Hypertensive disease in pregnancy in Botswana: prevalence and impact on perinatal outcomes

Abstract

Objectives:

Perinatal morbidity in sub-Saharan Africa has been attributed to infection, obstetric emergencies, and preterm birth, but less is known about hypertension in pregnancy. Our objective was to characterize the prevalence of hypertension in pregnancy and the impact of hypertension on perinatal outcomes in sub-Saharan Africa.

Study Design:

We performed surveillance of obstetric records at eight of the largest public hospitals in Botswana. Women were included in this analysis if they were HIV-uninfected and had singleton gestations and at least one prenatal blood pressure measurement.

Main Outcome Measures:

We measured stillbirth, preterm birth, small for gestational age, and neonatal death in women with and without hypertension in pregnancy.

Results:

We included 14,170 pregnancies. Hypertension occurred in 3,156 (22.2%) women, with 602 (19.1%) defined as severe. Severe hypertension increased risk of stillbirth (RR 4.4; 95% CI 3.2–6.2), preterm birth (RR 2.5; 95% CI 2.2–2.8), small for gestational age (RR 2.7; 95% CI 2.3–3.1) and neonatal death (RR 5.1; 95% CI 2.9–5.6). Non-severe hypertension increased risk of stillbirth (RR 2.0; 95% CI 1.5–2.7), preterm birth (RR 1.2; 95% CI 1.1–1.3), and small for gestational age (RR 1.6; 95% CI 1.4–1.8). Perinatal outcomes were worse in women with hypertension who had spontaneous preterm birth compared to those who underwent iatrogenic preterm delivery.

Conclusions:

Hypertension in pregnancy is common in Botswana and leads to a large number of adverse outcomes. Improved management of hypertension in pregnancy may improve perinatal morbidity and mortality.

INTRODUCTION

Hypertensive disease in pregnancy significantly impacts maternal and perinatal outcomes worldwide, particularly in low and middle income countries (LMIC) [1]. A majority of adverse perinatal outcomes in LMICs are thought to be due to non-hypertensive etiologies, including infection [2–4], obstructed labor [5,6], and preterm or postterm pregnancy [3,7]. However, approximately one quarter of perinatal deaths are attributed to preeclampsia or eclampsia in LMICs [8] and infants born to mothers with preeclampsia have a three-fold higher rate of perinatal mortality in low income countries than in high income countries [1] This disparity is likely due to limitations in neonatal care for preterm infants and limited access to emergency obstetric services and close fetal monitoring to prevent stillbirth [6,9].

While eclampsia and preeclampsia have been identified as major causes of perinatal mortality worldwide, they are relatively uncommon, affecting approximately 4% of pregnant women in sub-Saharan Africa [10]. In contrast, non-preeclamptic hypertension is far less studied in LMICs, especially in sub-Saharan Africa where the attributable risk of chronic hypertension on stillbirth is highest [3]. Given that non-preeclamptic hypertension is far more common [11], may be difficult to identify and manage without strong primary health systems [12] and may have a large overall impact on perinatal outcomes [2,13], a better overall understanding of its impact on perinatal outcomes is needed in a region that ranks highest in perinatal mortality (28 deaths per 1000 births) [14].

Botswana represents a unique setting to study hypertension in pregnancy, as 99% of women deliver in an institutional setting [15], the majority of women receive antenatal care [2], and anti-hypertensive medications are widely available [16]. In this study, we document the prevalence of hypertensive disease in a nationally representative sample of pregnancies among HIV-uninfected women in Botswana and describe the impact of hypertensive disease on perinatal outcomes.

