The line shows the rate of SMM. The bars show the percent relative change in the number of pregnancies.
eTable 1. List of variables in this study
eTable 2. Frequency individual indicators of severe maternal morbidity, among 54,219 separate severe maternal morbidity events
eTable 3. Most common causes of severe maternal morbidity, within the antepartum period, within the delivery or peripartum period, and within the postpartum readmission and the number needed to experience harm (death) for each condition (i.e. the number of women with this condition needed to result in one additional death)
eTable 4. Most common causes of severe maternal morbidity by age category
eTable 5. Estimated regression coefficients for the multi-level mixed logistic regression models for the outcome of Severe Maternal Morbidity [Outcome = Severe Maternal Morbidity, main predictor = quintile of hospitals according to the number of pregnancy admission at each hospital]
eTable 6. Estimated regression coefficients for the multi-level mixed logistic regression models for the outcome of death [Outcome = death, main predictor = quintile of hospitals according to the number of pregnancy admission at each hospital]
eTable 7. Sensitivity analysis: Estimated regression coefficients for the multi-level mixed logistic regression models with different main predictors for the outcome of SMM and death [Outcome = (a) Severe Maternal Morbidity, (b) Death, main predictors = quintile of hospitals according to (a) the number of pregnancy admission to ICU, (b) pregnancy-related ICU admission rate at each hospital]
eTable 8. Sensitivity analysis: Estimated regression coefficients for the multi-level mixed logistic regression models for the outcome of Severe Maternal Morbidity, restricted to first pregnancy [Outcome = Severe Maternal Morbidity, main predictor = quintile of hospitals according to the number of pregnancy admission at each hospital]
eTable 9. Estimated regression coefficients for the multi-level mixed logistic regression models for the outcome of (a) SMM, (b) Death within multiple imputation datasets [Outcome = (a) Severe Maternal Morbidity, (b) Death, main predictor = quintile of hospitals according to the number of pregnancy admission at each hospital]
eTable 10. Sensitivity analysis: Estimated regression coefficients for the multi-level mixed logistic regression models for the outcome of Severe Maternal Morbidity, excluding those who had severe maternal morbidity during the episode of care and transfer on first hospital admission l during the episode [Outcome = severe maternal morbidity, main predictor = quintile of hospitals according to the number of pregnancy admission at each hospital]
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Aoyama K, Pinto R, Ray JG, et al. Association of Maternal Age With Severe Maternal Morbidity and Mortality in Canada. JAMA Netw Open. 2019;2(8):e199875. doi:10.1001/jamanetworkopen.2019.9875
在这项针对加拿大 310 万孕产妇进行的全国人口队列研究中，过去十年来严重的孕产妇发病率有所上升。这一趋势与孕产妇年龄以及孕产妇与高龄母亲的比例随时间推移而增加的情况相吻合。与 20 至 24 岁的成年人相比，极端年龄的孕产妇，特别是 45 岁或以上，与严重的孕产妇发病率和孕产妇死亡率相关。
Over the past 2 decades, there has been a trend toward increasing maternal age in many high-income countries. Maternal age may lead to greater attendant morbidity and mortality for Canadian mothers.
To investigate the association of maternal age, adjusting for patient-level and hospital-level factors, with severe maternal morbidity (SMM) and maternal death in Canada.
Design, Setting, and Participants
A nationwide population-based cohort study of all antepartum, peripartum, and postpartum women and adolescents seen at Canadian acute care hospitals from April 1, 2004, to March 31, 2015. All analyses were completed on September 13, 2018.
Maternal age at the index delivery.
Main Outcomes and Measures
Severe maternal morbidity and maternal death during pregnancy and within 6 weeks after termination of pregnancy.
During the study period, there were 3 162 303 new pregnancies (mean [SD] maternal age, 29.5 [5.6] years) and 3 533 259 related hospital admissions. There were 54 219 episodes of SMM (17.7 cases per 1000 deliveries) in the entire study period, with a 9.8% relative increase from 2004-2005 to 2014-2015, in addition to an increasing proportion of pregnancies to older mothers. Independent patient-level factors associated with SMM included increasing Maternal Comorbidity Index; maternal age 19 years or younger and 30 years or older, with the greatest risk experienced by women 45 years or older (odds ratio [OR], 2.69; 95% CI, 2.34-3.06 compared with maternal age 20-24 years); and lowest income quintile (OR, 1.19; 95% CI, 1.14-1.22 compared with highest income quintile). Hospital-level factors associated with SMM included specific provinces. Independent patient-level factors associated with maternal mortality included increasing Maternal Comorbidity Index, age 40 to 44 years (OR, 3.39; 95% CI, 1.68-6.82 compared with age 20-24 years), age 45 years or older (OR, 4.39; 95% CI, 1.01-19.10 compared with age 20-24 years), and lowest income quintile (OR, 4.14; 95% CI, 2.03-8.50 compared with highest income quintile). Hospital-level factors associated with maternal mortality included lowest hospital pregnancy volume.
