With one of the highest birth rates in the world, Uganda also grapples with high maternal mortality as demonstrated by a maternal mortality ratio (MMR) of 343 per 100,000 live births in 2015 compared with high-income countries (HIC) such as the United States with a MMR of 14 (https://data.worldbank.org/indicator/SH.STA.MMRT?view=chart). Obstacles to decreasing the MMR in Uganda include its large rural population (84%) and the amount of women who do not receive skilled care during delivery (43%) (https://data.worldbank.org/indicator/SP.RUR.TOTL.ZS). Over the past decade, the total number of institutional deliveries in Uganda has nearly doubled, with the most dramatic increase in numbers coming from the poorest rural population (https://data.worldbank.org/indicator/SP.RUR.TOTL.ZS). As technology availability spreads, more women take the opportunity to give birth in a safe, controlled medical environment, demonstrating that given the chance, women choose to receive adequate health care that optimizes the safety of their baby and themselves. With that said, referral from one level of care to another requires travel time and financial resources, neither of which is readily available in an emergency situation in rural Uganda. Given this, the impact of potential pregnancy complications is likely greater than in HIC, and early identification of risk factors that may require emergency obstetric provisions has the potential to impact maternal mortality.
Through use of screening obstetric ultrasound, pregnancies with risk factors for serious complications that would benefit from delivery at centers with emergency obstetric provision may be identified. The acceptability of obstetric ultrasound has been demonstrated in multiple low-resource settings1. Modern ultrasound technology is compact, portable, safe, relatively affordable, and suitable for practice at rural health centers2. In June 2010, Imaging the World (ITW), a US based nongovernmental organization, pioneered a low cost, high-quality antenatal ultrasound program designed for low-resource areas. The program was piloted at Nawanyago Health Centre III (NHC), a private rural health center in Kamuli District, Uganda; in the pilot, pregnant women were offered a screening obstetric ultrasound with a trained sonographer. Scans needing follow-up were triaged using a preordained algorithm3. During the study period, findings on screening obstetric ultrasound changed medical decision making in over 22% of cases3. In addition, with the implementation of this program, a “magnet effect” was seen at NHC with a 60% increase in monthly deliveries and antenatal care attendance3.
Screening point-of-care obstetric ultrasound can aid in accurate pregnancy dating and identification of pregnancy risk factors such as placental abnormalities, fetal malpresentation, multiple gestations, and ectopic pregnancy4. Such pregnancy risk factors are associated with hemorrhage, preterm labor, hypertensive disorders of pregnancy, and obstructed labor. When these pregnancy risk factors are identified before labor, women may be referred to appropriate facilities for delivery, and be made aware of signs of complications. Ultrasound, however, is rarely available in rural and resource-constrained areas.
While there is questionable benefit of antenatal screening ultrasound in high-resource urban settings5, limited evaluation has been done in low-resource, rural areas. Importantly, it is unknown whether the prevalence of high-risk features of pregnancy detectable by ultrasound is comparable to those found in the industrialized world. The goal of this study is to demonstrate the prevalence of 3 high-risk obstetric conditions, which include placental abnormality, multiple gestation, and fetal malpresentation, identified by point-of-care ultrasound in rural Uganda, and compare this with established rates in HIC. This comparison will explore the possible impact of screening obstetric ultrasound in a rural, low-resource setting where its occupants must travel a long way for care. With this understanding, future studies can be adequately powered to evaluate programs targeting pregnancies with risk factors and give the best outcomes to mothers and their babies.
Cohort data in this paper are extracted from a pilot study undertaken by ITW (www.imagingtheworld.com), a nongovernmental organization working in Uganda with Ugandan partners, from June 2010 to April 2012. Screening obstetric ultrasound was performed at Nawanyago Health Center III, a private, not for profit, peripheral health facility located 60 miles from Kampala, the capital city of Uganda.
