Keywords
COVID-19 - cesarean delivery prevention - singleton deliveries
Cesarean delivery (CD) rates have risen significantly in the United States over the
past decades.[1] While it is already established that cesarean deliveries are associated with higher
maternal morbidity, newer literature have also suggested increased neonatal morbidity
and financial consequences related to the rising rate of CD.[2]
[3]
[4] As many women will opt for elective repeat cesarean after a primary CD, campaigns
to prevent the primary CD have developed.[5]
Suggestions have included allowing for a longer second stage, implementing a CD “checklist”
to avoid nonindicated cesarean sections or incorporating more “midwifery-style care,”
such as positional changes during labor and intermittent fetal heart auscultation.[6]
[7]
[8] These interventions have met with varying levels of success, and preventing the
primary CD remains a major focus of investigation.
On March 23, 2020, in response to the novel coronavirus disease 2019 (COVID-19) pandemic,
a state-wide stay-at-home order was issued. Additionally, hospital-wide policy changes
occurred, including an increased emphasis on telemedicine and remote care in outpatient
obstetric (OB) clinics and limiting the number of support people for laboring women
on labor and delivery.
During the time of hospital protocol changes, quality improvement data for our institution
suggested a reduction in the rate of CD by several percentage points. Our primary
objective was to quantify the reduction in the rate of CD associated with the COVID-19
pandemic and identify factors potentially related to this decrease. Identification
of such factors could help inform counseling and management to reduce the CD rate,
as well as associated maternal and neonatal morbidity and health care costs, in the
long-term.
Materials and Methods
This was a retrospective cohort study comparing all nulliparous singleton deliveries
at a single large academic hospital in Boston, MA, during the months of a state-wide
stay-at-home recommendation during the COVID-19 pandemic (April through July 2020)
to those singleton deliveries at the same hospital and in the same months during the
same time period 1 year prior in a pre-COVID era (April through July 2019). The same
four months were chosen for both cohorts to account for any potential variation in
delivery patterns and patient characteristics throughout the calendar year. The study
was approved by the Partners Human Research Committee (protocol no.: 2020P001887).
All singleton deliveries occurring to nulliparous women were identified by review
of the electronic medical record system. Data regarding these deliveries, including
demographics and delivery type and indication for CD or induction of labor, were then
obtained from the electronic medical record system. “Trial of labor” was defined as
either spontaneous labor or induced labor.
CDs were grouped as occurring “during labor” or “not during labor.” CD during labor
included indications such as fetal intolerance of labor, failure to progress, worsening
maternal status during labor, and failed operative delivery. CD, not during labor,
included indications such as malpresentation/breech, prior uterine surgery, maternal
medical complication, fetal anomalies, and placenta or vasa previa. If the indication
for CD was not listed in the electronic medical record, individual chart review was
performed by the primary author (C.M.S.).
Additional chart review of operative reports, ultrasound imaging, and prenatal records
was performed for all women in the “cesarean delivery without a trial of labor” group
to determine whether these women met criteria for a trial of labor. For example, chart
review was performed to assess whether external cephalic version was attempted prior
to CD for breech presentation. Similarly, operative notes were reviewed where available
for women who underwent CD for a history of prior uterine surgery to determine if
future CD had been recommended. For any CD with an unclear contraindication to trial
of labor, the chart was then separately reviewed by a second author for clarity (S.E.L.).
Statistical Methods
All data were analyzed in SAS 9.4 (Cary, NC). Binary outcomes were compared with Chi-squared
or Fisher's exact testing, where appropriate. Continuous variables were compared with
nonparametric methods using Wilcoxon's testing.
Results
There were 1,909 nulliparous singleton deliveries during the two cohorts of interest:
890 in April to July of 2019 and 1,019 in April to July of 2020. Patient characteristics
including age, body mass index, race, ethnicity, and insurance type did not differ
significantly between the groups. Median gestational age at delivery was the same
in both groups (39.4 weeks, p = 0.12, [Table 1]).
Table 1
Patient characteristics
|
Year
|
p-Value
|
|
2019
|
2020
|
|
(n = 890)
|
(n = 1,019)
|
|
Age (IQR) in years
|
31 (29–34)
|
32 (29–34)
|
0.51
|
|
BMI (IQR) in kg/m2
|
29.4 (26.6–33.1)
|
29.4 (26.5–33.4)
|
0.96
|
|
Race
n (%)
|
|
White
|
585 (65.7)
|
650 (63.8)
|
0.62
|
|
Black
|
85 (9.6)
|
115 (11.3)
|
|
Asian
|
104 (11.7)
|
116 (11.4)
|
|
Other/unknown[a]
|
116 (13.0)
|
138 (13.5)
|
|
Ethnicity
n (%)
|
|
Hispanic
|
119 (13.4)
|
127 (12.5)
|
0.62
|
|
Not Hispanic
Unknown
|
749 (84.2)
22 (2.5)
|
855 (83.9)
37 (3.6)
|
|
Insurance type
n (%)
|
|
Public
|
113 (12.7)
|
119 (11.7)
|
0.50
|
|
Private
|
|
|
|
Abbreviations: BMI, body mass index; IQR, Interquartile range.
