A C TA Obstetricia et Gynecologica
AOGS M A I N R E SE A RC H A R TI C LE
Reducing maternal mortality from preeclampsia and eclampsia in low-resource countries – what should work? ROBERT L. GOLDENBERG1, BONNIE JONES2, JENNIFER B. GRIFFIN2, DORIS J. ROUSE2, BEENA D. KAMATH-RAYNE3, NEHAL TRIVEDI2 & ELIZABETH M. MCCLURE2 1
Department of Obstetrics and Gynecology, Columbia University, New York, NY, 2Social, Statistical and Environmental Sciences, RTI International, Durham, NC, and 3Perinatal Institute, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA
Key words Preeclampsia, eclampsia, maternal mortality, magnesium sulfate, sub-Saharan Africa Correspondence Robert L. Goldenberg, Department of Obstetrics & Gynecology, Columbia University Medical Center, 622 W. 168th Street, PH 16-66, New York, NY 10032, USA. E-mail: [email protected] Conflict of interest The authors have stated explicitly that there are no conflicts of interest in connection with this article. Please cite this article as: Goldenberg RL, Jones B, Griffin JB, Rouse DJ, Kamath-Rayne BD, Trivedi N, et al. Reducing maternal mortality from preeclampsia and eclampsia in low-resource countries – what should work? Acta Obstet Gynecol Scand 2015; 94: 148– 155. Received: 4 May 2014 Accepted: 13 October 2014 DOI: 10.1111/aogs.12533
Abstract Objective. Preeclampsia/eclampsia (PE/E) remains a major cause of maternal death in low-income countries. We evaluated interventions to reduce PE/Erelated maternal mortality in sub-Saharan Africa. Design. Mathematical model to assess impact of interventions on PE/E-related maternal morbidity and mortality. Setting. Sub-Saharan Africa countries. Population. Pregnant women in sub-Saharan Africa in 2012. Methods. A systematic literature review populated a decision-tree mathematical model with interventions to diagnose, prevent, and treat women with PE/E. The impact of increased use of interventions [diagnostics, transfer to a hospital, magnesium sulfate (MgSO4) use, cesarean section/ labor induction] on PE/E-related maternal mortality was analyzed. Main outcome measures. Prevalence of PE/E and PE/E-associated maternal mortality rates in sub-Saharan Africa. Results. Without interventions, an estimated 20 570 PE/ E-associated deaths would have occurred in sub-Saharan Africa in 2012. With current low rates of diagnosis, MgSO4 use, transfers and cesarean section/induction rates, about 17 520 maternal deaths were associated with PE/E in 2012. Higher use of MgSO4 would have prevented about 610 deaths. With high diagnostic levels, MgSO4 use, transfer and cesarean section/induction, mortality was reduced to 3750 annual deaths, saving about 13 770 maternal lives. If all MgSO4 use was removed from the model, 4060 maternal deaths would occur, increasing maternal deaths by only 310. Conclusions. In sub-Saharan Africa, our model suggests that increasing use of PE/E diagnostics, transfer to higher levels of care and increased hospitalization with cesarean section/induction of labor would substantially reduce maternal mortality from PE/E. Increasing use of MgSO4 would have a smaller impact on maternal mortality. HIC, high-income countries; LIC, low-income countries; MANDATE, Maternal and Neonatal Directed Assessment of Technology model; MgSO4, magnesium sulfate; PE/E, preeclampsia/eclampsia; SSA, subSaharan Africa. Abbreviations:
Introduction Preeclampsia/eclampsia (PE/E) is one of the major causes of maternal mortality in low-income countries (LIC), the leading cause of stillbirth, and an important contributor
148
to neonatal mortality (1–3). PE/E occurs in about 3–6% of all pregnancies. Large variations in the reported prevalence of preeclampsia occur worldwide (4,5); these variations may reflect actual differences, or differences in definitions or reporting practices. Untreated, between 2 and 10% of women with preeclampsia will develop
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
R.L. Goldenberg et al.
eclampsia; most seizures occur in women with severe preeclampsia. In LIC, risk of mortality with seizures, and especially with multiple seizures, is as high as 25–50% (6– 8). In high-income countries (HIC) the progression from preeclampsia to eclampsia and the case fatality rates from both conditions have substantially declined over time; there are now large differences between LIC and HIC in these outcomes. We have previously demonstrated that the more than 90% reduction in PE/E-associated maternal mortality in HIC from the early 1900s to about 1980 is explained by a reduction in the progression of preeclampsia to eclampsia, together with the reduction in case-fatality rates associated with eclampsia (7). PE/E is thought to account for about 15% of maternal mortality world-wide, 20–25% of fetal mortality, and with increased risk of asphyxia and preterm delivery for as much as 25% of neonatal mortality (1,2,8–10). While the impact of PE/E on adverse pregnancy outcomes has been much reduced in HIC since about 1935, there has been a much smaller reduction in these outcomes in LIC (7). Currently, about 98% of maternal, fetal, and neonatal mortality related to PE/E occurs in LIC, with much of that mortality in south Asia and sub-Saharan Africa (SSA) (10). A number of strategies have been advocated to reduce the risk of progression of preeclampsia to eclampsia as well as the risk of death from the condition. Diagnosis of preeclampsia with blood pressure and proteinuria determinations and timely delivery through cesarean section or induction of labor are both important in reducing PE/E-related maternal mortality (7,11–13). Magnesium sulfate (MgSO4) is one of the most common interventions proposed to delay progression of preeclampsia to eclampsia and reduce maternal mortality (14–20). To better understand the interventions most likely to reduce PE/E-associated maternal mortality in SSA, we created a model for this condition. Our objective was to examine maternal mortality rates associated with PE/E and the impact of interventions designed to reduce those rates.
Preeclampsia/eclampsia in low-income countries
MANDATE estimates the proportion of women treated in the home, clinic, and hospital setting as well as the impact of various interventions in each setting. Each intervention, whether existing or proposed, is characterized by its efficacy in diagnosing, preventing or treating a condition as well as its penetration (availability in a given setting) and utilization (the rate at which it is used correctly) in the home, clinic or hospital (22). In these analyses, interventions studied had evidence for substantially reducing maternal mortality and included improved diagnosis of preeclampsia (blood pressure and proteinuria determination), transfer of women with PE/E to a higher level of care, greater use of MgSO4 to prevent the first and subsequent seizures, and increased use of cesarean section or labor induction to deliver mothers diagnosed with PE/E. We first evaluated a hypothetical scenario in which no interventions were available, then a baseline scenario considering current use of interventions, as well as hypothesized scenarios where the penetration, utilization and, if appropriate, efficacy of various interventions were altered. We also assessed the impact of transfers from the home to clinic or hospital or from the clinic to the hospital. The numbers that populate the model are derived from numerous sources. A systematic literature review was conducted which included all English literature in PubMed since 1980, the Cochrane Library and the WHO database. For the literature review, we used the search terms “pregnancy”, “hypertension”, “preeclampsia”, “eclampsia, and “magnesium sulfate”. Additionally, published papers and reports including the Demographic and Household Surveys and United Nations data provided the rates of birth and utilization of various interventions and mortality rates. A modified algorithm using the GRADE system (Grading of Recommendations Assessment, Development and Evaluation) was applied, which used the estimates from the highest quality sources as primary, with support from other sources where no other data were available (23). We restricted this analysis to SSA, which covers countries in the World Bank’s SSA Region stretch-
Material and methods To help define which interventions would have the largest impact on maternal, fetal, and neonatal mortality in LIC, a model – Maternal and Neonatal Directed Assessment of Technology (MANDATE) – was created (21). This model uses decision tree and mathematical modeling to evaluate maternal, fetal, and neonatal mortality rates and the impact of interventions on these outcomes. For this study, we modeled the maternal mortality associated with PE/E in SSA. The MANDATE version used in this analysis is posted online (http://mnhtech.org).