The methodology of the pilot study was previously reported3. In brief, ultrasound services were integrated into existing antenatal care algorithms at NHC. Women presenting to NHC for their first antenatal visit (ANC) during the study period were offered an ultrasound scan; a repeat scan was offered at 32–34 weeks gestation. Only second and third trimester ultrasound scans were considered in this subanalysis because some first trimester abnormalities can resolve as pregnancy progresses. Assessment categories similar to the American College of Radiology BI-RADS categories used by radiologists on a regular basis for breast imaging assessment and management were used for triage6 (Fig. 1). Midwives were trained to generate volume ultrasound sweeps by moving the transducer over the pregnant abdomen using surface anatomical landmarks. These images were then compressed and sent over cellular networks to a trained radiologist for quality assurance. This technique has been validated for quality assurance. During the study period, however, an on-site trained sonographer also scanned each patient using traditional technique. The sonographer was trained through a diploma program and is able to operate independently in Uganda in much the same manner as US trained AIUM sonographers. The data presented in this manuscript is based on the on-site sonographer interpretation only.
Patients were asked to contribute a small scan fee of ∼US$2 to cover operating costs (power, gel, supplies). The fee was determined by the local nurse midwives, and anecdotally there were no barriers to affordability in the community.
IRB approval was obtained from the Mengo Hospital Research Review Committee. Permission was also obtained from the Uganda National Council for Science and Technology as well as the District Health Officer (Kamuli) and the Jinja Diocese. Written informed consent was obtained from all participants with appropriate local language translation. All ultrasound images and patient records were deidentified with the patients receiving an “ITW number” that functioned as a medical record number. All identifying information was kept safely at NHC and was coded anonymously for data analysis. Ultrasound results were recorded in the Uganda Ministry of Health Maternal Passport, which is distributed to all pregnant patients. For ethical reasons, fetal gender determination was not performed.
A retrospective cohort was developed using the dataset to evaluate the prevalence of high-risk obstetric conditions associated with the introduction of antenatal ultrasound scanning at the rural health center.
All analyses were performed using SAS/STAT software, Version 9.4 of the SAS system for Windows. Descriptive statistics were used to summarize the data; the prevalence of prenatal conditions are presented using counts, percentages and 95% confidence intervals, while continuous conditions are summarized using means, median, and ranges.
A total of 1087 women from the age of 14–42 presented for antenatal evaluation at the NHC during the study period (Fig. 2). Among these, 946 had at least 1 antenatal ultrasound during the second or third trimester of pregnancy. Missing data were evident in some data categories and accounts for different denominators used in certain categories. Sixty-one women (6.7%) needed additional imaging for conclusive diagnosis7.
Table 1 presents the baseline characteristics of the study population. The mean and median age was 26 and 25 years, respectively. Most women (62%) had their most recent ultrasound conducted during the third trimester of pregnancy. The mean number of antenatal visits was 3.3 which is on par with Uganda as a whole3.
Table 2 provides the prevalence of high-risk obstetric conditions detected by ultrasound examinations. The most frequent diagnosis was fetal malpresentation in 26% (CI, 22.6–29.3), followed by multiple gestation in 3.9% (CI, 2.6–5.2) of the women. The prevalence of abnormal placental position was 1.7% (97.5–99.2). Malpresentation in the third trimester was noted in 15% (11.7–18.6) of cases.
In rural Uganda where access to emergency obstetric services is often a distance over rough roads, identifying pregnancies with high-risk features is important. Placental abnormalities such as placenta previa, fetal malpresentation, and multiple gestation tangibly contribute to maternal morbidity and mortality.
The prevalence of placenta previa in the United States is up to 4%–6% of pregnancies during the second trimester8; at the time of delivery, however, the prevalence is 4 in 1000 deliveries (0.4%) (https://data.worldbank.org/indicator/SH.STA.MMRT?view=chart). This discrepancy is due to the migration of the placenta away from the cervical os as the uterus expands9. Risk factors for previa include increasing parity, prior uterine surgery including previous cesarean delivery, prior placenta previa and multiple gestations10. In the NHC cohort, the prevalence of abnormal placental location was 1.7%.
The study analyzed abnormal placentation only with regards to placental location and did not differentiate between low lying placenta and placenta previa, which may partly account for the higher prevalence in NHC cohort. The discrepancy between the rate of abnormal placentation in the NHC cohort and published data is also impacted by the gestational age at time of ultrasound. Approximately 40% of the NHC ultrasounds were performed in the second trimester when placental abnormalities are more common. As pregnancy progresses, placenta previa tends to resolve, resulting in the lower prevalence in the published data. Future work will include confirmatory transvaginal ultrasound.
Regardless, if >1% of women in this cohort have suspected abnormal placentation, then universal ultrasound screening for placental location would be beneficial. If ultrasound technology is considered medically necessary in the United States where the prevalence of placenta previa is so low, and most parts of the country have the luxury the latest technology and efficient transportation, then the need for ultrasounds in rural Uganda is even more justified with its higher prevalence rates and limited resources. The perinatal mortality rate attributable to placenta abnormalities is thought to be 3–4 times that of pregnancies without such risk factors11. In a resource constrained setting, emergent transport during labor or delivery is difficult or impossible to access. Early detection would allow referral to a center where surgical delivery and blood products are available, thus curbing maternal-fetal morbidity and mortality.
The rate of multiple pregnancies in the industrialized world has changed over time due in large part to the use of artificial reproductive technologies (ART). Twin births increased 2-fold from 1971 to 2009. In the United States, 36% of twin pregnancies and 77% of higher order multiples are conceived using ART12. In 2015, the twin birth rate in the United States was 33.5 per 1000 births (3.4%) and the rate of triplet and higher-order multiple births was 103.6 per 100,000 total births13 (www.cdc.gov/nchs/fastats/multiple.htm).
In the NHC cohort, the prevalence of multiples closely resembled the rate of the ART-enhanced US data. Only twin gestations were found in this cohort. Although ovulation induction and in vitro fertilization are available in Uganda, such technologies are far from commonplace and are not accessible in rural areas of Uganda. Historical information suggests that global rates of twinning are highest in Africa with the highest numbers seen in Central African Republic14. Genetics are thought to be driving this phenomenon. The same phenomenon is seen in Americans of African descent compared with those of European descent14.
Although advances have been made in decreasing maternal and perinatal mortality in multiple gestations in the United States, complications due to preterm birth, stillbirth and neonatal demise make multiple gestations a high-risk pregnancy15. In 2002 in the United States, 1.6% of singletons delivered before 32 weeks, whereas 11.9% of twins delivered before 32 weeks16. The stillbirth rate in twins is 50% higher than in singleton pregnancies and neonatal mortality is 4–5 times that of a singleton gestation17.
Maternal mortality in twin or higher order multiples is complicated by a 2–3 times higher rate of preeclampsia which is likely to be more severe than in singleton counterparts18. Twin pregnancies also have elevated risks of postpartum hemorrhage (due to atony from the distended uterus) and cesarean delivery. Morbidity and mortality in low and middle income countries associated with hemorrhage are 18% and 3%, respectively19. Timely access to emergency obstetric treatment and blood transfusion can minimize this the associated morbidity and mortality. Given these risks, early identification of a multiple gestation is beneficial to both the mother and the fetuses.
At NHC, midwives typically deliver twins if the presenting twin is cephalic. However, in the obstetric literature vaginal delivery is controversial if the second twin is malpresented; if this is undertaken the center should be able to perform a cesarean section. Indeed, anecdotally, midwives at NHC related that malpresentation in a second twin can often become complicated as there is no access to an operating theater locally. Ultrasound, however, may determine the presentation of the multiples close to the time of delivery. Given the high proportion of twin pregnancies in this cohort, ultrasound is an easy way to ensure a safe delivery.