METHODS

We performed prospective surveillance of obstetric records at eight of the largest public hospitals in Botswana: Princess Marina Hospital in Gaborone, Nyangabgwe Hospital in Francistown, Maun General Hospital in Maun, Sekgoma Memorial Hospital in Serowe, Scottish Livingstone Hospital in Molepolole, Selebi-Phikwe Government Hospital in Selebi-Phikwe, Mahalapwe Hospital in Mahalapye, and Ghanzi Primary Hospital in Ghanzi. All women who delivered at these hospitals during the surveillance period were included in this analysis if they were HIV-uninfected, had singleton gestations, and had at least one antenatal visit with a blood pressure (BP) measurement. Women who presented in pregnancy with a history of hypertension, but who did not develop elevated blood pressure during that pregnancy, were excluded from the analysis to avoid misclassification of women with chronic hypertension as normotensive. We restricted our analysis to HIV-uninfected women to allow a targeted investigation of hypertension, as prior studies in Botswana have suggested potential confounding of the relationship between hypertension and perinatal outcomes by maternal immune status and possibly antiretroviral therapy [2,17,18].

Standard maternal demographic and medical information was abstracted from the maternal obstetric record at the time of discharge from the postpartum ward. We defined booking weight as the first weight recorded in the antenatal record. We used booking weight greater than 90kg to approximate obesity. While pre-pregnancy body mass index would be more accurate, data on pre-pregnancy weight and height was unavailable for most women and prior studies have shown an association between booking weight itself and adverse perinatal outcomes [19]. Weight gain during pregnancy, defined as the last weight in the antenatal record minus the first weight in the antenatal record, was considered high if greater than 20kg. Factors relating to healthcare delivery, including gestational age at first antenatal care visit, number of antenatal care visits, hospital of delivery, and maternal history of preterm birth or stillbirth also were collected. Gestational age recorded by the delivering midwife was used for all analyses, and was most often based on the patient’s last menstrual period. We defined hypertension as at least one systolic BP (SBP) ≥140 or diastolic BP (DBP) ≥90 and severe hypertension as SBP ≥160 or DBP ≥110 during pregnancy. Women were considered to have chronic hypertension if they had a reported history of chronic hypertension, presented to antenatal care on anti-hypertensive medication, or developed hypertension before 20 weeks of gestation.

We assessed the risk of hypertension on the incidence of stillbirth (Apgar of 0 at 1, 5, and 10 minutes), preterm birth (<37 weeks of gestation), small for gestational age (SGA) (<10^th percentile of weight for gestational age based on standard World Health Organization (WHO) curves), admission to the neonatal unit, and neonatal death (in-hospital death within 28 days of life). Neonatal units in Botswana often are equipped with incubators, but rarely have ventilators or one-on-one nursing care. While not neonatal intensive care units, admission to these units represents an escalation of care. Preterm birth was defined as iatrogenic if the labor was induced or if the birth occurred by way of a planned cesarean section. Descriptive data are presented as proportion or median (interquartile range), due to a non-normal distribution. Categorical data were compared using the chi-square test, and continuous variables were compared using the Wilcoxon signed-rank test. Modified Poisson regression was used to calculate risk ratios (RR) and 95% confidence intervals (CI) [20]. We adjusted for factors that have previously been identified as confounders in the published literature.

Approval for this study was obtained from the Human Research Development Council in Botswana and the Harvard T.H. Chan School of Public Health institutional review board in Botswana.

Role of the funding source

The study sponsor did not have any role in the study design, analysis or interpretation of data, the writing of the report, or the decision to submit the paper for publication.

RESULTS

From August 2014 through July 2015, 14,305 HIV-uninfected women registered for antenatal care and delivered a singleton infant at one of the 8 sites included in this study, representing approximately 44% of all deliveries in Botswana to HIV-uninfected women receiving antenatal care [14]. We excluded 135 women who had a history of any type of hypertension but did not develop hypertension during the current pregnancy; thus, the included population was 14,170 women. The median maternal age at delivery was 25 (21.0–30.0) and the median number of antenatal care visits was 10 (7.0–12.0) (Table 1). The majority of women were citizens of Botswana, unmarried, non-smokers, and gained more than 20kg during pregnancy (Table 1). Women with hypertension were slightly older, and more likely to be married, multiparous, and obese than normotensive women (Table 1). While the proportion of women with a history of stillbirth and preterm birth was overall low, a higher proportion of women with hypertension compared with normotensive women had a history of stillbirth or preterm birth.