Key Message A prenatal care program consisting of testing for hypertension and proteinuria, and increased use of hospitalization for cesarean section/induction of labor would eliminate most maternal deaths associated with preeclampsia/eclampsia. MgSO4 use appears to have a much smaller impact on preclampsia/eclampsia-associated maternal mortality in these settings.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
149
Preeclampsia/eclampsia in low-income countries
R.L. Goldenberg et al.
ing from South Africa to Chad and Sudan in the north (24). Table 1 defines the key variables for the analysis. These include the number of births in SSA in 2012, the estimated overall incidence of preeclampsia, divided into mild/moderate and severe categories (25,26). Also included is the percent of women with untreated mild/ moderate and severe preeclampsia who die without progressing to eclampsia, the percent of women with mild/ moderate and severe preeclampsia who progress to eclampsia, and the case-fatality rates for women with untreated eclampsia. Table 2 lists the potentially important interventions related to preeclampsia and eclampsia included in MANDATE. These include diagnosis of preeclampsia by blood pressure and proteinuria, transfer to a higher level of care, treatment with MgSO4, and delivery by cesarean section/labor induction. Interventions are modeled using the constructs of penetration, utilization, and efficacy. Because there is evidence that various interventions might be used more frequently under certain conditions, the model may include a range of utilization rates. For example, women diagnosed with hypertension have an increased likelihood to receive a test for proteinuria, compared with women not diagnosed with hypertension (27). Efficacy was defined as the ability to prevent or diagnose a condition or decrease mortality due to that condition. MgSO4 was modeled with an efficacy to reduce eclampsia by 65% in women with mild/moderate preeclampsia without a prior seizure, an efficacy of 55% to prevent eclampsia in women with severe preeclampsia and a similar efficacy to reduce further seizures in women already having a seizure. Any impact of MgSO4 on maternal death had to result from a reduction in eclampsia.
Table 1. Of the estimated 33 000 000 pregnancies ending at 20 weeks or more in sub-Saharan Africa, 2012, estimates are presented of the incidence and case fatality rates for preeclampsia and eclampsiaa. Conditions
Rates
Preeclampsia –total incidence Preeclampsia –incidence severe Preeclampsia –incidence mild/moderate Rate severe preeclampsia progresses to eclampsia Rate mild/moderate preeclampsia progresses to eclampsia Case-fatality rates Mild/moderate preeclampsia not progressing to eclampsia Severe preeclampsia not progressing to eclampsia Eclampsia: one seizure Eclampsia: two or more seizures
5% 1.5% 3.5% 10% 0.2% 0.3% 0.7% 25% 50%
a The incidence and case-fatality rates assume no preventive or treatment interventions.
150
Table 2. Interventions for preeclampsia/eclampsia and estimated baseline rates of penetration, utilization and efficacy in the home, clinic and hospital in sub-Saharan Africa, 2012. Setting Home Diagnostics Blood pressure measurement Penetration 15% Utilization (if no proteinuria 20% diagnostic) Utilization (if proteinuria diagnostic 40% is positive) Efficacy to detect high blood 75% pressure Proteinuria Penetration 0% Utilization (if no blood pressure 0% diagnostic) Utilization (if blood pressure 0% diagnostic is positive) Efficacy to detect proteinuria 85% Prevention/treatment Antihypertensives Penetration 0% Utilization 0% Efficacy 0% Magnesium sulfate (MgSO4) Penetration 0% Utilization (mild to moderate 0% preeclampsia, or undiagnosed severe preeclampsia) Utilization (diagnosed severe 0% preeclampsia) Utilization (eclampsia) 0 Efficacy to prevent seizure (mild/ 65% moderate preeclampsia) Efficacy to prevent seizure (severe 55% preeclampsia or recurrent eclamptic seizures) Cesarean section/induction for prevention/treatment death Penetration 0% Utilization (mild to moderate 0% preeclampsia) Utilization (severe preeclampsia/ 0% eclampsia) Efficacy (mild/ moderate preeclampsia) 75% Efficacy (for severe preeclampsia/ 95% eclampsia) Transfer (for mild/moderate preeclampsia) Home to clinic 35% Home to hospital 10% Clinic to hospital 20% Transfer (for severe preeclampsia and eclampsia) Home to clinic 35% Home to hospital 20% Clinic to hospital 30%
Clinic
Hospital
85% 60%
85% 70%
85%
95%
75%
75%
60% 40%
90% 45%
75%
85%
85%
85%
25% 45% 0%
70% 65% 0%
15% 30%
50% 50%
40%
60%
50% 65%
75% 65%
55%
55%
of eclampsia/ 15% 30%
75% 40%
60%
80%
75% 95%
75% 95%
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
Preeclampsia/eclampsia in low-income countries
R.L. Goldenberg et al.
Furthermore, to discriminate the reduction in mortality, we have defined a range of efficacy rates for the same intervention (such as cesarean section) when given for different indications. As an example, we estimated that cesarean section/labor induction will have a greater efficacy to prevent maternal death in women with severe preeclampsia than with mild/moderate preeclampsia. As shown in Table 2, currently, the estimated penetration and utilization of diagnostics, MgSO4 and cesarean section or labor induction for delivery are generally low in SSA, with substantially lower rates in clinic and home compared to hospital.
Results Using the MANDATE model, we ran 10 scenarios for SSA (Table 3). The columns on the left present the interventions and the next to last column on the right presents the estimated maternal mortality. The last column provides the estimated number of lives saved compared with the baseline treatment scenario. Scenario 1 presents maternal mortality estimates if no interventions related to PE/E were available. This scenario suggests that without any treatment, there would be 20 570 maternal deaths from PE/E. Scenario 2 suggests that with the current treatments, including relatively low levels of use of preeclampsia diagnostics, MgSO4, cesarean section/induction and transfer of women with PE/E to a hospital, there were 17 520 maternal deaths in SSA from PE/E. This number approximates the current WHO estimates of maternal deaths from PE/E for SSA (10).
Scenario 3 suggests that if MgSO4 were available at all locations but had the current levels of low and incorrect use coupled with low levels of preeclampsia diagnostics and other interventions, only 260 additional lives would be saved. In scenario 4, the model estimates that even with universal availability and near perfect MgSO4 use, but still low levels of preeclampsia diagnostics and other interventions, only 610 additional lives over baseline treatment would be saved. Scenario 5 estimates the number of deaths due to PE/E and lives saved over baseline with near perfect diagnosis of PE/E but all other interventions, including MgSO4, transfer and cesarean section/induction kept at baseline levels. In this scenario, with better diagnosis of PE/E, MANDATE estimates that 3780 lives were saved compared with current treatments. Scenario 6 illustrates that when correct and near universal MgSO4 use is added to near perfect diagnosis of PE/E, there will be 7150 lives saved over baseline and an additional 3370 lives compared with scenario 5, which included near perfect diagnosis but current low levels of MgSO4 use. Scenario 7 adds high levels of transfer to scenario 5 in which there was high use of diagnostics but no other interventions added over baseline use. In this scenario, 7850 lives are estimated to be saved compared with baseline, and an additional 4070 lives saved compared with scenario 5. In scenario 8, where high levels of diagnostics, MgSO4, transfers and cesarean section/induction are used, there were only 3750 maternal deaths and 13 770 lives were saved compared with current treatments.
Table 3. The estimated impact of various treatment scenarios on the number of maternal deaths from preeclampsia/eclampsia and lives saved compared with current baseline treatments in sub-Saharan Africa, 2012.
