Maternal sudden death: A nationwide retrospective study

Abstract

Objective

To determine the prevalence of maternal sudden death (MSD) and to compare the characteristics of death between women with explained and unexplained sudden death.

Design

A national retrospective study in France.

Population

Maternal deaths related to an unexpected sudden cardiac arrest were extracted from the French National Confidential Enquiry into Maternal Deaths database for 2007–2012.

Methods

Maternal, pregnancy, sudden death characteristics and maternal investigations were compared between women with explained and unexplained cause of death.

Results

A total of 83 maternal sudden deaths and 4 949 890 live births occurred over the period studied, thus accounting for 16% of all maternal deaths (n = 510). Death was explained in 51 (61%) women and unexplained in 32 women (39%). Compared with women with unexplained death, women with explained death were more often found to have in‐hospital cardiac arrest (47% versus 12%, P < 0.01), witnessed cardiac arrest (86% versus 62%, P = 0.03) and in‐hospital death (82% versus 47%, P < 0.01). Postmortem investigations such as autopsy and/or CT scan (65% versus 31%, P < 0.01) were also more often carried out in women with explained death. The proportion of deaths for which the preventability factors could not be assessed was 58% among unexplained MSD and 7% among explained MSD.

Conclusion

Maternal sudden death is a rare event but accounts for a high proportion of all maternal deaths. This highlights the importance of providing training in diagnostic and management strategy for care providers. Systematic postmortem investigations are required to help understand causes and improve practices.

1. INTRODUCTION

Sudden death is a sudden, unexpected, non‐traumatic death without drowning, occurring within 1 hour of symptom onset (event witnessed) or within 24 hours of having been seen alive and symptom free (unwitnessed). ¹ The majority of sudden deaths in adults are caused by cardiac arrhythmias, including those resulting from acute myocardial infarction. ² However, due to the very low rate of autopsy performed (ranging from 10% to 23%), particularly when sudden death occurs out‐of‐hospital, the cause of death reported on the death certificate may be speculative. ³ , ⁴ In a recent prospective study in the general population, postmortem systematic investigation enabled identification of more than 40% of non‐cardiac and non‐arrhythmic causes of sudden death. ⁵

Despite maternal deaths being a rare occurrence in high‐resource countries, the maternal mortality ratio remains an important national health indicator. Accurate documentation of cause of death and events leading up to death, helps improve both understanding of these dramatic events and upstream management to prevent fatal outcome. This is currently recommended by confidential enquiries set up in several countries, including the UK and France. ⁶ , ⁷ Beyond the traditional categorisation of maternal deaths into direct and indirect causes, analysis of sudden maternal deaths is of interest, as these deaths have a common clinical presentation but multiple possible factors which may differ from those observed outside pregnancy. Documenting the profile and causes of sudden maternal deaths may reveal avenues for their prevention. To our knowledge, although the available literature provides information on selected subgroups of maternal sudden deaths such as cardiac sudden deaths, ⁸ to date no previous study has investigated all sudden maternal deaths. This overall category and the contribution of each cause of sudden death remains as yet undocumented.

In France, as in most high‐resource countries, cardiovascular conditions have become the first cause of maternal death, accounting for 14% of maternal mortality in the most recent period; yet, strictly speaking, the ‘undetermined’ cause‐of‐death category is of equal contribution, with 14% of maternal deaths recorded as having no established cause. ⁹ The fact that over half of these undetermined deaths are sudden, reinforces the importance of further investigation into sudden maternal deaths.

The aim of this study was to determine the contribution of maternal sudden deaths (MSD) to total maternal mortality, to identify causal conditions and to compare context of death characteristics for explained sudden deaths (those with a characterised cause of death) and unexplained sudden deaths.