Note: Data presented with IQR represents median value with associated IQR.
a Other includes Native American, Pacific Islander, and patients who reported multiple
races.
During the COVID era, the CD rate was lower than in the pre-COVID era (28.8 vs. 33.5%,
p = 0.03, [Table 2]). There was no difference in the rate of CD during labor (19.1 vs. 19.8%, p = 0.72); however, the rate of CD without trial of labor decreased (9.6 vs. 13.7%,
p < 0.01). This appeared to be driven by an increase in the proportion of women attempting
a trial of labor during the COVID-era (90.4 vs. 86.3%, p < 0.01). Additionally, the rate of labor induction increased (45.8 vs. 40.7%, p = 0.02). Of note, there was no difference in the rate of elective inductions of labor
despite the increase in overall induction rate (19.5 vs. 20.7%, p = 0.66).
Table 2
Delivery methods and trial of labor
|
2019 (n = 890)
n (%)
|
2020 (n = 1,019)
n (%)
|
p-Value
|
|
Gestational age at delivery (IQR) in weeks
|
39.43 (38.3–40.1)
|
39.4 (38.4–40.3)
|
0.12
|
|
Trial of labor, all
|
768 (86.3)
|
921 (90.4)
|
<0.01[a]
|
|
Cesarean delivery, all
|
298 (33.5)
|
293 (28.8)
|
0.03[b]
|
|
Cesarean delivery, during labor
|
176 (19.8)
|
195 (19.1)
|
0.72
|
|
Cesarean delivery, without labor
|
122 (13.7)
|
98 (9.6)
|
<0.01[a]
|
|
IOL total
|
362 (40.7)
|
467 (45.8)
|
0.02[b]
|
|
Proportion of IOL elective
|
20.7%
|
19.5%
|
0.66
|
|
Operative vaginal delivery
|
99 (11.1)
|
107 (10.5)
|
0.66
|
Abbreviations: IOL, labor induction; IQR, interquartile range.
a
p < 0.05.
There was no change in the rate of operative delivery between the two cohorts (11.1
vs. 10.5%, p = 0.66).
Subgroup analysis of all individual indications for CD without trial of labor, including
breech/malpresentation, placenta previa, prior uterine surgery, fetal anomaly, or
maternal medical showed no difference between the two groups ([Table 3]).
Table 3
Indications for cesarean delivery without trial of labor
|
2019 (n = 122)
n (%)
|
2020 (n = 98)
n (%)
|
p-Value
|
|
Breech/malpresentation
|
63 (51.6)
|
46 (41.9)
|
0.61
|
|
Previa (placenta or vasa)
|
8 (6.6)
|
7 (7.1)
|
|
Prior uterine surgery[a]
|
11 (9.0)
|
13 (13.3)
|
|
Fetal anomaly
|
10 (8.2)
|
5 (5.1)
|
|
Nonreassuring fetal testing[b]
|
4 (3.3)
|
4 (4.1)
|
|
Maternal medical condition
|
17 (13.9)
|
10 (10.2)
|
|
Elective
|
9 (7.4)
|
13 (13.3)
|
a All cases reviewed in this category were prior myomectomies.
b This category included exclusively patients in whom nonreassuring fetal testing outside
of labor was identified and the patient then proceeded directly to primary cesarean
delivery.
For these same indications, there was no difference in the proportion of women who
underwent CD but who were candidates for trial of labor. Among women who had a CD
for breech/malpresentation, there was no difference in the rates of attempted external
cephalic version (ECV). In both cohorts, there were relatively low rates of primary
CD for maternal request ([Table 4]).
Table 4
Candidates for labor among patients who underwent cesarean delivery without trial
of labor
|
2019
|
2020
|
p-Value
|
|
Breech, total
|
63
|
46
|
|
|
ECV attempted (%)
|
19 (30.2)
|
15 (32.6)
|
0.79
|
|
Placenta previa, total
|
8
|
7
|
|
|
Candidate for labor[a] (%)
|
2 (2.5)
|
1 (14.5)
|
0.60
|
|
Prior uterine surgery, total
|
11
|
13
|
|
|
Candidate for labor (%)
|
0 (0)
|
0 (0%)
|
–
|
|
Maternal medical condition, total
|
17
|
10
|
|
|
Candidate for labor (%)
|
6 (35.3)
|
3 (30.0)
|
0.78
|
|
Fetal abnormality, total
|
10
|
5
|
|
|
Candidate for labor
|
2 (20)
|
0 (0)
|
0.28
|
|
Nonreassuring fetal testing, total
Candidate for labor (%)
|
4
0 (0)
|
4
0 (0)
|
–
|
|
Elective, total
Candidate for labor (%)
|
9
9 (100)
|
13
13 (100)
|
–
|
|
Any unlabored cesarean delivery, total
|
122
|
98
|
|
|
Candidate for labor (%)
|
19 (15.6)
|
17 (17.4)
|
0.72
|
Abbreviation: ECV, external cephalic version.
