Eclampsia (seizures during pregnancy, birth or postpartum without previous occurrences)

  Yes

1.4

0.3

0.4

4.50 (1.90–10.7)

  No

98.6

99.7

99.6

1.00 (ref.)

  (n)a

(583)

(5,497)

(6,080)

 

 Hemorrhage (heavy bleeding during the first 3 days postpartum)

  Yes

18.5

13.7

14.1

1.43 (1.02–2.01)

  No

81.5

86.3

85.9

1.00 (ref.)

  (n)a

(587)

(5,520)

(6,107)

 

 Infection (high fever after delivery, with chills and smelly vaginal discharge)

  Yes

0.6

0.6

0.6

1.00 (0.35–2.84)

  No

99.4

99.4

99.4

1.00 (ref.)

  (n)a

(584)

(5,486)

(6,070)

 

Interventions

  Hysterectomy

0.04

0.2

0.2

0.20 (0.04–1.12)

  (n)a

(523)

(5,328)

(5,851)

 

  Admission to ICU

1.5

0.4

0.5

3.75 (1.27–11.0)

  (n)a

(584)

(5,517)

(6,101)

 

  Blood transfusion

0.9

0.6

0.6

1.43 (0.55–3.72)

  (n)a

(584)

(5,506)

(6,090)

 

  Interhospital transfer

7.7

1.6

2.2

5.07 (2.56–10.0)

  (n)a

(587)

(5,518)

(6,105)

 

  Mechanical ventilation

4.1

1.1

1.4

3.73 (1.79–7.74)

  (n)a

(586)

(5,517)

(6,103)

 

  Length of hospital stay >1 week

13.5

2.9

4.0

5.20 (2.88–9.41)

  (n)a

(585)

(5,514)

(6,099)

 

Any

  Yes

33.7

17.1

18.7

2.46 (1.74–3.48)

  No

66.3

82.9

81.3

1.00 (ref.)

  (n)a

(558)

(5,360)

(5,918)

 

Bold values indicate statistically significant

All design variables were considered in the analysis

aValid number of live birth children without expansion

Discussion

Using information from a secondary analysis of a population household survey, this study is perhaps the first attempt to use data from a DHS to estimate national data on preterm births and their regional differences. It also identified a mean prevalence of preterm births in Brazil, for the last 6 years preceding the survey among all women interviewed, of around 10 % of live births, with significant regional variations. The highest variation was in the south and the lowest was in the north-east. Over the study period, however, there were no significant changes within any region.

The analysis also allowed a correlation between the prevalence of preterm birth with maternal morbidity. The study found that women with preterm delivery were more often white, living in urban areas and had a history of hypertension. In addition, preterm births were associated with increased risk of neonatal death, low birth weight, hospitalization, longer hospital stay, twin pregnancies, Cesarean section, and fewer prenatal care visits.

Data obtained from this study indicated a prevalence of preterm birth of 9.9 % in Brazil. Although these data represent only self-reported information of the women interviewed (hence such information is not from medical records), these results, however, substantially exceeded the 6.5 % officially reported for the country [16]. Paradoxically, these data are very close to those reported by developed countries, in which the birth registries are conducted more systematically and therefore more reliably. This suggests that the real prevalence of preterm births is not exactly known in Brazil.

This enormous disparity between the SINASC data and hospital records had already been observed by other authors. By using a linkage method, Silva et al. [19] observed low concordance (Kappa = 0.09) among the preterm records when using the SINASC and a hospital survey. Generally, Brazilian studies have shown poor reliability in estimates of gestational age, by comparing the SINASC with other sources of data, with Kappa ranging from 0.09 to 0.83 [20].

Other studies also showed similar results, with the SINASC prevalence always lower than the prevalence obtained by other methods [19, 21, 22, 23]. However, more recent studies found only a minor difference between the SINASC and the method used, which may be a result of an improvement in the information reliability of SINASC [24]. The difference in this study between SINASC data and DHS data could be the result of the way the gestational age is represented in the declaration of live births for the SINASC. This document does not report the exact gestational age but relatively wide intervals of age (less than 22 weeks, 22–27, 28–31, 32–36, 37–41, 42 or more). This could be a source of error if a gestational age close to term, for instance a 36 weeks baby, was wrongly classified as at term, then considering as preterm delivery only the cases of extreme gestational ages. This would lead to an under-reporting of cases of late preterm birth from 34 to 36 weeks, which would be classified as term [25].