2. METHODS

We identified maternal deaths from the French National Confidential Enquiry into the Maternal Deaths database (Enquête Confidentielle sur les Morts Maternelles [ENCMM]). ¹⁰ This permanent database has since 1996 allowed collation of data on all pregnancy‐associated deaths, for women who were pregnant or within 1 year of the termination of pregnancy. Deaths are identified from three sources: (1) death certificates with any cause of death coded in the pregnancy chapter of the International Classification of Diseases, 10th Revision (ICD‐10), any mention of pregnancy or puerperium in the text, or when the pregnancy checkbox was ticked; (2) computer‐based national linkage of the death and birth registers to identify women who died within a year after pregnancy; and (3) hospital discharge database identifying hospitalisations of women with at least one diagnostic code in the ICD‐10 O00–O99 range or a code related to pregnancy, delivery or the postpartum period who died during the hospitalisation. For each identified pregnancy‐associated death, a team of assessors (an obstetrician or midwife and an anaesthesiologist) conducted a confidential enquiry using a standardised questionnaire to collect relevant clinical information about the woman and her death, via interviews and a review of all medical records and autopsy reports. When a confidential enquiry is not possible, information from the death certificate and hospital discharge summary is analysed. Deaths are then anonymously reviewed by the ENCMM national expert committee to reach a unanimous decision on the underlying cause of death, whether it was a maternal death (defined as a woman's death during pregnancy or within 1 year of its end, regardless of its duration and site, from any cause related to or aggravated by the pregnancy or its management, but not accidental or incidental), its preventability and on the adequacy of care provided. A maternal death is considered preventable if one or more changes in the care provided or in patient behaviour might have prevented the fatal outcome.

For this analysis, we reviewed all maternal deaths that occurred between 2007 and 2012 (most recent completed year at the time of our analysis) and selected those who met MSD criteria. MSD was defined as a maternal death related to a sudden unexpected abrupt loss of consciousness with cardiac arrest that occurs within 1 hour of the symptoms when it was witnessed, and within 24 hours of being seen alive and well when not witnessed. ² The variables of interest, i.e. maternal, pregnancy and sudden death characteristics, were extracted from the national confidential enquiry database. Maternal socio‐demographical and medical variables collected were age, parity, immigrant status, living with a partner, occupation, body mass index (BMI), smoking, previous medical history (including chronic hypertension, diabetes mellitus, epilepsy and cardiac diseases). Pregnancy variables recorded were in vitro fertilisation, multiple pregnancy and hypertensive disorders during pregnancy. Sudden death variables evaluated were cardiac arrest characteristics (including ante‐ or postpartum, presence of witnesses, in‐hospital occurrence, preceding symptoms) and investigations performed (including autopsy, body scan, cardiac ultrasound, laboratory tests for amniotic embolism).

Quality of care was reported as suboptimal, optimal or undetermined. Death preventability was classified as non‐preventable, probably preventable, possibly preventable or undetermined.

2.1. Statistical analyses

We calculated mortality ratio for MSD with exact Poisson 95% confidence interval (95% CI) using the available national birth data during the study period. All variables were described using mean ± SD for quantitative variables, or numbers with percentages for qualitative variables. In the univariate analysis, we compared maternal, pregnancy and sudden death characteristics and maternal investigations between women with explained and unexplained cause of death using Chi‐square and Fisher exact tests when appropriate for qualitative variables, and Student t‐tests for quantitative variables. We compared quality of care and death preventability between women with explained and unexplained cause of death using Chi‐square and Fisher exact tests. Analyses were performed using STATA 11 software (StataCorp, College Station, TX, USA).

3. RESULTS

From January 2007 till December 2012, 510 maternal deaths and 4 949 890 live births occurred in France for a total maternal mortality ratio of 10.3 per 100 000 live births (95% CI 9.4–11.2). During this 6‐year period, 83 maternal deaths (16.3%) were classified as MSD, thus there was a total maternal sudden mortality ratio of 1.7 par 100 000 live births (95% CI 1.3–2.1).

Of the 83 sudden maternal deaths, the causal condition was characterised for 51 (61.4%) women and remained unexplained for 32 women (38.6%). Explained causes of MSD were pulmonary thromboembolism (n = 19, 23%), amniotic fluid embolism (n = 14, 17%), cardiovascular diseases (n = 10, 12%, including two sudden arrhythmic deaths), epilepsy (n = 5, 6%), stroke (n = 2, 2%) and air embolism (n = 1, 1%) Maternal and pregnancy characteristics are reported in Table 1. For all sudden deaths, the median maternal age was 32.3 years, 31.6% of women were immigrants, 22.5% were primiparous and 27.2% had a previous medical history. There was no statistical difference in maternal or pregnancy characteristics between women with explained and unexplained cause of death (Table 1).