Scenario number 1 2 3 4 5 6 7 8 9 10
Baseline treatmenta
+ + + + + + + + +b
MgSO4 at high (99%) penetration only
MgSO4 at high (99%) penetration and utilization
+ +
+
+
+
+
+
High levels of diagnostics (99%)
+ + + + + +
High cesarean section and Induction (99%) in hospital
+ + +
High levels of transfer
Maternal Deaths (n)
Lives saved (n)
+ + + +
20 570 17 520 17 260 16 920 13 740 10 370 9670 3750 3960 4060
NA NA 260 610 3780 7150 7850 13 770 13 560 13 460
a
Baseline treatment includes estimated rates for currently available preeclampsia diagnosis, prevention/treatment and transfer, with rates as shown in Table 2. b All current baseline treatments except MgSO4 are provided. + Presence of indicated intervention; Absence of indicated intervention; NA, not applicable.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
151
Preeclampsia/eclampsia in low-income countries
In scenario 9, MgSO4 use is reduced back to baseline rate from the previous scenario, but still has high levels of diagnosis, transfer, and cesarean section/induction. There is only an increase of 210 deaths. In scenario 10, when all MgS04 use is removed, there is only an estimated increase of 100 maternal deaths from the low MgS04 use scenario and an increase of only 310 deaths from scenario 8, which had high use of MgSO4 in all settings. These last two scenarios suggest that at high penetration and utilization of the other interventions, MgSO4 will have only a small additional impact on the reduction in maternal mortality from PE/E. Further evaluation of four of the scenarios reveals their impact on the number of cases of eclampsia as well as the number of deaths from both preeclampsia and eclampsia (Table 4): scenario 1 – no treatment, scenario 2 – current baseline treatments, scenario 4 – current baseline treatments with high utilization and penetration of MgSO4, and scenario 8 – high levels of diagnosis, MgSO4 use, transfer, and cesarean section/induction. In this analysis, the number of preeclampsia cases does not change across the scenarios, since no intervention reduces its incidence. With current treatment, however, the number of eclampsia cases declines by 9700 from the no treatment scenario. Adding high treatment levels of MgSO4 everywhere (scenario 4) reduces the number of eclampsia cases by 1590. However, in scenario 8, when all interventions are provided at high levels, the number of eclampsia cases decreases an estimated 85% from current levels. In the no treatment scenario, 67% of the maternal PE/ E deaths are associated with eclampsia, and with current treatments, 63% are associated with eclampsia. With high MgSO4 use, 61% are associated with eclampsia. With all treatments available and used, we estimate that only 40% of the PE/E maternal deaths would be due to eclampsia. Across these scenarios, deaths due to preeclampsia decrease an estimated 67%, while the maternal deaths associated with eclampsia decrease by an estimated 90%. It should be noted that between scenario 2, with low
R.L. Goldenberg et al.
levels of MgSO4 use, and scenario 4, with high levels of MgSO4 use, while the total number of deaths attributable to both PE/E and to eclampsia alone decrease, the deaths attributable to preeclampsia rise slightly. This occurs because MgSO4 decreases the progression of preeclampsia to eclampsia, resulting in more preeclampsia and fewer eclampsia cases. Since a small but finite risk of death remains associated with preeclampsia, the deaths associated with preeclampsia increase slightly. However, since eclampsia carries a much higher risk of death and MgSO4 decreases the number of women with eclampsia, the overall mortality associated with PE/E decreases.
Discussion This analysis suggests that increasing MgSO4 use alone with other interventions remaining at current levels will have little added impact on maternal mortality from PE/ E. Similarly, high MgSO4 use in the presence of high levels of diagnosis, transfer and availability of cesarean section/labor induction, will have little impact on maternal mortality. However, with high levels of diagnosis, but lower levels of transport and cesarean section/labor induction, high MgSO4 use should contribute to a moderate reduction in PE/E-related maternal deaths. Previously, we noted that the greatest reductions in PE/ E deaths in the USA, and virtually all the reductions in PE/E mortality in the UK, occurred from 1935 to about 1980, and in the absence of MgSO4 (7). We attributed this reduction to the introduction of prenatal care with screening for preeclampsia, increased hospitalization for delivery, and increasing use of cesarean section and labor induction for women with PE/E. MANDATE suggests that this interpretation of historical data is correct and relevant for SSA. Furthermore, if the four interventions were scaled up to high levels, the prevalence of preeclampsia would not decrease but there should be a marked decrease in the progression of preeclampsia to eclampsia and a reduction in the progression of both
Table 4. Estimated cases of preeclampsia and eclampsia and maternal mortality associated with preeclampsia and eclampsia for select scenarios, sub-Saharan Africa, 2012.
Outcomes (n) Cases of Preeclampsia Cases of Eclampsia Maternal deaths from preeclampsia Maternal deaths from eclampsia Total maternal deaths
No treatment Scenario (1)
Baseline Scenario (2)
Baseline with high MgSO4 Scenario (4)
Best case Scenario (8)
1 630 870 49 890 6850 13 720 20 570
1 630 870 40 190 6510 11 010 17 520
1 630 870 38 600 6540 10 370 16 910
1 630 870 5630 2250 1500 3750
Scenario number corresponds to Table 3.
152
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
R.L. Goldenberg et al.
preeclampsia and eclampsia to death. This, too, is consistent with our evaluation of historical data in HIC (7). Since these interventions have a greater impact on deaths from eclampsia than preeclampsia, the model predicts that as the total number of PE/E deaths is reduced, a smaller proportion of those deaths would be associated with eclampsia. There are a number of components of hospital care that may have an impact on PE/E mortality, including the use of antihypertensives. However, we did not model the impact of these medications because we found no published studies documenting specific reductions in maternal mortality associated with this intervention. This finding is consistent with the WHO recommendations for PE/E, which concluded there was very low quality evidence for antihypertensives to treat women with severe PE/E (28,29) but, based on expert opinion, included them in the WHO Essential Medicines List (30). Therefore, while antihypertensives are frequently recommended for women with severe hypertension to prevent hemorrhagic stroke, at present we are unable to substantiate a measure of efficacy to model this intervention. Several publications have concluded that increased use of MgSO4 is likely to have a substantial impact on maternal mortality (14,15,18). These estimates are often derived from the original randomized trials of MgSO4, mostly done in HIC. For example, the Magpie Trial reported a nearly 60% reduction in seizures and a similar (but not statistically significant) reduction in maternal mortality (19). The mortality reduction was from a relatively low level to an even lower level. The effectiveness of MgSO4 to reduce mortality was evaluated predominantly in modern hospitals, with immediate access to cesarean section and induction of labor for eclampsia or worsening preeclampsia. There also was likely access to the other interventions that may prevent maternal death associated with PE/E, including antihypertensives. These estimates of reduced maternal mortality attributed to increasing MgSO4 use may not apply to many areas in SSA. Therefore, we believe that the large number of estimated maternal lives saved from PE/E in some of these reports is not due to treatment with MgSO4 alone, but from high levels of diagnosis, transport, and hospitalization, including cesarean section/labor induction for women with PE/E. The conclusions from the two types of analyses are therefore very different. From other analyses, one would assume that widespread use of MgSO4 alone with other treatments held constant will have a large impact on mortality from PE/E in SSA. From the MANDATE analysis, it is apparent that increasing coverage for diagnosis, transport, and hospitalization with cesarean section/labor induction are likely to have a much greater role in mortality reduction from PE/E than increasing use of MgSO4.
Preeclampsia/eclampsia in low-income countries
In effect, the model shows that widespread introduction of MgSO4 with the current level of other interventions will have little impact on PE/E maternal mortality. Only in settings with very high levels of diagnostics, but relatively low levels of transfer and access to cesarean section/ labor induction, does it appear that there will be a moderate reduction in maternal mortality from PE/E when MgSO4 is introduced. This would imply that increasing coverage of both diagnostics and MgSO4 together might be an effective strategy to reduce some of the PE/Erelated mortality. Therefore, the settings in which high levels of MgSO4 use are introduced, and the penetration and utilization of other interventions are all crucial to understanding the potential impact of MgSO4 on the progression of preeclampsia to eclampsia and on PE/Erelated mortality. Thus, our model provides a more nuanced analysis to answer questions related to the impact of MgSO4 on maternal mortality from PE/E. This study has some potential weaknesses. First, while efforts were made to use the best estimates, the available data were often disparate, non-representative or inexact. We are aware that other estimates could easily be justified. Perhaps more importantly, we created a mathematical model of two complicated medical conditions, preeclampsia and eclampsia, and are aware that this model presents only an approximation of clinical care. Modeling a medical condition and changes in survival rates associated with variations in treatment for that condition, may not realistically capture the essence of the disease or its management. Nevertheless, this model allows the estimation of the effects on maternal survival of single as well as multiple interventions at various levels of penetration and utilization in the home, clinic and hospital. The results with increasing levels of interventions were similar to those seen historically in HIC. Another limitation is that we evaluated the SSA as a whole, rather than by individual countries, and did not attempt to evaluate urban/rural or income differences in populations. However, MANDATE takes into account whether the care is provided in the home, clinic or hospital and uses our best estimates of penetration and utilization of each intervention. Finally, the model is not intended to be the sole information used to make local policy decisions about implementation of any specific intervention. Information about cost, available resources, feasibility, cultural acceptability, together with many other considerations will need to be considered. However, having estimates about the number of potential lives saved and relative differences with various interventions or groups of interventions is an important input. In summary, we evaluated several interventions aimed at reducing maternal mortality from PE/E in SSA, with a focus on MgSO4 use. Our analysis suggests that more
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
153
Preeclampsia/eclampsia in low-income countries
widespread use of diagnostics, hospitalization with cesarean section and labor induction for women with PE/E, and transport for those with PE/E to a center where treatment is available will have a greater impact on maternal mortality from PE/E than the use of MgSO4 alone or even in combination with other interventions. Putting emphasis and resources into increasing MgSO4 use while ignoring the other interventions even more likely to reduce maternal mortality from PE/E – such as better diagnostics, transport, and hospitalization with cesarean section/labor induction – will likely lead to disappointing reductions in maternal mortality. We emphasize that we are not disagreeing with the WHO recommendation regarding the use of MgSO4 as the anti-convulsant of choice for women with PE/E (29). However, in addition to the use of MgSO4, great attention should be given to ensuring that every woman has access to the other interventions that historically have achieved large reductions in PE/E-related maternal mortality.
Funding This study was funded by a grant from the Bill & Melinda Gates Foundation. References 1. Dolea C, AbouZahr C. Global burden of hypertensive disorders of pregnancy in the year 2000. Geneva: World Health Organization, 2003. 2. Duley L. The global impact of preeclampsia and eclampsia. Semin Perinatol. 2009;33:130–7. 3. Blencowe H, Cousens S, Oestergaard MZ, Chou D, Moller AB, Narwal R, et al. 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. 2012;379:2162–72. 4. Bhutta ZA, Chopra M, Axelson H, Berman P, Boerma T, Bryce J, et al. Countdown to 2015 decade report (2000– 10): taking stock of maternal, newborn, and child survival. Lancet. 2010;375:2032–44. 5. Ghulmiyyah L, Sibai B. Maternal mortality from preeclampsia/eclampsia. Semin Perinatol. 2012;36:56–9. 6. Adam GK, Bakheit KH, Adam I. Maternal and perinatal outcomes of eclampsia in Gadarif Hospital. Sudan. J Obstet Gynaecol. 2009;29:619–20. 7. Goldenberg RL, McClure EM, Macguire ER, Kamath BD, Jobe AH. Lessons for low-income regions following the reduction in hypertension-related maternal mortality in high-income countries. Int J Gynaecol Obstet. 2011;113:91–5. 8. O’Connor HD, Hehir MP, Kent EM, Foley ME, Fitzpatrick C, Geary MP, et al. Eclampsia: trends in incidence and outcomes over 30 years. Am J Perinatol. 2013;30:661–4.
154
R.L. Goldenberg et al.
9. Swain S, Ojha KN, Prakash A, Bhatia BD. Maternal and perinatal mortality due to eclampsia. Indian Pediatr. 1993;30:771–3. 10. Khan KS, Wojdyla D, Say L, G€ ulmezoglu AM, Van Look PF. WHO analysis of causes of maternal death: a systematic review. Lancet. 2006;367:1066–74. 11. Ferrazzani S, Luciano R, Garofalo S, D’Andrea V, De Carolis S, De Carolis MP, et al. Neonatal outcome in hypertensive disorders of pregnancy. Early Hum Dev. 2011;87:445–9. 12. Chu K, Cortier H, Maldonado F, Mashant T, Ford N, Trelles M. Cesarean section rates and indications in sub-Saharan Africa: a multi-country study from Medecins sans Frontieres. PLoS ONE. 2012;7:e44484. 13. Jim B, Sharma S, Kebede T, Acharya A. Hypertension in pregnancy: a comprehensive update. Cardiol Rev. 2010;18:178–89. 14. McDonald SD, Lutsiv O, Dzaja N, Duley L. A systematic review of maternal and infant outcomes following magnesium sulfate for preeclampsia/eclampsia in real-world use. Int J Gynaecol Obstet. 2012;118:90–6. 15. Langer A, Villar J, Tell K, Kim T, Kennedy S. Reducing eclampsia-related deaths – a call to action. Lancet. 2008;371:705–6. 16. United Nations. Life-saving commodities. Available at: http://www.everywomaneverychild.org/resources/ un-commission-on-life-saving-ommodities/ life-saving-commodities (accessed 12 June 2013). 17. Okereke E, Ahonsi B, Tukur J, Ishaku SM, Oginni AB. Benefits of using magnesium sulphate (MgSO₄) for eclampsia management and maternal mortality reduction: lessons from Kano State in Northern Nigeria. BMC Res Notes. 2012;5:421. 18. Duley L, G€ ulmezoglu AM, Henderson-Smart DJ, Chou D. Magnesium sulphate and other anticonvulsants for women with preeclampsia. Cochrane Database Syst Rev. 2010;11: CD000025. 19. Smyth RM, Spark P, Armstrong N, Duley L. Magpie Trial in the UK: methods and additional data for women and children at 2 years following pregnancy complicated by preeclampsia. BMC Pregnancy Childbirth. 2009;9:15. 20. Dhakal G, Subedi M, Paudel K. Magnesium sulphate in management of severe preeclampsia and eclampsia. J Nepal Health Res Counc. 2012;10:113–7. 21. McClure EM, Rouse DJ, Macguire ER, Jones B, Griffin JB, Jobe AH, et al. The MANDATE model for evaluating interventions to reduce postpartum hemorrhage. Int J Gynaecol Obstet. 2013;121:5–9. 22. Maternal and Neonatal Directed Assessment of Technology model. Available at: http://mnhtech.org (accessed 12 June 2013). 23. Rudan I, Gibson J, Kapiriri L, Lansang MA, Hyder AA, Lawn J, et al. Setting priorities in global child health research investments: assessment of principles and practice. Croat Med J. 2007;48:595–604.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
R.L. Goldenberg et al.