Conclusions and Relevance
In Canada, maternal age and SMM have increased over the past decade. Results of this study suggest that province of residence, maternal comorbidity, residence income quintile, and extremes of maternal age, especially those 45 years or older, were associated with SMM and mortality. These findings are relevant to prospective parents, their health care team, and public health planning.
The World Health Organization estimates that more than 300 000 women die each year from pregnancy or childbirth-related complications.1 The Global Strategy for Women’s, Children’s and Adolescents’ Health (2016-2030), seeks to end all preventable deaths of women, children, and adolescents.1 While maternal mortality has decreased in the developed world over the past 25 years,2 a recent nationwide cohort study found that severe maternal morbidity (SMM) in Canada may be increasing.3,4 The latest rate of SMM in Canada is around 17 cases per 1000 deliveries.3
In many high-income countries, adolescent (age range, 10-19 years) birth rates are declining, whereas there has been a trend toward increasing maternal age.5 Hence, more attention has shifted toward ascertaining adverse outcomes in older mothers (≥35 years). Older maternal age is associated with increased childbirth-related complications.6-9 However, little information is available on associations of increasing maternal age with SMM and maternal mortality, especially in the Canadian context.
Therefore, we performed a nationwide population-based cohort study of pregnancies in Canada. We explored trends in maternal age and the association between maternal age and outcomes for pregnant and postpartum women and adolescents.
All pregnant and postpartum (within 6 weeks of delivery) women and adolescents, including those with ectopic pregnancy or spontaneous abortion, seen at acute care hospitals from April 1, 2004, to March 31, 2015, were included in this nationwide population-based cohort study. All analyses were completed on September 13, 2018. Variables included maternal demographics (age; comorbid conditions as classified by the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Canada [ICD-10CA]; socioeconomic status; and residence location), pregnancy-specific variables (gestational week at admission and parity), hospital-level and system-level variables (province, territory, and urban vs rural location), and outcome-related variables (vital status at hospital discharge, transfers between hospitals, and readmissions) (eTable 1 in the Supplement). A previously described algorithm captures all hospital admissions and discharges over a single pregnancy as an episode of care.3 When more than 1 admission per episode existed, variables of the first admission of the episode were used.
The main exposure of interest was maternal age at the index delivery. Maternal age was categorized into 5-year age bins and used as a categorical variable in the analyses.
The primary outcome of this study was SMM during an episode of care (pre–life-threatening condition) defined by the Canadian Perinatal Surveillance System.12 Severe maternal morbidity originally consisted of 34 conditions (eg, eclampsia and postpartum hemorrhage requiring transfusion), diagnosed with the ICD-10CA and the Canadian Classification of Health Interventions. In this study, HIV and blood transfusion were removed from the list because these do not reflect life-threatening conditions.3 Also, chronic congestive heart failure and preexisting hypertension and renal disease were removed from the list because these conditions were not acute SMM.
Maternal death (during pregnancy and within 6 weeks after termination of pregnancy) was the secondary outcome. This was determined based on either the ICD-10CA codes or death as vital status at hospital discharge.
