The global network: a prospective study of stillbirths in developing countries

Article Outline

• Abstract
• Methods
• Results
• Comment
• Acknowledgments
• References
• Copyright

Objective

Our goal was to determine stillbirth rates in a multisite population-based study in community settings in the developing world.

Study Design

Outcomes of all community deliveries in 5 resource-poor countries (Democratic Republic of Congo, Guatemala, India, Zambia, and Pakistan) and in 1 mid-level country (Argentina) were evaluated prospectively over an 18-month period. Births of >1000 g with no signs of life were defined as stillbirth.

Results

Outcomes of 60,324 deliveries were included. Stillbirth rates ranged from 34 per 1000 in Pakistan to 9 per 1000 births in Argentina. Increased stillbirth rates were associated significantly with lower skilled providers, out-of-hospital births, and low cesarean section rates. Maceration was present in 17.2% of stillbirths.

Conclusion

The stillbirth rates among births of ≥1000 g in these developing countries were substantially higher than reported stillbirth rates in developed countries (3-5/1000 births). Because most developed countries define stillbirth as ≥20 weeks of gestation or ≥500 g and because almost one-half of all stillbirths are <1000 g, the developing/developed country difference is actually larger than apparent from this study. Maceration was uncommon, which indicates that most of the deaths probably occurred during labor. The low rates of physician attendance, hospital delivery, and cesarean section deliveries suggest that stillbirth rates could be reduced by access to higher quality institutional deliveries.

Key words: developing country, intrapartum stillbirth, stillbirth

Stillbirths generally account for one-half of all perinatal deaths, with an estimated 4 million occurring worldwide each year. More than 97% of these stillbirths take place in developing countries.¹ For many reasons, stillbirths have been understudied, underreported, and rarely have been considered in attempts to improve adverse pregnancy outcomes in developing countries.¹, ²

Recent estimates suggest that stillbirth rates of >30 per 1000 births are common among the least developed countries, especially in Sub-Saharan Africa and Southeast Asia. By comparison, rates of 3-5 per 1000 deliveries have been documented in the United States and other developed countries, and rates of 10-15 per 1000 are reported in mid-level countries, such as those in South and Central America.³, ⁴ Although the World Health Organization has attempted to standardize the definition of stillbirth by recommending 1000 g as the lower limit for international comparisons (corresponding to approximately 28 weeks of gestation), the lower limit of the gestational age or birthweight that is reported varies widely. In developed countries, stillbirth has been defined generally as fetal loss beyond 20 weeks of gestation; however, some developed countries (such as Sweden) still use 28 weeks of gestation as the lower cutoff. In less developed countries, a gestational age of 28 weeks or a birthweight of 1000 g is often the lower cutoff that is used.⁵

The timing of stillbirth in relation to delivery also varies from developed to developing countries. Stillbirths that occur more than 12-24 hours before delivery have skin that is “macerated,”² although those stillbirths that occur in the intrapartum period or immediately before or during delivery are generally normal in appearance and are often called fresh stillbirths. In developed countries, intrapartum stillbirths comprise less than 10% of all stillbirths; in some of the least developed countries, up to one-half of all stillbirths are thought to occur intrapartum.², ⁶ When intrapartum stillbirths occur, they likely represent inadequate access to or poor quality of essential obstetric care.⁷, ⁸

Because data on stillbirths are not collected routinely in many countries and most of the stillbirth research has been hospital-based, much is still unknown about the prevalence, timing, and circumstances that are associated with stillbirths in developing countries, where over one-half of all deliveries occur at home. Understanding the burden of stillbirth has important programmatic and resource implications, which are of particular concern in very low-resource settings. Our goal in this study was to determine population-based stillbirth rates and to characterize healthcare at delivery in prospective, well-defined community-based birth cohorts in developing country settings. On the basis of a review of previous studies of stillbirth,⁴ we hypothesized that home birth and delivery with unskilled attendant (traditional birth attendant or family) would be associated with higher rates of stillbirth.

Back to Article Outline

Methods 

The study was conducted as part of the Global Network for Women’s and Children’s Health Research (Global Network), a National Institutes of Health–funded, multisite research network that represents partnerships of US and international investigators. Prospective data registries were created to establish baseline delivery rates as part of a larger study of neonatal resuscitation in developing countries that was conducted in 6 countries: Argentina, Democratic Republic of Congo, Guatemala, India (1 site in Orissa and 1 site in Belgaum), Pakistan, and Zambia. The study was reviewed and approved by the institutional ethics review committees of all participating foreign sites, the partner institutions in the United States, and the data center at Research Triangle Institute. Consent was obtained at the community level; women provided informed verbal consent.