Table 1:

Maternal characteristics

	Total n=14,170	Hypertensive n=3,156	Normotensive n=11,014
Maternal age	24.0 (21.0–29.0)	26.0 (22.0–32.0)	24.0 (21.0–29.0)
Married	1,494 (10.5)	453 (14.4)	1,041 (9.5)
Secondary school education or higher	13,035 (92.0)	2,914 (92.3)	10,121 (91.9)
Salaried employment	8,907 (62.9)	1,860 (58.9)	7,047 (64.0)
Citizen of Botswana	13,712 (96.8)	3,005 (95.2)	10,707 (97.2)
Nulliparous	6651 (46.9)	1,395 (44.2)	5,256 (47.7)
Number of antenatal care visits	10.0 (7.0–12.0)	11.0 (8.0–13.0)	10.0 (7.0–12.0)
History of preterm birth
Yes	477 (3.4)	146 (4.6)	331 (3.0)
No	7,240 (51.1)	1,663 (52.7)	5,577 (50.6)
Unknown	6,453 (45.5)	1,347 (42.7)	5,106 (46.4)
History of stillbirth
Yes	318 (2.2)	98 (3.1)	220 (2.0)
No	7,462 (52.7)	1,721 (54.5)	5,741 (52.1)
Unknown	6,390 (45.1)	1,337 (42.4)	5,053 (45.9)
Booking weight > 90 kg
Yes	923 (8.5)	456 (14.5)	467 (4.2)
No	13,197 (93.1)	2,686 (85.1)	10,501 (95.3)
Unknown	60 (0.4)	14 (0.4)	46 (0.4)
Weight gain during pregnancy > 20 kg
Yes	12,754 (90.0)	2,828 (89.6)	9,926 (90.1)
No	467 (3.3)	129 (4.1)	338 (3.1)
Unknown	949 (6.7)	199 (6.3)	750 (6.8)
Smoking during pregnancy
Yes	158 (1.1)	38 (1.2)	120 (1.1)
No	13,485 (95.2)	3,006 (95.3)	10,479 (95.1)
Unknown	527 (3.7)	112 (3.6)	415 (3.8)

Data presented as median (interquartile range) or n (%)

Proportions may not sum to 100% due to rounding

Prevalence of hypertension in pregnancy

Overall, 3,156 (22.2%) women had hypertension during pregnancy; the majority (n=2477, 78.5%) developed hypertension after 20 weeks of gestation, while the remaining 21.5% (n=679) had chronic hypertension. Among women with hypertension, 602 (19.1%) had severe hypertension; the majority (n=505, 83.9%) of these women had chronic hypertension. Data on urine protein were only collected for 10% of women with hypertension, limiting our ability to assess for preeclampsia.

Incidence of adverse perinatal outcomes

Overall, 2,294 (16.2%) babies were born preterm, 1,510 (10.7%) were SGA, 281 (2.0%) were stillborn, and 136 (1.0%) died in the first 28 days. Women with severe hypertension experienced the highest incidence of poor perinatal outcomes, though women with non-severe hypertension also experienced a higher incidence of adverse perinatal outcomes than normotensive women (Figure 1).