As the study collected information from the past last 5–6 years, a potential recall bias in interviewing the women cannot be overlooked. This bias, however, would not explain the lower prevalence in the official registry. Recall bias could theoretically reduce the prevalence of the condition for the years more distant from the interview, and this may be so, as there was a trend toward a rise in the estimated prevalence of preterm births over recent years, as was observed by other authors [4, 5, 20]. In this study, such an increase was observed when comparing the first triennium (2001–2003) to the second triennium (2004–2006/7).

There were some marked regional variations, with higher prevalence in southern and southeastern regions, in both the SINASC data and the DHS data. It can be observed, however, that there was an increase in the overall preterm birth rate in both the DHS and in the SINASC data, the latter showing an increase from 5 % in 1994 to 5.4 % in 1998, 5.6 % in 2000, 6.6 % in 2005, and 6.7 % in 2007. Such an increase in the prevalence may represent an improvement in the reliability of data collection [20, 25]. Another relevant explanation could be the improvements in obstetric care with more at risk infants being delivered live preterm. The following conditions related to the possibility of live preterm delivery may be recognized: (a) improved access to health services and consequent reduction in delay to identify and treat risk factors; and (b) improvement in conditions for obstetric and perinatal care, consequently with a higher rate of therapeutic preterm after maternal complications [13].

Brazil has been experiencing an increase in the coverage of prenatal care and an improvement in obstetric and perinatal care, with easier access to such services [18]. Therefore, a rise in the prevalence of preterm births could possibly be a consequence of better treatment for maternal complications and more babies surviving in utero. Some of the findings in this study concerning the characteristics and risk factors for preterm births support this hypothesis. White women living in urban areas, who are users of health insurance, are part of a population with greater access to health services. In addition, women with private insurance are known to be at higher risk for preterm birth for several reasons, including wealth, higher access to infertility treatments and also the consequent multiple pregnancies. The lower number of prenatal visits, associated with preterm, could be a consequence of shorter pregnancies.

Other factors associated with preterm births (twins, hypertension, Cesarean section) could be related to an increase in morbid conditions of pregnant women, except for the variable ‘Cesarean scheduled’, which appears as a protective factor for preterm birth, probably because there is no medical indication to schedule an elective Cesarean section before term. Thus, maternal morbidity is the major factor which explains the increase in prevalence of preterm births. In fact, eclampsia increased the risk of preterm birth by 4.5 times. Other indicators of severe maternal morbidity, e.g., hemorrhage, admission to intensive care units, inter-hospital transfer, need for mechanical ventilation, and prolonged hospitalization also increased the risk. The presence of any indicator of severe maternal morbidity or of a life-threatening condition during gestation increased the risk of preterm nearly 2.5 times in the study.

Other authors have already reported an association between clinical and gestation conditions and preterm birth [4, 10, 26]. However, this seems to be the first study to identify an association of life-threatening conditions and severe maternal morbidity (maternal near miss), as recently defined by WHO [27], with preterm births in a sample representative of the whole female population in a large country.

As far as we know, this study is also the first to use a DHS to analyze data related to preterm, and this may bring some limitations to the study. The accuracy of self-reported data for the diagnosis of preterm birth is not yet known. Likewise, as the information depends on the interviewees’ memory, a recall bias may occur, although it is unlikely for such a major life event. In addition, the sampling process selected nearly 6,000 women for a population representative of 17 million people. As the primary objective of the study was not to identify the prevalence of preterm births, the power of this analysis may have been reduced, generating some distortions.

This study, however, presented results which suggest a trend for a rise in the prevalence of preterm births in Brazil. Furthermore, it identified factors associated with preterm birth, and, in particular, showed for the first time that preterm delivery was more frequent among women who have any type of severe maternal morbidity. Strategies used to promote improvements in healthcare for a population should prioritize high-risk situations which occur more frequently and/or with greater severity, or both. In order to do so, knowledge of the prevalence of adverse clinical conditions which affects the target population, is vital. For this, further detailed studies on the topic are warranted [28].

Finally, although there are concerns regarding the validity of self-reporting of preterm births, it may be a useful method for epidemiological analysis, especially in developing countries, where official detailed collection of such data is still limited. If the validity can be confirmed, the possibility of assessing the prevalence of preterm births in the last two or three decades in all countries that implemented a DHS would be presented. In addition, if these analyses could also identify possible risk factors associated with preterm birth, then this could help in regional policy-making to reduce the recent observed increase in the global prevalence of preterm delivery.

Notes

Acknowledgments

This manuscript was performed with the support from the grant 2009/53245-5 from Fapesp.

Conflict of interest

The authors declare they have no conflict of interests.