a Candidacy for labor determined by chart review of operative reports, ultrasound images,
and other documentation for criteria including distance of placental edge to cervical
os, extent of prior uterine surgery, and other characteristics.
Discussion
Our study found that the CD rate decreased nearly 5% points at one institution during
the COVID-19 pandemic; however, this decrease was driven entirely by a decrease in
the rate of CD without a trial of labor. There was no change in the rate of CD after
a trial of labor despite an increase in the rate of labor induction during the COVID-19
pandemic. Among CDs performed without a trial of labor, there was no change in the
distribution of indication, nor was there any change in the proportion of women who
may have otherwise been considered a candidate for labor.
We initially postulated that due to the COVID-19 pandemic, more mothers may have chosen
the concrete certainty of an induction date rather than awaiting spontaneous labor,
and that this increase in labor induction rate was directly tied to the falling CD
rate. In other words, it seemed a real world manifestation and perhaps even confirmation
of the ARRIVE trial.[9] Another theory was that due to the COVID-19 pandemic, OB providers, including nurses,
physicians, and midwives, simply entered labor and delivery rooms with less frequency
in an attempt to maintain as much social distancing as possible, and that perhaps
this decrease in “interventional-ism” in labor resulted in the decreased CD rate.
However, when specifics of the Cesarean deliveries were more carefully examined, these
initial hypotheses do not seem to hold true.
When it became clear that the driving force behind the decrease in CD rates was indeed
among CD without a trial of labor, we hypothesized that rates of nonindicated CD may
have decreased during the COVID-19 pandemic due to more stringent restrictions on
surgical procedures. However, this was ultimately not found to be the case, as despite
review of these cases by at least one obstetrician, and in more complex cases two
obstetricians, there were very few marginally indicated CD cases identified in the
prepandemic time period. Attempts to avoid the primary CD in the 2019 cohort were
admirable but did not support a theory of less-strict implementation of CD criteria
prior to the COVID-19 pandemic.
At the conclusion of this analysis, it appears that there must be, as of yet, unrecognized
factors at play, either related or unrelated to the COVID-19 pandemic, and likely
patient-driven, as no changes in labor practices were identified. One potential explanation
could be that women who required scheduled primary CDs and were otherwise healthy,
uncomplicated patients may have sought care at smaller hospitals closer to home to
avoid spending time in a large urban hospital during the COVID-19 pandemic; however,
this seems an incomplete explanation as there were more total deliveries in the 2020
pandemic cohort as compared with the year prior. Further monitoring of the CD rate
over the coming months at our institution will be critical to assess whether the decrease
persists or whether the rate returns to the prepandemic mean.
Reassuringly, despite an increase in the rate of labor inductions, there was no change
in the rate of CD after trial of labor. In a population diverse in age, ethnicity,
body mass index, and medical complexity, this is a reassuring validation of recent
data, showing that induction of labor in a low-risk population of nulliparous women
is associated with lower rates of CD.[9] Moreover, our study population's heterogeneity and varying levels of maternal medical
risk adds to its generalizability, suggesting that induction of labor in higher risk
populations is not associated with higher rates of CD.
Limitations
The major limitation of this study was the short time period used in the analysis.
This was specifically chosen in an attempt to isolate the true impact of the COVID-19
pandemic, as after July 2020, while the pandemic persisted, it briefly lessened in
severity in our area, leading to resumption of nearer normal hospital activity. However,
as demonstrated in this investigation, studying large-scale metrics, such as CD rates,
would likely be more accurate over longer time periods to decrease the impact of variation
around the mean. Additionally, our study was not powered to detect small differences
in the individual indications for CD, limiting our conclusion.
Conclusion
Despite initially promising findings of a sharp decrease in primary CD during the
COVID-19 pandemic, analysis ultimately demonstrated that this decrease was entirely
driven by a decrease in CD without trial of labor, although we could not isolate specific
changes in indication. Encouragingly there was no change in the rate of CD after trial
of labor despite a marked rise in the rate of labor induction. In a population diverse
in age, ethnicity, body mass index, and medical complexity, increased rates of labor
induction are not associated with increased rates of CD.