The outcomes of all deliveries in the communities, defined as a distinct geographic region whose birth attendants did not overlap with other communities, were collected. All birth attendants (n = 3676) were trained prospectively to collect data and assess basic clinical variables and outcomes, which included differentiation of stillbirths and neonatal deaths at birth, type of stillbirth, and assessment of gestational age. Birth attendants were trained to identify maceration using pictures to standardize reporting of this condition. Data collection was overseen by trained community coordinators (nurses or physicians) who oversaw data collection of all birth attendants in the community.

Each Global Network site included 10-28 communities, with approximately 300-500 deliveries per community annually. The sites that were studied were distinct geographic entities and included rural areas in Orissa, India; Thatta, Pakistan; Kafue, Zambia; and Equateur, Democratic Republic of Congo, all with very limited access to health care services, to Belgaum, India which had more access to healthcare, to the most developed geographic area, in Argentina.

Women were registered by 24-28 weeks of pregnancy. After delivery, the community coordinator collected the data that were recorded by the birth attendant. Data included basic information on maternal demographics and neonatal and maternal outcomes at delivery. A stillbirth was defined as any delivery of ≥1000 g, corresponding to approximately 28 weeks of gestation, in which no signs of life (breathing, crying, heartbeat, movement) were evident. The type of delivery attendant included physician, nurse or nurse-equivalent, traditional birth attendant (TBA), family, or unattended. Location of delivery included hospital, health center, home (including the TBA’s home), or other (in transit). Prenatal care was defined as at least 1 visit with a health provider. Finally, the birthweight was taken within 48 hours of delivery with scales that were provided for the study.

All data were entered centrally at each study site; data edits, which included inter- and intraform consistency checks, were performed at entry, with additional edits performed by the data center. The data were analyzed with SAS software (version 9.0; SAS Institute Inc, Cary, NC). Relative risks were calculated with Cochran-Mantel-Haenszel for the prospectively identified variables that were associated with stillbirth. Reference categories were defined as those categories that were associated with the lowest stillbirth rates.

Back to Article Outline

Results 

From March 2005 to December 2006, 60,324 deliveries were recorded in 103 communities in the participating Global Network sites; consent was obtained from 60,154 women (99.7%) whose pregnancy outcomes were included in this study (Table 1). Most women (89.0%) received at least 1 prenatal care visit. In Argentina, 68.9% of the deliveries were conducted by a physician; in 3 countries (Guatemala, Democratic Republic of Congo, and Zambia), <1% of deliveries were conducted by a physician. Most deliveries (66.3%) were conducted in a home setting (family or birth attendant’s home). The site of delivery ranged from 100% of deliveries in a hospital or health clinic in Argentina to 99.9% of the deliveries in a home setting in Guatemala. Cesarean delivery rates ranged from 19.1% in Argentina to 0 in the communities in Guatemala and Orissa, India. Birthweights were available for 76% of stillbirths and 91% of the live births.

TABLE 1. All deliveries: characteristics by study site

Site	Total deliveries (n)	Prenatal care (%)	Birth attendant type (%)			Delivery location (%)		Cesarean delivery (%)	Birth weight in grams (mean ± SD)	≥2500 g (%)
			Physician	Nurse	TBA/family or unattended	Hospital or clinic	Home
Argentina	6,837	90.7	68.9	31	0	100.0	0	19.1	3303±549	93.8
Guatemala	6,320	85.1	0.2	34.8	54.3	0.1	99.9	0	3204±537	92.5
Democratic Republic of Congo	7,952	95.3	0.1	21.6	78.3	25.6	74.4	0.1	3045±567	88.1
Zambia	6,946	94.7	0.5	34.8	64.7	34.5	65.5	0.3	2994±556	86.2
India
Orissa	10,320	87.5	3.5	42.7	53.8	5.5	94.5	0	2607±278	85.7
Belgaum	11,994	98.0	31.4	30.5	38.1	51.8	48.2	2.9	2694±415	79.8
Pakistan	9,785	71.5	6.2	17.7	76.1	22.8	77.2	0.1	2717±481	77.7
Total	60,154	89.0	15.8	26.7	57.5	33.7	66.3	2.8	2901±542	85.6

