Keywords
mean arterial blood pressure - mean uterine artery pulsatility index - Doppler ultrasonography
- fetal ultrasonography - pregnancy
Introduction
The Samrakshan protocol utilizes maternal clinical and demographic details, mean arterial
blood pressure (MAP), and mean uterine artery (UtA) pulsatility index (PI) measurements
to assess the risk for preterm preeclampsia (PE) and fetal growth restriction (FGR)
in pregnant women in India.[1]
[2]
[3] MAP and mean UtA PI measurements are noninvasive, painless tests that are integrated
with routine antenatal assessments. Several studies have confirmed the utility of
mean UtA PI measurements during pregnancy to assess uteroplacental circulation and
the risk for PE and FGR.[4]
[5]
[6]
[7]
[8]
[9]
[10] Previous studies have reported that earlier reduction in UtA flow resistance results
in better placentation and higher birthweights and that late normalization of UtA
blood flow was associated with higher perinatal mortality.[11]
[12]
[13]
[14] We aimed to determine the sequential changes in the MAP and mean UtA PI between
two gestational week intervals, the 11 to 14 gestation week period and the 19–24 + 6
gestation week period, in a population of Asian Indian pregnant women.
Methods
The study protocol adhered to the tenets of the Declaration of Helsinki and all patient
data were anonymized before analysis to protect patient confidentiality. The study
sample was an opportunistic, purposive sample from fetal radiology centers that are
participating in the Samrakshan program of the Indian Radiological and Imaging Association.[1] Fetal radiologists certified by the Fetal Medicine Foundation performed all fetal
Doppler assessments. The sample size for the study was estimated as a minimum of 957
first-trimester pregnant women using the likelihood ratio test for proportions. The
assumptions underlying the sample size estimate included a 1: 2 ratio for pregnant
women at high risk and low risk for preterm PE in the 11 to 14 gestation weeks assessment,
an alpha of 0.05, 80% power and an anticipated delta of 0.07. The least required sample
size was revised upwards to 1,149 after factoring in a maximum anticipated 20% loss
to follow-up between the 11 to 14 weeks and 19–24 + 6 weeks assessment. We estimated
that study recruitment must include at least 1,264 pregnant women (based on an anticipated
exclusion of 10% for maternal comorbidities and missing data) to achieve the minimum
sample size of 1,149. The study data were prospectively collected between March and
July 2022.
The first-trimester assessments that included the collection of clinical and demographic
details, measurement of MAP, and Doppler assessment of the uterine artery (UtA PI)
were performed between 11 and 14 gestation weeks. Clinical and demographic details
that were collected in the first trimester included maternal age, parity, prior obstetric
history, prior and current comorbidity, type of conception, personal risk behaviors,
and height and weight measurements to estimate the body mass index (BMI). Dating of
the pregnancy was done at the first-trimester visit to assign an expected date of
delivery and the crown-rump length (CRL) was measured. The study included singleton
live pregnancies and fetuses with a CRL between 45 and 84 mm at 11 to 14 gestation
weeks. The details were entered in the Fetal Medicine Foundation online calculator
for risk assessment of PE (available at https://fetalmedicine.org/research/assess/preeclampsia/first-trimester) to determine a customized risk for the development of preterm PE and FGR for each
woman. Pregnant women at high risk for the development of preterm PE (based on a 1
in 150 cutoff) were recommended low dose aspirin 150 mg once daily at bedtime to be
continued up to 36 weeks of gestation, development of preterm PE or childbirth, whichever
is earlier.
