Objective
The purpose of this study was to evaluate the utility of ultrasound surveillance in obese women.
Study Design
This is a retrospective cohort of all obese women who underwent sonography at a single center from 2005-2013. Inclusion criteria were body mass index ≥30 kg/m 2 , singleton, ≥1 ultrasound scan performed at <20 weeks of gestation, and ≥1 ultrasound scan performed at ≥24 weeks of gestation. Pregnancies with medical complications, fetal anomalies, or preterm premature rupture of membranes were excluded. Outcomes that were considered were small for gestational age, large for gestational age (LGA), macrosomia, oligohydramnios, and polyhydramnios. We calculated the number needed to screen (NNS) and 95% confidence interval (CI) for scans that were performed during 3 gestational age ranges (24–31 weeks 6 days, 32–35 weeks 6 days, and ≥36 weeks).
Results
Two thousand two sonograms were performed in 1164 obese women at ≥24 weeks of gestation. Small for gestational age was diagnosed in 59 pregnancies (5.1%); 7 pregnancies (0.6%) were diagnosed at <32 weeks of gestation (NNS, 159; 95% CI, 69–490). LGA was diagnosed in 38 cases (3.3%); only 1 case was identified at <32 weeks of gestation and was not LGA at birth. For every 29 (95% CI, 19–46) scans performed at >36 weeks of gestation, 1 case of macrosomia was identified. Amniotic fluid abnormalities were diagnosed in 44 pregnancies (3.8%; oligohydramnios, 19; polyhydramnios, 25); 34.1% abnormalities were diagnosed at <32 weeks of gestation (NNS: oligohydramnios, 113; 95% CI, 55–282 for oligohydramnios; polyhydramnios, 100; 95% CI, 50–230). At ≥36 weeks of gestation, 7 (95% CI, 6–8) scans were needed to diagnose any fluid or growth abnormality.
Conclusion
In obese women without comorbidities, few sonographic diagnoses of amniotic fluid or fetal growth abnormalities are made at <32 weeks of gestation. Therefore, if a policy of serial sonographic surveillance is used, we suggest ultrasound scans for fluid and growth in obese women to begin at ≥32 weeks of gestation.
Obesity, which affects more than one-third of all adult women in the United States, poses significant risks to pregnancy, including increased risks of stillbirth, growth disorders, and placental dysfunction. In unexplained stillbirths, 52% of cases were growth restricted. Prenatal identification of fluid and growth abnormalities enables the managing physician to initiate antenatal testing, with the goal of reducing the risk of stillbirth. Obesity also increases the risk of large-for-gestational age infants and macrosomia. The finding of macrosomia on ultrasound scanning may affect labor and delivery management, which includes the decision to perform a prelabor cesarean delivery.
Routine screening for fluid and growth abnormalities occurs at every prenatal visit with fundal height measurements. However, obesity, particularly when a pannus is present, may hamper fundal height measurements by limiting the ability to palpate the pubic symphysis and uterine fundus and by potentially falsely increasing the measurements because of inclusion of the pannus. Consequently, some investigators recommend the performance of serial sonograms for fluid and growth in obese women. This approach typically adds 2-4 sonograms to the care of approximately 30% of all obstetric patients, but, the utility of this approach has never been evaluated systematically. Therefore, we aimed to evaluate the utility of sonographic surveillance in the obese obstetric population.
Materials and Methods
This was a retrospective cohort study that was performed at a single tertiary care center from January 2005 to June 2013. Institutional review board approval was obtained from the University of Alabama at Birmingham.
We identified all women with a body mass index (BMI) ≥30 kg/m 2 who underwent obstetric ultrasound scanning and who delivered at the University of Alabama at Birmingham. Women were identified with the University of Alabama at Birmingham obstetric automated record system, which tracks all obstetric patients who use the University of Alabama at Birmingham health system and records information such as physical examination (including height and weight), medical and obstetric complications, gestational age, and delivery date. BMI was determined by documented height and weight at the first visit <20 weeks’ gestation. These records were combined with our ultrasound database, which has been in place since January 2005. Data that were obtained from the ultrasound database include date of ultrasound scan, fetal biometric measurements, estimated fetal weight, and amniotic fluid index (AFI). During the time period of this study, it was our practice to obtain at least 1 ultrasound scan for fluid and growth in obese women at >28 weeks’ gestation.
Women were included if they had a singleton pregnancy, had at least 1 ultrasound scan for dating at ≤20 weeks’ gestation, and at least 1 ultrasound scan at ≥24 weeks’ gestation with fluid and biometry measurements. Women were excluded if they had another indication for ultrasound scanning, which included maternal medical problems (eg, chronic hypertension, diabetes mellitus [pregestational and gestational], lupus) and pregnancy complications (eg, preterm premature rupture of membranes, fetal anomalies). Because the diagnosis of a growth-restricted fetus on ultrasound scan may prompt an evaluation for preeclampsia, women with preeclampsia were not excluded from this study. For women with >1 pregnancy recorded during the study ascertainment period, only the first pregnancy during the time period was considered.