A total of 1472 stillbirths were recorded (Table 2). The mean stillbirth rate was 24 per 1000 deliveries, which ranged from 9 per 1000 in Argentina to 34 per 1000 deliveries in Pakistan. Signs of maceration were reported in 17.2% of stillbirths (range between sites was 3.6%-45.8%). The mean birthweight for the stillbirths was 2221 g ± 744 g. In comparison, the mean birthweight for live births was 2918 g ± 520 g (P ≤ .001); 63.6% of the stillbirths were ≥2000 g.

TABLE 2. Stillbirths: characteristics by study site

Site	Stillbirths (n)	Stillbirth rate per 1000 births (%)	Macerated stillbirth (%)	Birth weight in grams (mean±SD)	Birthweight in grams (%)
					1000-1499	1500-1999	2000-2500	>2500
Argentina	59	9	45.8	2323±971	14.5	18.4	10.5	31.6
Guatemala	105	17	8.6	2540±808	13.0	13.0	19.0	51.0
Democratic Republic of Congo	240	30	21.3	2242±834	23.7	15.7	13.2	44.5
Zambia	201	29	33.3	1985±791	19.3	10.9	7.9	17.8
India
Orissa	280	27	3.6	2333±371	13.3	13.3	33.3	52.0
Belgaum	257	21	26.8	2072±546	14.1	20.3	35.0	28.2
Pakistan	330	34	6.1	2257±512	11.1	11.1	27.8	50.0
Total	1472	24	17.2	2221±744	18.9	17.5	22.0	41.6

Women who were >35 years of age at delivery, who had no formal education, who were primiparous or multiparous (≥4th pregnancy) had a higher relative risk of stillbirth (Table 3). In addition, women who had no prenatal care, who had a lower level of care provider at delivery, and who delivered out of hospital were more likely to have a stillbirth than women without these characteristics. Of the perinatal characteristics, infants who were male, preterm, and <2500 g all had a higher risk of stillbirth. Less than 1% of all stillbirths had documented congenital abnormalities at the time of delivery.

TABLE 3. Characteristics by stillbirth

Characteristic	Total (n)	Stillbirths per 1000 (n)	Relative risk (95% CI)
Maternal
Age(y)
<25	27,814	22	0.9(0.8, 1.0)
25-35⁎	27,739	25	1.0
>35	3,625	36	1.5(1.2, 1.8)
Education
No formal education	26,849	30	1.6(1.4, 1.8)
Any formal education⁎	32,639	19	1.0
Living children(n)
0	14,999	29	1.4(1.2, 1.5)
1-4⁎	37,874	21	1.0
>4	4,011	29	1.4(1.1, 1.6)
Prenatal Care
≥1 Visit⁎	53,248	22	1.0
No prenatal care	6,590	44	2.0(1.8, 2.3)
Birth attendant
Physician⁎	9,486	19	1.0
Nurse/midwife	16,036	25	1.3(1.1, 1.6)
TBA/family/unattended	34,563	26	1.3(1.1, 1.6)
Delivery location
Home/other	39,839	26	1.2(1.1, 1.4)
Clinic/hospital⁎	20,282	22	1.0
Infant
Gender
Male	31,497	28	1.2(1.0, 1.3)
Female⁎	28,554	25	1.0
Gestational age(wk)
<37	7,002	45	2.9(2.5, 3.4)
≥37⁎	32,305	15	1.0
Birthweight(g)
<2500	8,089	84	4.6(4.4, 4.9)
≥2500⁎	47,217	7	1.0

Back to Article Outline

Comment 

The major strength of this study was that we prospectively collected population-based delivery outcomes for distinct, geographically defined communities in 6 countries that represented different levels of care. Data collectors received standardized formal training and ongoing oversight by community coordinators, who verified all pregnancy outcome data. We are not aware of any multicountry study of stillbirth with this level of data standardization or study oversight. In addition, most previous studies of stillbirths in developing countries have neither been prospective nor population-based.