In the United States, 5% of infants at term have an abnormal lie or presentation10. The most common is breech presentation, accounting for 3%–4% of all deliveries at term. It is estimated that at <28 weeks, 25% of fetuses are breech and 7% at 32 weeks gestation20. As the pregnancy progresses, most fetuses will convert to cephalic presentation. Causes of persistent breech presentation include: prematurity, fetal anomalies such as trisomies, uterine anomalies, and fibroids10.
The NHC data found that 26% of fetuses were malpresented in the second trimester with 15% malpresented in third trimester scans. Overall, it appears that while malpresentation in the second trimester is comparable to US data, malpresentation is >50% more common in the third trimester at NHC. The reason for this is unknown though uterine fibroids, which are more common in those with African ancestry21, is a leading cause of persistent fetal malpresentation.
The optimal mode of delivery for breech presenting pregnancies is debatable in HIC. While it was initially thought that cesarean delivery was the safest method of delivering a breech22, it is now suggested that cesarean and vaginal delivery produce similar outcomes23, but provision for cesarean delivery should be available. Complications of breech vaginal delivery include cord prolapse, head or arm entrapment, and obstructed labor11. Identification of fetal malpresentation with ultrasound allows for delivery preparations to be made. This is essential if transport to a higher level of care needs to be undertaken.
There are certain limitations to address in this context. First, pregnancy dating was based on stated last menstrual period in most cases. Ultrasound dating was performed but generally after 20 weeks gestation; reliable ultrasound dating is difficult after 20 weeks of gestation. Therefore, there may be irregularities in terms of gestational age in the cohort. Midwives have mitigated this by tracking fundal height of those enrolled.
Second, the preponderance of second trimester ultrasounds in the cohort may have inflated the observed proportion of placental abnormalities. In cases of suspected placental abnormality, a mid-third trimester follow-up ultrasound is necessary. In this study protocol, follow-up ultrasounds were obtained at 32 weeks gestation, which may still be too early to be pertinent at the time of delivery. Furthermore, the true diagnosis of a placental abnormality such as previa is by transvaginal ultrasound which was not done in our cohort.
In addition, some women will choose to not seek preventive or necessary care despite knowing of a high-risk feature of pregnancy. Thus increasing access to health care does not guarantee an improved health outcome for the mother or fetus. Travel distance alone can be enough of a deterrent, as well as religious factors, pressures of family, and lack of education on the importance of medical intervention. This, however, does not mean that ultrasound should be withheld for those who may not act on the information.
The aim of this study was to establish the prevalence of 3 high-risk features of pregnancy (placental abnormality, multiple pregnancy, fetal malpresentation) in our cohort. While it is impossible to identify all pregnancies that will end with a complication, these 3 features contribute tangibly to maternal and perinatal morbidity and mortality. Identification of high-risk pregnancies allows for prompt referral to a center with emergency obstetric services, which may decrease maternal morbidity and mortality.
As the prevalence of pregnancy complications in Uganda were similar to or higher than rates found in industrialized countries, it can be extrapolated that their contribution to maternal morbidity and mortality is at least similar, if not greater. This prevalence information can be used to design an appropriately powered study to evaluate changes in maternal and perinatal morbidity with the implementation of ultrasound.
Sources of funding
Supported by the University of Vermont College of Medicine, The Fineberg Foundation, and the Bill & Melinda Gates Foundation.
Conflict of interest disclosure
K.D. is a member of the Philips Healthcare Medical Advisory Board. The remaining authors declare that they have no financial conflict of interest with regard to the content of this report.
The authors acknowledge Imaging the World Africa (Kampala, Uganda), the Ernest Cook Ultrasound Research and Education Institute (Kampala, Uganda) for assistance of the study.
1. Gonzaga MA, Kiguli-Malwadde E, Francis B, et al. Current knowledge, attitudes and practices of expectant women toward routine sonography in pregnancy at Naguru health centre, Uganda
. Pan Afr Med J 2009;3:18–26.