Descriptive statistics included counts and proportions, means and standard deviations, and medians and interquartile ranges, as appropriate. First, we calculated relative changes over the entire study period by comparing the rates in 2014-2015 relative to 2004-2005 based on the following equation: [(rates in 2014-2015 − rates in 2004-2005) / (rates in 2004-2005)] × 100. Second, we used standardized differences to compare the characteristics of patients who did or did not develop SMM.13 Third, the number needed to be exposed for 1 additional person to be harmed was calculated for each SMM indicator (ie, the number of women and adolescents with this condition needed to result in 1 additional death).14 Fourth, we examined the association between maternal age and SMM and the association between maternal age and maternal death independently using a multilevel logistic regression model with random intercepts to account for clustering of patients within hospitals, which generated odds ratios (ORs) and 95% CIs.15
Patient-level covariates included age, parity, socioeconomic status by household postal code income quintile, urban or rural residence location, and whether the patient was transferred or not between hospitals. Hospital-level covariates included province or territory, urban or rural hospital location, and hospital pregnancy volume.16 We used the Maternal Comorbidity Index to perform additional risk adjustment among mothers with different baseline characteristics in all models.17 The original Maternal Comorbidity Index consisted of 19 comorbidities (eg, drug abuse, sickle cell disease, and chronic congestive heart failure). Maternal age was removed from the Maternal Comorbidity Index because age range of the Maternal Comorbidity Index does not include younger than 35 years, but age and the Maternal Comorbidity Index were separately entered into the model. The Maternal Comorbidity Index was eventually categorized into 3 groups (0, 1, and ≥2). Because 6 conditions of the Maternal Comorbidity Index had potential overlap with SMM outcome conditions, only preexisting conditions were included in the Maternal Comorbidity Index, and those occurring after pregnancy onset were considered as outcomes for SMM to ensure a valid temporal association.3,17 We used listwise deletion for records with missing data and compared characteristics of women and adolescents with missing and nonmissing data to ensure generalizability.18
We repeated the models for primary and secondary outcomes by exploring hospital pregnancy volume as a function of the number of pregnancy-related intensive care unit (ICU) admissions and, separately, the pregnancy-related ICU admission rate at each hospital (the numerator is the number of ICU admissions, and the denominator is the number of pregnancy admissions, both over the entire study period).
Three post hoc sensitivity analyses were performed. First, we restricted the cohort of analysis to the first pregnancy episode experienced by a woman or adolescent because our primary analysis was done in the cohort that included more than 1 pregnancy episodes experienced by the same woman or adolescent. Second, because of a high missing rate in income quintile, we attempted to confirm findings using listwise deletion for records with missing data by a different approach. To address data missing at random for 3 variables—income quintile (13.9% missing), residence rurality (0.8%), and hospital rurality (0.6%)—all analyses were redone with 20 multiple imputation data sets using fully conditional specification and discriminant function methods to impute these categorical values, which affords good efficiency of the coefficients (ie, smaller standard error of the point estimates) and sufficient power to detect differences.19-22 Third, because we wanted to explore the association of hospital-to-hospital transfer status on SMM and because SMM could occur before a patient was transferred from one hospital to another, we excluded from the cohort those who had SMM and for whom there was a transfer from one hospital to another that occurred on the first hospital admission during an episode of pregnancy.
All analyses were performed using SAS statistical software (version 9.4; SAS Institute Inc) and Excel for Macintosh (version 15.3.9; Microsoft Corp). All statistical tests were 2-sided, and P < .05 was considered statistically significant.
During the study period, there were 3 162 303 new pregnancies (mean [SD] maternal age, 29.5 [5.6] years) and 3 533 259 related hospital admissions among 2 035 453 mothers, resulting in 3 026 323 live births, 29 067 stillbirths, and 106 913 abortions or ectopic pregnancies. There were 54 219 episodes of SMM (17.7 cases per 1000 deliveries among 3 055 390 live births and stillbirths) in the entire study period (Table 1 and Table 2). In this study, 2.4% of all SMM events happened during the antepartum period, including ectopic pregnancies or abortions. Women and adolescents who developed SMM had a higher preexisting Maternal Comorbidity Index (Table 1). The most common reasons underlying SMM included postpartum hemorrhage requiring blood products (5.5 cases per 1000 deliveries), sepsis (3.8 cases per 1000 deliveries), and cardiac failure (1.5 cases per 1000 deliveries) (eTable 2 in the Supplement). The most common causes of SMM and the accompanying number needed to have that diagnosis for 1 additional person to die are listed in eTable 3 in the Supplement. Respiratory and cardiac illnesses among pregnant and postpartum women and adolescents were associated with the greatest individual risk of harm; however, other conditions, such as sepsis and hemorrhage, were highly prevalent. Postpartum hemorrhage requiring transfusion and sepsis were the first and second most common SMM across all ages, but the rate varied depending on age group (eTable 4 in the Supplement). In adolescents (age range, 10-19 years), the third most common SMM was eclampsia. In older mothers (≥35 years), the third most common SMM was cardiac failure, with the highest rate (2.3 cases per 1000 deliveries) compared with other age groups.
The lowest incidence of SMM was among women in their twenties, with greater morbidity among younger and older mothers (Figure and Table 3). Severe maternal morbidity increased over the study period (a 9.8% relative increase) from 17.2 cases per 1000 deliveries in 2004-2005 to 18.9 cases per 1000 deliveries in 2014-2015, corresponding to an increasing proportion of pregnancies in older (≥40 years) women in Canada over the study period (Figure).