Figure 1:

Incidence of adverse perinatal outcomes, comparing normotensive, non-severe hypertensive and severely hypertensive women

After adjusting for maternal age, marital status, salaried employment, and obesity, women with any hypertension were more than twice as likely as normotensive women to deliver stillborn infants (RR: 2.5; 95% CI: 1.9–3.1) and have an infant admitted to the neonatal unit (RR: 2.4; 95% CI: 1.9–3.0). Women with hypertension also were more likely than normotensive women to deliver preterm (RR: 1.4; 95% CI: 1.3–1.6), deliver an SGA infant (RR: 1.8; 95% CI: 1.6–2.0), and experience neonatal death (RR: 1.6; 95% CI: 1.1–2.3). When stratified by severity of hypertension, the adjusted risk ratios were significantly elevated for all adverse outcomes among women with severe hypertension compared to normotensive women. Similar results were seen for all outcomes except neonatal death when comparing women with non-severe hypertension to normotensive women; however, the magnitude of effect was less than among women with severe hypertension (Table 2). Chronic hypertension conferred a particularly high risk of stillbirth (RR 2.8; 95%CI 2.7–5.4) and neonatal death (RR 2.4; 95%CI 1.4–4.2) compared with women without any hypertension and those with only gestational hypertension (Table 3).

Table 2:

Risk of adverse perinatal outcomes among hypertensive women compared with normotensive women

	Any hypertension		Non-severe hypertension		Severe hypertension
	Crude RR (95% CI)	Adjusted* RR (95% CI)	Crude RR (95% CI)	Adjusted* RR (95% CI)	Crude RR (95% CI)	Adjusted* RR (95% CI)
Stillbirth	2.5 (2.0–3.2)	2.5 (1.9–3.1)	2.0 (1.5–2.6)	2.0 (1.5–2.7)	4.7 (3.4–6.6)	4.4 (3.2–6.2)
Preterm birth	1.4 (1.3–1.5)	1.4 (1.3–1.6)	1.2 (1.0–1.3)	1.2 (1.1–1.3)	2.3 (2.1–2.6)	2.5 (2.2–2.8)
SGA	1.7 (1.5–1.8)	1.8 (1.6–2.0)	1.5 (1.3–1.7)	1.6 (1.4–1.8)	2.5 (2.1–2.9)	2.7 (2.3–3.1)
Neonatal death	1.8 (1.2–2.5)	1.6 (1.1–2.3)	1.3 (0.8–1.9)	1.2 (0.8–1.9)	4.0 (2.5–6.6)	5.1 (2.9–5.6)
NNU admission	2.4 (2.0–3.1)	2.4 (1.9–3.0)	2.0 (1.6–2.6)	2.0 (1.6–2.6)	4.3 (3.1–5.9)	3.5 (2.2–5.8)

^*models adjusted for maternal age, marital status, salaried employment, and booking weight

SGA: small for gestational age; NNU: neonatal unit

Table 3:

Risk of adverse perinatal outcomes, stratified by gestational and chronic hypertension, comparing infants born to hypertensive versus normotensive mothers

	Gestational hypertension		Chronic hypertension
	Crude RR (95% CI)	Adjusted* RR (95% CI)	Crude RR (95% CI)	Adjusted* RR (95% CI)
Stillbirth	2.2 (1.7–2.8)	2.1 (1.6–2.7)	3.9 (2.8–5.5)	2.8 (2.7–5.4)
Preterm birth	1.2 (1.1–1.3)	1.2 91.1–1.4)	2.1 (196–2.4)	2.3 (2.1–2.6)
SGA	1.7 (1.5–1.8)	1.8 (1.6–2.0)	1.7 (1.4–2.1)	1.9 (1.6–2.3)
Neonatal death	1.5 (1.0–2.3)	1.5 (0.98–2.2)	2.6 (1.5–4.5)	2.4 (1.4–4.2)
NNU Admission	2.1 (1.9–2.7)	2.1 (1.6–2.7)	3.6 (2.6–5.0)	3.6 (2.6–5.1)

^*models adjusted for maternal age, marital status, salaried employment, and booking weight

SGA: small for gestational age; NNU: neonatal unit

Among the women with hypertension who delivered preterm (n=649), 43.6% (n=283) delivered before 34 weeks of gestation and 56.4% (n=366) delivered from 34 to 37 weeks of gestation. Results were similar among women with severe hypertension (n=209); 49.8% (n=104) delivered preterm prior to 34 weeks of gestation, and 50.2% (n=105) delivered from 34 to 37 weeks. When stratifying by gestational age at delivery, risk ratios for all adverse perinatal outcomes were similar between preterm and term births, with the exception of SGA, which was more than twice as high among the preterm births (Table 4).

