[Maternal mortality in France, 2016-2018, frequency, causes and women's profile]

OBJECTIVE

To describe, for the 2016-2018 period, the frequency, causes and risk factors of maternal deaths in France.

METHOD

Data from the National Confidential Enquiry into Maternal Deaths for 2016-2018.

RESULTS

For 2016-2018, 272 maternal deaths occurred in France up to 1 year after the end of pregnancy, i.e a maternal mortality ratio of 11.8 per 100,000 live births (95 % CI 10.4-13.3), and 8.5 (IC 95 % 7.4-9.8) for maternal mortality up to 42 days. Compared to women aged 20-24, the risk of maternal death is multiplied by 2.6 for women aged 35-39, by 5 for women aged 40 and over. Obese women are twice as frequent among maternal deaths (26 %) than in the general population of parturients (11 %). There are territorial disparities -the maternal mortality ratio in the French overseas departments is 2 times higher than in metropolitan France (significant difference but smaller than in 2013-2015)-, and social disparities -the mortality of migrant women remains higher than that of women born in France, particularly for women born in sub-Saharan Africa whose risk is 3 times higher than that of native women. One in three women who died (34 %) had socio-economic vulnerability versus 22 % in the overall population of parturients. Among causes of maternal deaths, the predominant role of psychiatric conditions (mostly suicides) is confirmed for the period 2016-2018, leading cause of maternal mortality considered up to 1 year (17 %), MMR of 1.9/100,000 NV. i.e. approximately one death from psychiatric causes every 3 weeks. Cardiovascular diseases are the second leading cause of maternal mortality up to one year (14 %) and the leading cause up to 42 days (16 %), with 1.3 deaths per 100,000 NV. Amniotic fluid embolism ranks as the third cause (8 %) (2nd cause, 11 %, for MM limited to 42 days), i.e. MMR of 0.9 per 100,000 NV. After a regular decline over the last decade, maternal mortality from obstetric hemorrhage is at a stable level compared to the previous triennium 2013-2015, MMR of 0.9/100,000 NV, i.e 5th cause of MM up to one year (7 %) and 4th cause of MM up to 42 days.

CONCLUSION

The overall national maternal mortality ratio does not show a downward trend, even with constant surveillance method. Territorial inequalities persist but change in their magnitude and in the regions concerned. The profile of the causes of maternal mortality up to one year of the pregnancy end shows the leading role of suicides and cardiovascular diseases, which illustrates that the health of pregnant women or those who have recently given birth is not limited to the obstetric domain, and highlights the importance of multidisciplinarity in the management and organization of care for women in this period.

[National confidential enquiry into maternal deaths in France, a 25-year enhanced surveillance system, essential for the reliable characterization of maternal deaths]

Although maternal mortality is rare in wealthy countries, it remains a fundamental indicator of maternal health. It is considered to be a "sentinel event", the occurrence of which indicates dysfunctions, often cumulative, in the healthcare system. In addition to the classic epidemiological surveillance findings - number of deaths, maternal mortality ratio, distribution of medical causes, sub-groups of women at risk - its study, through a precise analysis of the history of each woman who died, enables to highlight areas for improvement in the content or organisation of care, the correction of which will make it possible to prevent not only deaths but also upstream morbid events involving the same mechanisms. To achieve this dual epidemiological and clinical audit objective, an ad hoc "enhanced" system is needed. France has had such a system since 1996, the Enquête Nationale Confidentielle sur la Mortalité Maternelle (ENCMM), under the joint supervision of Santé Publique France and Inserm. The ENCMM method aims to identify maternal deaths exhaustively and reliably up to 1 year after the end of pregnancy, and to document each death as fully as possible. The 1st step is the multi-source identification (direct declaration, death certificates, linkage with birth certificates, hospital stay database) of women who died during pregnancy or in the year following its end. The 2nd step is the collection of detailed information for each death by a pair of clinical assessors. The 3rd step is the review of these anonymised documents by the National Expert Committee on Maternal Mortality, which establishes the maternal nature of the death (causal link with pregnancy) and, with a stated aim of improvement rather than judgement, assesses the adequacy of care and the preventability of the death. The summary of the information gathered for maternal deaths in the 2016-2018 period is presented in the other articles of this special issue.

[Maternal mortality, adequation of care and preventability, France 2016-2018]

OBJECTIVE

To describe, for the 2016-2018 period, the adequacy of care provided to women who died of maternal death in France, as well as the possible preventability of these deaths.

METHOD

National data from the Enquête Nationale Confidentielle sur les Morts Maternelles for 2016-2018. For each maternal death identified, the National Expert Committee on Maternal Mortality (CNEMM) assesses the adequacy of the care provided and whether the death was probably, possibly or not preventable.