TABLE 1.

Maternal and pregnancy characteristics of women with maternal sudden deaths, overall and for women with unexplained versus explained cause of death

Socio‐demographical factors (n a )	All maternal sudden deaths (n = 83)	Women with unexplained sudden death (n = 32)	Women with explained sudden death (n = 51)	P‐value
Age (n = 83) (years)
Mean ± SD	32.3 ± 5.7	32.7 ± 6.5	32 ± 5.2	0.60
Age ≥35	34 (40.9%)	16 (50%)	18 (35.3%)	0.18
Immigrant women (n = 79)	25 (31.6%)	9 (30.0%)	16 (32.6%)	0.81
BMI (n = 71) (kg/m2)
Mean ± SD	27.7 ± 8	28.5 ± 8.8	27.3 ± 7.6	0.57
BMI ≥30	18 (25.3%)	6 (24.0%)	12 (26.1%)	0.84
Without a partner (n = 79)	18 (22.8%)	9 (29.0%)	9 (18.7%)	0.53
Unemployed (n = 64)	16 (25%)	8 (33.3%)	8 (20.0%)	0.23
Primiparous (n = 80)	18 (22.5%)	4 (12.5%)	14 (27.4%)	0.24
Previous medical history (n = 81)	22 (27.2%)	6 (18.8%)	16 (31.4%)	0.40
Smoking (n = 79)	9 (11.4%)	5 (16.5%)	4 (8%)	0.50
Pregnancy characteristics
In vitro fertilisation (n = 71)	6 (8.4%)	2 (7.4%)	4 (9.1%)	0.58
Multiple pregnancy (n = 78)	4 (5.1%)	1 (3.2%)	3 (6.4%)	0.48
Hypertensive disorders during pregnancy (n = 76)	12 (15.8%)	4 (13.8%)	8 (17.0%)	0.91

Note: Data are presented as mean ± standard deviation or n (%).

Abbreviation: BMI, body mass index.

^a n, number of women with available data.

Compared with women with unexplained death, women with explained death had a greater number of in‐hospital cardiac arrests (47% versus 12%, P < 0.01), witnessed cardiac arrests (86% versus 62%, P = 0.03) and in‐hospital deaths (82% versus 47%, P < 0.01). Investigations such as a body scan (29% versus 0%), an autopsy and/or CT scan (65% versus 31%, P < 0.01), and laboratory tests for AFE (14% versus 0%) were also more often carried out for women with explained death than for those with unexplained death (Table 2).

TABLE 2.

Sudden death characteristics and maternal investigations, overall and for women with unexplained versus explained cause of death

Characteristics (n a )	All maternal sudden deaths (n = 83)	Women with unexplained sudden death (n = 32)	Women with explained sudden death (n = 51)	P‐value
Antepartum cardiac arrest (n = 83)	52 (62.6%)	19 (59.3%)	33 (64.7%)	0.62
In‐hospital cardiac arrest (n = 83)	28 (33.7%)	4 (12.5%)	24 (47.1%)	<0.01
Witnessed cardiac arrest (n = 83)	65 (78.3%)	21 (65.2%)	44 (86.3%)	0.03
Symptoms before cardiac arrest (n = 76)	53 (69.7%)	17 (60.7%)	36 (70.5%)	0.11
In‐hospital death (n = 82)	57 (69.5%)	15 (46.8%)	42 (82.5%)	<0.01
Autopsy carried out (n = 83)	34 (41.4%)	10 (31.3%)	24 (47.1%)	0.15
Body scan carried out (n = 83)	15 (18.1%)	0 (0%)	15 (29.4%)
Autopsy and/or CT Scan (n = 83)	43 (51.8%)	10 (31.3%)	33 (64.7%)	<0.01
Cardiac US (n = 83)	5 (6.0%)	1 (3.1)	4 (7.8%)	0.35
Laboratory tests for AFE (n = 83)	7 (8.6%)	0	7 (13.7%)

Abbreviations: AFE, amniotic fluid embolism; CT‐Scan, computerised tomodensitometry; US, ultrasound.

^a n, number of women with available data.

