Evidence-based practice inquiry

Article8.pdf

Home >Nursing homework help >Evidence-based practice inquiry

Received: 19 May 2019 | Revised: 4 November 2019 | Accepted: 6 November 2019 DOI: 10.1111/aogs.13769

O R I G I N A L R E S E A R C H A R T I C L E

Pregnancy complications and risk of preterm birth according to maternal age: A population-based study of delivery hospitalizations in Alberta

Abbreviations: aRD, adjusted risk difference; aRR, adjusted risk ratio; DAD, discharge abstract database; iPTB, iatrogenic preterm birth; PAF, population-attributable fraction; PTB, preterm birth; RD, risk difference; RERI, relative excess risk due to interaction; RR, risk ratio; sPTB, spontaneous preterm birth.

1Department of Community Health Sciences, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada 2Department of Obstetrics & Gynecology, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada 3Department of Pediatrics, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada 4Department of Medicine, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada

Correspondence Natalie V. Scime, University of Calgary, Owerko Center in the Child Development Center, 2500 University Drive NW, Calgary, T2N 1N4 Alberta, Canada. Email: [email protected]

Funding information No external funding was sought for this research. NVS is supported by a Canadian Institutes of Health Research Canada Graduate Scholarship Doctoral Award. AM is supported by a Canadian Institutes of Health Research New Investigator Award.

Abstract Introduction: Pregnancy-related medical complications are associated with a 2- to 5-fold increased risk of preterm birth (PTB), but the nature of this etiologic relation in context with maternal factors remains poorly understood. Previous studies have gen- erally treated maternal age as a confounder but overlooked its potential as an effect modifier, whereby the magnitude of the effect of complications on PTB could dif- fer significantly across age groups. We investigated whether advanced maternal age (≥35 years) modified the association between pregnancy complications and PTB, and compared population-attributable fractions of PTB from complications in women older vs younger than 35 years. Material and methods: We analyzed population-based, cross-sectional data from the Alberta Discharge Abstract Database for women aged 18-50 years with singleton live births in hospital between 2014 and 2017 (n = 152 246). Complications were preec- lampsia, gestational diabetes, and placental disorders identified using diagnostic codes. Outcomes were spontaneous (sPTB) or iatrogenic (iPTB) PTB before 37 weeks of gestation. We estimated risk ratios and risk differences using modified Poisson and log binomial regression, respectively, adjusting for confounders (pregnancy history, comorbidities). Population-attributable fractions estimates were calculated from risk ratios. Age modification was tested using interaction terms and Z-tests. Results: Prevalence of advanced maternal age was 19.2%. Pregnancy complications and s/iPTB were more common among women aged ≥35 years. Age modified the risk of PTB from preeclampsia only, with risk differences of 9.9% (95% CI 7.2%-12.6%) in older women vs 6.1% (95% CI 4.8%-7.4%) in younger women (P-interaction = 0.012) for sPTB, and 29.5% (95% CI 26.0%-33.1%) vs 20.8% (95% CI 18.9%-22.6%, P-interaction <0.001) for iPTB. Population-attributable fractions of s/iPTB types for all complica- tions were consistently 2%-5% larger in women aged ≥35 years, and significantly larger for preeclampsia (sPTB: 5.1% vs 2.7%, P = 0.002; iPTB: 18.8% vs 14.0%, P < 0.001) and placental disorders (sPTB: 12.5% vs 8.7%, P < 0.001; iPTB: 13.2% vs 8.9%, P < 0.001).

www.wileyonlinelibrary.com/journal/aogs

mailto:

https://orcid.org/0000-0002-5811-7661

mailto:[email protected]

460 | SCIME Et al.

1 | I N T R O D U C T I O N

In the context of childbearing, advanced maternal age is typically defined as 35 years or older.1 The number of women becoming pregnant after 35 years for personal (eg, relationship stability), educational, or financial reasons continues to rise in high-income countries.1 However, delaying childbearing is not without impor- tant health considerations. Older women are at greater risk of obstetric interventions and adverse pregnancy outcomes.2 High- quality observational studies indicate that advanced maternal age is an independent risk factor for preterm birth (PTB),2-4 after controlling for confounders like parity, previous PTB, and socio- economic status. When born preterm, infants often have substan- tially higher odds of neonatal morbidity and mortality, suboptimal development and health problems into childhood, and behaviural and school difficulties into adolescence.5 On a societal level, the economic burden of PTB in Canada from birth to 10 years of age is Can$587.1 million.6

Pregnancy-related complications also occur more frequently among women aged ≥35 years. Preeclampsia occurs in roughly 3% of older women and 2% of younger women, representing a 1.5-fold relative increase in this condition with advanced age.7 Gestational diabetes is over twice as prevalent in women aged ≥35 years com- pared with women aged <35 years, with reports of 11% and 5%, re- spectively.8 Despite differences in presentation, placental disorders like placenta previa, placenta accreta, and placental abruption occur more often among older women with adjusted odds ratios ranging from 2.7 to 3.4 and a baseline risk in younger women of between 2-7 cases per 1000 births.9-11 Importantly, preeclampsia,12 gestational diabetes,13 and placental disorders11 increase a woman’s risk of PTB, regardless of her age.

The separate effects of maternal age and pregnancy complica- tions on PTB have been extensively studied. However, effect modi- fication from maternal age in the association between complications and PTB has been poorly described, as age is often treated as a con- founder and controlled for during analysis. Furthermore, it is unclear whether the proportion of PTBs attributable to complications dif- fers in older vs younger women. Thorough knowledge on the impact of advanced maternal age may enhance the precision of PTB risk

management at the individual (clinical) and population (policy/public health) levels.

This study aimed to assess whether advanced maternal age mod- ifies the association between pregnancy complications and PTB, and compare population-attributable fractions (PAFs) of PTB from complications in women younger and older than 35 years. We hy- pothesized that advanced maternal age potentiates the risk of PTB associated with pregnancy complications, and that PAFs of PTBs from complications may be higher among older women.