Table 4:

Risk of adverse perinatal outcomes, stratified by preterm and term birth, comparing infants born to hypertensive versus normotensive mothers

	Preterm birth		Term birth
	Crude RR (95% CI)	Adjusted* RR (95% CI)	Crude RR (95% CI)	Adjusted* RR (95% CI)
Stillbirth	2.1 (1.6–2.8)	1.9 (1.5–2.5)	1.8 (1.1–2.9)	1.7 (1.1–2.7)
SGA	3.1 (2.5–3.7)	2.8 (2.3–3.4)	1.3 (1.2–1.5)	1.4 (1.3–1.6)
Neonatal death	1.4 (1.0–2.1)	1.2 (0.78–1.7)	1.6 (0.8–3.2)	-
NNU admission	2.0 (1.5–2.6)	1.8 (1.4–2.4)	2.0 (1.3–3.0)	1.9 (1.3–2.9)

^*models adjusted for maternal age, marital status, salaried employment, and booking weight

SGA: small for gestational age; NNU: neonatal unit

Among women with hypertension who delivered preterm, 483 (74.4%) were spontaneous and 163 (25.1%) were iatrogenic. In contrast, only 10.8% (n=177) of the normotensive preterm births were iatrogenic. After controlling for maternal age, women with hypertension who had iatrogenic preterm deliveries were significantly less likely to have an infant who was stillborn (RR: 0.40; 95% CI: 0.27–0.59) or admitted to the neonatal unit (RR: 0.39; 95% CI: 0.27–0.57) compared with women with hypertension who spontaneously delivered preterm. These same infants were also less likely to be SGA (RR: 0.87; 95% CI: 0.67–1.12) but more likely to die within 28 days of life (RR: 2.0; 95% CI: 0.79–4.7) compared to those borne to mothers who did not undergo iatrogenic preterm delivery, though these risk ratios were not statistically significant.

DISCUSSION

To our knowledge, this is the largest study to date to evaluate the epidemiology of hypertensive disorders in pregnancy and adverse perinatal outcomes in sub-Saharan Africa. We found that hypertension and severe hypertension in pregnancy are common in Botswana. We also found that women with chronic hypertension were more likely to develop severe hypertension than women with gestational hypertension. Infants borne to mothers with hypertension in this study were at greatly increased risk of stillbirth, SGA, preterm birth and neonatal death, with the highest risk occurring among those infants whose mothers had severe hypertension. Risk of stillbirth, SGA, and neonatal death was increased among hypertensive women who delivered at term and those who delivered preterm, indicating that hypertension is a risk factor for these outcomes independent of gestational age at delivery. Finally, spontaneous preterm birth was associated with worse perinatal outcomes than iatrogenic preterm birth in this cohort.

Compared with the United States and other high-income countries where hypertensive disease complicates 5–10% of pregnancies [21,22], we found a markedly higher prevalence of hypertension in Botswana. The prevalence of hypertension in pregnancy was higher than what has been seen in women from Tanzania and Zambia (1.7%) [23], Zimbabwe (19.4%) [24], and South Africa (12–18%) [11,25]. Differences in hypertension between Botswana and other countries do not appear to be due to a higher prevalence of traditional risk factors, as the proportion of women with history of tobacco use and/or obesity was lower than in the US [21] and South Africa [26]. Differences from other African studies could be due to our improved ability to measure hypertension in Botswana in this large, nationally-representative population, where most women access antenatal clinic care. The higher prevalence of hypertension also might be due to our definition, which required only one elevated blood pressure measurement in pregnancy; we believe, however, that it is not an overly sensitive definition for this condition. Genetic or other unmeasured differences between populations also are possible explanations for the higher prevalence of hypertension in Botswana.