Table 3 shows results regarding the quality of care and death preventability as assessed by the national expert committee. Overall, the quality of care was considered suboptimal in 17 (23%) women and death was considered probably (n = 1) or possibly (n = 12) preventable in 13 women (17%). The proportion of deaths for which preventability factors could not be assessed was 58% among unexplained MSD and 7% among explained MSD.

TABLE 3.

Quality of care and death preventability, overall and for women with unexplained versus explained cause of death

Characteristics (n a )	All maternal sudden deaths (n = 83)	Women with unexplained sudden death (n = 32)	Women with explained sudden death (n = 51)	P‐value
Quality of care (n = 75)
Suboptimal	17 (22.6%)	2 (7.7%)	15 (30.6%)	0.03
Optimal	47 (62.7%)	18 (69.2%)	29 (59.2%)
Undetermined	11 (14.7%)	6 (23.1%)	5 (10.2%)
Preventable death (n = 75)
No	41 (54.7%)	10 (38.5%)	31 (63.3%)	<0.01
Yes	1 (1.3%)	0	1 (2.0%)
Possibly	12 (16.0%)	1 (3.8%)	11 (22.5%)
Undetermined	21 (28.0%)	15 (57.7%)	6 (6.7%)

Note: Data are presented as n (%).

^a n, number of women with available data.

4. DISCUSSION

4.1. Main findings

Maternal sudden death is a rare event, but it accounted for a high proportion of all maternal deaths during the study period. More than one‐third of maternal sudden deaths remained unexplained, particularly in women with out‐of‐hospital cardiac arrest and when postmortem investigations were not carried out.

4.2. Strengths and limitations

Our study has several strengths. To our knowledge, it is the first nationwide investigation of all cases of MSD over a given period. Previous studies have reported on subgroups of maternal sudden deaths due to one specific cause (such as cardiac conditions ⁸ or amniotic fluid ¹¹ ), whereas our study included all sudden deaths. Data were collected from the confidential enquiry into maternal deaths database, enabling full identification and expertise of deaths associated with pregnancies, thus ensuring the accuracy and completeness of case identification.

However, our study also has limitations. First, it is a retrospective study with maternal and pregnancy data extracted from patient medical records, which may be incomplete, as reflected in the proportion of missing data in our analysis. This is notably restrictive for social determinants, which tend to be limited or poorly documented in medical records. Secondly, our results refer to a not‐so‐recent period, reflecting as in many European countries, the time lag before finalised birth and death data for a given year, are made available for research. ¹² Thirdly, the low study case numbers led to limited power for demonstration of statistically significant differences between the two groups concerning maternal and pregnancy characteristics.

4.3. Interpretation

4.3.1. An important contribution of maternal sudden deaths to overall maternal mortality

Our results show that approximately one in six maternal deaths are sudden deaths. This high proportion, occurring both in and outside the hospital, indicates the importance for both obstetric and primary care providers to be prepared for this eventuality and be trained in the diagnostic and resuscitation strategies required should this event occur. ¹³ This concerns not only obstetric care providers but also primary care providers, as shown by the high proportion of these MSDs occurring outside the hospital.

4.3.2. A high proportion of unexplained sudden deaths

We report a high rate (38.6%) of unexplained MSD, higher than the overall rates of unexplained maternal deaths reported in high‐resource countries (ranges from 6% to 14%). ¹⁴ , ¹⁵ , ¹⁶ MSD were more often unexplained when cardiac arrest and/or death occurred out‐of‐hospital, possibly reflecting difficulties in obtaining postmortem examinations. Similarly, Bagnall et al. reported a higher rate of unexplained sudden death in children and young adults when death occurred at night. ¹⁷