Independent patient-level factors associated with SMM included increasing Maternal Comorbidity Index; maternal age 19 years or younger and 30 years or older, with the greatest risk experienced by women 45 years or older (OR, 2.69; 95% CI, 2.34-3.06 compared with maternal age 20-24 years); transfer between hospitals (OR, 2.01; 95% CI, 1.90-2.14); and lowest income quintile (OR, 1.19; 95% CI, 1.14-1.22 compared with highest income quintile) (Table 4). Hospital-level factors associated with SMM included specific provinces (OR, 0.78; 95% CI, 0.68-0.89 for British Columbia compared with Ontario) (Table 4 and eTable 5 in the Supplement).
During the study period, 192 mothers died, with a maternal mortality rate of 0.01%, or 6.2 per 100 000 deliveries (Table 2). Independent patient-level factors associated with maternal mortality included increasing Maternal Comorbidity Index, age 40 to 44 years (OR, 3.39; 95% CI, 1.68-6.82 compared with age 20-24 years), age 45 years or older (OR, 4.39; 95% CI, 1.01-19.10 compared with age 20-24 years), transfer between hospitals (OR, 11.00; 95% CI, 7.10-17.10), and lowest income quintile (OR, 4.14; 95% CI, 2.03-8.50 compared with highest income quintile) (Table 4). Hospital-level factors associated with maternal mortality included lowest hospital pregnancy volume and specific provinces (OR, 0.48; 95% CI, 0.28-0.84 for British Columbia compared with Ontario) (Table 4 and eTable 6 in the Supplement).
We performed sensitivity analyses by replacing quintile of hospital pregnancy volume with pregnancy-related ICU admission volume in one model and the rate in another model and found similar outcomes (eTable 7 in the Supplement). Restricting the cohort of analyses to the first pregnancy (57.9%), the incidence of SMM was 17.0 cases per 1000 deliveries; other findings did not change (eTable 8 in the Supplement). Using multiple imputation to explore the consequences of any missing data yielded similar findings (eTable 9 in the Supplement). Last, excluding those who had SMM before transfer at the first admission hospital during an episode (152 of 421 [36.1%] patients), findings were similar (eTable 10 in the Supplement).
In this nationwide population-based cohort study, we found that the incidence of SMM during the pregnancy and postpartum period increased between 2004-2005 and 2014-2015, and this trend coincided with an increase over time in maternal age and in the proportion of pregnancies to older mothers. Increasing maternal age was an independent characteristic of SMM and mortality throughout the study period.
Recent population-based studies3,4 have demonstrated annual increases in SMM over the past decade in Canada of approximately 1.3% (95% CI, 0.6%-2.0%). The United States has experienced a similar trend in maternal age and SMM23,24; however, direct country-to-country comparisons are challenging due to differences in available data and definitions.25,26 Other jurisdictions have also reported higher SMM for maternal age younger than 14 years and older than 35 years.27 Although the number of previous pregnancies (ie, parity) could potentially confound the association of SMM with age, our subgroup analysis of only first-time mothers showed that the association of maternal age with SMM appears consistent for primiparous and multiparous women and adolescents. Associations between hospitals with the lowest pregnancy volume and SMM and maternal death were perhaps due to health care clinician volume of experience, such as the volume-outcomes association previously described in traumatic injury and ischemic cardiac disease.28,29
Body mass index and assisted reproductive technology (ART) use are known factors associated with SMM.30,31 Beyond age 40 years, conception is unlikely to happen without ART use. Hence, ART use may partly account for the greater risk of SMM among women 40 years or older in the present study, although these data were not reported in the CIHI DAD. Moreover, mothers 34 years or older in Canada have shown increased obstetric acute renal failure.32 Thus, baseline variables associated with aging (eg, lower cardiac output, hypertension, and atherosclerosis) could make older mothers vulnerable to significant physiological changes occurring during pregnancy, thereby increasing the risk of developing SMM. Moreover, placental insufficiency (ie, reduced uterine artery blood flow) is associated with preeclampsia.33 Increasing oxidative stress, an imbalance between free radical generation and antioxidant defense, with aging may be a key factor of placental insufficiency and subsequent development of SMM.34