RESULTS

For 2016-2018, 272 maternal deaths (considered up to 1 year after the end of pregnancy) occurred in France. Of these, 265 had sufficient information collected by the confidential survey and could be assessed by the CNEMM. In total, care was judged to be "sub-optimal" for 66% of deaths assessed for all causes, a proportion similar to that for previous periods. In addition to the obstetric and anaesthetic care provided at the time of the acute complication, which was judged to be sub-optimal for 45% (obstetric care) and 38% (anaesthetic care) of maternal deaths, this report highlights the scope for improvement in other types of care, more related to prevention and screening: "sub-optimal" preconception care for 51% of the women who died for whom it was justified, particularly notable for deaths linked to a preexisting condition (52%) and for suicides (67%); prenatal surveillance judged to be "sub-optimal" in 30% of cases, a sub-optimality also more frequent among deaths linked to a preexisting condition (35%) and suicides (34%). In all, 59.7% of maternal deaths assessed were judged to be "probably" (17%) or "possibly" (42.7%) preventable, a profile that remained stable. Suicide and other psychiatric causes, the leading cause of maternal death, were considered to be potentially preventable in 79% of cases. Deaths from haemorrhage remained largely preventable (95%, the highest proportion by cause). The factor most often implicated was inadequate care, and preventability linked to this factor was identified in 53% of deaths, all causes combined. Gap in organization of care was a preventability factor identified in 24% of deaths, and poor interaction between the woman and the healthcare system in 22% of deaths.

CONCLUSION

This proportion of more than half of potentially preventable maternal deaths shows that a reduction in maternal mortality in France is still possible and must be achieved, the objective being to prevent all preventable deaths. Analysis of the factors involved, overall and by cause of death, suggests areas for improvement.

Binational confidential enquiry of maternal deaths due to postpartum hemorrhage in France and the Netherlands: Lessons learned through the perspective of a different context of care

OBJECTIVE

To learn lessons for maternity care by scrutinizing postpartum hemorrhage management (PPH) in cases of PPH-related maternal deaths in France and the Netherlands.

METHODS

In this binational Confidential Enquiry into Maternal Deaths (CEMD), 14 PPH-related maternal deaths were reviewed by six experts from the French and Dutch national maternal death review committees regarding cause and preventability of death, clinical care and healthcare organization. Improvable care factors and lessons learned were identified. CEMD practices and PPH guidelines in France and the Netherlands were compared in the process.

RESULTS

For France, new insights were primarily related to organization of healthcare, with lessons learned focusing on medical leadership and implementation of (surgical) checklists. For the Netherlands, insights were mainly related to clinical care, emphasizing hemostatic surgery earlier in the course of PPH and reducing the third stage of labor by prompter manual removal of the placenta. Experts recommended extending PPH guidelines with specific guidance for women refusing blood products and systematic evaluation of risk factors. The quality of CEMD was presumed to benefit from enhanced case finding, also through non-obstetric sources, and electronic reporting of maternal deaths to reduce the administrative burden.

CONCLUSION

A binational CEMD revealed opportunities for improvement of care beyond lessons learned at the national level.

Timing and Preventability of Cardiovascular-Related Maternal Death

OBJECTIVE

To describe the clinical profile, management, and potential preventability of maternal cardiovascular deaths.

METHODS

We conducted a retrospective, descriptive study of all maternal deaths resulting from a cardiovascular disease during pregnancy or up to 1 year after the end of pregnancy in France from 2007 to 2015. Deaths were identified through the nationwide permanent enhanced maternal mortality surveillance system (ENCMM [Enquête Nationale Confidentielle sur les Morts Maternelles]). Women were classified into four groups based on the assessment of the national experts committee: those who died of a cardiac condition and those who died of a vascular condition and, within these two groups, whether the condition was known before the acute event. Maternal characteristics, clinical features and components of suboptimal care, and preventability factors, which were assessed with a standard evaluation form, were described among those four groups.

RESULTS

During the 9-year period, 103 women died of cardiac or vascular disease, which corresponds to a maternal mortality ratio from these conditions of 1.4 per 100,000 live births (95% CI 1.1-1.7). Analyses were conducted on 93 maternal deaths resulting from cardiac (n=70) and vascular (n=23) disease with available data from confidential inquiry. More than two thirds of these deaths occurred in women with no known pre-existing cardiac or vascular condition. Among the 70 deaths resulting from a cardiac condition, 60.7% were preventable, and the main preventability factor was a lack of multidisciplinary prepregnancy and prenatal care for women with a known cardiac disease. For those with no known pre-existing cardiac condition, preventability factors were related mostly to inadequate prehospital care of the acute event, in particular an underestimation of the severity and inadequate investigation of the dyspnea. Among the 23 women who died of a vascular disease, three had previously known conditions. For women with no previously known vascular condition, 47.4% of deaths were preventable, and preventability factors were related mostly to wrong or delayed diagnosis and management of acute intense chest or abdominal pain in a pregnant woman.