We also found that unexplained MSD were characterised by a low rate (31%) for performed autopsies and/or postpartum CT body scans, similar to that observed in adult sudden cardiac deaths, ¹⁸ , ¹⁹ and this despite French adherence in 1999 to European Union recommendations on the Harmonisation of Medico‐Legal Autopsy Rules, concerning the necessity for a medical‐legal autopsy in all cases of unexpected, unexplained or alleged unnatural deaths. ²⁰ In 2010, the Royal College of Pathologists also published guidelines for maternal autopsy procedures, recommending that an autopsy should be performed on as many maternal deaths as possible. ²¹ Currently, the UK follows a systematic forensic autopsy protocol and reports the highest autopsy rate (84%) of maternal deaths worldwide. ²² The availability of this standardised autopsy data enables better identification of the causes of deaths and their remediable factors, consequently reducing maternal mortality. ²² In France, however, an overall autopsy rate in maternal deaths of approximately 29%, has remained low and unchanged over the last decade. ¹⁴ Possible reasons for this low rate include a persistent cultural reluctance to offer and accept autopsy, coupled with insufficient training for the care providers involved, a lack of professionals nationwide trained in maternal autopsy, and administrative barriers and costs associated with the transfer of the body and the autopsy procedure itself. National regulations are required to establish procedure, the financing of maternal autopsy and the introduction of systematic performing of postmortem standardised histopathological examination, notably for young, previously healthy women dying at home.

Although we are unable to demonstrate directly that a better understanding of the causal condition of these sudden deaths would lead to care improvement, a standard approach to maternal death review and surveillance allows an understanding of the chain of events that lead to death, provides identification and the proposal of measures for prevention of future deaths, in addition to possible corresponding non‐fatal morbid events; the key first step being to establish cause of death. Our data show that the preventability of death could not be determined for a majority (58%) of unexplained MSD, compared with only a small number (7%) among explained MSD.

4.3.3. Causes of maternal sudden death

In our study, the main causes of death among explained MSD were pulmonary embolism, amniotic fluid embolism and cardiovascular conditions, the same as those reported in fatal cases of cardiac arrest during pregnancy in the CAPS study in the UK. ²³ Interestingly, we report only two cases of sudden arrhythmic cardiac death. This low number contrasts with cardiac autopsy findings in a series of 80 maternal sudden cardiac deaths in UK showing that the main cause of death was sudden arrhythmic cardiac death (53.8%), followed by cardiomyopathies (13.8%), dissection of aorta or its branches (8.8%), congenital heart disease (2.5%) and valvular disease (3.7%). ⁸ In the 2009–2014 report on maternal deaths in the UK and Ireland, ²⁴ maternal sudden arrhythmic cardiac death was also the most frequent cause of cardiac deaths (38%). The diagnosis of sudden arrhythmic cardiac death is retained when a good quality, complete autopsy, associated with toxicology, microbiology and biochemistry investigations, rules out other causes; without such an examination, this diagnosis cannot be made. This highlights the need for systematic postmortem examinations in cases of MSD. It is likely that the low rate of autopsies in France prevents the identification of sudden arrhythmic deaths among unexplained MSD. Beyond understanding cause of death, identification of constitutional disease in the deceased woman opens opportunities for family investigation and preventive measures for surviving relatives.

5. CONCLUSION

Maternal sudden death is a rare event, but it accounts for a high proportion of all maternal deaths. This underlines the importance for providing training for care providers in diagnostic and management strategy. National regulations are required to facilitate performance of systematic postmortem investigations to allow greater understanding of maternal sudden deaths and their prevention.

AUTHOR CONTRIBUTIONS

CDT coordinateD the ENCMM system and data collection. JL and EV conceived and designed the analysis, extracted the data and performed the analyses. SB, JL, CDT, EM and EV wrote the paper.

FUNDING INFORMATION

The Enquête Nationale Confidentielle sur les Morts Maternelles (ENCMM, French National Confidential Enquiry into Maternal Deaths) is funded by the French Institute for Public Health (Santé Publique France) and by the French National Institute of Health and Medical Research (Inserm). The authors received no funding for this specific analysis.

CONFLICT OF INTERESTS

None declared. Completed disclosure of interest forms are available to view online as supporting information.

ETHICS APPROVAL

Ethical approval for the ENCMM was granted by the French Commission on Information Technology and Liberties (CNIL) on 26 June 2018, DR‐2018‐157. The present retrospective study was approved by the local institutional review board (Comité d'Ethique de la Recherche Non‐Interventionnelle du CHU de Rouen) (E2022‐29).

Supporting information

Appendix S1:

Braund S, Leviel J, Morau E, Deneux‐Tharaux C, Verspyck E, ENCMM study group . Maternal sudden death: A nationwide retrospective study. BJOG. 2023;130(3):257–263. 10.1111/1471-0528.17294

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.