The mean stillbirth rate of 24.0 per 1000 deliveries is >5-fold higher than stillbirth rates in most developed countries. Stillbirths of <1000 g were not included in this study but are included in the US rates and in the US account for 50% of the stillbirths.⁹, ¹⁰ In addition, women who did not register and experienced a stillbirth before 28 weeks of gestation may never have reported their loss. Thus, the disparity between the developed/developing country stillbirth rates is even larger than indicated by the comparison described earlier.¹¹, ¹² Within our study, stillbirth rates ranged from 9 per 1000 in the Argentinian communities to 34 per 1000 in the Pakistani communities.

Similar to studies in developed countries, maternal age of >35 years and lower socioeconomic status were associated with higher stillbirth rates.¹³ In addition, higher stillbirth rates were associated with less prenatal care, unattended deliveries or deliveries by TBAs, out-of-hospital births, and lower rates of cesarean section delivery. Cesarean section delivery rates of at least 5% are considered necessary to reduce stillbirth and prevent maternal death.⁸, ¹⁴, ¹⁵ Although cesarean section delivery may be a proxy for many healthcare quality factors, in this study, the site with the highest cesarean section delivery rate, Argentina, also had the lowest stillbirth rate. As another example, the Asian sites that had no access to cesarean section delivery (Orissa, India; and Thatta, Pakistan) had significantly higher stillbirth rates than did Belgaum, India, which had a 3% cesarean section delivery rate.

Because previous studies have also reported an association between lower level providers and adverse pregnancy outcomes, ensuring increased access to skilled delivery attendants has been used in an attempt to improve adverse pregnancy outcomes.¹ However, because skilled providers are unavailable in many of the least developed geographic areas, also studies have examined a strategy of training traditional birth attendants. For example, a cluster-randomized trial in Pakistan found that training traditional birth attendants in basic delivery skills significantly reduced the stillbirth rates (50 per 1000 in the intervention clusters vs 71 per 1000 in the control clusters).¹⁶

Most studies of stillbirth in developing countries have not included the birthweight, which is an important proxy for viability, especially where reliable gestational age dating is unavailable. Birthweight of stillbirths has been difficult to collect, often because of cultural barriers.³ A few hospital-based studies have reported birthweight for stillbirths in less developed countries¹⁷, ¹⁸; however, population-based stillbirth birthweights are not available. We found that the mean birthweight for stillbirths was lower than that of the live births, but more than one-half of the stillbirths were ≥2000 g and thus were likely to represent near-term or term deliveries. Furthermore, in this study, most of the stillbirths were fresh and are likely to have occurred during labor.

The acquisition of more knowledge about stillbirths is important because of its significant contribution to adverse pregnancy outcomes. In this study, the mean stillbirth rate of >24 per 1000 represents a >5-fold increase compared with developed country rates. Importantly, in the less developed communities, where nearly all deliveries occurred in home settings without trained health providers, rates were as high as 34 per 1000, compared with the rates in Argentina of 9 per 1000, where nearly all deliveries occurred in hospital settings. Although our data suggest that higher quality of healthcare at delivery, especially access to high level healthcare providers and cesarean delivery, is associated with lower stillbirth rates, more research on the specific causes of these stillbirths would assist in planning appropriate interventions. The fact that most of the stillbirths were fresh and that many were term or near-term suggests that stillbirth rates could be reduced substantially by higher quality intrapartum care.

Back to Article Outline

Acknowledgments 

First Breath Study Group: Argentina — Fernando Althabe and Jose M Belizan, University of Buenos Aires; Pierre Buekens, Tulane School of Public Health and Tropical Medicine; Democratic Republic of Congo — John Ditekemena and Antoinette Tshefu, Kinshasa School of Public Health; Carl Bose, University of North Carolina at Chapel Hill; Zambia — Margaret Mbelenga, Center for Infectious Disease Research in Zambia; Elwyn Chomba, University of Zambia; Waldemar A Carlo, University of Alabama at Birmingham; Guatemala — Manolo Mazariegos, Center for Studies of Sensory Impairment, Aging and Metabolism; Ana Garces, San Carlos University; Michael Hambidge, University of Colorado; Orissa, India — Arjit Mohapatra and Sailajanandan Parida, SCB Medical College; Pinaki Panigrahi, University of Maryland, Baltimore; Belgaum, India — Shivaprasad S. Goudar and Bhalchandra S. Kodkany, Jawaharlal Nehru Medical College; Richard J. Derman, University of Missouri–Kansas City; Pakistan — Syed Rafat Ali Jafri, Imtiaz Jehan, and Omrana Pasha, Aga Khan University; United States — Robert L. Goldenberg, Drexel University; Hillary Harris, Elizabeth M. McClure, and Ty Hartwell, Research Triangle Institute; Linda L. Wright, Macaya Douoguih, Anne Willoughby, National Institutes of Child Health and Human Development, Bethesda, MD.