CONCLUSION

Most maternal deaths attributable to cardiac or vascular diseases were potentially preventable. The preventability factors varied according to the cardiac or vascular site and whether the condition was known before pregnancy. A more granular understanding of the cause and related risk factors for maternal mortality is crucial to identify relevant opportunities for improving care and training health care professionals.

Ultra-high dose of intravenous immunoglobulin restores strength and motor function in a patient with refractory multifocal motor neuropathy

INTRODUCTION

Multifocal motor neuropathy (MMN) is a chronic progressive immune-mediated neuropathy, predominantly involving upper limbs asymmetrically with electrophysiologic evidence of motor conduction block. The treatment of choice is immunoglobulin (Ig). Nevertheless, some patients may become resistant to treatment. We describe a patient with history of MMN who became resistant to gammaglobulin treatment but markedly improved using ultra-high doses of intravenous immunoglobulin.

CASE REPORT

A 36-year-old woman with diagnosis of MMN. After 5 years of clinical stability under subcutaneous Ig (2g/kg/month) the patient developed bilateral weakness involving both hands. Treatment was switched to intravenous Ig 2g/kg/month, nevertheless, she progressed and became totally dependent for activities of daily living. We started ultra-high dose intravenous immunoglobulin 5 g/kg/month, with good response. She became independent for activities of daily living and returned to work. The only treatment related adverse event was headache during infusion.

CONCLUSION

Ultra-high dose intravenous Ig seems to be a useful therapy in aggressive MMN with severe disability despite conventional treatment. A low cardiovascular risk score (QRISK2 less than 10%) and a daily intravenous Ig lower than 35 g reduce the risk of severe complications related to intravenous Ig.

Non-convulsive status epilepticus as the initial manifestation in a family with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL)

Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is an autosomal dominant small-vessel disease caused by mutations of the NOTCH3 gene. It typically presents with migraine, recurrent brain ischaemia, and cognitive disorders. Seizures rarely present as the initial manifestation, with non-convulsive status epilepticus being even less frequent. We present a series of 3 related patients with this arteriopathy, 2 of whom presented status epilepticus as a manifestation of the disease.

Maternal mortality in eight European countries with enhanced surveillance systems: descriptive population based study

OBJECTIVE

To compare maternal mortality in eight countries with enhanced surveillance systems.

DESIGN

Descriptive multicountry population based study.

SETTING

Eight countries with permanent surveillance systems using enhanced methods to identify, document, and review maternal deaths. The most recent available aggregated maternal mortality data were collected for three year periods for France, Italy, and the UK and for five year periods for Denmark, Finland, the Netherlands, Norway, and Slovakia.

POPULATION

297 835 live births in Denmark (2013-17), 301 169 in Finland (2008-12), 2 435 583 in France (2013-15), 1 281 986 in Italy (2013-15), 856 572 in the Netherlands (2014-18), 292 315 in Norway (2014-18), 283 930 in Slovakia (2014-18), and 2 261 090 in the UK (2016-18).

OUTCOME MEASURES

Maternal mortality ratios from enhanced systems were calculated and compared with those obtained from each country's office of vital statistics. Age specific maternal mortality ratios; maternal mortality ratios according to women's origin, citizenship, or ethnicity; and cause specific maternal mortality ratios were also calculated.

RESULTS

Methods for identifying and classifying maternal deaths up to 42 days were very similar across countries (except for the Netherlands). Maternal mortality ratios up to 42 days after end of pregnancy varied by a multiplicative factor of four from 2.7 and 3.4 per 100 000 live births in Norway and Denmark to 9.6 in the UK and 10.9 in Slovakia. Vital statistics offices underestimated maternal mortality by 36% or more everywhere but Denmark. Age specific maternal mortality ratios were higher for the youngest and oldest mothers (pooled relative risk 2.17 (95% confidence interval 1.38 to 3.34) for women aged <20 years, 2.10 (1.54 to 2.86) for those aged 35-39, and 3.95 (3.01 to 5.19) for those aged ≥40, compared with women aged 20-29 years). Except in Norway, maternal mortality ratios were ≥50% higher in women born abroad or of minoritised ethnicity, defined variously in different countries. Cardiovascular diseases and suicides were leading causes of maternal deaths in each country. Some other conditions were also major contributors to maternal mortality in only one or two countries: venous thromboembolism in the UK and the Netherlands, hypertensive disorders in the Netherlands, amniotic fluid embolism in France, haemorrhage in Italy, and stroke in Slovakia. Only two countries, France and the UK, had enhanced methods for studying late maternal deaths, those occurring between 43 and 365 days after the end of pregnancy.

CONCLUSIONS

Variations in maternal mortality ratios exist between high income European countries with enhanced surveillance systems. In-depth analyses of differences in the quality of care and health system performance at national levels are needed to reduce maternal mortality further by learning from best practices and each other. Cardiovascular diseases and mental health in women during and after pregnancy must be prioritised in all countries.

Maternal mortality in women with pre-viable premature rupture of membranes: An analysis from the French confidential enquiry into maternal deaths

INTRODUCTION

Pre-viable premature rupture of membranes (pre-viable PROM) is a rare event occurring in less than 1% of pregnancies. Nevertheless, it can be responsible for severe maternal complications, the risk of which needs to be balanced with the possibility to prolong the pregnancy up to viable gestational age. Maternal sepsis was reported in 1%-5% of women who received conservative management and prophylactic antibiotics, but information on maternal mortality is lacking. Our objective was to identify maternal deaths in women who had pre-viable PROM, describe the characteristics of the women, explore preventability factors within the care they received, and estimate the lethality of pre-viable PROM.

MATERIAL AND METHODS

We identified all maternal deaths associated with pre-viable PROM from the 2001-2015 French National Confidential Enquiry into Maternal Deaths (NCMM). Data on women's characteristics and the care they received were extracted from the ENCMM database. The lethality was determined after estimating the total number of pregnant women with pre-viable PROM from the national hospital discharge database.

RESULTS

Between 2001 and 2015, we identified seven maternal deaths associated with pre-viable PROM, representing 0.6% of all maternal deaths over this period (ie, maternal mortality ratio 0.06/100 000 live births). Six maternal deaths were attributed to sepsis after genital infection by Gram-negative bacilli and one to postpartum hemorrhage due to placenta accreta. Four of these seven cases were considered preventable. The main preventability factors were delayed diagnosis, delayed fetal extraction, and inappropriate antibiotic treatment. The estimated lethality was 4.5/10 000 women with pre-viable PROM.

CONCLUSIONS

Maternal death associated with pre-viable PROM is rare but possible. Most of these deaths seem preventable, with areas for improvement related to earlier diagnosis and better treatment of uterine infections, which can evolve rapidly.

[Enhanced system for maternal mortality surveillance in France, context and Methods]

Maternal mortality, despite its rarity in rich countries, remains a fundamental indicator of maternal health. It is considered as a "sentinel event", consequence of dysfunctions of the health care system, often cumulative. In addition to the classical epidemiological surveillance outcomes-number of deaths, maternal mortality ratio and identification of the subgroups of women at higher risk-its study allows an accurate analysis of each deceased woman's trajectory to identify opportunities for improvements in the content or organization of care; the correction of which will make it possible to prevent deaths but also upstream morbid events affected by the same dysfunctions. To achieve this dual epidemiological and clinical audit objective, an ad hoc enhanced system is needed. France has had such a system since 1996, the National Confidential enquiry into maternal deaths (ENCMM), coordinated by the Inserm Epopé team. The methodology has been adapted over time to improve completeness and better document cases. The first step is the multi-source identification (direct declaration, death certificate, birth certificates, hospital discharge data) of women who died during pregnancy or within one year of its end, in metropolitan France and overseas departments. The second step is the collection of detailed information for each death by a pair of clinical assessors. Recent evolutions aim to better document the social context of women as well as the background of women who have died of suicide. Psychiatrists have been included among the assessors. The third stage is the review of these anonymized documents by the National Committee of Experts on Maternal Mortality, which judges whether the death is maternal (causal link) and makes a judgment on the adequacy of care and avoidability of death. A psychiatrist is now associated to the CNEMM for the assessment of maternal suicides. The synthesis of the information thus collected for maternal deaths in the period 2013-2015 is presented in these articles of this special issue.

[Maternal Mortality, Frequency, causes, women's profile and preventability of deaths in France, 2013-2015]

OBJECTIVE

To describe, for the 2013-2015 period, the frequency, causes, risk factors, adequacy of care and preventability of maternal deaths in France.

METHOD

Data from the National Confidential Enquiry into Maternal Deaths for 2013-2015.