References

  1. 1.
    Black, R. E., Morris, S. S., & Bryce, J. (2003). Where and why are 10 million children dying every year? Lancet, 361(9376), 2226–2234.PubMedCrossRefGoogle Scholar
  2. 2.
    United Nations. (2000). The millennium development goals. Source: http://www.un.org/millenniumgoals/poverty.shtml. Accessed on 18 March 2011.
  3. 3.
    Parikh, N. I., Cnattingius, S., Ludvigsson, J. F., & Ingelsson, E. (2011). Birth characteristics and subsequent risks of maternal cardiovascular disease: Effects of gestational age and fetal growth. Circulation, 124(25), 2839–2846.PubMedCrossRefGoogle Scholar
  4. 4.
    Blencowe, H., Cousens, S., Oestergaard, M. Z., Chou, D., Moller, A. B., Narwal, R., et al. (2012). National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: A systematic analysis and implications. Lancet, 379(9832), 2162–2172.PubMedCrossRefGoogle Scholar
  5. 5.
    Martin, J. A., Hamilton, B. E., Sutton, P. D., Ventura, S. J., Menacker, F., Kirmeyer, S., et al. (2007). Births: final data for 2005. National Vital Statistics Report, 56(6), 1–103.Google Scholar
  6. 6.
    Ashton, D. M., Lawrence, H. C., Adams, N. L, I. I. I., & Fleischman, A. R. (2009). Surgeon general’s conference on the prevention of preterm birth. Obstetrics and Gynecology, 113, 925–930.PubMedGoogle Scholar
  7. 7.
    Beck, S., Wojdyla, D., Say, L., Betran, A. P., Merialdi, M., Requejo, J. H., et al. (2010). The worldwide incidence of preterm birth: A systematic review of maternal mortality and morbidity. Bulletin of the World Health Organization, 88(1), 31–38.PubMedCrossRefGoogle Scholar
  8. 8.
    WHO. (1977). WHO: recommended definitions, terminology and format for statistical tables related to the perinatal period and use of a new certificate for cause perinatal deaths. Modifications recommended by FIGO as amended October 14, 1976. Acta Obstetricia et Gynecologica Scandinavica, 56, 247–253.Google Scholar
  9. 9.
    Marlow, N., Wolke, D., Bracewell, M. A., Samara, M., & EPICure Study Group. (2005). Neurologic and developmental disability at six years of age after extremely preterm birth. New England Journal of Medicine, 352, 9–19.PubMedCrossRefGoogle Scholar
  10. 10.
    Goldenberg, R. L., Culhane, J. F., Iams, J. D., & Romero, (2008). Epidemiology and causes of preterm birth. Lancet, 371(9606), 75–84.PubMedCrossRefGoogle Scholar
  11. 11.
    Yücesoy, G., Ozkan, S., Bodur, H., Tan, T., Calişkan, E., Vural, B., et al. (2005). Maternal and perinatal outcome in pregnancies complicated with hypertensive disorder of gestation: A seven year experience of a tertiary care center. Archives of Gynecology and Obstetrics, 273(1), 43–49.PubMedCrossRefGoogle Scholar
  12. 12.
    Ferrazzani, S., Luciano, R., Garofalo, S., D’Andrea, V., De Carolis, S., De Carolis, M. P., et al. (2011). Neonatal outcome in hypertensive disorders of gestation. Early Human Development, 87(6), 445–449.PubMedCrossRefGoogle Scholar
  13. 13.
    Yang, X., & Zeng, W. (2011). Clinical analysis of 828 cases of iatrogenic preterm births. Journal of Obstetrics and Gynaecology Research, 37(8), 1048–1053.PubMedCrossRefGoogle Scholar
  14. 14.
    Fuchs, K., & Wapner, R. (2006). Elective cesarean section and induction and their impact on late preterm births. Clinics in Perinatology, 33(4), 793–801.PubMedCrossRefGoogle Scholar
  15. 15.
    Malloy, M. H. (2009). Impact of cesarean section on intermediate and late preterm births: United States, 2000–2003. Birth, 36(1), 26–33.PubMedCrossRefGoogle Scholar
  16. 16.
    Brazil Ministry of Health. (2011). DATASUS/SINASC—information system on live births. Source: http://tabnet.datasus.gov.br/cgi/tabcgi.exe?sinasc/cnv/nvuf.def. Accessed on 29 May 2011.
  17. 17.