2. Moore CL, Copel JA. Point-of-care ultrasonography. N Engl J Med 2011;364:749–57.
3. Ross AB, DeStigter KK, Rielly M, et al. A low-cost ultrasound program leads to increased antenatal clinic visits and attended deliveries at a health care clinic in rural Uganda
. PloS one 2013;8:e78450.
4. Harris RD, Marks WM. Compact ultrasound for improving maternal and perinatal care in low-resource
settings: review of the potential benefits, implementation challenges, and public health issues. J Ultrasound Med 2009;28:1067–76.
5. Bricker L, Neilson JP. Routine ultrasound in late pregnancy (after 24 weeks gestation). Cochrane Database Syst Rev 2000;2:CD001451.
6. D’Orsi CJ SE, Mendelson EB, Morris EA, et al. ACR BI-RADS® Atlas, Breast Imaging Reporting and Data System. Reston, VA: American College of Radiology; 2013.
7. Ross AB, DeStigter KK, Coutinho A, et al. Ancillary benefits of antenatal ultrasound: an association between the introduction of a low-cost ultrasound program and an increase in the numbers of women receiving recommended antenatal treatments. BMC Pregnancy Childbirth 2014;14:424–32.
8. Faiz AS, Ananth CV. Etiology and risk factors for placenta previa: an overview and meta-analysis of observational studies. Obstet Gynecol 2003;13:175–90.
9. Rosati P, Guariglia L. Clinical significance of placenta previa detected at early routine transvaginal scan. J Ultrasound Med 2000;19:581–5.
10. Gabbe SG, Niebyl JR, Simpson JL. Fifth Edition Obstetrics Normal and Problem Pregnancies. Philadelphia, PA: Elsevier; 2007.
11. Crane JM, van den Hof MC, Dodds L, et al. Neonatal outcomes with placenta previa. Obstet Gynecol 1999;93:541–4.
12. Kulkarni AD, Jamieson DJ, Jones HW Jr, et al. Fertility treatments and multiple births in the United States. N Engl J Med 2013;369:2218–2225.
14. Bulmer MG. The twinning rate in Europe and Africa. Ann Hum Genet 1960;24:121–5.
15. Ananth CV, Joseph Ks K, Smulian JC. Trends in twin neonatal mortality rates in the United States, 1989 through 1999: influence of birth registration and obstetric intervention. Am J Obstet Gynecol 2004;190:1313–21.
17. Pharoah PO, Cooke T. Cerebral palsy and multiple births. Arch Dis Child Fetal Neonatal Ed 1996;75:F174–7.
18. Day MC, Barton JR, O’Brien JM, et al. The effect of fetal number on the development of hypertensive conditions of pregnancy. Obstet Gynecol 2005;106 (pt 1):927–31.
19. Sheldon WR, Blum J, Vogel JP, et al. Postpartum haemorrhage management, risks, and maternal outcomes: findings from the World Health Organization Multicountry Survey on Maternal and Newborn Health. BJOG 2014;121 (suppl 1):5–13.
20. Collea JV. Current management of breech presentation. Clin Obstet Gynecol 1980;23:525–31.
21. Marshall LM, Spiegelman D, Barbieri RL, et al. Variation in the incidence of uterine leiomyoma among premenopausal women by age and race. Obstet Gynecol 1997;90:967–73.
22. Hannah ME, Hannah WJ, Hewson SA, et al. Planned caesarean section versus planned vaginal birth for breech presentation at term: a randomised multicentre trial. Term Breech Trial Collaborative Group. Lancet 2000;356:1375–83.
23. Whyte H, Hannah ME, Saigal S, et al. Outcomes of children at 2 years after planned cesarean birth versus planned vaginal birth for breech presentation at term: the International Randomized Term Breech Trial. Am J Obstet Gynecol 2004;191:864–71.