RESULTS

For the period 2013-2015, 262 maternal deaths occurred in France, one every four days, i.e a maternal mortality ratio of 10.8 per 100,000 live births (95 % CI 9.5-12.1), stable compared to 2010-2012. Compared to women aged 25-29, the risk is multiplied by 1.9 for women aged 30-34, by 3 for women aged 35-39 and by 4 for women aged 40 and over. Obese women are twice as frequent among maternal deaths (24 %) than in the general population of parturients (11 %). There are territorial disparities - the maternal mortality ratio in the French overseas departments is 4 times higher than in metropolitan France -, and social disparities - the mortality of migrant women remains higher than that of women born in France, particularly for women born in sub-Saharan Africa whose risk is 2.5 times higher than that native women. A striking result of the 2013-2015 period is the preponderance of suicides and cardiovascular diseases, the two leading causes of maternal mortality, each responsible for approximately one death per month. These two etiologies are not only the most frequent, but also among those with the highest proportion of preventable deaths, 91.3 % and 65.7 % respectively. Another important result is the continued decrease in mortality from obstetric haemorrhage, halved in 15 years, particularly the decrease n hemorrhages due to uterine atony. Overall, 57.8 % of maternal deaths are considered probably or possibly preventable and in 66 % of cases, the care provided was not optimal.

CONCLUSION

While the overall maternal mortality ratio remains stable overall, and territorial and social inequalities unchanged, the profile of the causes of maternal mortality is changing. Some developments are a success, such as the continued decrease in maternal mortality due to haemorrhage, the result of the general mobilisation of health actors on this issue. Others point to new priorities for mobilisation, in particular on the mental and cardiovascular health of women during pregnancy or in the year following childbirth. In order to go further in understanding the mechanisms involved, and to identify precise avenues for prevention, it is necessary to analyse in detail the stories of each maternal death in order to identify the opportunities for improvement repeatedly found in the series of deaths. This is what the following articles in this issue propose, with an analysis by cause of death.

Understanding maternal mortality in women with obesity and the role of care they receive: a national case-control study

OBJECTIVE

Obesity has significant implications for the health of pregnant women. However, few studies have quantified its association with maternal mortality or examined the relevant underlying causes and the role of care, although this remains the most severe maternal outcome. Our objectives were to quantify the risk of maternal death by prepregnancy body mass index and to determine whether obesity affected the quality of care of the women who died.

DESING

This is a national population-based case-control study in France. Cases were 364 maternal deaths from the 2007-2012 National Confidential Enquiry. Controls were 14,681 parturients from the nationally representative 2010 perinatal survey. We studied the association between categories of prepregnancy BMI and maternal death by multivariable logistic regression, estimating adjusted odds ratios and 95% confidence intervals, overall and by specific causes of death. Individual case reviews assessed the quality of care provided to the women who died, by obesity status.

RESULTS

Compared with women with normal BMI, underweight women (<18.5 kg/m²) had an adjusted OR of death of 0.75 (95% CI, 0.42-1.33), overweight women (25-29.9 kg/m²) 1.65 (95% CI, 1.24-2.19), women with class 1 obesity (30-34.9 kg/m²) 2.22 (95% CI, 1.55-3.19) and those with class 2-3 obesity (≥35 kg/m²) 3.40 (95% CI, 2.17-5.33). Analysis by cause showed significant excess risk of maternal death due to cardiovascular diseases, venous thromboembolism, hypertensive complications and stroke in women with obesity. Suboptimal care was as frequent among women with (35/62, 57%) as without obesity (136/244, 56%), but this inadequate management was directly related to obesity among 14/35 (40%) obese women with suboptimal care. Several opportunities for improvement were identified.

CONCLUSIONS

The risk of maternal death increases with BMI; it multiplied by 1.6 in overweight women and more than tripled in pregnant women with severe obesity. Training clinicians in the specificities of care for pregnant women with obesity could improve their outcomes.

Non-convulsive status epilepticus as the initial manifestation in a family with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL)

Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is an autosomal dominant small-vessel disease caused by mutations of the NOTCH3 gene. It typically presents with migraine, recurrent brain ischaemia, and cognitive disorders. Seizures rarely present as the initial manifestation, with non-convulsive status epilepticus being even less frequent. We present a series of 3 related patients with this arteriopathy, 2 of whom presented status epilepticus as a manifestation of the disease.

Delivery Hospital Characteristics and Postpartum Maternal Mortality: A National Case-Control Study in France

BACKGROUND

The variability in resources for managing critical events among maternity hospitals may impact maternal safety. Our main objective was to assess the risk of postpartum maternal death according to hospitals' organizational characteristics. A secondary objective aimed to assess the specific risk of death due to postpartum hemorrhage (PPH).