24. United Nations. Available at: http://www.afro.who.int/ (accessed 28 October 2013). 25. United Nations. Birth registry for Sub-Saharan Africa, 2012. Available at: www.data.un.org (accessed 12 December 2013) 26. Goudar SS, Carlo WA, McClure EM, Pasha O, Patel A, Esamai F, et al. The Maternal and Newborn Health Registry Study of the Global Network for Women’s and Children’s Health Research. Int J Gynaecol Obstet. 2012;118:190–3. 27. Villar J, Say L, Shennan A, Lindheimer M, Duley L, Conde-Agudelo A, et al. Methodological and technical issues related to the diagnosis, screening, prevention, and
Preeclampsia/eclampsia in low-income countries
treatment of preeclampsia and eclampsia. Int J Gynaecol Obstet. 2004;85(Suppl 1):S28–41. 28. Abalos E, Duley L, Steyn DW. Antihypertensive drug therapy for mild to moderate hypertension during pregnancy. Cochrane Database System Rev. 2007;1: CD002252. 29. World Health Organization. WHO recommendations for prevention and treatment of preeclampsia and eclampsia. Geneva: WHO, 2011. 30. Lalani S, Firoz T, Magee LA, Sawchuck D, Payne B, Gordon R, et al. Pharmacotherapy for preeclampsia in low and middle income countries: an analysis of essential medicines lists. J Obstet Gynaecol Can. 2013;35:215–23.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
155
AOGS M A I N R E SE A RC H A R TI C LE
Reducing maternal mortality from preeclampsia and eclampsia in low-resource countries – what should work? ROBERT L. GOLDENBERG1, BONNIE JONES2, JENNIFER B. GRIFFIN2, DORIS J. ROUSE2, BEENA D. KAMATH-RAYNE3, NEHAL TRIVEDI2 & ELIZABETH M. MCCLURE2 1
Department of Obstetrics and Gynecology, Columbia University, New York, NY, 2Social, Statistical and Environmental Sciences, RTI International, Durham, NC, and 3Perinatal Institute, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA
Key words Preeclampsia, eclampsia, maternal mortality, magnesium sulfate, sub-Saharan Africa Correspondence Robert L. Goldenberg, Department of Obstetrics & Gynecology, Columbia University Medical Center, 622 W. 168th Street, PH 16-66, New York, NY 10032, USA. E-mail: [email protected] Conflict of interest The authors have stated explicitly that there are no conflicts of interest in connection with this article. Please cite this article as: Goldenberg RL, Jones B, Griffin JB, Rouse DJ, Kamath-Rayne BD, Trivedi N, et al. Reducing maternal mortality from preeclampsia and eclampsia in low-resource countries – what should work? Acta Obstet Gynecol Scand 2015; 94: 148– 155. Received: 4 May 2014 Accepted: 13 October 2014 DOI: 10.1111/aogs.12533
Abstract Objective. Preeclampsia/eclampsia (PE/E) remains a major cause of maternal death in low-income countries. We evaluated interventions to reduce PE/Erelated maternal mortality in sub-Saharan Africa. Design. Mathematical model to assess impact of interventions on PE/E-related maternal morbidity and mortality. Setting. Sub-Saharan Africa countries. Population. Pregnant women in sub-Saharan Africa in 2012. Methods. A systematic literature review populated a decision-tree mathematical model with interventions to diagnose, prevent, and treat women with PE/E. The impact of increased use of interventions [diagnostics, transfer to a hospital, magnesium sulfate (MgSO4) use, cesarean section/ labor induction] on PE/E-related maternal mortality was analyzed. Main outcome measures. Prevalence of PE/E and PE/E-associated maternal mortality rates in sub-Saharan Africa. Results. Without interventions, an estimated 20 570 PE/ E-associated deaths would have occurred in sub-Saharan Africa in 2012. With current low rates of diagnosis, MgSO4 use, transfers and cesarean section/induction rates, about 17 520 maternal deaths were associated with PE/E in 2012. Higher use of MgSO4 would have prevented about 610 deaths. With high diagnostic levels, MgSO4 use, transfer and cesarean section/induction, mortality was reduced to 3750 annual deaths, saving about 13 770 maternal lives. If all MgSO4 use was removed from the model, 4060 maternal deaths would occur, increasing maternal deaths by only 310. Conclusions. In sub-Saharan Africa, our model suggests that increasing use of PE/E diagnostics, transfer to higher levels of care and increased hospitalization with cesarean section/induction of labor would substantially reduce maternal mortality from PE/E. Increasing use of MgSO4 would have a smaller impact on maternal mortality. HIC, high-income countries; LIC, low-income countries; MANDATE, Maternal and Neonatal Directed Assessment of Technology model; MgSO4, magnesium sulfate; PE/E, preeclampsia/eclampsia; SSA, subSaharan Africa. Abbreviations:
Introduction Preeclampsia/eclampsia (PE/E) is one of the major causes of maternal mortality in low-income countries (LIC), the leading cause of stillbirth, and an important contributor
148
to neonatal mortality (1–3). PE/E occurs in about 3–6% of all pregnancies. Large variations in the reported prevalence of preeclampsia occur worldwide (4,5); these variations may reflect actual differences, or differences in definitions or reporting practices. Untreated, between 2 and 10% of women with preeclampsia will develop
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
R.L. Goldenberg et al.
eclampsia; most seizures occur in women with severe preeclampsia. In LIC, risk of mortality with seizures, and especially with multiple seizures, is as high as 25–50% (6– 8). In high-income countries (HIC) the progression from preeclampsia to eclampsia and the case fatality rates from both conditions have substantially declined over time; there are now large differences between LIC and HIC in these outcomes. We have previously demonstrated that the more than 90% reduction in PE/E-associated maternal mortality in HIC from the early 1900s to about 1980 is explained by a reduction in the progression of preeclampsia to eclampsia, together with the reduction in case-fatality rates associated with eclampsia (7). PE/E is thought to account for about 15% of maternal mortality world-wide, 20–25% of fetal mortality, and with increased risk of asphyxia and preterm delivery for as much as 25% of neonatal mortality (1,2,8–10). While the impact of PE/E on adverse pregnancy outcomes has been much reduced in HIC since about 1935, there has been a much smaller reduction in these outcomes in LIC (7). Currently, about 98% of maternal, fetal, and neonatal mortality related to PE/E occurs in LIC, with much of that mortality in south Asia and sub-Saharan Africa (SSA) (10). A number of strategies have been advocated to reduce the risk of progression of preeclampsia to eclampsia as well as the risk of death from the condition. Diagnosis of preeclampsia with blood pressure and proteinuria determinations and timely delivery through cesarean section or induction of labor are both important in reducing PE/E-related maternal mortality (7,11–13). Magnesium sulfate (MgSO4) is one of the most common interventions proposed to delay progression of preeclampsia to eclampsia and reduce maternal mortality (14–20). To better understand the interventions most likely to reduce PE/E-associated maternal mortality in SSA, we created a model for this condition. Our objective was to examine maternal mortality rates associated with PE/E and the impact of interventions designed to reduce those rates.
Preeclampsia/eclampsia in low-income countries
MANDATE estimates the proportion of women treated in the home, clinic, and hospital setting as well as the impact of various interventions in each setting. Each intervention, whether existing or proposed, is characterized by its efficacy in diagnosing, preventing or treating a condition as well as its penetration (availability in a given setting) and utilization (the rate at which it is used correctly) in the home, clinic or hospital (22). In these analyses, interventions studied had evidence for substantially reducing maternal mortality and included improved diagnosis of preeclampsia (blood pressure and proteinuria determination), transfer of women with PE/E to a higher level of care, greater use of MgSO4 to prevent the first and subsequent seizures, and increased use of cesarean section or labor induction to deliver mothers diagnosed with PE/E. We first evaluated a hypothetical scenario in which no interventions were available, then a baseline scenario considering current use of interventions, as well as hypothesized scenarios where the penetration, utilization and, if appropriate, efficacy of various interventions were altered. We also assessed the impact of transfers from the home to clinic or hospital or from the clinic to the hospital. The numbers that populate the model are derived from numerous sources. A systematic literature review was conducted which included all English literature in PubMed since 1980, the Cochrane Library and the WHO database. For the literature review, we used the search terms “pregnancy”, “hypertension”, “preeclampsia”, “eclampsia, and “magnesium sulfate”. Additionally, published papers and reports including the Demographic and Household Surveys and United Nations data provided the rates of birth and utilization of various interventions and mortality rates. A modified algorithm using the GRADE system (Grading of Recommendations Assessment, Development and Evaluation) was applied, which used the estimates from the highest quality sources as primary, with support from other sources where no other data were available (23). We restricted this analysis to SSA, which covers countries in the World Bank’s SSA Region stretch-
Material and methods To help define which interventions would have the largest impact on maternal, fetal, and neonatal mortality in LIC, a model – Maternal and Neonatal Directed Assessment of Technology (MANDATE) – was created (21). This model uses decision tree and mathematical modeling to evaluate maternal, fetal, and neonatal mortality rates and the impact of interventions on these outcomes. For this study, we modeled the maternal mortality associated with PE/E in SSA. The MANDATE version used in this analysis is posted online (http://mnhtech.org).
Key Message A prenatal care program consisting of testing for hypertension and proteinuria, and increased use of hospitalization for cesarean section/induction of labor would eliminate most maternal deaths associated with preeclampsia/eclampsia. MgSO4 use appears to have a much smaller impact on preclampsia/eclampsia-associated maternal mortality in these settings.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
149
Preeclampsia/eclampsia in low-income countries
R.L. Goldenberg et al.
ing from South Africa to Chad and Sudan in the north (24). Table 1 defines the key variables for the analysis. These include the number of births in SSA in 2012, the estimated overall incidence of preeclampsia, divided into mild/moderate and severe categories (25,26). Also included is the percent of women with untreated mild/ moderate and severe preeclampsia who die without progressing to eclampsia, the percent of women with mild/ moderate and severe preeclampsia who progress to eclampsia, and the case-fatality rates for women with untreated eclampsia. Table 2 lists the potentially important interventions related to preeclampsia and eclampsia included in MANDATE. These include diagnosis of preeclampsia by blood pressure and proteinuria, transfer to a higher level of care, treatment with MgSO4, and delivery by cesarean section/labor induction. Interventions are modeled using the constructs of penetration, utilization, and efficacy. Because there is evidence that various interventions might be used more frequently under certain conditions, the model may include a range of utilization rates. For example, women diagnosed with hypertension have an increased likelihood to receive a test for proteinuria, compared with women not diagnosed with hypertension (27). Efficacy was defined as the ability to prevent or diagnose a condition or decrease mortality due to that condition. MgSO4 was modeled with an efficacy to reduce eclampsia by 65% in women with mild/moderate preeclampsia without a prior seizure, an efficacy of 55% to prevent eclampsia in women with severe preeclampsia and a similar efficacy to reduce further seizures in women already having a seizure. Any impact of MgSO4 on maternal death had to result from a reduction in eclampsia.
Table 1. Of the estimated 33 000 000 pregnancies ending at 20 weeks or more in sub-Saharan Africa, 2012, estimates are presented of the incidence and case fatality rates for preeclampsia and eclampsiaa. Conditions
Rates
Preeclampsia –total incidence Preeclampsia –incidence severe Preeclampsia –incidence mild/moderate Rate severe preeclampsia progresses to eclampsia Rate mild/moderate preeclampsia progresses to eclampsia Case-fatality rates Mild/moderate preeclampsia not progressing to eclampsia Severe preeclampsia not progressing to eclampsia Eclampsia: one seizure Eclampsia: two or more seizures
5% 1.5% 3.5% 10% 0.2% 0.3% 0.7% 25% 50%
a The incidence and case-fatality rates assume no preventive or treatment interventions.
150
Table 2. Interventions for preeclampsia/eclampsia and estimated baseline rates of penetration, utilization and efficacy in the home, clinic and hospital in sub-Saharan Africa, 2012. Setting Home Diagnostics Blood pressure measurement Penetration 15% Utilization (if no proteinuria 20% diagnostic) Utilization (if proteinuria diagnostic 40% is positive) Efficacy to detect high blood 75% pressure Proteinuria Penetration 0% Utilization (if no blood pressure 0% diagnostic) Utilization (if blood pressure 0% diagnostic is positive) Efficacy to detect proteinuria 85% Prevention/treatment Antihypertensives Penetration 0% Utilization 0% Efficacy 0% Magnesium sulfate (MgSO4) Penetration 0% Utilization (mild to moderate 0% preeclampsia, or undiagnosed severe preeclampsia) Utilization (diagnosed severe 0% preeclampsia) Utilization (eclampsia) 0 Efficacy to prevent seizure (mild/ 65% moderate preeclampsia) Efficacy to prevent seizure (severe 55% preeclampsia or recurrent eclamptic seizures) Cesarean section/induction for prevention/treatment death Penetration 0% Utilization (mild to moderate 0% preeclampsia) Utilization (severe preeclampsia/ 0% eclampsia) Efficacy (mild/ moderate preeclampsia) 75% Efficacy (for severe preeclampsia/ 95% eclampsia) Transfer (for mild/moderate preeclampsia) Home to clinic 35% Home to hospital 10% Clinic to hospital 20% Transfer (for severe preeclampsia and eclampsia) Home to clinic 35% Home to hospital 20% Clinic to hospital 30%
Clinic
Hospital
85% 60%
85% 70%
85%
95%
75%
75%
60% 40%
90% 45%
75%
85%
85%
85%
25% 45% 0%
70% 65% 0%
15% 30%
50% 50%
40%
60%
50% 65%
75% 65%
55%
55%
of eclampsia/ 15% 30%
75% 40%
60%
80%
75% 95%
75% 95%
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
Preeclampsia/eclampsia in low-income countries
R.L. Goldenberg et al.
Furthermore, to discriminate the reduction in mortality, we have defined a range of efficacy rates for the same intervention (such as cesarean section) when given for different indications. As an example, we estimated that cesarean section/labor induction will have a greater efficacy to prevent maternal death in women with severe preeclampsia than with mild/moderate preeclampsia. As shown in Table 2, currently, the estimated penetration and utilization of diagnostics, MgSO4 and cesarean section or labor induction for delivery are generally low in SSA, with substantially lower rates in clinic and home compared to hospital.
Results Using the MANDATE model, we ran 10 scenarios for SSA (Table 3). The columns on the left present the interventions and the next to last column on the right presents the estimated maternal mortality. The last column provides the estimated number of lives saved compared with the baseline treatment scenario. Scenario 1 presents maternal mortality estimates if no interventions related to PE/E were available. This scenario suggests that without any treatment, there would be 20 570 maternal deaths from PE/E. Scenario 2 suggests that with the current treatments, including relatively low levels of use of preeclampsia diagnostics, MgSO4, cesarean section/induction and transfer of women with PE/E to a hospital, there were 17 520 maternal deaths in SSA from PE/E. This number approximates the current WHO estimates of maternal deaths from PE/E for SSA (10).
Scenario 3 suggests that if MgSO4 were available at all locations but had the current levels of low and incorrect use coupled with low levels of preeclampsia diagnostics and other interventions, only 260 additional lives would be saved. In scenario 4, the model estimates that even with universal availability and near perfect MgSO4 use, but still low levels of preeclampsia diagnostics and other interventions, only 610 additional lives over baseline treatment would be saved. Scenario 5 estimates the number of deaths due to PE/E and lives saved over baseline with near perfect diagnosis of PE/E but all other interventions, including MgSO4, transfer and cesarean section/induction kept at baseline levels. In this scenario, with better diagnosis of PE/E, MANDATE estimates that 3780 lives were saved compared with current treatments. Scenario 6 illustrates that when correct and near universal MgSO4 use is added to near perfect diagnosis of PE/E, there will be 7150 lives saved over baseline and an additional 3370 lives compared with scenario 5, which included near perfect diagnosis but current low levels of MgSO4 use. Scenario 7 adds high levels of transfer to scenario 5 in which there was high use of diagnostics but no other interventions added over baseline use. In this scenario, 7850 lives are estimated to be saved compared with baseline, and an additional 4070 lives saved compared with scenario 5. In scenario 8, where high levels of diagnostics, MgSO4, transfers and cesarean section/induction are used, there were only 3750 maternal deaths and 13 770 lives were saved compared with current treatments.
Table 3. The estimated impact of various treatment scenarios on the number of maternal deaths from preeclampsia/eclampsia and lives saved compared with current baseline treatments in sub-Saharan Africa, 2012.