2 | M AT E R I A L A N D M E T H O D S

We performed a cross-sectional analysis of in-hospital deliver- ies in Alberta from April 1, 2014 to March 31, 2017, and followed the RECORD reporting guidelines.14 Data were obtained from ma- ternal delivery hospitalization records in the provincial Discharge Abstract Database (DAD). Data in the DAD is abstracted from medical charts by trained personnel using standardized protocols, and includes patient demographics, up to 25 diagnostic codes from the International Classification of Diseases 10th Revision (ICD-10), and up to 20 procedure codes from the Canadian Classification for Health Interventions. The DAD has excellent coverage of all provin- cial births except for home births (<1% of births in Alberta).

We included women aged 18-50 years with singleton live births (Z37.0, Z38) with >22 weeks completed gestation who delivered in

Conclusions: Of the pregnancy complications studied, advanced maternal age only modified the association between PTB and preeclampsia, such that older women with preeclampsia have a higher risk for s/iPTB than younger counterparts. Pregnancy complications contribute to a sizable proportion of PTBs in Alberta, especially among women aged ≥35 years. Findings may inform clinical risk assessment and population- level policy targeting PTB.

K E Y W O R D S

Alberta, discharge abstract database, effect modification, maternal age, pregnancy complications, premature birth

Key Message

Advanced maternal age (≥35 years) modifies the risk of spontaneous and iatrogenic preterm birth associated with preeclampsia, but not gestational diabetes or placental disorders. The population fractions of preterm birth at- tributable to these pregnancy complications are consist- ently higher among older women compared with younger women.

| 461SCIME Et al.

an Albertan hospital. Data on stillbirths were not available. Multiple births were excluded given that they are strongly associated with PTB and assisted reproductive technology,15 the latter of which is poorly captured in the DAD. Pregnancy complications included preeclamp- sia, gestational diabetes, and placental disorders, and were measured by the presence of corresponding ICD-10 codes in any diagnosis field. Cases of preeclampsia, including super-imposed preeclampsia, eclampsia, and HELLP (hemolysis, elevated liver enzymes, low plate- lets) syndrome, were identified using an algorithm (O11, O14-15) pre- viously published for describing international trends in hypertensive disorders of pregnancy.16 Gestational diabetes is diagnosed between 24 and 28 weeks of gestation with universal screening in Alberta, and cases were identified using an algorithm (O24.4, O24.8) validated in the Alberta DAD with a sensitivity of 86% and specificity of 99%.17 Placental disorders were operationalized as the presence of placenta previa (O44), placenta accreta (O43.2), or placental abruption (O45).

Gestational age at birth is recorded in the DAD as weeks of com- pleted gestation, which was dichotomized at <37 weeks to identify cases of PTB. We distinguished the type of PTB by using diagnostic codes to determine labor onset and procedure codes to ascertain induction use or cesarean delivery;18,19 women who were induced before delivery or had a scheduled cesarean delivery (without ex- periencing labor) at <37 weeks were classified as iatrogenic (iPTB), and the remaining women who delivered at <37 weeks were coded as spontaneous (sPTB).

Maternal age at delivery was categorized into younger age (<35 years) and advanced age (≥35 years) according to the clinical definition.1 To minimize misclassification, concordance for each woman’s age group was assessed from admission to discharge. Forty-one women turned 35 during the delivery hospitalization, as evidenced by classification of younger age upon admission and older age upon discharge; given that we did not have wom- en’s dates of birth to discern age at delivery, these records were excluded.

All analyses were adjusted for parity and past PTBs (primipa- rous, multiparous without past PTB, multiparous with past PTB), and maternal comorbidity. Comorbidities were measured using a modified version of the Obstetric Comorbidity Index, which is a weighted algorithm for 20 conditions such as physical illness, pregnancy complications, and substance abuse. This index was developed in 2013 using US Medicaid data, and subsequently val- idated in 2015 using the Alberta DAD for ability to predict mater- nal end-organ damage, prolonged length of stay, and death.20,21 To eliminate overlap with our study variables, we adapted the index to exclude preeclampsia, placenta previa, and advanced maternal age as outlined in the Supplementary material (Table S1). Multiple gestation was also excluded given our focus on singletons only. Summary index scores were calculated by summing the weights (corresponding to risk estimates) for each present condition,20 then categorizing scores into 0 or ≥1. We used this composite co- morbidity variable instead of individual comorbidity variables to avoid data sparsity given the highly imbalanced prevalence for many of these conditions.

2.1 | Statistical analyses

Records with extreme (ie, ≤22 or ≥43 weeks) or missing gestational age values were excluded. All analyses were stratified by maternal age group. Obstetric characteristics for the study population were summarized using descriptive statistics and compared using chi- squared tests or t tests. First, Poisson regression with robust error variance was used to estimate risk ratios (RR), and log binomial re- gression was used to estimate risk differences (RD) for pregnancy complications and PTB; crude and multivariable adjusted estimates are reported. We ran three models for each complication-PTB type association separately on the multiplicative and additive scales: 2 age-stratified models, and one model with the full sample includ- ing a maternal age interaction term (to assess effect modification). Next, PAFs were estimated using adjusted RRs from each model with the PUNAF command, and were compared using z-tests (see Supplementary material, Table S2 for details and formulae). All estimates were calculated with 95% CI and statistical significance was set to P < 0.05. Data cleaning and analyses were performed in Stata version 15 (Stata Corp., College Station, TX, USA).

We repeated the analyses using maternal age categories of <40 and ≥40 years to determine whether our results were sensitive to the maternal age cut-point selected. To explore effect modifica- tion using an alternative approach to interaction terms, we then deconstructed our full sample models (that included a complica- tion*age term) for both age cut-points into joint and main effects on the multiplicative scale and calculated the relative excess risk due to interaction (RERI). The RERI is a measure of modification on the additive scale, whereby a value of 0 indicates no modifi- cation and a value of >1 indicates positive effect modification.22 We also conducted a sensitivity analysis to examine the potential role of unmeasured confounding (eg, smoking, body mass index, socio-economic status) by calculating e-values for the associations studied. An e-value is the minimum strength of association, on the RR scale, that an unmeasured confounder would need to have with both the pregnancy complication and PTB, conditional on the mea- sured covariates, to fully explain away the observed association.23

2.2 | Ethical approval

This study was approved by the Conjoint Health Research Ethics Board at the University of Calgary (REB13-0760) on 18 November 2018.

3 | R E S U LT S

Overall, 153 580 singleton live births occurred in an Alberta hospital between 1 April 2014 and 31 March 2017. After excluding records of women aged <18 years, with indiscernible age at delivery, or records containing implausible or missing values, 152 246 women were in- cluded in this analysis (see Figure 1). With respect to maternal age,

462 | SCIME Et al.

123 062 women (80.8%) were <35 years and 29 184 women (19.2%) were ≥35 years. Obstetric characteristics for the study population are displayed in Table 1. Individual pregnancy complications were significantly more prevalent among women aged ≥35 years. Older women were substantially more likely to have at least 1 pregnancy complication compared with younger women (18.1% vs 9.2%, re- spectively) and to have sPTB (4.6% vs 3.8%) or iPTB (3.6% vs 2.5%). Women aged ≥35 years were significantly more likely to be mul- tiparous, have past PTB, experience an obstetric comorbidity, and deliver via cesarean section. The most common individual obstetric comorbidities (and only comorbidities with a prevalence exceeding 1%) were previous cesarean delivery and gestational hypertension (see Supplementary material, Table S1). Age-related differences in prolonged maternal length of stay were observed for cesarean birth (>4 days), but not vaginal birth (>2 days).

Table 2 displays the prevalence of PTB among exposure and ma- ternal age groups, and Table 3 displays the measures of association for pregnancy complications and PTB across maternal age groups. Adjustment for confounders generally attenuated the RRs and RDs for both types of PTB from pregnancy complications. Significant inter- action terms on the additive scale indicated age modification for risk of both sPTB and iPTB from preeclampsia, and for risk of iPTB from pla- cental disorders. The adjusted RDs (aRD) were higher with advanced maternal age for preeclampsia (sPTB: 9.9% for women ≥35 years vs 6.1% for women <35 years, P = 0.012; iPTB: 29.5% vs 20.8%, P < 0.001) and placental disorders (iPTB: 14.1% vs 11.0%; P = 0.021). Maternal age did not modify the association between gestational diabetes and sPTB or iPTB. Overall, placental disorders conferred the highest risk of sPTB (adjusted RR [aRR] 5.1, 95% CI 4.8-5.5) whereas preeclampsia conferred the highest risk of iPTB (aRR 8.9, 95% CI 8.3-9.6).

Table 4 displays the population fractions of PTB attributable to pregnancy complications across maternal age groups. The pro- portions of sPTB and iPTB exposed to each complication (which was used in estimating PAFs) were consistently higher among women aged ≥35 years. Pregnancy complications invariably contributed to larger PAFs of sPTB and iPTB in older women. For sPTB, age-related differences in PAFs were significant for preeclampsia (<35 PAF 2.7% vs ≥35 PAF 5.1%, P = 0.002) and placental disorders (<35 PAF 8.7% vs ≥35 PAF 12.5%, P < 0.001), but not gestational diabetes (<35 PAF 1.4% vs ≥35 PAF 3.8%, P = 0.057). Trends were similar for iPTB, whereby age-related differences in PAFs were significant for preeclampsia (<35 PAF 14.0% vs ≥35 PAF 18.8%, P < 0.001) and placental disorders (<35 PAF 8.9% vs ≥35 PAF 13.2%, P < 0.001), but not gestational diabetes (<35 PAF 4.3% vs ≥35 PAF 6.6%, P = 0.121). Overall, placental disorders were attributable to the largest proportion of sPTB (PAF 9.6%, 95% CI 8.8%-10.3%) and preeclampsia was attributable to the largest proportion of iPTB (PAF 15.2%, 95% CI 14.2%-16.2%).

Sensitivity analyses using a maternal age cut-point of 40 years are included in the Supplementary material (Table S3). Findings were similar with the exception of the direction of age modifica- tion in placental disorders; the aRR for iPTB was lower with ad- vanced maternal age of 40 years or older (3.4 for women ≥40 vs 5.6 for women <40; P = 0.004). Alternative effect modification analyses using joint and main effects and RERI are presented in the Supplementary material (Table S4). The main effects for all 3 pregnancy complications were consistently higher than that of ad- vanced maternal age, and the joint effects of both exposures were often larger than either main effect. RERIs for both sPTB and iPTB

F I G U R E 1 Flow diagram of study population. aWomen were initially classified as younger age (<35 y) at admission and re-classified as advanced age (≥35 y) upon discharge. Given that we did not have access to women’s dates of birth, these observations were excluded to minimize misclassification of maternal age

Total delivery hospitalization records for singleton live births in an Alberta

hospital between 1 April 2014 to 31 March 2017

n = 153 580

Total delivery hospitalization records analyzed

n = 152 246

Maternal age n = 41 excluded due to inability to

discern age at deliverya n = 1131 excluded <18 years

Extreme or missing values n = 162 excluded

| 463SCIME Et al.

were small to moderate in size (range −0.6 to 2.2) and generally indicated positive effect modification, but were non-significant at both age cut-points with the exception of preeclampsia, maternal age ≥35 years, and iPTB (RERI 1.8, 95% CI 0.3-3.2). Sensitivity anal- yses examining the role of unmeasured confounding using e-values are included in the Supplementary material (Table S5). The find- ings were consistent across maternal age groups: very strong con- founding (as indicated by large e-values) would be required to fully explain the associations with PTB types for preeclampsia and pla- cental disorders, whereas weak confounding (as indicated by small e-values) could explain away the associations between PTB types and gestational diabetes.

4 | D I S C U S S I O N

Using 3 years of population-based data from Alberta, we investi- gated whether advanced maternal age modified the association between pregnancy complications and PTB, and compared PAFs of PTB from complications in women younger or older than 35 years. As expected, pregnancy complications were more prevalent among women aged ≥35 years. There was no evidence of age modifica- tion for risk of sPTB or iPTB from gestational diabetes, but some evidence of age modification for risk of sPTB and iPTB from preec- lampsia, with larger aRDs among older women. Advanced maternal age also modified the risk of iPTB from placental disorders; however,

T A B L E 1 Obstetric characteristics for women aged 18-50 years with singleton live births in an Alberta hospital between April 2014 and March 2017

Characteristic

Maternal age

P-value

Overall n = 152 246 % (95% CI)

<35 years n = 123 062 % (95% CI)

≥35 years n = 29 184 % (95% CI)

Pregnancy complications

Preeclampsia 1.8 (1.8-1.9) 1.7 (1.6-1.8) 2.3 (2.2-2.5) <0.001

Gestational diabetes 7.3 (7.2-7.5) 5.8 (5.7-6.0) 13.5 (13.2-13.9) <0.001

Placental disorders 2.3 (2.2-2.4) 2.0 (2.0-2.1) 3.4 (3.2-3.6) <0.001

Placenta previa 0.7 (0.7-0.7) 0.6 (0.5-0.6) 1.3 (1.2-1.5)

Placenta accreta 0.2 (0.2-0.2) 0.2 (0.2-0.2) 0.4 (0.3-0.4)

Placental abruption 1.4 (1.4-1.5) 1.3 (1.3-1.4) 1.8 (1.6-1.9)

Birth outcomes

Spontaneous PTB 4.0 (3.9-4.1) 3.8 (3.7-3.9) 4.6 (4.4-4.9) <0.001

Iatrogenic PTB 2.8 (2.7-2.8) 2.5 (2.5-2.6) 3.6 (3.4-3.9) <0.001

Gestational age (weeks), mean (SD) 38.8 (1.8) 38.9 (1.8) 38.6 (1.9) <0.001

Low birthweight 5.3 (5.1-5.4) 5.0 (4.8-5.1) 6.5 (6.2-6.8) <0.001

Birthweight (g), mean (SD) 3329 (538) 3340 (532) 3285 (562) <0.001

Parity and past PTB < 37 weeks

Primiparous 41.2 (40.9-41.4) 44.9 (44.6-45.2) 25.5 (25.0-26.0) <0.001

Multiparous and no past PTB 53.2 (52.9-53.4) 50.0 (49.8-50.3) 66.5 (65.9-67.0)

Multiparous and past PTB 5.6 (5.5-5.8) 5.1 (5.0-5.2) 8.0 (7.7-8.3)

Maternal comorbidity scorea

0 80.8 (80.6-81.0) 83.1 (82.8-83.3) 71.3 (70.8-71.8) <0.001

≥1 19.2 (19.0-19.4) 16.9 (16.7-17.2) 28.7 (28.2-29.2)

Mode of delivery and maternal LOS

Spontaneous vaginal 59.5 (59.2-59.7) 61.4 (61.1-61.7) 51.4 (50.8-52.0) <0.001

Prolonged LOS (>2 days) 9.0 (8.8-9.2) 8.9 (8.7-9.1) 9.3 (8.8-9.8) 0.150

Operative vaginalb 11.4 (11.2-11.6) 11.8 (11.6-12.0) 9.7 (9.4-10.1)

Prolonged LOS (>2 days) 19.0 (18.3-19.5) 18.9 (18.3-19.6) 19.1 (17.7-20.6) 0.803

Cesarean section 29.1 (28.9-29.4) 26.8 (26.6-27.1) 38.9 (38.3-39.5)

Prolonged LOS (>4 days) 6.5 (6.2-6.7) 6.2 (5.9-6.5) 7.2 (6.8-7.7) <0.001

Abbreviations: LOS, length of stay at delivery hospitalization in days from date of admission to date of discharge; PTB, preterm birth. Prolonged LOS was derived from population norms for average LOS in Canada (Metcalfe et al, 2016, http://doi:10.1136/bmjop en-2016-012007). aMaternal comorbidity scores were measured using the obstetric comorbidity index.20,21 bOperative vaginal included forceps- or vacuum-assisted delivery.

http://doi:10.1136/bmjopen-2016-012007

464 | SCIME Et al.

the nature of this difference in risk varied according to advanced maternal age cut-point. Overall, pregnancy complications were strongly and significantly associated with risk of PTB, with placental disorders and preeclampsia conferring the highest risk. PAFs of both sPTB and iPTB were consistently higher among older women, and these age differences were statistically significant for preeclamp- sia and placental disorders but not gestational diabetes. Placental disorders contributed to approximately 1 in 8 sPTB among women aged ≥35 years vs 1 in 12 sPTB among women aged <35 years, and preeclampsia contributed to nearly 1 in 5 iPTB among women aged ≥35 years vs 1 in 7 iPTB among women aged <35 years.

Our research provides the novel contribution of quantifying age modification of the association between pregnancy complica- tions and PTB, which has rarely been examined in previous stud- ies. Findings indicate that, among women with gestational diabetes, advanced maternal age itself does not appear to elevate the risk of sPTB or iPTB beyond what is expected in younger women. This in- formation may improve risk estimation and communication about PTB, if included with other known risk factors (eg, parity, lifestyle). Qualitative studies with older mothers indicate that such informa- tion and reassurance from clinicians help to offset concerns about perinatal health, and more appropriately frame personal perceptions of risk.24

Results suggest that the risks of sPTB and iPTB from preeclamp- sia could be partly dependent on maternal age. In our study, the aRDs for both PTB types given a preeclampsia diagnosis were higher among older women, with an excess risk of approximately 4% for sPTB (9.9% vs 6.1%) and 9% for iPTB (29.5% vs. 20.8%). Our alterna- tive analyses supported the presence of modest positive age modifi- cation for preeclampsia and sPTB (RERI 0.6, 95% CI −0.1 to 1.2) and iPTB (RERI 1.8, 95% CI 0.3-3.2); however, only the latter value was statistically significant. Our findings are somewhat consistent with one previous Finnish study indicating that advanced maternal age is a risk factor for PTB among primiparous women with preeclampsia,25 in which the analysis was restricted to affected women (age as the

exposure). This differed from our approach of analyzing the entire population, using PTB types, and stratifying by age. Taken together, maternal age appears to potentiate the risk of PTB associated with preeclampsia, with stronger effect modification occurring for iPTB than for sPTB. For sPTB, the magnitude of age modification may be small but aligns with biological sciences research on aging-related impairments in reproductive functioning. For example, analyses of both human and rodent data suggest that advanced maternal age is associated with uterine dysfunction (eg, reduced decidualization) and altered utero-placental vascular function (eg, enhanced myo- genic response), even when samples are restricted to pregnancies with normal outcomes.26-28 Consequently, older mothers may expe- rience more severe physiologic responses to the incomplete spiral artery remodeling, imbalance of angiogenic factors, oxidative stress, and inflammation that characterizes preeclampsia,29 possibly wors- ening ischemia and/or triggering sPTB.30 Further research on the po- tential mechanisms for age modification of the preeclampsia-sPTB relation would be valuable. For iPTB, care-related mechanisms could involve a cascade of interventions or different treatment algorithms for women presenting with both older age and a preeclampsia diagnosis.

Other observed age-related differences in risk of PTB are less straightforward. Although statistically significant, the additive age modification detected in our main analysis between placental dis- orders and iPTB (adjusted RD 14.1% for ≥35 vs RD 11.0% for <35), with an excess risk of approximately 3% is small and may reflect normal variation in individual clinician’s perceptions of maternal and fetal risk. The age modification detected in our sensitivity analysis indicating that risk of iPTB from placental disorders is in fact higher among women aged <40 years (aRR 5.6) than women aged ≥40 years (RR 3.4) may be due to the higher underlying (unexposed) risk for PTB in older women and the underestimated risk among those ex- posed (given the absence of stillbirth data). Future studies includ- ing all births would help to clarify the role of maternal age in this scenario.

T A B L E 2 Prevalence of singleton live birth PTB at <37 weeks among women exposed and unexposed to pregnancy complications stratified by maternal age

Pregnancy complication

Spontaneous PTB Iatrogenic PTB

Overall n (%)

<35 years n (%)

≥35 years n (%)

Overall n (%)

<35 years n (%)

≥35 years n (%)

Preeclampsia

Exposed 322 (11.6) 217 (10.3) 105 (15.5) 718 (25.9) 492 (23.5) 226 (33.4)

Unexposed 5726 (3.8) 4487 (3.7) 1239 (4.3) 3474 (2.3) 2636 (2.2) 838 (2.9)

Gestational diabetes

Exposed 592 (5.3) 359 (5.0) 233 (5.9) 538 (4.8) 326 (4.5) 212 (5.4)

Unexposed 5456 (3.9) 4345 (3.7) 1111 (4.4) 3654 (2.6) 2802 (2.4) 852 (3.4)

Placental disorders

Exposed 720 (20.6) 508 (20.3) 212 (21.5) 516 (14.8) 341 (13.6) 175 (17.7)

Unexposed 5328 (3.6) 4196 (3.5) 1132 (4.0) 3676 (2.5) 2787 (2.3) 889 (3.2)

Abbreviation: PTB, preterm birth.

| 465SCIME Et al.

T A

B L

E 3

pa ct

o f p

re gn

an cy

c om

pl ic

at io

ns o

n ri

sk o

f P TB

a t

<3 7

w ee

ks a

m on

g si

ng le

to n

liv e

bi rt

hs in

A lb

er ta

a cc

or di

ng t

o m

at er

na l a

ge y

ou ng

er o

r ol

de r

th an

3 5

ye ar

P re

gn an

cy c

om pl

ic at

io n

Sp on

ta ne

ou s

P TB

Ia tr

og en

ic P

<3 5

ye ar

s ≥3

5 ye

ar s

P- va

lu e

fo r

in te

ra ct

io n

O ve

ra ll

<3 5

ye ar

s ≥3

5 ye

ar s

P- va

lu e

fo r

in te

ra ct

io n

O ve

ra ll

P re

ec la

m ps

R R

(9 5%

C I)

2. 8

(2 .5

-3 .2

) 3.

6 (3

.0 -4

.3 )

0 (2

.7 -3

.4 )

10 .8

(9 .9

-1 1.

7 )

11 .4

(1 0.

0 -1

2. 9)

.1 (1

0. 4-

12 .0

)

aR R

(9 5%

C I)

2. 5

(2 .2

-2 .8

) 2.

9 (2

.4 -3

.5 )

0. 26

1 2.

6 (2

.3 -2

.9 )

9. 0

(8 .3

-9 .9

) 8

.9 (7

.7 -1

0. 2)

0. 46

4 8

.9 (8

.3 -9

.6 )

R D

(9 5%

C I)

6. 6

(5 .3

-7 .9

) 11

.2 (8

.4 -1

3. 9)

—

21 .3

(1 9.

5- 23

.1 )

3 0.

4 (2

6. 9-

3 4.

—

aR D

(9 5%

C I)

6. 1

(4 .8

-7 .4

) 9.

9 (7

.2 -1

2. 6)

0. 01

2 —

20 .8

(1 8

.9 -2

2. 6)

29 .5

(2 6.

0 -3

3. 1)

<0 .0

01 —

G es

ta ti

on al

d ia

be te

R R

(9 5%

C I)

1. 3

(1 .2

-1 .5

) 1.

3 (1

.2 -1

.5 )

4 (1

.3 -1

.5 )

1. 9

(1 .7

-2 .1

) 1.

6 (1

.4 -1

.8 )

9 (1

.7 -2

.0 )

aR R

(9 5%

C I)

1. 2

(1 .1

-1 .4

) 1.

3 (1

.1 -1

.5 )

0. 79

8 1.

2 (1

.2 -1

.4 )

1. 7

(1 .5

-1 .9

) 1.

5 (1

.3 -1

.7 )

0. 10

8 1.

6 (1

.5 -1

.8 )

R D

(9 5%

C I)

1. 2

(0 .7

-1 .8

) 1.

5 (0

.7 -2

.3 )

5 (1

.0 -1

.9 )

2. 1

(1 .6

-2 .6

) 2.

0 (1

.3 -2

.7 )

2 (1

.8 -2

.6 )

aR D

(9 5%

C I)

0. 9

(0 .4

-1 .3

) 1.

2 (0

.5 -1

.9 )

0. 46

2 1.

0 (0

.6 -1

.4 )

1. 8

(1 .3

-2 .3

) 1.

6 (0

.9 -2

.2 )

0. 68

6 1.

7 (1

.3 -2

.1 )

P la

ce nt

al d

is or

de rs

R R

(9 5%

C I)

5. 8

(5 .4

-6 .3

) 5.

4 (4

.7 -6

.1 )

8 (5

.4 -6

.2 )

5. 9

(5 .3

-6 .5

) 5.

6 (4

.8 -6

.5 )

0 (5

.5 -6

.5 )

aR R

(9 5%

C I)

5. 2

(4 .8

-5 .7

) 4.

8 (4

.2 -5

.5 )

0. 16

9 5.

1 (4

.8 -5

.5 )

5. 4

(4 .8

-6 .0

) 5.

1 (4

.4 -6

.0 )

0. 46

5 5.

3 (4

.9 -5

.8 )

R D

(9 5%

C I)

16 .8

(1 5.

2- 18

.4 )

17 .5

(1 4.

9- 20

.0 )

.1 (1

5. 7-

18 .4

) 11

.3 (1

0. 0

-1 2.

7 )

14 .6

(1 2.

2- 17

.0 )

—

aR D

(9 5%

C I)

16 .1

(1 4.

5- 17

.6 )

16 .8

(1 4.

2- 19

.3 )

0. 69

3 16

.3 (1

4. 9-

17 .6

) 11

.0 (9

.7 -1

2. 4)

14 .1

(1 1.

7- 16

.5 )

0. 02

1 —

N ot

e: P

-v al

ue s

fo r

ri sk

e st

im at

es c

or re

sp on

d to

in te

ra ct

io n

te rm

s fo

r ag

e m

od if

ic at

io n

in a

m ul

ti va

ri ab

le m

od el

. W he

re a

ge d

if fe

re nc

es a

re s

ta ti

st ic

al ly

s ig

ni fi

ca nt

, o ve

ra ll

es ti

m at

es a

re n

ot r

ep or

te d.

M od

el s

ar e

ad ju

st ed

f or

p ar

it y

an d

pr ev

io us

p re

te rm

b ir

th a

nd p

re se

nc e

of o

bs te

tr ic

c om

or bi

di ti

es (e

g, p

re -e

xi st

in g

he ar

t di

se as

e, d

ru g

ab us

e) . I

n th

e ab

se nc

e of

a ge

m od

if ic

at io

n, m

od el

s fo

r th

e ov

er al

l sa

m pl

e ar

e ad

di ti

on al

ly a

dj us

te d

fo r

ad va

nc ed

m at

er na

l a ge

≥ 35

y ea

rs .

A bb

re vi

at io

ns : a

R D

, a dj

us te

d ri

sk d

if fe

re nc

e; a

R R

, a dj

us te

d ri

sk r

at io

; P TB

, p re

te rm

b ir

th ; R

D , c

ru de

r is

k di

ff er

en ce

; R R

, c ru

de r

is k

ra ti

466 | SCIME Et al.

Our findings support that pregnancy complications are inde- pendent risk factors for singleton sPTB and iPTB, and have a larger main effect on risk of PTB than advanced maternal age. Our ratio estimates align with those reported in large, contemporary studies ranging from approximately 2-7 for preeclampsia,12,31 4-11 for pla- cental disorders,11 and 1-2 for gestational diabetes depending on whether any PTB or PTB types were used.13,32,33 Several Doppler and histological studies have characterized how preeclampsia and placental disorders lead to sPTB, implicating a common pathophys- iology of the placenta.34 However, biologically plausible pathways from gestational diabetes to sPTB have yet to be described. For all pregnancy complications, heightened concern for maternal and fetal well-being and clinical protocols for managing high-risk pregnancy are likely contributors to iPTB.35

In addition to quantifying individual risks, our estimation of PAFs provides a population-level perspective to understanding how com- plications contribute to PTB across clinically relevant maternal age groups. In theory, PAFs represent the proportion of PTBs that would be eliminated if the studied risk factor were eliminated.36 Our pop- ulation-based findings are therefore useful for modifiable factors and can inform targeted policy action, public health intervention, or research. We observed PAFs for sPTB and iPTB that were consis- tently and often significantly larger among older mothers regardless of trends in aRRs across age groups, a finding that aligns with our hypothesis. For example, with iPTB the aRR for placental disorders was higher among women aged <35 (5.4) than women aged ≥35 (5.1), yet the PAF was higher among women aged ≥35 (13.2%) vs women aged <35 (8.9%). This is because PAFs contextualize risk estimates with the proportion of PTBs exposed to complications, which we have shown are consistently higher among older women.

Few studies have reported PAFs for pregnancy complications and PTB, as the predominant focus has been on behavioral fac- tors (eg, substance use).37 Comparisons to previous research are hindered by study variation. A Canadian study reported a PAF of 1.91% for gestational diabetes, which is marginally smaller than the PAFs of 2.0% for sPTB and 4.9% for iPTB reported in our study.33 However, their population included multiple gestations and stillbirths and likely has a different distribution of maternal and obstetric characteristics compared with our sample of single- ton live births. A US study of singleton live births reported PAFs of iPTB of 19.1% for preeclampsia, 10.5% for placental abruption, and 4.8% for placenta previa,12 which are somewhat similar to our findings (approximately 14%-19% for preeclampsia and 8%-13% for placental disorders depending on maternal age group), but the au- thors defined PTB as <35 weeks compared with our definition of <37 weeks. Although our estimates appear to align with published PAF patterns, these examples highlight that careful attention to population composition and outcome definitions is warranted when comparing PAFs.

The presence of preeclampsia and placental disorders contrib- uted to a significantly higher proportion of sPTB and iPTB in women aged ≥35 years compared with women aged <35 years. Research and policy aimed at identifying and preventing these conditions, T

A B

L E

PA Fs

o f p

re gn

an cy

c om

pl ic

at io

ns o

n P

TB a

t <3

7 w

ee ks

a m

on g

si ng

le to

n liv

e bi

rt hs

in A

lb er

ta a

cc or

di ng

t o

m at

er na

l a ge

y ou

ng er

o r

ol de

r th

an 3

5 ye

ar s

P re

gn an

cy c

om pl

ic at

io n

Sp on

ta ne

ou s

P TB

Ia tr

og en

ic P

<3 5

ye ar

s ≥3

5 ye

ar s

P- va

lu e

fo r

Z- te

st O

ve ra

ll <3

5 ye

ar s

≥3 5

ye ar

s P-

va lu

e fo

r Z-

te st

O ve

ra ll

P re

ec la

m ps

% P

TB e

xp os

ed 4.

6 7.

5. 3

15 .7

21 .2

PA F

% (9

5% C

I) 2.

7 (2

.1 -3

.3 )

5. 1

(3 .7

-6 .5

) 0.

0 02

3. 2

(2 .7

-3 .8

) 14

.0 (1

2. 8-

15 .2

) 18

.8 (1

6. 6-

21 .0

) <0

.0 01

15 .2

(1 4.

2- 16

.2 )

G es

ta ti

on al

d ia

be te

% P

TB e

xp os

ed 7.

6 17

9. 8

10 .4

19 .9

PA F

% (9

5% C

I) 1.

4 (0

.7 -2

.2 )

3. 8

(1 .5

-6 .0

) 0.

05 7

2. 0

(1 .2

-2 .7

) 4.

3 (3

.2 -5

.4 )

6. 6

(3 .9

-9 .3

) 0.

12 1

4. 9

(3 .8

-6 .0

)

P la

ce nt

al d

is or

de rs

% P

TB e

xp os

ed 10

.8 15

11 .9

10 .9

16 .4

PA F

% (9

5% C

I) 8

.7 (7

.9 -9

.6 )

12 .5

(1 0.

6- 14

.3 )

<0 .0

01 9.

6 (8

.8 -1

0. 3)

8 .9

(7 .8

-9 .9

) 13

.2 (1

1. 1-

15 .4

) <0

.0 01

10 .0

(9 .0

-1 1.

N ot

e: P

-v al

ue s

w er

e co

m pu

te d

us in

g Z-

te st

s to

c om

pa re

t he

a ge

-s tr

at if

ie d

PA Fs

. A

bb re

vi at

io ns

: P A

F, p

op ul

at io

n at

tr ib

ut ab

le f

ra ct

io n;

P TB

, p re

te rm

b ir

th .

| 467SCIME Et al.

particularly in older women, could meaningfully reduce the popu- lation burden of PTB given the moderate size of these PAFs. PTB is heterogeneous and multiple causative factors may be influencing pregnancy outcome alongside complications.38 Future research ex- ploring PAFs for combinations of medical, lifestyle, and social factors across maternal age groups may yield valuable information given the advent of precision public health in PTB prevention.

Our use of the DAD as a data source presents both limitations and strengths. Non-differential misclassification bias from under-re- porting of pregnancy complications in the DAD may have occurred. Validation studies indicate that pregnancy complications are gen- erally under-estimated when using delivery records compared with multiple healthcare contacts.39 The inability to include stillbirths in our sample likely introduced some selection bias, leading to an under-estimated risk of PTB in exposed groups. This bias would be more pronounced in the advanced maternal age group, because stillbirth occurs more frequently in women aged ≥35 years.4 Some degree of unmeasured confounding is expected given that the DAD poorly captures assisted reproductive technology and lacks qual- ity data on non-medical factors that influence perinatal health. Our sensitivity analysis of e-values provides some reassurance that our findings related to preeclampsia and placental disorders (but not gestational diabetes) are unlikely to be fully explained by unmea- sured confounding, because the minimum strength of association (RRs ranging from 4 to 17) that would nullify the observed associ- ation is implausible. Notwithstanding these limitations, use of the DAD enabled us to capture all in-hospital deliveries in the prov- ince. Few women were excluded for poor data quality. Our large sample size allowed for high precision and power to detect effect modification, reducing the likelihood of type II error and improving generalizability.

5 | C O N C L U S I O N

The risks of sPTB and iPTB associated with preeclampsia are sig- nificantly higher among older mothers, which may prove useful when estimating and communicating obstetric risk for older women. Maternal age does not appear to meaningfully modify the risk of PTB associated with gestational diabetes or placental disorders; however, additional research on age modification in the context of placental disorders is warranted. Pregnancy complications contribute to a siz- able proportion of live PTBs in Alberta, particularly iPTBs, and are attributable to a higher proportion of PTBs among older compared with younger women. Our findings may be used to inform clinical risk assessment and public health policy pertaining to PTB.

A C K N O W L E D G E M E N T S We thank Stephanie Garies and Bing Li for their assistance in obtain- ing and managing this dataset.

C O N F L I C T O F I N T E R E S T None.

O R C I D Natalie V. Scime https://orcid.org/0000-0002-5811-7661

R E F E R E N C E S 1. Johnson JA, Tough S. No-271-delayed child-bearing. J Obstet

Gynaecol Canada. 2017;39(11):e500-e515. 2. Joseph KS, Allen AC, Dodds L, Turner LA, Scott H, Liston R.

The Perinatal effects of delayed childbearing. Obstet Gynecol. 2005;105(6):1410-1418.

3. Waldenstrom U, Cnattingius S, Vixner L, Norman M. Advanced maternal age increases the risk of very preterm birth, irre- spective of parity: a population-based register study. BJOG. 2017;124:1235-1244.

4. Kenny LC, Lavender T, McNamee R, O’Neill SM, Mills T, Khashan AS. Advanced maternal age and adverse pregnancy outcome: evi- dence from a large contemporary cohort. PLoS ONE. 2013;8(2):1-9.

5. Saigal S, Doyle LW. An overview of mortality and sequelae of preterm birth from infancy to adulthood. Lancet. 2008;371(9608):261-269.

6. Johnston KM, Gooch K, Korol E, et al. The economic burden of pre- maturity in Canada. BMC Pediatr. 2014;14(1):93.

7. Khalil A, Syngelaki A, Maiz N, Zinevich Y, Nicolaides KH. Maternal age and adverse pregnancy outcome: a cohort study. Ultrasound Obstet Gynecol. 2013;42(6):634-643.

8. Lai FY, Johnson JA, Dover D, Kaul P. Outcomes of singleton and twin pregnancies complicated by pre-existing diabetes and ges- tational diabetes: a population-based study in Alberta, Canada, 2005–11. J Diabetes. 2016;8(1):45-55.

9. Cleary-Goldman J, Malone FD, Vidaver J, et al. Impact of mater- nal age on obstetric outcome. Obstet Gynecol. 2005;105(5 Pt 1):983-990.

10. Faiz A, Ananth C. Etiology and risk factors for placenta previa: an overview and meta-analysis of observational studies. J Matern Neonatal Med. 2003;13(3):175-190.

11. Vahanian S, Lavery J, Ananth C, Vintzileos A. Placental implantation abnormalities and risk of preterm delivery: a systematic review and metaanalysis. Am J Obstet Gynecol. 2015;213(4 Suppl):S78-S90.

12. Ananth CV, Vintzileos AM. Maternal-fetal conditions necessitat- ing a medical intervention resulting in preterm birth. Am J Obstet Gynecol. 2006;195(6):1557-1563.

13. Billionnet C, Mitanchez D, Weill A, et al. Gestational diabetes and adverse perinatal outcomes from 716 152 births in France in 2012. Diabetologia. 2017;60:636-644.

14. Benchimol EI, Smeeth L, Guttmann A, et al. The reporting of stud- ies conducted using observational routinely-collected health data (RECORD) statement. PLoS Medicine. 2015;12(10):1-22.

15. Gunby J, Bissonnette F, Librach C, Cowan L. Assisted reproductive technologies (ART) in Canada: 2007 results from the Canadian ART Register. Fertil Steril. 2011;95(2):542-547.

16. Roberts CL, Ford JB, Algert CS, et al. Population-based trends in pregnancy hypertension and pre-eclampsia: an international com- parative study. BMJ Open. 2011;2(1):e00101.

17. Bowker SL, Savu A, Lam NK, Johnson JA, Kaul P. Validation of ad- ministrative data case definitions for gestational diabetes mellitus. Diabet Med. 2017;34(1):51-55.

18. Gregory KD, Korst LM, Gornbein JA, Platt LD. Using administrative data to identify indications for elective primary cesarean delivery. Health Serv Res. 2002;37(5):1387-1401.

19. Liu S, Liston RM, Joseph KS, Heaman M, Sauve R, Kramer MS. Maternal mortality and severe morbidity associated with low-risk planned cesarean delivery versus planned vaginal delivery at term. Can Med Assoc J. 2007;176(4):455-460.

20. Bateman BT, Mhyre JM, Hernandez-Diaz S, et al. Development of a comorbidity index for use in obstetric patients. Obstet Gynecol. 2013;122(5):957-965.

https://orcid.org/0000-0002-5811-7661

468 | SCIME Et al.

21. Metcalfe A, Lix LM, Johnson JA, et al. Validation of an ob- stetric comorbidity index in an external population. BJOG. 2015;122:1748-1755.

22. Van Der Weele TJ, Knol MJ. A tutorial on interaction. Epidemiol Method. 2014;3(1):33-72.

23. Van Der Weele TJ, Ding P. Sensitivity analysis in observa- tional research: Introducing the E-value. Ann Intern Med. 2017;167(4):268-274.

24. Bayrampour H, Heaman M, Duncan KA, Tough S. Advanced ma- ternal age and risk perception: a qualitative study. BMC Pregnancy Childbirth. 2012;12:100.

25. Lamminpää R, Vehviläinen-Julkunen K, Gissler M, Heinonen S. Preeclampsia complicated by advanced maternal age: a regis- try-based study on primiparous women in Finland 1997–2008. BMC Pregnancy Childbirth. 2012;12:47.

26. Woods L, Perez-Garcia V, Kieckbusch J, et al. Decidualization and placentation are a major cause of age-related reproductive decline. Nat Commun. 2017;8:352.

27. Lean SC, Heazell AEP, Dilworth MR, Mills TA, Jones RL. Placental dysfunction underlies increased risk of fetal growth restriction and stillbirth in advanced maternal age women. Sci Rep. 2017;7:9677.

28. Care AS, Bourque SL, Morton JS, Hjartarson EP, Davidge ST. Pregnancy and aging effect of advanced maternal age on preg- nancy outcomes and vascular function in the rat. Hypertension. 2015;65:1324-1330.

29. Warrington JP, George EM, Palei AC, Spradley FT, Granger JP. Recent advances in hypertension recent advances in the under- standing of the pathophysiology of preeclampsia. Hypertension. 2013;62:666-673.

30. Romero R, Espinoza J, Kusanovic JP, et al. The preterm parturition syndrome. BJOG. 2006;113(Suppl. 3):17-42.

31. Davies E, Bell J, Bhattacharya S. Preeclampsia and preterm deliv- ery: a population-based case-control study. Hypertens Pregnancy. 2016;35(4):510-519.

32. Hedderson MM, Ferrara A, Sacks DA. Gestational diabetes mel- litus and lesser degrees of pregnancy hyperglycemia: association with increased risk of spontaneous preterm birth. Obstet Gynecol. 2003;102(4):850-856.

33. Metcalfe A, Sabr Y, Hutcheon JA, et al. Trends in obstetric interven- tion and pregnancy outcomes of Canadian women with diabetes in pregnancy from 2004 to 2015. J Endocr Soc. 2017;1(12):1540-1549.

34. Ananth C, Vintzileos A. Medically indicated preterm birth: recogniz- ing the importance of the problem. Clin Perinatol. 2008;35:53-67.

35. MacDorman MF, Declercq E, Zhang J. Obstetrical intervention and the singleton preterm birth rate in the United States from 1991– 2006. Am J Public Health. 2010;100(11):2241-2247.

36. Rockhill B, Newman B, Weinberg C. Use and misuse of population attributable fractions. Am J Public Health. 1998;88(1):15-19.

37. Dzakpasu S, Fahey J, Kirby RS, et al. Contribution of prepregnancy body mass index and gestational weight gain to adverse neona- tal outcomes: population attributable fractions for Canada. BMC Pregnancy Childbirth. 2015;15(1):21.

38. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet. 2008;371(9606):75-84.

39. Lain SJ, Hadfield RM, Raynes-Greenow CH, et al. Quality of data in perinatal population health databases: a systematic review. Med Care. 2012;50(4):e7-e20.

S U P P O R T I N G I N F O R M AT I O N Additional supporting information may be found online in the Supporting Information section.

How to cite this article: Scime NV, Chaput KH, Faris PD, Quan H, Tough SC, Metcalfe A. Pregnancy complications and risk of preterm birth according to maternal age: A population-based study of delivery hospitalizations in Alberta. Acta Obstet Gynecol Scand. 2020;99:459–468. https ://doi.org/10.1111/ aogs.13769

https://doi.org/10.1111/aogs.13769