METHODS

This national population-based case-control study included all 2007-2009 postpartum maternal deaths from the national confidential enquiry (n = 147 cases) and a 2010 national representative sample of parturients (n = 14,639 controls). To adjust for referral bias, cases were classified by time when the condition/complication responsible for the death occurred: postpartum maternal deaths due to conditions present before delivery (n = 66) or during or after delivery (n = 81). Characteristics of delivery hospitals included 24/7 on-site availability of an anesthesiologist and an obstetrician, level of perinatal care, number of deliveries annually, and their teaching and profit status. In teaching and other nonprofit hospitals in France, obstetric care is organized on the principle of collective team-based management, while in for-profit hospitals, this organization is based mostly on that of "one woman-one doctor." Logistic regression models were used to estimate adjusted odds ratios (aORs) and 95% confidence intervals (CIs) for postpartum maternal death.

RESULTS

The risk of maternal death from prepartum conditions was lower for women who gave birth in for-profit compared with teaching hospitals (aOR, 0.3; 95% CI, 0.1-0.8; P = .02) and in hospitals with <1500 vs ≥1500 annual deliveries (aOR, 0.4; 95% CI, 0.1-0.9; P = .02). Conversely, the risk of postpartum maternal death from complications occurring during or after delivery was higher for women who delivered in for-profit compared with teaching hospitals (aOR, 2.8; 95% CI, 1.3-6.0; P = .009), as was the risk of death from PPH in for-profit versus nonprofit hospitals (aOR, 2.8; 95% CI, 1.2-6.5; P = .019).

CONCLUSIONS

After adjustment for the referral bias related to prepartum morbidity, the risk of postpartum maternal mortality in France differs according to the hospital's organizational characteristics.

Maternal admissions to Intensive Care Units in France: trends in rates, causes and severity

Introduction

Maternal intensive care unit admission (ICU) is an indicator of severe maternal morbidity. This study aimed to estimate rates of maternal ICU admission during or following pregnancy in France, and to describe the characteristics of women admitted, the severity of their condition, associated diagnoses, regional disparities, and temporal trends between 2010 and 2014.

Methods

Women hospitalised in France in ICU during pregnancy or up to 42 days after pregnancy between 2010 and 2014 were identified using the national hospital discharge database (PMSI-MCO). The Simplified Acute Physiology Score (SAPS II) was used to estimate the severity. Trends in incidence rates were quantified using percentages of average annual variation based on a Poisson regression model.

Results

In total, 16,011 women were admitted to ICU, representing an overall rate of 3.97 ‰ deliveries. The average annual decrease in this rate between 2010 and 2014 was 1.7% (IC95%:-2, 00%; -1, 45%; p &lt; 0.0001) on average per year. The SAPS II score increased significantly from 18.4 in 2010 to 21.5 in 2014. Obstetrical hemorrhage (39.8%) and hypertensive complications during pregnancy (24.8%) were the most common reasons for admission. Within mainland France, we found notable disparities in maternal ICU admission rates between regions, from lowest in Pays-de-la-Loire region (2.69‰) to highest in Ile-de-France (5.05‰).

Conclusions

The rate of maternal ICU admission decreased from 2010 to 2014 in France, with a concomitant increase in case severity. Additional studies are needed to understand the territorial disparities identified in our study.

Key messages

The decreasing incidence of maternal ICU admission could be due to organisational changes with increased admission to intermediate care units. These changes have to be understood to accurately use maternal CU admission for maternal health surveillance.

[Maternal mortality among women with Marfan syndrome or vascular Ehlers-Danlos syndrome in France, 2001-2012]

OBJECTIVE

To describe maternal deaths in France associated with Marfan's syndrome or vascular Ehlers-Danlos syndrome.

STUDY DESIGN

A retrospective descriptive study based on data from the national confidential enquiry into maternal deaths, in France, during 2001-2012. Characteristics of the patients, their pregnancies and details of their deaths were analysed. The specific maternal mortality ratio by Marfan's syndrome or vascular Ehlers-Danlos syndrome was estimated.

RESULTS

Among 973 maternal deaths that occurred during the study period, five (0.4%) had a Marfan's syndrome (n=3) or a vascular Ehlers-Danlos syndrome (n=2), confirmed or suspected. The maternal mortality ratio due to Marfan's syndrome or vascular Ehlers-Danlos syndrome between 2001 and 2012 was 0.04/100,000 live births (IC 95% [0.011-0.2]). Three maternal deaths were caused by aortic dissections and two by other arterial ruptures. The deaths have occurred after 37 weeks of pregnancy for 4 patients, and at fifteen days of post-partum for one patient. The median age of death was 30 years. Three patients were nulliparous. Marfan's syndrome and vascular Ehlers-Danlos syndrome were not identified before the death of these five patients.