MAP was measured at each trimester of pregnancy after explaining the procedure to
the woman. The blood pressure was measured in both upper arms simultaneously using
calibrated digital oscillometry machines with the pregnant woman seated in a comfortable
upright position, legs uncrossed and placed flat on the floor, and forearms resting
at the level of the heart on a table in front, and the digital monitor facing away
from the woman.[2] The blood pressure measurement was repeated after at least 1 minute.[2] The blood pressure measurements were done in a silent environment avoiding any distraction
for the pregnant woman during the measurement.[2]
The mean UtA PI was measured using the transabdominal approach.[3] A mid sagittal section of the uterus and cervix was obtained, and the transducer
tilted gently sideways while using color flow mapping. The pulsed wave Doppler sampling
gate was set at approximately 2 mm and positioned on the UtA (ascending or descending
branch) at the point closest to the internal cervical os. The insonation angle was
less than 30 degrees and as close to 0 as possible with a peak systolic velocity more
than 60cm/sec. The pulsatility of the right and left UtA was measured and the PI was
measured when at least three identical waveforms were obtained. An abnormal UtA PI
was defined as mean UtA PI more than 95th percentile.
The second-trimester assessments were done between 19 and 24 + 6 weeks of pregnancy.
These assessments included the collection of clinical and demographic details including
self-reported use of low-dose aspirin, a targeted assessment for fetal anomalies (TIFFA
scan), measurement of MAP and Doppler studies of the UtA, umbilical artery, and ductus
venosus.
The data were entered into an MS Office Excel spreadsheet and subsequently exported
to STATA version 12.0 (StataCorp, College Station, Texas, United States) for further
analysis. Continuous data were expressed as mean (standard deviation) and categorical
data as frequencies and proportions. The median and interquartile range (IQR) was
estimated for continuous variables. The Shapiro–Wilk test was used to determine the
normality of the data distribution and nonparametric tests were used for analysis
as MAP and mean UtA PI were not normally distributed. The data of pregnant women who
did not provide information on the use of low dose aspirin or reported irregular use
of low dose aspirin and women with chronic hypertension or polycystic ovarian syndrome
were excluded from further analysis. The nonparametric K sample equality of medians
test was used to compare the medians of MAP and mean UtA PI and the Wilcoxon test
was used for the paired comparison of the change in MAP and mean UtA PI between the
first- and second-trimester. The 95th percentile of the mean UtA PI based on the global reference for each gestational
week was used to determine abnormal UtA PI (>95th percentile) at each gestational week of interest.[15] A multivariate regression model that adjusted for maternal age, parity, and BMI
was used to determine associations for MAP and mean UtA PI and the β coefficient and
95% confidence intervals (CI) around the point estimates are reported. A p-value less
than 0.05 was considered statistically significant.
Results
The initial study dataset included 1,274 pregnant women with ultrasound scans over
2 gestational week intervals, 11 to 14 gestation weeks, and 19-24 + 6 gestation weeks.
The 11 to 14 weeks assessment identified 484 pregnant women at high risk for preterm
PE who were recommended low-dose aspirin. We excluded the data of 42 (3.30%) women
with chronic hypertension and 12 (0.94%) women with polycystic ovarian disease from
the analysis. The data of 48 (9.92%) of the 484 high-risk women who did not provide
information on the use of low dose aspirin and 14 (2.89%) women who reported irregular
use of low dose aspirin in the second-trimester follow-up visit were not considered
for further analysis. The final cohort for analysis included 1,163 pregnant women
including 390 women identified as high-risk for preterm PE in the first-trimester
ultrasound assessment ([Table 1]) and with sequential ultrasound scans in the first and second trimester of pregnancy.
Table 1
Clinicodemographic details of the 1,163 pregnant women in the 11 to 14 gestation weeks'
assessment
Parameter
|
Distribution
|
Age in years, median, IQR
|
28, 24 to 30.7 years
|
Body mass index, median, IQR
|
23.53, 20.57 to 26.84
|
Spontaneous conception, n, %
|
1137, 97.76%
|
Nulliparous, n, %
|
685, 58.90%
|
Multiparous, n, %
|
478, 41.10%
|
Mean arterial blood pressure, median, IQR
|
81.67, 76.83 to 86
|
Mean uterine artery PI, median, IQR
|
1.7, 1.35 to 2.13
|
Abbreviations: IQR, interquartile range; PI, pulsatility index.