Previous studies have identified hypertension as a major contributing factor to maternal and infant mortality in other regions of the world [27,28], but our findings highlight its importance to perinatal mortality in sub-Saharan Africa. While few studies in sub-Saharan Africa have examined hypertension in pregnancy, two African studies that specifically evaluated non-proteinuric hypertension did not find hypertension to be significantly associated with perinatal mortality [13,29]. This may be due to under diagnosis of gestational hypertension in the other studies, which lacked information on all elevated blood pressures in pregnancy [21]. Our findings could be due to our inability to evaluate urine protein and exclude women with preeclampsia. However, in practice, preeclampsia is difficult to diagnose in settings that lack consistent access to monitoring urine protein and the ability to rapidly obtain important bloodwork. Our results suggest that antenatal care programs in resource-limited settings must focus on improving treatment and monitoring of all women with hypertension and not just women with preeclampsia or eclampsia.

A concerning result from our analysis is the increased risk for both stillbirth and neonatal death among women with any hypertension regardless of gestational age at delivery. Hypertensive disease can lead to stillbirth through various mechanisms, primarily through placental insufficiency and abruption mediated by severely elevated blood pressures [1], but our finding that the risk stillbirth is increased even for women with non-severe hypertension adds to mounting evidence in sub-Saharan Africa that attending to hypertensive disease overall is critical for reaching worldwide goals to decrease stillbirth [3,24]. Neonatal death is typically associated with prematurity, but the increased risk of neonatal death among those with chronic or severe hypertension, even among term deliveries, suggests that hypertension itself also may impact neonatal survival in ways not yet understood

Hypertension is commonly associated with preterm birth [1], and we expected this would be due to iatrogenic preterm delivery for severe hypertension in this population. We were surprised by the high incidence of spontaneous preterm delivery among women with hypertension in Botswana, and the mechanisms responsible for this deserve further study.

This study has several strengths including the large, nationally-representative population, prospective data collection that included all blood pressure measurements in pregnancy in a population where over 99% of women deliver at health care facilities. This study also has several limitations. First, the data were abstracted from medical records, which are prone to recording errors. Nevertheless, we expect this would lead to non-differential misclassification as the data on hypertension were collected prospectively without knowledge of the outcome. In addition, the gestational age was an estimate determined either from reported last menstrual period or by clinical estimation at the time of presentation to antenatal care. While many women had ultrasound this was rarely performed in the first trimester, which would be the gold standard [30]. While there has been some association between increased adverse perinatal outcomes and uncertain last menstrual period [31], we did not expect this effect to differ with respect to hypertension. We also excluded women with HIV from the analysis, a purposeful exclusion to explore the role of hypertension without the confounding effect of HIV positivity and antiretroviral use. Nevertheless, a quarter of the pregnant population in Botswana is HIV-infected and thus further analysis on the role of hypertension on adverse outcomes in this group is of great public health significance. Finally, we chose to study all hypertension rather than exclusively preeclampsia given the difficulties of defining and diagnosing preeclampsia during outpatient prenatal care in Botswana. Many clinics lack laboratory testing (hemoglobin, platelets, liver function, kidney function and urine protein), and complete documentation of clinical exam, particularly the neurological exam, is rare. While this is a significant limitation of our study design, it also reflects the clinical reality of many resource-limited settings.

In conclusion, our study highlights the high prevalence of hypertension in pregnancy and its association with poor perinatal outcomes among HIV-uninfected women in Botswana. We believe detailed investigation into the extent of hypertension in pregnancy is warranted elsewhere in sub-Saharan Africa to determine the generalizability of these findings, as improved management of hypertension in pregnancy may be needed to reach world-wide goals of reducing perinatal morbidity and mortality.