Scenario number 1 2 3 4 5 6 7 8 9 10
Baseline treatmenta
+ + + + + + + + +b
MgSO4 at high (99%) penetration only
MgSO4 at high (99%) penetration and utilization
+ +
+
+
+
+
+
High levels of diagnostics (99%)
+ + + + + +
High cesarean section and Induction (99%) in hospital
+ + +
High levels of transfer
Maternal Deaths (n)
Lives saved (n)
+ + + +
20 570 17 520 17 260 16 920 13 740 10 370 9670 3750 3960 4060
NA NA 260 610 3780 7150 7850 13 770 13 560 13 460
a
Baseline treatment includes estimated rates for currently available preeclampsia diagnosis, prevention/treatment and transfer, with rates as shown in Table 2. b All current baseline treatments except MgSO4 are provided. + Presence of indicated intervention; Absence of indicated intervention; NA, not applicable.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
151
Preeclampsia/eclampsia in low-income countries
In scenario 9, MgSO4 use is reduced back to baseline rate from the previous scenario, but still has high levels of diagnosis, transfer, and cesarean section/induction. There is only an increase of 210 deaths. In scenario 10, when all MgS04 use is removed, there is only an estimated increase of 100 maternal deaths from the low MgS04 use scenario and an increase of only 310 deaths from scenario 8, which had high use of MgSO4 in all settings. These last two scenarios suggest that at high penetration and utilization of the other interventions, MgSO4 will have only a small additional impact on the reduction in maternal mortality from PE/E. Further evaluation of four of the scenarios reveals their impact on the number of cases of eclampsia as well as the number of deaths from both preeclampsia and eclampsia (Table 4): scenario 1 – no treatment, scenario 2 – current baseline treatments, scenario 4 – current baseline treatments with high utilization and penetration of MgSO4, and scenario 8 – high levels of diagnosis, MgSO4 use, transfer, and cesarean section/induction. In this analysis, the number of preeclampsia cases does not change across the scenarios, since no intervention reduces its incidence. With current treatment, however, the number of eclampsia cases declines by 9700 from the no treatment scenario. Adding high treatment levels of MgSO4 everywhere (scenario 4) reduces the number of eclampsia cases by 1590. However, in scenario 8, when all interventions are provided at high levels, the number of eclampsia cases decreases an estimated 85% from current levels. In the no treatment scenario, 67% of the maternal PE/ E deaths are associated with eclampsia, and with current treatments, 63% are associated with eclampsia. With high MgSO4 use, 61% are associated with eclampsia. With all treatments available and used, we estimate that only 40% of the PE/E maternal deaths would be due to eclampsia. Across these scenarios, deaths due to preeclampsia decrease an estimated 67%, while the maternal deaths associated with eclampsia decrease by an estimated 90%. It should be noted that between scenario 2, with low
R.L. Goldenberg et al.
levels of MgSO4 use, and scenario 4, with high levels of MgSO4 use, while the total number of deaths attributable to both PE/E and to eclampsia alone decrease, the deaths attributable to preeclampsia rise slightly. This occurs because MgSO4 decreases the progression of preeclampsia to eclampsia, resulting in more preeclampsia and fewer eclampsia cases. Since a small but finite risk of death remains associated with preeclampsia, the deaths associated with preeclampsia increase slightly. However, since eclampsia carries a much higher risk of death and MgSO4 decreases the number of women with eclampsia, the overall mortality associated with PE/E decreases.
Discussion This analysis suggests that increasing MgSO4 use alone with other interventions remaining at current levels will have little added impact on maternal mortality from PE/ E. Similarly, high MgSO4 use in the presence of high levels of diagnosis, transfer and availability of cesarean section/labor induction, will have little impact on maternal mortality. However, with high levels of diagnosis, but lower levels of transport and cesarean section/labor induction, high MgSO4 use should contribute to a moderate reduction in PE/E-related maternal deaths. Previously, we noted that the greatest reductions in PE/ E deaths in the USA, and virtually all the reductions in PE/E mortality in the UK, occurred from 1935 to about 1980, and in the absence of MgSO4 (7). We attributed this reduction to the introduction of prenatal care with screening for preeclampsia, increased hospitalization for delivery, and increasing use of cesarean section and labor induction for women with PE/E. MANDATE suggests that this interpretation of historical data is correct and relevant for SSA. Furthermore, if the four interventions were scaled up to high levels, the prevalence of preeclampsia would not decrease but there should be a marked decrease in the progression of preeclampsia to eclampsia and a reduction in the progression of both
Table 4. Estimated cases of preeclampsia and eclampsia and maternal mortality associated with preeclampsia and eclampsia for select scenarios, sub-Saharan Africa, 2012.
Outcomes (n) Cases of Preeclampsia Cases of Eclampsia Maternal deaths from preeclampsia Maternal deaths from eclampsia Total maternal deaths
No treatment Scenario (1)
Baseline Scenario (2)
Baseline with high MgSO4 Scenario (4)
Best case Scenario (8)
1 630 870 49 890 6850 13 720 20 570
1 630 870 40 190 6510 11 010 17 520
1 630 870 38 600 6540 10 370 16 910
1 630 870 5630 2250 1500 3750
Scenario number corresponds to Table 3.
152
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
R.L. Goldenberg et al.
preeclampsia and eclampsia to death. This, too, is consistent with our evaluation of historical data in HIC (7). Since these interventions have a greater impact on deaths from eclampsia than preeclampsia, the model predicts that as the total number of PE/E deaths is reduced, a smaller proportion of those deaths would be associated with eclampsia. There are a number of components of hospital care that may have an impact on PE/E mortality, including the use of antihypertensives. However, we did not model the impact of these medications because we found no published studies documenting specific reductions in maternal mortality associated with this intervention. This finding is consistent with the WHO recommendations for PE/E, which concluded there was very low quality evidence for antihypertensives to treat women with severe PE/E (28,29) but, based on expert opinion, included them in the WHO Essential Medicines List (30). Therefore, while antihypertensives are frequently recommended for women with severe hypertension to prevent hemorrhagic stroke, at present we are unable to substantiate a measure of efficacy to model this intervention. Several publications have concluded that increased use of MgSO4 is likely to have a substantial impact on maternal mortality (14,15,18). These estimates are often derived from the original randomized trials of MgSO4, mostly done in HIC. For example, the Magpie Trial reported a nearly 60% reduction in seizures and a similar (but not statistically significant) reduction in maternal mortality (19). The mortality reduction was from a relatively low level to an even lower level. The effectiveness of MgSO4 to reduce mortality was evaluated predominantly in modern hospitals, with immediate access to cesarean section and induction of labor for eclampsia or worsening preeclampsia. There also was likely access to the other interventions that may prevent maternal death associated with PE/E, including antihypertensives. These estimates of reduced maternal mortality attributed to increasing MgSO4 use may not apply to many areas in SSA. Therefore, we believe that the large number of estimated maternal lives saved from PE/E in some of these reports is not due to treatment with MgSO4 alone, but from high levels of diagnosis, transport, and hospitalization, including cesarean section/labor induction for women with PE/E. The conclusions from the two types of analyses are therefore very different. From other analyses, one would assume that widespread use of MgSO4 alone with other treatments held constant will have a large impact on mortality from PE/E in SSA. From the MANDATE analysis, it is apparent that increasing coverage for diagnosis, transport, and hospitalization with cesarean section/labor induction are likely to have a much greater role in mortality reduction from PE/E than increasing use of MgSO4.
Preeclampsia/eclampsia in low-income countries
In effect, the model shows that widespread introduction of MgSO4 with the current level of other interventions will have little impact on PE/E maternal mortality. Only in settings with very high levels of diagnostics, but relatively low levels of transfer and access to cesarean section/ labor induction, does it appear that there will be a moderate reduction in maternal mortality from PE/E when MgSO4 is introduced. This would imply that increasing coverage of both diagnostics and MgSO4 together might be an effective strategy to reduce some of the PE/Erelated mortality. Therefore, the settings in which high levels of MgSO4 use are introduced, and the penetration and utilization of other interventions are all crucial to understanding the potential impact of MgSO4 on the progression of preeclampsia to eclampsia and on PE/Erelated mortality. Thus, our model provides a more nuanced analysis to answer questions related to the impact of MgSO4 on maternal mortality from PE/E. This study has some potential weaknesses. First, while efforts were made to use the best estimates, the available data were often disparate, non-representative or inexact. We are aware that other estimates could easily be justified. Perhaps more importantly, we created a mathematical model of two complicated medical conditions, preeclampsia and eclampsia, and are aware that this model presents only an approximation of clinical care. Modeling a medical condition and changes in survival rates associated with variations in treatment for that condition, may not realistically capture the essence of the disease or its management. Nevertheless, this model allows the estimation of the effects on maternal survival of single as well as multiple interventions at various levels of penetration and utilization in the home, clinic and hospital. The results with increasing levels of interventions were similar to those seen historically in HIC. Another limitation is that we evaluated the SSA as a whole, rather than by individual countries, and did not attempt to evaluate urban/rural or income differences in populations. However, MANDATE takes into account whether the care is provided in the home, clinic or hospital and uses our best estimates of penetration and utilization of each intervention. Finally, the model is not intended to be the sole information used to make local policy decisions about implementation of any specific intervention. Information about cost, available resources, feasibility, cultural acceptability, together with many other considerations will need to be considered. However, having estimates about the number of potential lives saved and relative differences with various interventions or groups of interventions is an important input. In summary, we evaluated several interventions aimed at reducing maternal mortality from PE/E in SSA, with a focus on MgSO4 use. Our analysis suggests that more
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
153
Preeclampsia/eclampsia in low-income countries
widespread use of diagnostics, hospitalization with cesarean section and labor induction for women with PE/E, and transport for those with PE/E to a center where treatment is available will have a greater impact on maternal mortality from PE/E than the use of MgSO4 alone or even in combination with other interventions. Putting emphasis and resources into increasing MgSO4 use while ignoring the other interventions even more likely to reduce maternal mortality from PE/E – such as better diagnostics, transport, and hospitalization with cesarean section/labor induction – will likely lead to disappointing reductions in maternal mortality. We emphasize that we are not disagreeing with the WHO recommendation regarding the use of MgSO4 as the anti-convulsant of choice for women with PE/E (29). However, in addition to the use of MgSO4, great attention should be given to ensuring that every woman has access to the other interventions that historically have achieved large reductions in PE/E-related maternal mortality.
Funding This study was funded by a grant from the Bill & Melinda Gates Foundation. References 1. Dolea C, AbouZahr C. Global burden of hypertensive disorders of pregnancy in the year 2000. Geneva: World Health Organization, 2003. 2. Duley L. The global impact of preeclampsia and eclampsia. Semin Perinatol. 2009;33:130–7. 3. Blencowe H, Cousens S, Oestergaard MZ, Chou D, Moller AB, Narwal R, et al. 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. 2012;379:2162–72. 4. Bhutta ZA, Chopra M, Axelson H, Berman P, Boerma T, Bryce J, et al. Countdown to 2015 decade report (2000– 10): taking stock of maternal, newborn, and child survival. Lancet. 2010;375:2032–44. 5. Ghulmiyyah L, Sibai B. Maternal mortality from preeclampsia/eclampsia. Semin Perinatol. 2012;36:56–9. 6. Adam GK, Bakheit KH, Adam I. Maternal and perinatal outcomes of eclampsia in Gadarif Hospital. Sudan. J Obstet Gynaecol. 2009;29:619–20. 7. Goldenberg RL, McClure EM, Macguire ER, Kamath BD, Jobe AH. Lessons for low-income regions following the reduction in hypertension-related maternal mortality in high-income countries. Int J Gynaecol Obstet. 2011;113:91–5. 8. O’Connor HD, Hehir MP, Kent EM, Foley ME, Fitzpatrick C, Geary MP, et al. Eclampsia: trends in incidence and outcomes over 30 years. Am J Perinatol. 2013;30:661–4.
154
R.L. Goldenberg et al.
9. Swain S, Ojha KN, Prakash A, Bhatia BD. Maternal and perinatal mortality due to eclampsia. Indian Pediatr. 1993;30:771–3. 10. Khan KS, Wojdyla D, Say L, G€ ulmezoglu AM, Van Look PF. WHO analysis of causes of maternal death: a systematic review. Lancet. 2006;367:1066–74. 11. Ferrazzani S, Luciano R, Garofalo S, D’Andrea V, De Carolis S, De Carolis MP, et al. Neonatal outcome in hypertensive disorders of pregnancy. Early Hum Dev. 2011;87:445–9. 12. Chu K, Cortier H, Maldonado F, Mashant T, Ford N, Trelles M. Cesarean section rates and indications in sub-Saharan Africa: a multi-country study from Medecins sans Frontieres. PLoS ONE. 2012;7:e44484. 13. Jim B, Sharma S, Kebede T, Acharya A. Hypertension in pregnancy: a comprehensive update. Cardiol Rev. 2010;18:178–89. 14. McDonald SD, Lutsiv O, Dzaja N, Duley L. A systematic review of maternal and infant outcomes following magnesium sulfate for preeclampsia/eclampsia in real-world use. Int J Gynaecol Obstet. 2012;118:90–6. 15. Langer A, Villar J, Tell K, Kim T, Kennedy S. Reducing eclampsia-related deaths – a call to action. Lancet. 2008;371:705–6. 16. United Nations. Life-saving commodities. Available at: http://www.everywomaneverychild.org/resources/ un-commission-on-life-saving-ommodities/ life-saving-commodities (accessed 12 June 2013). 17. Okereke E, Ahonsi B, Tukur J, Ishaku SM, Oginni AB. Benefits of using magnesium sulphate (MgSO₄) for eclampsia management and maternal mortality reduction: lessons from Kano State in Northern Nigeria. BMC Res Notes. 2012;5:421. 18. Duley L, G€ ulmezoglu AM, Henderson-Smart DJ, Chou D. Magnesium sulphate and other anticonvulsants for women with preeclampsia. Cochrane Database Syst Rev. 2010;11: CD000025. 19. Smyth RM, Spark P, Armstrong N, Duley L. Magpie Trial in the UK: methods and additional data for women and children at 2 years following pregnancy complicated by preeclampsia. BMC Pregnancy Childbirth. 2009;9:15. 20. Dhakal G, Subedi M, Paudel K. Magnesium sulphate in management of severe preeclampsia and eclampsia. J Nepal Health Res Counc. 2012;10:113–7. 21. McClure EM, Rouse DJ, Macguire ER, Jones B, Griffin JB, Jobe AH, et al. The MANDATE model for evaluating interventions to reduce postpartum hemorrhage. Int J Gynaecol Obstet. 2013;121:5–9. 22. Maternal and Neonatal Directed Assessment of Technology model. Available at: http://mnhtech.org (accessed 12 June 2013). 23. Rudan I, Gibson J, Kapiriri L, Lansang MA, Hyder AA, Lawn J, et al. Setting priorities in global child health research investments: assessment of principles and practice. Croat Med J. 2007;48:595–604.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
R.L. Goldenberg et al.
24. United Nations. Available at: http://www.afro.who.int/ (accessed 28 October 2013). 25. United Nations. Birth registry for Sub-Saharan Africa, 2012. Available at: www.data.un.org (accessed 12 December 2013) 26. Goudar SS, Carlo WA, McClure EM, Pasha O, Patel A, Esamai F, et al. The Maternal and Newborn Health Registry Study of the Global Network for Women’s and Children’s Health Research. Int J Gynaecol Obstet. 2012;118:190–3. 27. Villar J, Say L, Shennan A, Lindheimer M, Duley L, Conde-Agudelo A, et al. Methodological and technical issues related to the diagnosis, screening, prevention, and
Preeclampsia/eclampsia in low-income countries
treatment of preeclampsia and eclampsia. Int J Gynaecol Obstet. 2004;85(Suppl 1):S28–41. 28. Abalos E, Duley L, Steyn DW. Antihypertensive drug therapy for mild to moderate hypertension during pregnancy. Cochrane Database System Rev. 2007;1: CD002252. 29. World Health Organization. WHO recommendations for prevention and treatment of preeclampsia and eclampsia. Geneva: WHO, 2011. 30. Lalani S, Firoz T, Magee LA, Sawchuck D, Payne B, Gordon R, et al. Pharmacotherapy for preeclampsia in low and middle income countries: an analysis of essential medicines lists. J Obstet Gynaecol Can. 2013;35:215–23.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 148–155
155