MAP was significantly higher (K sample Equality of Medians test Fisher's exact p-value <0.001) in high-risk (median: 86, IQR: 81.42–90.8) compared to low-risk pregnant
women (median: 79, IQR: 75.5–84) in the 11 to 14 weeks screening. Two (0.26%) of the
773 low-risk women and 10 (2.56%) of the 390 high-risk women had a MAP more than or
equal to 100. The 11 to 14 weeks MAP was significantly associated with increasing
maternal age (β coefficient: 0.16, 95% CI: 0.13, 0.20, p < 0.001) and increasing BMI (β coefficient: 0.03, 95% CI: 0.02, 0.04, p < 0.001) but was not associated with parity (β coefficient: 0.61, 95% CI: −0.27,
1.49, p = 0.18). MAP decreased (β coefficient: −0.008, 95% CI: −0.01, −0.001, p = 0.02) with increasing gestational age in the 11 to 14 weeks assessment.
Mean UtA PI was significantly higher (K sample Equality of Medians test Fisher's exact
p-value <0.001) in high-risk (median: 2.02, IQR: 1.72–2.4) compared to low-risk pregnant
women (median: 1.50, IQR :1.25–1.92) in the 11 to 14 weeks screening. Two hundred
and nineteen (18.83%) women were identified with abnormal mean UtA PI in the 11 to
14 weeks assessment. One hundred and twenty-eight (58.45%) of the 219 women with abnormal
mean UtA PI and 262 (27.75%) of the 944 women with normal mean UtA PI in the 11 to
14 weeks assessment were categorized as high-risk for preterm PE. Abnormal mean UtA
PI decreased with increasing maternal age (β coefficient: −2.20, 95% CI: −2.88, −1.52,
p < 0.001), increasing gestational weeks (β coefficient: −0.31, 95% CI: −0.43, −0.19,
p < 0.001) and increasing BMI (β coefficient: −1.16, 95% CI: −1.83, −0.50, p = 0.001) but was not associated with parity (β coefficient: 0.03, 95% CI: −0.03,
0.09, p = 0.31) in the first trimester.
[Table 2] presents the changes in MAP from the baseline first-trimester to the second-trimester
assessment. The MAP showed a significant reduction from baseline to the second trimester
in the subgroup analyses except for obese pregnant women. The mean UtA PI showed a
significant reduction from baseline to the second trimester in all subgroup analyses
(see [Table 3]). Both MAP (mean difference: 5.14, p < 0.001) and mean UtA PI (mean difference: 0.14, p < 0.001) remained significantly higher at the second-trimester assessment in pregnant
women on daily low-dose aspirin (high risk for preterm PE) compared to women not using
low-dose aspirin (low risk for preterm PE). High-risk pregnant women taking low-dose
aspirin daily showed a larger reduction in mean UtA PI compared to high-risk pregnant
women that did not report the use of low-dose aspirin (0.89 vs. 0.62, p <0.001).
Table 2
Changes in MAP from 11 to 14 to 19–24 + 6 gestation weeks by subgroups
|
11–14 weeks MAP
Mean (SD)
|
19-24 + 6 weeks MAP, Mean (SD)
|
Wilcoxon test p-Value
|
Nulliparous (n = 685)
|
81.39 (7.19)
|
79.76 (7.55)
|
<0.001
|
Multiparous (n = 478)
|
82.00 (7.99)
|
78.92 (8.00)
|
<0.001
|
Normal BMI (n = 612)
|
80.73 (7.28)
|
78.19 (7.53)
|
<0.001
|
Lean BMI (n = 99)
|
77.09 (9.53)
|
74.86 (8.14)
|
<0.001
|
Overweight BMI (n = 320)
|
83.47 (6.62)
|
81.33 (7.18)
|
<0.001
|
Obese BMI (n = 132)
|
84.87 (6.60)
|
83.86 (6.37)
|
0.08
|
No aspirin use (n = 773)
|
79.28 (6.97)
|
77.68 (7.56)
|
<0.001
|
Daily aspirin use (n = 390)
|
86.32 (6.34)
|
82.82 (6.94)
|
<0.001
|
Abbreviations: BMI, body mass index; MAP, mean arterial blood pressure; SD, standard
deviation.
Table 3
Changes in mean UtA PI from 11 to 14 to 19–24 + 6 gestation weeks by subgroups
|
11–14 weeks mean UtA PI
Mean (SD)
|
19–24 + 6 weeks mean UtA PI
Mean (SD)
|
Wilcoxon test p-Value
|
Nulliparous (n = 685)
|
1.73 (0.55)
|
1.00 (0.33)
|
<0.001
|
Multiparous (n = 478)
|
1.77 (0.52)
|
1.05 (0.35)
|
<0.001
|
Normal BMI (n = 612)
|
1.77 (0.52)
|
1.03 (0.32)
|
<0.001
|
Lean BMI (n = 99)
|
1.95 (0.52)
|
0.99 (0.43)
|
<0.001
|
Overweight BMI (n = 320)
|
1.67 (0.55)
|
0.99 (0.33)
|
<0.001
|
Obese BMI (n = 132)
|
1.66 (0.52)
|
1.07 (0.39)
|
<0.001
|
No aspirin use (n = 773)
|
1.60 (0.49)
|
0.97 (0.33)
|
<0.001
|
Daily aspirin use (n = 390)
|
2.05 (0.49)
|
1.11 (0.34)
|
<0.001
|
Abbreviations: BMI, body mass index; SD, standard deviation; UtA PI, uterine artery
pulsatility index.
Seventy-seven (35.16%) of the 219 pregnant women (6.62% of 1,163 women, 95% CI: 5.33,
8.20) had an abnormal mean UtA PI at both gestation age intervals. One hundred (10.59%)
of the 944 pregnant women with normal mean UtA PI in the 11 to 14 weeks assessment
had abnormal mean UtA PI in the 19–24 + 6 weeks assessment. On further analysis, we
observed that the mean UtA PI shifted from abnormal at 11 to 14 weeks to normal at
19–24 + 6 weeks in 59.38% of high-risk women. Among high-risk women, 40.63% had an
abnormal UtA PI at both first- and second-trimester assessments. The proportion of
women that shifted from a normal first-trimester UtA PI to an abnormal second-trimester
UtA PI was significantly higher in the high-risk group compared to the low-risk group
(13.74% vs. 9.38%, p = 0.048).
Multiparity (β coefficient: −0.19, 95% CI: −0.34, −0.04, p = 0.01) reduced risk and higher second-trimester MAP (β coefficient: 0.01, 95% CI:
0.004, 0.02, p = 0.007) increased risk for abnormal mean UtA PI in both trimesters in a multivariate
regression model that adjusted for maternal age, parity, BMI, and second-trimester
MAP. Maternal age (p = 0.30), BMI (p = 0.91), parity (p = 0.17), or second-trimester MAP (p = 0.72) was not significantly associated in the multivariate regression model with
a shift from normal mean UtA PI in the first-trimester assessment to an abnormal mean
UtA PI in the second trimester.
Discussion
We found that MAP and mean UtA PI measures decreased significantly from the first
to the second trimester of pregnancy in this study. Both MAP and mean UtA PI were
significantly higher in high-risk pregnant women compared to low-risk pregnant women
in the first trimester. We found that MAP measures were higher with increasing maternal
age and BMI in the first trimester and reduced with an increase in gestational age.
MAP reduced significantly between trimesters in both high-and-low risk subgroups and
probably reflected the natural decrease of blood pressure in the early trimesters
of pregnancy. Consistent with prior knowledge, we found that mean UtA PI decreased
with older maternal age and higher BMI in the first trimester.[15] The reduction in mean UtA PI with increasing gestational age is consistent with
existing knowledge of a significant reduction within-and-between trimesters of pregnancy.[13]
[16]
[17]
[18] The high-risk subgroup that received daily low-dose aspirin had MAP and mean UtA
PI measures that were significantly higher than the low-risk subgroup in the second-trimester
assessment. The reduction in mean UtA PI was significantly more among those taking
low-dose aspirin daily compared to those not taking low-dose aspirin in the high-risk
subgroup.
We found that the high-risk subgroup had a higher proportion of pregnant women that
shifted from an abnormal first trimester mean UtA PI to a normal second trimester
mean UtA PI (or normalized late) and a higher proportion of women with a UtA mean
PI shifting from normal to abnormal between the two intervals studied. These results
are consistent with a previous study that reported a similar pattern for complicated
and normal pregnancies between the first and second trimesters.[16] Previous studies have reported that delayed normalization of the UtA PI is associated
with an intermediate risk for adverse perinatal outcomes.[13]
[16] The subgroup of 77 pregnant women (6.62% of the overall study cohort, 95% CI: 5.33,
8.20) that had an abnormal mean UtA PI at both gestation age intervals is an important
group for surveillance when we consider the effects of delayed normalization of the
mean UtA PI. The estimate of 6.62% (95% CI: 5.33, 8.20) is consistent with the estimated
magnitude of PE in India that ranges from 5 to 10%. We found that nulliparity and
higher MAP increased the risk for abnormal mean UtA PI in the second trimester in
the high-risk subgroup that is consistent with the known risk factors for PE. We did
not find any specific associations for the transition from a normal first-trimester
mean UtA PI to an abnormal second-trimester mean UtA PI in the low-risk subgroup and
hence cannot comment on follow-up surveillance strategies for specific low-risk subgroups.
However, we feel it is important to repeat the UtA PI measurements in the second trimester
even for low-risk pregnant women as 1 in 10 low-risk pregnant women transitioned to
an abnormal UtA PI in the second trimester.
The quantum of reduction in MAP between the gestation intervals was statistically
significant. Previous studies have reported that maternal predisposing factors can
cause vasoactive and atherosclerotic changes that affect the transformation of spiral
arteries and influence the risk for PE.[19]
[20] Any decrease in MAP must therefore be useful to minimize the risk for PE. Underlying
maternal predisposing factors may contribute to the higher MAP and mean UtA PI measures
in the second trimester in high-risk pregnant women even though there is a significant
decrease from the first-trimester measures.
The inability to perform sequential assessments within trimesters may be considered
a limitation of our study but is a pragmatic reality of clinical practice in India
where it is not feasible to perform weekly ultrasound measurements routinely in a
clinical setting. We did not collect information on childbirth outcomes as part of
this study and hence cannot comment on the possible association of late normalization
of mean UtA PI with adverse perinatal outcomes in this study. The large number of
paired assessments at both gestational age intervals by radiologists certified for
ultrasound assessments is a strength of the study. Since this was a real-life practice-based
study, we had not masked the radiologist to the baseline measures. However, we do
not anticipate the lack of masking to significantly impact the study results as both
Doppler and MAP measures are objective measures.
Our results cannot be generalized to the larger population of pregnant women in India
due to the diversity and variations in population characteristics within and between
states and settings in India. However, our results suggest that a sequential assessment
of the MAP and mean UtA PI in the first and second trimesters of pregnancy will be
useful to identify a subgroup of women with abnormal mean UtA PI at both trimesters
that may need more intense surveillance and follow-up till childbirth. We do not provide
any specific recommendations on possible targets or management of MAP based on our
results as further studies on the normative distribution of MAP in pregnant Asian
Indian women, the potential associations of MAP in pregnancy in India, and the effectiveness
of possible management strategies including optimal thresholds are required.