While substantially lower than in the United States, Canada’s maternal mortality rate is greater than that in many high-income countries in Europe.2 Postpartum hemorrhage, cerebrovascular and cardiovascular disease related to pregnancy-induced hypertension, and sepsis are the most common underlying etiologies.16 Focusing on these causes and characteristics of SMM and mortality identified herein may be one mechanism to further reduce mortality for Canadian mothers. While not the most common causes of SMM, respiratory illness and cardiac illness are associated with a high risk of harm and death and should be an important focus for quality improvement initiatives, especially in older mothers (aged ≥35 years). Hemorrhage and sepsis are both more prevalent across all ages, despite a lower individual risk of harm or death, and are thus also important diagnoses as a focus of quality-of-care improvement. Because the absolute number of maternal deaths is very small, establishment of a program of surveillance and follow-up, such as the Confidential Enquiries Into Maternal Deaths in the United Kingdom,35 may be a feasible and appropriate approach. Compared with our results herein, the United Kingdom has a similar maternal mortality rate and accompanying increasing maternal age.36
The findings of an increasing risk of SMM and mortality with increasing maternal age may be one of many important public health and education considerations in the timing of pregnancy across the reproductive life span. Outreach to disadvantaged communities and individuals, early recognition and appropriate management of acute illness, facilitation of transfer to tertiary care for high-risk pregnant mothers, and interventions to detect and prevent progression of illness in at-risk mothers are important considerations for primary and specialist clinicians.37,38
This study has some limitations. First, the CIHI DAD does not record certain variables that might influence maternal outcome (eg, body mass index, immigration status, race/ethnicity, frequency of visits to antenatal care, or validated measures of ART use).30,31 Second, although the DAD does not record deliveries that occur outside of hospitals, this represents only 2.1% of all deliveries in Canada.39 Third, the DAD does not identify deaths that have occurred outside of hospitals; however, this is rare for pregnant and postpartum women40,41 and is unlikely to have altered our findings. Fourth, 75% of therapeutic abortions normally occur in outpatient settings42 and are not captured in the DAD; however, subsequent complications of therapeutic abortions (eg, abortion-related hemorrhage and infection) requiring hospital admission are accounted for in our analyses.
Our study also has several strengths. First, this is the most contemporary nationwide population-based cohort study to date examining maternal outcomes in the context of increasing maternal age in Canada. Second, although most SMM occurred during the index delivery hospitalization and up to 6 weeks’ postpartum,43 creation of an episode of care enabled us to more sensitively capture SMM that occurred during the antepartum period. In the present study, 2.4% of all SMM events happened during the antepartum period, including ectopic pregnancies or abortions. Third, we used multilevel regression models to account for both patient-based and hospital-based factors associated with clinical outcomes, in addition to risk adjustment according to the Maternal Comorbidity Index.17 Sensitivity and subgroup analyses demonstrated our findings to be robust.44-47
In Canada, maternal age and the incidence of SMM have increased over time. This study found that, province of residence, maternal comorbidity, income quintile, and extremes of maternal age, especially those 45 years and older compared with those aged 20 to 24 years, were associated with SMM and mortality. These findings have direct relevance for prospective parents, clinicians, and public health professionals.
Accepted for Publication: June 28, 2019.
Published: August 23, 2019. doi:10.1001/jamanetworkopen.2019.9875
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Aoyama K et al. JAMA Network Open.
Corresponding Author: Kazuyoshi Aoyama, MD, PhD, Department of Anesthesia and Pain Medicine, The Hospital for Sick Children, 555 University Ave, Room 2211, Toronto, ON M5G 1X8, Canada (firstname.lastname@example.org).
Author Contributions: Drs Aoyama, Pinto, and Fowler had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Aoyama, Ray, Scales, Lapinsky, Hladunewich, Fowler.
Acquisition, analysis, or interpretation of data: Aoyama, Pinto, Ray, Hill, Scales, Seaward, Fowler.
Drafting of the manuscript: Aoyama, Fowler.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Aoyama, Fowler.
Obtained funding: Aoyama, Fowler.
Administrative, technical, or material support: Aoyama, Lapinsky, Fowler.
Supervision: Pinto, Scales, Lapinsky, Hladunewich, Seaward, Fowler.
Conflict of Interest Disclosures: Dr Aoyama reported receiving grants from the Canadian Institutes for Health Research and reported receiving fellowship funding from the Canadian Institutes for Health Research to pursue the PhD stream of Clinical Epidemiology & Health Care Research at the Institute of Health Policy, Management and Evaluation, University of Toronto. Dr Scales reported receiving grants from the Canadian Institutes for Health Research. No other disclosures were reported.
Funding/Support: Drs Aoyama, Ray, Scales, Lapinsky, Hladunewich, Seaward, and Fowler received a Canadian Institutes for Health Research Transitional Open Operating Grant (342397) for this project.
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We acknowledge data release from the Canadian Institute for Health Information.
Create a personal account or sign in to: