Preterm Birth and Stillbirth Rates During the COVID-19 Pandemic: a Population-Based Cohort Study

RESEARCH

Preterm birth and stillbirth rates during the COVID-19 pandemic: a population-based cohort study

Prakesh S. Shah MD, Xiang Y. Ye MSc, Jie Yang PhD, Michael A. Campitelli MPH

n Cite as: CMAJ 2021 August 3;193:E1164-72. doi: 10.1503/cmaj.210081

ABSTRACT

BACKGROUND: Conflicting reports have birth at 22–28, 29–32 and 33–36 weeks’ (range 0.48%–0.70%). From January to emerged for rates of preterm births and gestation, and stillbirths at term and pre- December 2020, the stillbirth rate was stillbirths during the COVID-19 pan- term gestation. We used Laney control P′ 0.53%, with no special cause variation. demic. Most of these reports did not charts for the 18-year study period (6-mo We did not find any special cause varia- account for natural variation in these observation periods) and interrupted tion for preterm birth or stillbirth sub- rates. We aimed to evaluate variations time-series analyses for monthly rates for groups. We found no changes in slope or in preterm birth and stillbirth rates the most recent 4 years. gap between prepandemic and pan- before and during the COVID-19 pandemic periods using interrupted time- demic in Ontario, Canada. RESULTS: We evaluated 2 465 387 preg- series analyses. nancies, including 13 781 that resulted METHODS: We conducted a retrospective in stillbirth. The mean preterm birth rate INTERPRETATION: In Ontario, Canada, cohort study using linked population for our cohort was 7.96% (range 7.32%– we found no special cause variation health administrative databases of preg- 8.59%). From January to December (unusual change) in preterm birth or nant people giving birth in any hospital in 2020, we determined that the preterm stillbirth rates, overall or by subgroups, Ontario between July 2002 and Decem- birth rate in Ontario was 7.87%, with no during the first 12 months of the COVID- ber 2020. We calculated preterm birth special cause variation. The mean still- 19 pandemic compared with the previ- and stillbirth rates. We assessed preterm birth rate for the cohort was 0.56% ous 17.5 years.

reterm birth (birth before 37 weeks’ gestation) is a lead- and reduced maternity services.15 Less stringent fetal surveil- ing cause of mortality and morbidities in the neonatal lance from reduced attendance at medical appointments for period,1 childhood and adulthood.2 Stillbirth has devas- fear of infection, cancellation of face-to-face appointments and Ptating consequences for families.3 The causes of both preterm reduced staffing for maternity services are possible reasons for birth and stillbirth are multifactorial. During the pandemic, increased rates of stillbirths. Thus, it is important to evaluate reports described reductions in preterm birth rates in Den- preterm births and stillbirths simultaneously to understand the mark,4 the Netherlands,5 Ireland6 and the United States.7 At the true impact.16 same time, increases in stillbirth rates were reported from the Some previous reports compared preterm birth and stillbirth United Kingdom,8 Italy,9 Nepal10 and India,11 with or without rates during the pandemic to similar time periods in the past few changes in rates of preterm births. Meta-analyses have years. However, within a jurisdiction, these rates are known to emerged with differing conclusions.12,13 Some speculated rea- fluctuate between epochs17 and, thus, it is preferable to evaluate sons for reductions in preterm births included reductions in rates over longer periods to establish whether observed varia- physical activity during pregnancy, reduced stress related to tions are usual (common cause variation) or unusual (special work–life balance, less exposure to infection, fewer medical cause variation). Our objective was to evaluate whether the interventions, reduced travel and pollution,14 and improved COVID-19 pandemic affected preterm birth or stillbirth rates in hygiene and rest. Proposed reasons for increases in preterm Ontario by comparing rates for the early COVID-19 pandemic birth rates include higher stress due to worry about the pan- time period with rates from the previous 17.5 years to identify demic, employment or financial challenges, home schooling patterns of variation.

E1165 37 weeks ≥ gestation as preterm as preterm gestation

International Statistical Classi- We used International Statistical to describe and detect common cause to describe and detect common cause 22 23 During the entire study period, if we identi- charts 24,25 ISSUE 30 ISSUE | fied special cause variation, we calculated a new mean, UCL and and variation cause special the of point start the from LCL adjusted the chart. However, for the pandemic period (January– of outside point estimate of rule only 1 used we 2020), December the control limit to detect special cause variation. Identification of change in slope To examine recent change, we estimated the rates of preterm birth and stillbirth for each month of the 4 latest study years (January 2017 to December 2020). We chose a 1-month interval to study change in trend in relation to change during the pandemic period to ensure that we had at least 8 data points during indi- the in of events number the and that period pandemic the vidual cell exceeded the threshold for data release. The 4-year fication of Diseases and Related Health Problems, 10th Revision fication of Diseases and Related Health Problems, and Z37.7 to identify (ICD-10) codes Z37.1, Z37.3, Z37.4, Z37.6 (at stillbirths term evaluated also We stillbirths. (at < 37 weeks’0 days’ gestation) and preterm stillbirths gesta- from maternal delivery tion). The data for stillbirths were derived records and represent a pregnancy-level denominator. (usual) and special cause (unusual) variations by plotting crude by plotting crude (usual) and special cause (unusual) variations control limit (LCL). rates, the upper control limit (UCL) and lower because control chart We chose 6 months as the time period analysis requires at least 24 data cause points to identify special variation. Because we had data for 18.5 years, including 1 year of using variation assess to us allowed this period, pandemic the 37 datastan- using variation cause special identified We points. dard definitions. Statistical analysis 1A, in Appendix data analysis (details We used 2 approaches for available at www.cmaj.ca/lookup/doi/10.1503/cmaj.210081/ tab-related-content). Identification of special cause variation stillbirth for each of We estimated the rates of preterm birth and 2002 to July– the 6-month periods from July–December were large, we used December 2020. Because the sample sizes Laney control P′ Outcomes Preterm birth 21 weeks occuring between live births as those We classified 6 days’ and 36 weeks 0 days’ gestation births. Although the limit of viability has changed during the the during changed has viability of limit the Although births. to in our study, we used this definition years that we reviewed of the rate of preterm birth consistent. keep the calculation 717 However, we excluded recorded as births at infants from the subgroup analyses. We planned 21 weeks’ gestation of preterm births at 22–28, 29–32 and to look at subgroups 33–36 weeks’ gestation. Stillbirth - as fetal death before the complete expul We defined stillbirth at least after conception of products of or extraction sion 20 weeks of pregnancy. VOLUME 193 VOLUME |

18 AUGUST 3, 2021 AUGUST | CMAJ We defined a rural region as jurisdic- 20 18 We used data from the Statistics Canada Postal 21 of 60–75 on a scale of 0–100 (where 100 represents100 (where 0–100 of scale 60–75 on a of 19 We divided the data records from July 2002 and December and December We divided the data records from July 2002 2020 into 6-month observation periods. The first case of SARS- 2020 into 6-month observation periods. on Jan. 25,CoV-2 infection was recorded in Ontario 2020, and implemented as of strict lockdown measures in Canada were Mar. 18, Since that time, Canada has had a Lockdown Strin- 2020. gency index tions outside the commuting zone of centres with populations of 10 000 or more. Code Conversion file for neighbourhood income quintiles, which divides areas into quintiles based on neighbourhood income, with quintile 1 representing lowest area-level income and quintile 5 representing highest area-level income. SARS-COV-2 positivity rates during the pandemic were higher in 4 of the 34 public health units of Ontario, and we compared outcome rates in these loca- tions to the rest of the province. The 4 public health units were 1 109 909),(population York Region 2 731 571), (population Toronto Peel Region (population 1 381 744) and Ottawa (population 934 243). We merged the remaining public health units into an “other” region (population 8 412 533).

Data sources for Health Information’s Discharge The Canadian Institute and administra- Abstract Database (CIHI-DAD) contains clinical including Ontario, in admissions inpatient all for data tive in hospital and their records for pregnant people who deliver 98% of all births newborns. Hospital births accounted for about Population and setting pregnant people aged 13–59 yearsWe included resident who in Ontario and matched them to their delivered in any hospital Ontario contributes about 130 000– live or stillborn offspring. - out of a total about 350 000 births in Can 140 000 births per year is covered under the provincial health ada. Facility-based care residents. insurance plan for all Study design cohort study retrospective a population-based, We conducted of province the in database administrative an from data using to December 2020. from July 2002 Ontario, Canada, Methods the strictest lockdown). We designated January to December 2020the strictest lockdown). We designated January and COVID-19 wereas the “pandemic period” because SARS-CoV-2 Nov. 17,mentioned in the Canadian media from 2019, with the the precedingfirst recorded case in Wuhan, China. We considered 35 periods (each of 6 mo) “free of pandemic effects,” and these calculated the rates ofwere used to derive baseline variability. We each public health unitpreterm birth and stillbirth using data from on the number of births among pregnant people living in the region, which included gestational age. To understand differences peo- pregnant assessed we settings, different across outcomes in ple living in rural versus urban areas and according to their neighbourhood income quintile. in the province. The Ontario Mother–Baby data set determinis in the province. The Ontario Mother–Baby infant birth records tically links maternal hospital delivery and number, with linkage using a unique maternal–newborn chart rates exceeding 98%. RESEARCH 0.05 asstatisticallysignificant. tistical analyses.Weconsidereda2-sidedp 4.0.1 (www.r-project.org)toperformdatamanagementandsta- nonpandemic period. rates overtime(slopes)betweenthepandemicperiodand difference intheratesofchangepretermbirthandstillbirth start ofthepandemicperiod(December2019)andiftherewasa to evaluateiftherewasasuddenchangeatthecut-offfor E1166 July toDecember 2002,and2003.1denotesJanuary to June2003.Note:LCL =lowercontrol limit,UCL =uppercontrollimit. Figure 1:Controlchartshowingvariationin rateofpretermbirthsfromJuly2002toDecember2020in6-month epochs. Forexample,2002.2denotes (raw datainAppendix1C).Wefoundthattherewasnospecial the variationofpretermbirthratesoverentirestudyperiod ranging from 7.32% to8.59%overstudyperiods. Figure 1 shows The meanpretermbirthratefortheentirecohortwas7.96%, Preterm birthrate 3344 (SD584) grams. weeks,andthemeanbirthweightwas 38.7 (SD2.0) We found that the mean gestational age (GA) at birth was total births was 66 259 (range 62 574 to 69 839) per half year. (Appendix 1B).Acrossourstudyperiod,themeannumberof births wereincludedintheanalysisofspecialcausevariation The remaining 2 465 387 records ofdelivery andassociated facilities. Ofthese,8897recordsmetourexclusioncriteria. identified 2 474 284maternal-newbornrecordsfrominpatient During the study period from July 2002 to December 2020, we Results Sinai Hospital,Toronto(approvalno.20–0205-C). The studywasapprovedbytheResearchEthicsBoardatMount Ethics approval in slope.Weusedtheinterruptedtime-seriesanalysismethod period We usedSASversion 9.4(SASInstituteInc.)andRversion provided us with 48 data pointsforanalyzingthechange provided uswith48 data

Preterm birth rate, % 10 5 6 7 8 9

2002.2 2003.1 2003.2 2004.1 2004.2 2005.1 2005.2 CMAJ 2006.1 2006.2 valueoflessthan

| 2007.1 AUGUST 3, 2021 2007.2 2008.1 2008.2 2009.1 2009.2

2010.1 26

| 2010.2 Period VOLUME 193 Mean 2 slopeswasalmostsignificant. thedifferencebetween except forGAat33–36 weeks: in theratesofpretermbirthsbetweenperiodssubgroups, groups (Appendix1D)alsodidnotshowanysignificantchange Appendix 1C). Interrupted time-series analysis of the GA sub- 3; raw data in compared with previous periods (Figure tify anyspecialcausevariationsduringthepandemicperiod and werecalibratedchartsfromthosepoints.Wedidnotiden- period 2013(July–December)and2017(July–December), showedspecialcausevariations(decrease)in 29–32 weeks GA(mean6.42%). 33–36 weeks’ becauseof2specialcausevariations)andat there are3 rates 0.57%), at29–32weeks’GA(mean1.01%,0.92%and0.88%; (Figure 2). the visualperceptionofreductioninslopeafterJanuary2020 (gap) betweentheprepandemicandpandemicperiods,despite cant decreaseorincreaseintherateofchange,trends,cut-off during the period of January 2017 to December 2020, no signifi- pandemic period.Ourinterruptedtime-seriesanalysisshowed, cause variation in preterm birth rate during the 12 months of the 2011.1 demic periods(Figure 5). slope, trendsorcut-off(gap)betweentheprepandemicandpan- interrupted time-series analysis showed no significant change in cause variationinstillbirthrateforthepandemicperiod.Our We found no special the study period (raw data in Appendix 1E). showsthevariationsinratesofstillbirthsover period. Figure 4 the change. We readjusted the charts from the subsequent 2012 period,withameanrateof0.59%beforeand0.52%after tion intheformofreducedstillbirthrateJuly–December 0.70% overthestudyperiod.Weidentifiedspecialcausevaria- The meanstillbirthratewas0.56%andrangedfrom0.48%to Stillbirth rate 2011.2 We evaluated preterm birth rates at 22–28 weeks’ GA(mean We evaluatedpretermbirthratesat22–28 weeks’ 2012.1 2012.2 2013.1 |

ISSUE 30 2013.2 2014.1 LC L 2014.2 2015.1 2015.2 2016.1

2016.2 UC 2017.1 L

2017.2 ThecontrolchartforGAat 2018.1 2018.2

2019.1 Rate 2019.2 2020.1 2020.2

RESEARCH

E1167

2020.11

2020.9

7 2020.

2020.5

2020.3

1 2020.

1 2019.1

Gap = 0.02 (p = 0.94)

9 Slope 2 = –0.056 (p = 0.11) Slope 2 = –0.056 2019.

2019. 30 ISSUE (Appendix 1P). The control chart did not identify identify (Appendix 1P). The control chart did not |

2019.5 2019.3

iod 2019.1 r

Interpretation We identified no special cause variation in rates of preterm births and stillbirths during the first 12 months of the COVID-19 pandemic compared with the previous 17.5 years in Ontario. We identified no special cause variation in subgroups of very preterm, moderately preterm and late preterm births; term Stillbirth rate people pregnant for stillbirths of rate mean the that found We 0.56% (Appendix 1N).from both rural and urban regions was rates showed special The control chart for urban stillbirth July–December 2012 cause variation (a decrease) in the that point. The con- period, and we adjusted the charts from for pregnant variation chart did not identify special cause trol the pandemic period people in rural or urban regions during (Appendix 1O). rates by neighbourhood income Mean stillbirth 0.48% and 0.44%, quintiles 1–5 were 0.70%, 0.59%, 0.55%, respectively although the special cause variation in any of the quintiles, rates of stillbirth decreased as neighbourhood income increased (Appendix 1Q). The control charts for regions of high SARS-CoV-2 positivity showed special cause variation (decreases) in the July–December 2012 period within the Toronto Public Health Unit region, and we adjusted the charts from that point. There was no special cause variation identi- period pandemic the during regions 4 these in fied (Appendix 1R and S). Data on monthly rates of preterm births and stillbirths during the last 4 years of the study are pre- sented in Appendix 1T and U, respectively. 2018.11

VOLUME 193 VOLUME |

2018.9

Time pe

2018.7

2018.5 2018.3

respectively respectively 2018.1

AUGUST 3, 2021 AUGUST 2017.11 |

Slope 1 = 0.007 (p = 0.31) Slope 1 = 0.007 2017.9

CMAJ 2017.7

2017.5

Comparison of slopes: Slope 2 v. Slope 1 = –0.06 (p = 0.08) Comparison of slopes: Slope 2017.3 2017.1

9 8 7 6

% rate, h rt bi m Preter We evaluated stillbirth at term gestation and identified a spe- We evaluated stillbirth at term gestation cial cause variation in the form of reduced stillbirth rate in the stillbirth rate in the cial cause variation in the form of reduced change the before rate mean The period. 2014 July–December UCL mean, the readjusted We 0.13%. was after and 0.16% was mean stillbirth rate at and LCL from the subsequent period. The identify any special preterm gestation was 0.41%. We did not the pandemic period cause variation in preterm stillbirth during analy- time-series interrupted of results 1E and F). The (Appendix significant changes in sis (Appendix 1G) also did not show any any parameters.

Subgroup analysis Preterm birth rate for 7.64% was birth preterm of rate mean the that found We mothers in rural regions and 7.99% for mothers in urban regions cause special identify not did chart control The 1H). (Appendix variation in either rural or urban mothers (Appendix 1I). The mean rates of preterm birth by neighbourhood income quintiles 1–5 7.55%, and 7.67% 7.94%, 8.12%, 8.39%, were (Appendix 1J). The control chart for income quintile 1 showed special cause variation (an increase) in the July–December 2016 period (with a mean of 8.22% before and 8.92% after the period), and we adjusted the charts (Appendix 1K). The control chart did not identify special cause variation in any other quintiles, although the rates of preterm births decreased as neighbourhood income increased. We identified no special cause variation in any income quintile group during the pandemic period (Appendix 1K) and no special cause variation in the 4 regions of interest with high rates of SARS-CoV-2 positivity (Appendix 1L and M). Interrupted time-series analysis of the rate of preterm births, by month, from January 2017 to December 2020. For example, 2017.1 denotes Figure 2: Interrupted time-series analysis of the rate of from and period) (nonpandemic 2019 December to 2017 January from rate birth preterm in trends of slopes the are 2 Slope and 1 Slope 2017. January respectively. Gap is the change in the rate at the turning point between the 2 periodsJanuary to December 2020 (pandemic period), (nonpandemic v. December 2019, which separated the pandemic period from the prepandemic period. pandemic period). The vertical broken line represents RESEARCH E1168 for furtherevaluation. time pointslongbeforethe pandemic period and areofinterest Some of the special cause variations identified involved rates at preterm birthanditssubgroups orstillbirthanditssubgroups. statistically significantdifferences inslopesofrateschangefor addition, ourinterruptedtime-series analysisdidnotidentify living in a public health unit with high COVID-19 prevalence. In location: ruralorurbanarea,neighbourhood income quintileor variations based on the following features of maternal residence stillbirths; andpretermstillbirths.Wefoundnospecialcause limit, UCL =uppercontrollimit. lowercontrol gestationalageatbirth,LCL = example, 2002.2denotesJulytoDecember2002,and2003.1 JanuarytoJune2003.Note:GA = Figure 3:Controlchart showing variation inrateofpreterm births fromJuly2002toDecember 2020 in6-monthepochs,bygestationalagegroup.For

Preterm birth rate, % 6. 7. 0. 1. 2. 3. 4. 5. 5 5 5 5 5 5 5 5 5 6 7 8 0 1 2 3 4

2002.2 2003.1 2003.2 2004.1 2004.2 Mean Mean Mean 2005.1 2005.2 CMAJ 2006.1 2006.2

| 2007.1 AUGUST 3, 2021 2007.2 2008.1 2008.2 LC LC LC

2009.1 L L L 2009.2

Time period 2010.1

| 2010.2 VOLUME 193 rates of preterm neonates (< 28 weeks’ gestation[n rates ofpreterm neonates(< 28 weeks’ 31 180 singletonlivebirthsinDenmark reportedsignificantlylower rates in2018and2019.Anationwide prevalencestudyinvolving births duringtheCOVID-19pandemic periodwhencomparedwith Ireland ing January2001toApril2019. Incontrast,anothercentrein compared with 8.18 per 1000 births (n reduced to2.17per1000births(n pital inIreland 2011.1 2011.2 Findings fromdifferentjurisdictionshaveconflicted.Onehos- 2012.1 27 2012.2 reportednodifferencesinperinatal deathsorpreterm 2013.1 UC UC UC | ISSUE 30

2013.2 L L L 6 reported that the rate of very low birthweight was reportedthattherateofverylowbirthweightwas 2014.1 2014.2 2015.1 2015.2 2016.1 2016.2 2017.1 = 3) during January–April 2020 3)duringJanuary–April2020 =

2017.2 GA =22–28wk GA =29–32wk GA =33–36wk

varied from 9 to 18/yr) dur- 2018.1 2018.2 2019.1 2019.2 2020.1

= 1] during the 1]duringthe = 2020.2 RESEARCH

E1169

2020.2

2020.1

2020.11

2019.2

9 2020. 2019.1

Rate

2018.2

7 2020.

2018.1

5 2020.

2017.2

2017.1 3 2020.

2016.2

L 2020. 2016.1

UC 2019.11 2015.2

Gap = –0.014 (p = 0.79)

2015.1 9 Slope 2 = 0.0007 (p = 0.92) 2019.

2014.2

7 2019.

2014.1 30 ISSUE

5 2019. 2013.2

2013.1

LC 2019.

2012. od

2019.1 ri 1 2012.

2018.11 2

2011. reported reductions in preterm birth rate using various time win- dows surrounding Mar. 9, 2020 (odds ratio [OR] 0.77, 95% confi- dence interval [CI] 0.66–0.91 for the 2 mo before and after Mar. 9, 2020; OR 0.85, 95% CI 0.73–0.98 for the 3 mo before and after

2011. 193 VOLUME | 2018.9

2 2010. Time Pe

2018.7

1 2010.

2 2009. Mean 2018.5 Time period

1 2009.

2018.3

2 2008.

2018.1 1 2008.

2 2007.

2017.11 3, 2021 AUGUST |

1 2007. Slope 1 = –0.0002 (p = 0.89)

2017.9

2 2006.

2017.7 1 2006. CMAJ

2 2005. 2017.5

1 2005.

Comparison of slopes: Slope 2 v. Slope 1 = 0.0009 (p = 0.90) 2017.3 2 2004.

2017.1 1 2004.

2 2003.

1 2003.

An analysis to investigate the association

4 2 2002.

0.9 0.8 0.7 0.6 0.5 0.4 0.3

% rate, th Stillbir

0 1 9 8 7 6 5 4 3 2 1

= 57]).

0. 0. 0. 0. 0. 0. 0. 0. 0. Stillbirth rate, rate, Stillbirth % between the national implementation of COVID-19 mitigation between the national implementation of COVID-19 mitigation measures and the incidence of preterm birth in the Netherlands

lockdown between Mar. 12 and Apr. 14, 2020, compared with rates for 2015–2019 [n Interrupted time-series analysis of the rate of stillbirths, by month, from January 2017 to December 2020. For example, 2017.1 to December 2020. For example, 2017.1 Figure 5: Interrupted time-series analysis of the rate of stillbirths, by month, from January 2017 from January 2017 to December 2019 (nonpandemic denotes January 2017. Slope 1 and Slope 2 are the slopes of trends in the rate of stillbirths in the rate at the turning point of the 2 periodsperiod) and January to December 2020 (pandemic period), respectively. Gap is the change (nonpandemic v. pandemic period). The vertical broken line represents December 2019,- which separated the pandemic period from the pre pandemic period. Control chart showing variation in rate of stillbirths from July 2002 to December 2020 in 6-month epochs. For example, 2002.2 denotes July to Figure 4: Control chart showing variation in rate of stillbirths June 2003. We found special cause variation in period 2012.2, with 8 pointsDecember 2002, and 2003.1 denotes January to below the mean; thus, the LCL = lower control limit, UCL = upper control limit. mean was adjusted from that point onward. Note: RESEARCH 0.4). (7.41%v.7.34%;p the same4monthsoverprevious4 years birth duringApril–July2020comparedwithanaggregatedratefor birth ratesinCaliforniareportednodifferenceofpreterm that assessedtheimpactofCOVID-19pandemiconpreterm cant whenMar.15or23,2020,wereusedasanchors.Astudy from MarchtoMay2020. but athreefoldincreaseinstillbirths(1.07%v.3.23%;p was a reduction in late preterm births (5.93% v. 4.62%; and reportedthatcomparedwiththesameperiodin2019,there natal datafromaregionalhospitaldischargeabstractdatabase Mar. 24, 2020). Mar. 24, to 2020)compared withtheprelockdownperiod(Jan. 1 June 2, (3.15% v.2.25%;p (0.20% v.0.13%;p pregnant peoplewithsignificantincreasesinmaternalmortality reported a43.2%reductioninratesofhospitaladmissionsfor tive analysisinvolvingpregnantpeoplefrom4centresinIndia 28 studies concluded that there was no change in stillbirth rates; 28 studies and May 15, 2020. and May 15, births andstillbirthsintheirpatientpopulationbetweenMar. 15 system inNewYorkCityreportednochangeratesofpreterm periods). Ina2021researchletter,researchersatlargehealth study period(1.2%,2.7%and2.1%inthesamerespective gestationduringthe2020 tion inpretermbirthsat< 34 weeks’ and nochangeinpretermbirth rates. in 2021. A review of 12 studies reported an increase in stillbirths births. births, whileothershaveidentified reductionsinearlypreterm the samemonthsin2018and2019. during thepandemicperiod(March–June2020)comparedwith reported nochangeintheincidenceofpretermbirthsorstillbirths pregnancy cohortofallbirthsin2hospitalsPhiladelphia compared withprepandemicratesusingdatafromtheGeobirth term birthandstillbirthchangedduringtheCOVID-19pandemic term birth,spontaneouspretermmedicallyindicatedpre - E1170 after Mar. 9, 2020). 95%CI0.73–0.97forthe4mobeforeand Mar. 9,2020;andOR 0.84, preterm birth. respectively), but no significant differences forratesof 1000 births, 2020)(9.31v.6.93per 2019,toJan. 31, prepandemic period(Oct. 1, 2020)comparedwiththe toJune 14, 19 pandemicperiod(Feb. 1 gestation)wassignificantlyhigherduringtheCOVID- ≥ 24 weeks’ London, UK,foundthattherateofstillbirths(fetaldeath in institutionalbirthsduringthelockdownperiod. (p 1000 births births compared with 13 per ity increased to 40 per 1000 duringprelockdown(p 1000 births duringthelockdownperiodcomparedwith14per 1000 births in Nepalreportedthatthestillbirthrateincreasedto21per 7.2%; (6.7%v. 0.07), andthesameperiodsinpreceding10 years 27, 2020,comparedwiththesameperiodin2019(6.7%v.8.0%;p no differencesinratesofpretermbirthsduringMarch20toJune outcomes ofsingletonpregnanciesat1centreinIsraelreported Overall, somestudieshavereported concerningrisesinstill- 28 A study that compared maternal, obstetrical and neonatal p These weresummarizedrecently in2systematicreviews = 0.4). = = 0.002), associated with a decline of more than 50% 0.002),associatedwithadeclineofmorethan50% = 29 8 11 In 2020,researchers in Rome, Italy, evaluated peri- However, this study also found a significant reduc- A2021studythatevaluatedwhetherratesofpre- 5 = 0.02) during the lockdown period (Mar. 25 to to 0.02)duringthelockdownperiod(Mar. 25 = However, the odds were not statistically signifi- = 0.01)andintrauterinedeaths orstillbirths = 30 Aretrospectivereportfromonecentrein 9 A 2020 prospective observational study A2020prospectiveobservationalstudy = 0.0002),andneonatalmortal- = CMAJ 12 Anotherreviewof | AUGUST 3, 2021 10 Aretrospec- p = 0.002) 0.002) = < 0.001) 0.001) < = = = = | VOLUME 193 in regionalornationalstudies. mates ofpretermbirthsinsingle-centreorlocalstudiesbutnot however, theauthorsdidfindareductioninunadjustedesti- pandemic speculations thatfewerpretermbabieswerebornduringthe remains unknown. ous obstetricconcerns;however,inmanyinstancesthecause predisposition, environmentalfactors,race/ethnicityandprevi- mation, stress, medical- or pregnancy-induced disorders, genetic have beenreportedinpreviousstudiesincludeinfection,inflam- parators, highlightingtheneedformorestudieslikeours. been inconsistent,withindividualstudieshavingvariablecom- rates overtime;however,theprogresshasbeenslow. the possiblemechanisms. criticized, because thetimingofdeclaration ofapandemic a lockdowndatetodefinethe start ofthisperiodcouldalsobe pandemic periodcouldbechallenged; however,webelieveusing or stillbirth.OurchoiceofJan. 1,2020,asthestartdatefor whether thereweredifferences inthereasonsforpretermbirth the pandemic).Welackedcase-level medicaldetailstodiscern practice formidwiveshomebirthshasnotchangedduring births (however,morehomeislesslikely,asthescope of restrictions), delayedregistrationbyhospitalsorincreasedhome Canada forchildbirth(butwhocouldnotbecauseoftravel adian peoplelivingoverseaswhowouldusuallyhavereturned to gesting areducedbirthrate.ThiscouldbeduetobirthsCan 2 pandemicperiodsaveragedtoabout63370per6months,sug - (range 62 574to69 840),whilethenumberoflivebirthsin births per6-monthperiodintheprepandemicwas66 425 ects with a high degree of reliability. The mean number of total this database has been well-curated and used in multiple proj- linkage database,whichispronetoregistrationerrors,although Our studywasaretrospective evaluation usinganadministrative Limitations dence andaproxymeasureforsocioeconomicstatus. database, andtheabilitytoassesseffectsofmaternalresi- methodologically rigorousevaluationusingavalid,multicentre health carewithrelativelyfewbarrierstoaccesscare;the population-based cohort;thecontextofprovinciallyfunded studies hadinformationonstillbirth. with ourfindingsandotherreports;however,noneofthose Denmark, 1centreinIreland,IsraelandtheNetherlandscontrast ours. Reducedratesofpretermbirththatwerereportedin observed inmostotherhigh-incomecountries ities; however,thefindingsofnochangeinstillbirthrates hypothesized tobeduereducedaccesstertiarycarefacil increased stillbirthratesobservedinIndiaandNepalwere changes inpretermbirthorstillbirthrates.Thefindingsof would berequiredand,whenapplied,didnotshowunusual have shownthatcarefulevaluationusingappropriatetechniques ever, longitudinaldataassessmentsandresultsfromourstudy an opportunitytoexplorethepreventionofpretermbirth.How- Some contributingfactorsforpretermbirthandstillbirththat Some progresshasbeenmadetowardreducingstillbirth Strengths ofourstudyincludedthegeographicallydefined, 33 weretrue,thenthepandemiccouldhaveprovided | ISSUE 30 32 Severaltheorieshavebeensuggestedabout 13 Therefore,findingssofarhave 27,30,31 aresimilarto 3 Ifthe RESEARCH

E1171 Conversion File (PCCF), reference guide 2017. Ottawa: Statistics Statistics 2017. Ottawa: Conversion File (PCCF), reference guide ISSUE 30 ISSUE OM | pollution assessed using TROPOMI and OMI observations. Geophys Res using TROPOMI and pollution assessed 2 Yang J, D’Souza R, Kharrat A, et al. COVID-19 pandemic and population-level population-level pandemic and al. COVID-19 A, et D’Souza R, Kharrat Yang J, systematic review and meta-analysis. neonatal outcomes: a living pregnancy and Scand 2021 June 6Acta Obstet Gynecol of print]. doi: 10.1111/ [Epub ahead aogs.14206. outbreak coronavirus of Impact al. et T, Stavrakou S, Compernolle M, Bauwens on N0 International perinatal outcomes in the pandemic study [iPOP]. Available: International perinatal outcomes in the pandemic study [iPOP]. Available: https://www.ipopstudy.com/ (accessed 2020 Dec. 14). -born-premature-during-covid-19-1.5056815 (accessed 2020 Dec. 14). 2020 Fewer babies are being born premature during COVID-19. CTV News 2020 Aug. 8. Available: https://www.ctvnews.ca/health/fewer-babies-are-being Goldenberg RL, Rouse DJ. Prevention of premature birth. N Engl J Med 1998;339:313-20. . Knoxville . Knoxville Wheeler DJ. Understanding variation: the key to managing chaos (TN): SPC Press; 1993. Knoxville Wheeler DJ, Chambers DS. Understanding statistical process control. (TN): SPC Press; 1992. analysis: an alternative Thistlethwaite D, Campbell D. Regression-discontinuity to the ex post facto experiment. J Educ Psychol 1960;51:309-17. McDonnell S, McNamee E, Lindow SW, et al. The impact of the COVID-19 pandemic on maternity services: a review of maternal and neonatal outcomes before, during and after the pandemic. Eur J Obstet Gynecol Reprod Biol 2020;255:172-6. for racialpreterm birth rates Chang S-C, Carpenter AM, et al. Singleton Main EK, and ethnic groups during the coronavirus disease 2019 pandemic in California. Am J Obstet Gynecol 2021;224:239-41. Meyer R, Bart Y, Tsur A, et al. A marked decrease in preterm deliveries during the coronavirus disease 2019 pandemic. Am J Obstet 2021;224:234-7. Janssen O, Thompson M, Milburn S, et al. The impact of perinatal SARS-CoV2 infection during the peripartum period. Am J Obstet Gynecol MFM 2021;3:100267. Handley SC, Mullin AM, Elovitz MA, et al. Changes in preterm birth phenotypes and stillbirth at 2 Philadelphia hospitals during the SARS-CoV-2 pandemic, March–June 2020. JAMA 2021;325:87-9. Canada; 2017. Available: https://www150.statcan.gc.ca/n1/pub/92-154-g/92 -154-g2017001-eng.htm (accessed 2020 Dec. 14). of rural. Rural and Smalldu Plessis V, Beshiri R, Bollman RD, et al. Definitions Town Canada Analysis Bulletin. Ottawa: Agricultural Division, Statistics Canada; https://www150.statcan.gc.ca/n1/pub/21 Available: 21-006XIE. no Cat 2001. -006-x/21-006-x2001003-eng.pdf (accessed 2020 Dec. 14). Vital Statistics — Stillbirth Database. Deaths 2004. Ottawa: Statistics Canada; https://www150.statcan.gc.ca/n1/ Available: 84F0211XWE. no Cat 2004. Dec. 14). pub/84f0211x/2004000/4068009-eng.htm (accessed 2020 2002;14:531-7. Laney DB. Improved control charts for attributes. Qual Eng checkedLevels /en/cv!recreate.action?pid=1310042901&selectedNodeIds=1D7,4D2& =1D1,1D2,2D1,2D2&refPeriods=19910101,20180101&dimensionLayouts=layout2, 2020 Dec. 14). layout3,layout3,layout2,layout2&vectorDisplay=false (accessed (UK): Our World in COVID-19: Government Response Stringency Index. Oxford Data. Available: https://ourworldindata.org/grapher/covid-stringency-index? tab=chart&stackMode=absolute&time=2020-01-22.latest&country=~CAN®ion =World (accessed 2020 Dec. 14). Postal Code Chawanpaiboon S, Vogel JP, Moller AB, et al. Global, regional, and national JP, Moller AB, et al. Global, regional, and national Chawanpaiboon S, Vogel birth in 2014: a systematic review and modelling estimates of levels of preterm 2019;7:e37-46. analysis. Lancet Glob Health by place of birth (hospital or non-hospital). Live births and fetal deaths (stillbirths), Available: https://www150.statcan.gc.ca/t1/tbl1 Ottawa: Statistics Canada; 2020. Homer CSE, Leisher SH, Aggarwal N, et al. Counting stillbirths and COVID N, et al. Counting stillbirths and COVID Homer CSE, Leisher SH, Aggarwal a more urgent time. Lancet Glob Health 2021;9: 19-there has never been e10-e11. Jardine J, Relph S, Magee LA, et al. Maternity services in the UK during the LA, et al. Maternity services in the UK during the Jardine J, Relph S, Magee pandemic: a national survey of modifications to coronavirus disease 2019 standard care. BJOG 2021;128:880-9. 2020 May 8 [Epub ahead of print]. doi: e2020GL087978. Lett 2020 May 8 [Epub

14. 13. 34. 33. 25. 26. 27. 28. 29. 30. 31. 32. 21. 22. 23. 24. 19. 20. 18. 17. 16. 15. VOLUME 193 VOLUME |

- 34 AUGUST 3, 2021 AUGUST | CMAJ De Curtis M, Villani L, Polo A. Increase of stillbirth and decrease of late preterm Ed Neonatal Fetal Child Dis infants during the COVID-19 pandemic lockdown. Arch Liu L, Oza S, Hogan D, et al. Global, regional, and national causes of child mor- Liu L, Oza S, Hogan D, et al. Global, regional, and national priorities: an updated tality in 2000–13, with projections to inform post-2015 systematic analysis. Lancet 2015;385:430-40. birth. Early Hum Dev outcomes after preterm adult health An overview of C. Crump 2020;150:105187. preterm birth and still- Lawn JE, Gravett MG, Nunes TM, et al. Global report on and opportunities to birth (1 of 7): definitions, description of the burden improve data. BMC Pregnancy Childbirth 2010;10 Suppl 1(Suppl 1):S1. doi: 10.1186/1471-2393-10-S1-S1. Danish premature birth Hedermann G, Hedley PL, Bækvad-Hansen M, et al. 2021;106: rates during the COVID-19 lockdown. Arch Dis Child Fetal Neonatal Ed 3-95. of COVID-19 mitigation Been JV, Burgos Ochoa L, Bertens LCM, et al. Impact quasi-experimental measures on the incidence of preterm birth: a national study. Lancet Public Health 2020;5:e604-11. Philip RK, Purtill H, Reidy E, et al. Unprecedented reduction in births of very low birthweight (VLBW) and extremely low birthweight (ELBW) infants during the COVID-19 lockdown in Ireland: a ‘natural experiment’ allowing analysis of data from the prior two decades. BMJ Glob Health 2020;5:e003075. Preston E. During coronavirus lockdowns, some doctors wondered: Where are the preemies? The New York Times 2020 July 19. Available: https://www. nytimes.com/2020/07/19/health/coronavirus-premature-birth.html [login required to access content] (accessed 2020 Dec. 14). Khalil A, von Dadelszen P, Draycott T, et al. Change in the incidence of stillbirth and preterm delivery during the COVID-19 pandemic. JAMA 2020;324:705-6. 2021;106:456. on response pandemic COVID-19 the of Effect al. et MV, Kinney R, Gurung A, Kc a pro- in Nepal: and neonatal mortality outcomes care, stillbirth, intrapartum spective observational study. Lancet Glob Health 2020;8:e1273-81. Kumari V, Mehta K, Choudhary R. COVID-19 outbreak and decreased hospital isation of pregnant women in labour. Lancet Glob Health 2020;8:e1116-7. Chmielewska B, Barratt I, Townsend R, et al. Effects of the COVID-19 pandemic on maternal and perinatal outcomes: a systematic review and meta-analysis. Lancet Glob Health 2021;9:e759-72. The effects of pandemic-related mitigation measures and and measures mitigation pandemic-related of effects The

1. 3. 4. 5. 6. 7. 8. 9. 2.

References 10. 11. 12. Conclusion hospital-based of cohort population-based rigorous a In no special cause vari- births in Ontario, Canada, we identified of preterm birth or ation (unusual change) in the incidences stillbirth or their subgroups during the first year of the COVID- previous the from data with compared period pandemic 19 17.5 years. ple, the restrictions could have had beneficial effects; in other could have had beneficial effects; in other ple, the restrictions - may have been counterproductive. Inter areas or people, they currently underway to understand the national efforts are on pregnancy and childbirth. global impact of COVID-19 and the country’s response does not necessarily coincide with with coincide necessarily not does response country’s the and News about in its population. stress onset of pandemic-related of 2019. Our from the end was common starting the pandemic well. monthly change as analyses evaluated discontinuity effects in have differential with them could compliance peo certain in and areas some In settings. local and regional These studies could provide hints into the associations hints into the associations could provide These studies - mitiga with compliance infection, population-wide between tion measures, and outcomes. RESEARCH E1172 adian PretermBirthNetwork. Canadian Institutes of Health Research for development of the Can (APR-126340). PrakeshShahhasreceivedfundingsupportfromthe and PolicyResearchfromtheCanadianInstitutesofHealth Applied Research Chair in Reproductive and Child Health Services response toCOVID-19anditsrelatedimpacts.PrakeshShahholdsan form, aProvinceofOntarioinitiativetosupportOntario’songoing Funding: ThisstudywassupportedbytheOntarioHealthDataPlat- by-nc-nd/4.0/ adaptations aremade.See:https://creativecommons.org/licenses/ commercial (i.e.,researchoreducationaluse),andnomodifications provided thattheoriginalpublicationisproperlycited,usenon- licence, which permits use, distribution and reproduction in any medium, with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) Content licence:ThisisanOpenAccessarticledistributedinaccordance priately resolved. ensuring thatquestionsrelatedtoitsaccuracyorintegrityareappro- lished and agreed to be accountable for all aspects of the work in revised themanuscript,gavefinalapprovalofversiontobepub- design, andcollectedinterpretedthedata.Allofauthors intrepreted thedata.MichaelCampitellicontributedtostudy Jie Yangcontributed tothestudydesign, andanalyzed and lyzed andintrepretedthedata,draftedwork.XiangYe Contributors: PrakeshShahcontributedtothestudydesign,ana- Pediatrics (Shah),UniversityofToronto;ICES(Campitelli),Toronto,Ont. Research Centre(Shah,Ye,Yang),MountSinaiHospital;Departmentof Affiliations: Department of PediatricsandMaternal-Infant Care This articlehasbeenpeerreviewed. Competing interests:Nonedeclared.

CMAJ | AUGUST 3, 2021 | VOLUME 193 Correspondence to:PrakeshShah,[email protected] Accepted: June25,2021 herein arethoseoftheauthors,andnotnecessarilyCIHI. ever, the analyses, conclusions, opinions and statements expressed provided bytheCanadianInstituteforHealthInformation(CIHI);how- Parts of this material are based on data and information compiled and ICES oranyofitsfunderspartnersisintendedshouldbeinferred. and conclusionsreportedarethoseoftheauthors.Noendorsementby Research (CIHR)andtheGovernmentofOntario.Theopinions,results, (SPOR), theOntarioSPORSupportUnit,CanadianInstitutesofHealth funders andpartners:Canada’sStrategyforPatient-OrientedResearch ICES DataRepository,whichismanagedbywithsupportfromits tracted ICESData&AnalyticServicesandusedde-identifieddatafromthe Province ofOntarioisintendedorshouldbeinferred.Thisstudycon- No endorsementbytheOntarioHealthDataPlatform,itspartnersor are thoseoftheauthorsandindependentfromfundingsources. Disclaimer: Theopinions,resultsandconclusionsreportedinthispaper Toronto, foreditorialsupportinthepreparationofthismanuscript. Maternal-Infant CareResearchCentre(MiCare)atMountSinaiHospital, also thankDr.HeatherMcDonaldKinkaid,ascientificwriteratthe University, Montréal,forhissuggestionsregardingtheanalyses.They Acknowledgements: TheauthorsthankDr.MarcBeltempoofMcGill PHIPA guidelines. tiality agreementsbetweenICESandindependentinvestigatorsasper and PrivacyCommissioner.Accesstorawdataisgovernedbyconfiden- and proceduresarereviewedapprovedbytheOntarioInformation a requirementofhavingthisstatusinPHIPA,ICESpolicies,practices of Ontario’sPersonalHealthInformationProtectionAct,2004(PHIPA).As Data sharing:ICESisnamedasaprescribedentityunderSection45(1) | ISSUE 30