    Barros, F. C., Barros, A. J., Villar, J., Matijasevich, A., Domingues, M. R., & Victora, C. G. (2011). How many low birthweight babies in low- and middle-income countries are preterm? Revista de Saude Publica, 45(3), 607–616.PubMedGoogle Scholar
  18. 18.
    Brazil Ministry of Health. (2009). National survey on demography and health of women and childrenPNDS 2006. Dimensions of the reproductive process and child health. 1st. Edn. Brasília, DF. Source: http://bvsms.saude.gov.br/bvs/publicacoes/pnds_crianca_mulher.pdf Accessed on 27 April 2011.
  19. 19.
    Silva, A. A., Ribeiro, V. S., Borba, A. F, Jr, Coimbra, L. C., & Silva, R. A. (2001). Evaluation of data quality from the information system on live births in 1997–1998. Revista de Saude Publica, 35(6), 508–514.PubMedCrossRefGoogle Scholar
  20. 20.
    Silveira, M. F., Santos, I. S., Barros, A. J., Matijasevich, A., Barros, F. C., & Victora, C. G. (2008). Increase in preterm births in Brazil: Review of population-based studies. Revista de Saude Publica, 42(5), 957–964.PubMedCrossRefGoogle Scholar
  21. 21.
    Bettiol, H., Rona, R. J., Chinn, S., Goldani, M., & Barbieri, M. A. (2000). Factors associated with preterm births in Southeast Brazil: A comparison of two birth cohorts born 15 years apart. Paediatric and Perinatal Epidemiology, 14, 30–38.PubMedCrossRefGoogle Scholar
  22. 22.
    Almeida, M. F., Alencar, G. P., França, I, Jr, Novaes, H. M., Siqueira, A. A., Schoeps, D., et al. (2006). Validation of birth certificates based on data from a case–control study. Cadernos Saúde Pública, 22, 643–652.CrossRefGoogle Scholar
  23. 23.
    Barros, A. J., Santos, I. S., Victora, C. G., Albernaz, E. P., Domingues, M. R., Timm, I. K., et al. (2006). The 2004 Pelotas birth cohort: Methods and description. Revista de Saúde Pública, 40, 402–413.PubMedCrossRefGoogle Scholar
  24. 24.
    Drumond, E. F., Machado, C. J., & França, E. (2008). Underreporting of live births: Measurement procedures using the hospital information system. Revista de Saude Publica, 42(1), 55–63.CrossRefGoogle Scholar
  25. 25.
    Silveira, M. F., Santos, I. S., Matijasevich, A., Malta, D. C., & Duarte, E. C. (2009). Preterm births in Brazil from 1994 to 2005 according to the information system on live birth (SINASC). Cadernos de Saude Publica, 25(6), 1267–1275.PubMedCrossRefGoogle Scholar
  26. 26.
    Meis, P. J., Goldenberg, R. L., Mercer, B. M., Iams, J. D., Moawad, A. H., Miodovnik, M., et al. (1998). The preterm prediction study: Risk factors for indicated preterm births. American Journal of Obstetrics and Gynecology, 178(3), 562–567.PubMedCrossRefGoogle Scholar
  27. 27.
    Say, L., Souza, J. P., Pattinson, R. C., & WHO working group on Maternal Mortality and Morbidity classifications. (2009). Maternal near miss – towards a standard tool for monitoring quality of maternal health care. Best Practice & Research Clinical Obstetrics & Gynaecology, 23(3), 287–296.CrossRefGoogle Scholar
  28. 28.
    Passini, R, Jr, Tedesco, R. P., Marba, S. T., Cecatti, J. G., Guinsburg, R., Martinez, F. E., et al. (2010). Brazilian multicenter study on prevalence of preterm birth and associated factors. BMC Pregnancy and Childbirth, 10, 22.PubMedCrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2012

Authors and Affiliations

  • Ricardo P. Tedesco
    • 1
    • 2
  • Renato PassiniJr.
    • 1
  • José G. Cecatti
    • 1
    • 3
    • 5
  • Rodrigo S. Camargo
    • 1
    • 2
  • Rodolfo C. Pacagnella
    • 1
    • 4
  • Maria H. Sousa
    • 3
  1. 1.Department of Obstetrics and Gynecology, School of Medical SciencesUniversity of CampinasSão PauloBrazil
  2. 2.Medical School of JundiaíSão PauloBrazil
  3. 3.Campinas Centre for Studies in Reproductive Health, Campinas (CEMICAMP)São PauloBrazil
  4. 4.Department of MedicineFederal University of São Carlos (UFSCAR)São CarlosBrazil
  5. 5.DTG/CAISM/UNICAMPCampinas, São PauloBrazil

Personalised recommendations