CONCLUSION

Five patients with, or suspected to have, Marfan's syndrome or vascular Ehlers-Danlos syndrome were identified. Early diagnosis of these syndromes in pregnant women before life threatening events is very important, especially to refer them to appropriate care.

[Enhanced system for maternal mortality surveillance in France, context and methods]

Maternal mortality, despite its rarity in rich countries, remains a fundamental indicator of maternal health. It is considered as a "sentinel event", consequence of dysfunctions of the health care system, often cumulative. In addition to the classical epidemiological surveillance outcomes-number of deaths, maternal mortality ratio and identification of the subgroups of women at risk-its study allows an accurate analysis of each deceased woman's trajectory to identify opportunities for improvements in the content or organization of care; the correction of which will make it possible to prevent deaths but also upstream morbid events affected by the same dysfunctions. To achieve this dual epidemiological and clinical audit objective, an ad hoc enhanced system is needed. France has had such a system since 1996, the National Confidential enquiry into maternal deaths (ENCMM), coordinated by the Inserm Epopé team. The first step is the multi-source identification (direct declaration, death certificate, birth certificates, hospital discharge data) of women who died during pregnancy or within one year of its end. The second step is the collection of detailed information for each death by a pair of clinical assessors. The third stage is the review of these anonymized documents by the National Committee of Experts on Maternal Mortality, which judges whether the death is maternal (causal link) and makes a judgment on the adequacy of care and avoidability of death. The synthesis of the information thus collected for maternal deaths in the period 2010-2012 is the subject of the last report.

[Epidemiology of maternal mortality in France, 2010-2012]

OBJECTIVE

To describe, for the period 2010-2012, the frequency, the causes, the risk factors, the adequacy of care and the avoidability of maternal deaths in France.

METHOD

Data from the National Confidential Enquiry into Maternal Deaths for 2010-2012.

RESULTS

For the period 2010-2012, 256 maternal deaths occurred in France, a maternal mortality ratio of 10.3 per 100,000 live births (95% CI: 9.1-11.7), stable compared to 2007-2009. Compared to women aged 25-29, the risk is multiplied by 2.4 for women aged 35-39, and by 3 for women over 40 years. There are territorial disparities: 1 out of 7 maternal deaths occurs in the French overseas departments, and the maternal mortality ratio in those departments is 4 times that of metropolitan, France; and social disparities: the mortality of migrant women remains 2.5 times higher than that of women born in France, particularly for women born in sub-Saharan Africa whose RMM is 3.5 times that native women. A major finding is the 1/3 decrease in direct maternal mortality over the last 10 years, mainly due to for the first time the statistically significant decrease in mortality from obstetric hemorrhage, the frequency of which was divided by 2 in 10 years. However, almost all of the remaining deaths from hemorrhage are considered preventable and this is still the leading cause of maternal mortality in France (11% of deaths). Overall, 56% of these maternal deaths are considered "avoidable" or "possibly avoidable" and in 59% of cases the care provided was not optimal.

CONCLUSION

Direct maternal mortality and in particular maternal mortality from hemorrhage has decreased significantly over the past 10 years, indicating improved obstetric care. However, territorial and social inequalities persist, and the majority of deaths remain preventable, which shows that the identification of opportunities for improvement must continue. To go even further in understanding the mechanisms involved, and to identify precise avenues of prevention, it is necessary to analyze in detail the stories of each maternal death in order to identify the repetitive elements in the series of deaths. This is what the following articles in this issue propose, with an analysis by cause of death, according to the idea that the same cause produces the same effects.

Caesarean Delivery and Postpartum Maternal Mortality: A Population-Based Case Control Study in Brazil

BACKGROUND

Cesarean delivery rates continue to increase worldwide and reached 57% in Brazil in 2014. Although the safety of this surgery has improved in the last decades, this trend is a concern because it carries potential risks to women's health and may be a modifiable risk factor of maternal mortality. This paper aims to investigate the risk of postpartum maternal death directly associated with cesarean delivery in comparison to vaginal delivery in Brazil.

METHODS

This was a population-based case-control study performed in eight Brazilian states. To control for indication bias, deaths due to antenatal morbidity were excluded. We included 73 cases of postpartum maternal deaths from 2009-2012. Controls were selected from the Birth in Brazil Study, a 2011 nationwide survey including 9,221 postpartum women. We examined the association of cesarean section and postpartum maternal death by multivariate logistic regression, adjusting for confounders.

RESULTS

After controlling for indication bias and confounders, the risk of postpartum maternal death was almost three-fold higher with cesarean than vaginal delivery (OR 2.87, 95% CI 1.63-5.06), mainly due to deaths from postpartum hemorrhage and complications of anesthesia.