Back to Article Outline

References 

1. Neonatal and perinatal mortality: country, regional and global estimates 2006. Geneva: World Health Organization; 2006;
   • View In Article
2. Lawn J, Shibuya K, Stein C. No cry at birth: global estimates of intrapartum stillbirth and intrapartum-related neonatal deaths. Bull World Health Org. 2005;83:409–417
   • View In Article
3. Stanton C, Lawn JE, Rahman H, Wilczynska-Ketende K, Hill K. Stillbirth rates: delivering estimates in 190 countries. Lancet. 2006;367:1487–1494
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (151 KB)
   • CrossRef
4. McClure EM, Nalubamba-Phiri M, Goldenberg RL. Stillbirth in developing countries. Int J Gynaecol Obstet. 2006;94:82–90
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (126 KB)
   • CrossRef
5. Goldenberg RL, Thompson C. The infectious origins of stillbirth. Am J Obstet Gynecol. 2003;189:861–873
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (193 KB)
   • CrossRef
6. Smith G. Predicting antepartum stillbirth. Curr Opin Obstet Gynecol. 2006;18:625–630
   • View In Article
   • MEDLINE
7. Kiely JL, Paneth N, Susser M. Fetal death during labor: an epidemiologic indicator of level of obstetric care. Am J Obstet Gynecol. 1985;153:721–726
   • View In Article
   • MEDLINE
8. Ronsmans C, De Brouwere V, Dubourg D, Dieltiens G. Measuring the need for life-saving obstetric surgery in developing countries. BJOG. 2005;111:1027–1030
   • View In Article
   • MEDLINE
   • CrossRef
9. Phelan ST, Goldenberg RL, Alexander G, Cliver SP. Perinatal mortality and its relationship to the reporting of low-birthweight infants. Am J Public Health. 1998;88:1236–1239
   • View In Article
   • MEDLINE
   • CrossRef
10. Kramer MS, Liu S, Luo Z, et al. Analysis of perinatal mortality and its components: time for a change?. Am J Epidemiol. 2002;156:493–497
    • View In Article
    • MEDLINE
    • CrossRef
11. Silver RM, Varner MW, Reddy U, et al. Work-up of stillbirth: a review of the evidence. Am J Obstet Gynecol. 2007;196:433–444
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (171 KB)
    • CrossRef
12. Goldenberg RL, Koski JF, Boyd BW, Cutter GR, Nelson KG. Fetal deaths in Alabama 1974 to 1983: a birth weight-specific analysis. Obstet Gynecol. 1987;70:831–835
    • View In Article
    • MEDLINE
13. Fretts R. Etiology and prevention of stillbirth. Am J Obstet Gynecol. 2005;193:1923–1935
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (386 KB)
    • CrossRef
14. Stanton C, Dubourg D, DeBrouwere V, Pujades M, Ronsmans C. Reliability of data on caesarean sections in developing countries. Bull World Health Org. 2005;83:449–455
    • View In Article
15. Ronsmans C, Holtz S, Stanton C. Socioeconomic differentials in caesarean rates in developing countries: a retrospective analysis. Lancet. 2006;368:1516–1523
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (110 KB)
    • CrossRef
16. Jokhio A, Winter HR, Cheng KK. An intervention involving traditional birth attendants and perinatal and maternal mortality in Pakistan. N Engl J Med. 2005;352:2091–2099
    • View In Article
    • CrossRef
17. Buchmann E, Pattinson RC. Babies who die from labour-related intrapartum hypoxia: a confidential enquiry in South African public hospitals. Trop Doct. 2006;36:8–10
    • View In Article
    • MEDLINE
    • CrossRef
18. Perinatal Mortality Study Group. Perinatal mortality rate: hospital-based study during 1998-2001 at Hacettepe University. J Perinat Med. 2003;31:435–440
    • View In Article
    • MEDLINE
    • CrossRef