CONCLUSION

Cesarean delivery is an independent risk factor of postpartum maternal death. Clinicians and patients should consider this fact in balancing the benefits and risks of the procedure.

Pitfalls of national routine death statistics for maternal mortality study

BACKGROUND

The lessons learned from the study of maternal deaths depend on the accuracy of data. Our objective was to assess time trends in the underestimation of maternal mortality (MM) in the national routine death statistics in France and to evaluate their current accuracy for the selection and causes of maternal deaths.

METHODS

National data obtained by enhanced methods in 1989, 1999, and 2007-09 were used as the gold standard to assess time trends in the underestimation of MM ratios (MMRs) in death statistics. Enhanced data and death statistics for 2007-09 were further compared by characterising false negatives (FNs) and false positives (FPs). The distribution of cause-specific MMRs, as assessed by each system, was described.

RESULTS

Underestimation of MM in death statistics decreased from 55.6% in 1989 to 11.4% in 2007-09 (P < 0.001). In 2007-09, of 787 pregnancy-associated deaths, 254 were classified as maternal by the enhanced system and 211 by the death statistics; 34% of maternal deaths in the enhanced system were FNs in the death statistics, and 20% of maternal deaths in the death statistics were FPs. The hierarchy of causes of MM differed between the two systems. The discordances were mainly explained by the lack of precision in the drafting of death certificates by clinicians.

CONCLUSION

Although the underestimation of MM in routine death statistics has decreased substantially over time, one third of maternal deaths remain unidentified, and the main causes of death are incorrectly identified in these data. Defining relevant priorities in maternal health requires the use of enhanced methods for MM study.

Understanding regional differences in maternal mortality: a national case-control study in France

OBJECTIVES

To assess the risk of postpartum maternal death associated with region, and to examine whether the quality of care received by the women who died differed by region.

DESIGN

A national case-control study.

SETTING

France.

POPULATION

Selected from recent nationwide surveys, 328 postpartum maternal deaths from 2001 through 2006 as cases; and a representative sample (n = 14 878) of women who gave birth in 2003 as controls.

METHODS

Crude and adjusted odds ratios (aOR) of maternal death associated with region were calculated with logistic regression, and the quality of care for women who died was compared according to region with chi-square tests or Fisher's exact tests.

MAIN OUTCOME MEASURES

Risk of postpartum maternal death associated with region, and quality of care.

RESULTS

After adjustment for maternal age and nationality, the risk of maternal death was higher in the Ile-de-France region (aOR 1.6, 95% CI 1.2-2.0) and the overseas districts (aOR 3.5, 95% CI 2.4-5.0) than in the group for the rest of continental France. In both regions, the excess risk of death from haemorrhage, amniotic fluid embolism and hypertensive disorders was significant. In continental France, after further controlling for women's obstetric characteristics, the risk of maternal death in Ile-de-France remained higher (aOR 1.8. 95% CI 1.3-2.6). The women in the cases groups received suboptimal care more frequently in Ile-de-France than in the other continental regions (64% versus 43%, P = 0.01).

CONCLUSIONS

These results suggest that quality of care and organisation of health services may play a role in the differential risk of maternal mortality between regions in France. Research on severe maternal morbidity and its determinants is needed to clarify the mechanisms involved.

[The confidential enquiries into maternal deaths, 1996-2006 in France: what consequences for the obstetrical care?]

The national confidential enquiry into maternal deaths (ENCMM) and its committee (CNEMM) have the target to study all maternal deaths occurring in France, in order to expertise the care provided. The current report covers the 1996--year of the ENCMM establishment--to 2006 years. After being informed of the potential maternal deaths by the Epidemiological center on medical causes of deaths (CépiDC), and agreement from the medical doctors concerned, two assessors (one anesthetist and one obstetrician) gather the medical or obstetrical information near the team involved in the care of the women, by the mean of a detailed and specific questionnaire. The completely anonymous files are expertised by the CNEMM. Maternal mortality rates have been calculated by periods, the distribution of the obstetrical causes and the characteristics of the dead women were calculated too. The substandard care and the avoidability of deaths were estimated by subgroup. Since 1996 to 2006, 729 maternal deaths were included of which 553 were expertised. The majority of maternal deaths were due to direct obstetrical causes (73%) mainly haemorrhages (22%), amniotic fluid embolism (12%), complications of hypertension (10 %), and venous thrombo-embolism (around 10 % each). Half of maternal deaths were considered preventable by the CNEMM, particularly haemorrhage and sepsis. The factors of avoidability are delay to treat (31%) inadapted therapeutics (28%), even professional default (20%) no diagnosis (15%) or reluctant patient (7%). Seven deaths are discussed in a specific section including a detailed description of, and recommendations on how the quality of care may be improved.