Outcomes that were considered included the sonographic diagnosis of small for gestational age (SGA), large for gestational age (LGA), macrosomia, oligohydramnios, or polyhydramnios. SGA was defined as an estimated fetal weight as <10th percentile; LGA was defined as >90th percentile on the Alexander growth standard. Macrosomia was defined as estimated fetal weight of >4000 g. Oligohydramnios was defined as an AFI of <5 cm; polyhydramnios was defined as an AFI of >25 cm.
Women who underwent 1, 2, and ≥3 sonograms were compared with the use of univariate statistics. Three gestational age ranges for ultrasound scanning were stratified for analysis: 24 weeks to 31 weeks 6 days’ gestation, 32 weeks to 35 weeks 6 days’ gestation, and ≥36 weeks’ gestation. If a woman underwent >1 ultrasound scan in a given gestational age range, each ultrasound scan was counted, but each subject was counted as only 1 case (ie, if a woman had 2 scans between 24 weeks to 31 weeks 6 days’ gestation and was diagnosed with SGA, it counted for 2 ultrasound scans but only 1 case of SGA). For each gestational age range, we determined the number of sonograms that were performed and the number of fluid and growth abnormalities that were identified. The number of SGA, LGA, and macrosomic infants who were identified on ultrasound scanning was compared with SGA, LGA, and macrosomia at birth to calculate the test characteristics (sensitivity, specificity, positive predictive value, negative predictive value) of ultrasound scanning at each gestational age range. We then determined the number of sonograms that were needed to be performed at each gestational age range (number needed to screen; NNS) to identify 1 case of each outcome. A stratified analysis was performed by BMI category (30-34.9 kg/m 2 , 35-39.9 kg/m 2 , and ≥40 kg/m 2 ). Additionally, we calculated the ongoing risk of an abnormal ultrasound after normal sonogram results at 24 weeks to 31 weeks 6 days’ gestation and 32 weeks to 35 weeks 6 days’ gestation.
Charts of women who had an abnormal ultrasound finding (SGA, LGA, macrosomia, oligohydramnios, and/or polyhydramnios) were reviewed to determine whether (1) the abnormal ultrasound finding affected patient treatment and (2) the abnormal ultrasound finding would have been suspected clinically (ie, based on fundal height or other indications for ultrasound scanning).
All analyses were performed with SAS software (version 9.2; SAS Institute Inc, Cary, NC).
Results
Of 9040 singleton gestations that were delivered to obese women at our institution during the 8.5-year time period, 1164 women were included (2095 women were excluded for maternal medical or pregnancy problems; 1365 women were excluded for not being the first pregnancy in the ascertainment period, and 4416 women were excluded for not having received prenatal care at our institution or not having had sonograms during the appropriate gestational age time periods at our institution). These 1164 women underwent 2002 sonograms with biometry, and AFI measurements were performed at ≥24 weeks’ gestation. Women who underwent 1, 2, and ≥3 sonograms were similar with respect to race, parity, education, marital status, and tobacco use ( Table 1 ). Women who underwent >1 sonogram at >24 weeks’ gestation were slightly older and more likely to have private insurance, to have class III obesity, and to deliver a heavier baby.
| Variable | Ultrasound scans | P value | ||
|---|---|---|---|---|
| 1 (n = 608) | 2 (n = 336) | ≥3 (n = 220) | ||
| Maternal age, y a | 25.6 ± 5.3 | 26.9 ± 6.1 | 27.7 ± 6.0 | < .01 |
| Maternal race, n | .09 | |||
| Black | 431 (71.2%) | 227 (68.2%) | 150 (68.5%) | |
| White | 84 (13.9%) | 58 (17.4%) | 43 (19.6%) | |
| Hispanic | 87 (14.4%) | 42 (12.6%) | 26 (11.9%) | |
| Other | 3 (0.5%) | 6 (1.8%) | 0 | |
| Nulliparous, n | 271 (44.6%) | 157 (46.7%) | 83 (37.7%) | .10 |
| Private pay, n | 19 (3.1%) | 31 (9.2%) | 19 (8.6%) | < .01 |
| College education, n | 73 (17.9%) | 43 (21.5%) | 17 (20.7%) | .53 |
| Married, n | 85 (14.7%) | 58 (17.7%) | 36 (16.5%) | .47 |
| Alcohol use, n | 101 (18.2%) | 80 (26.9%) | 34 (19.4%) | .01 |
| Tobacco use, n | 158 (27.8%) | 105 (34.0%) | 56 (29.8%) | .16 |
| Illicit drugs use, n | 81 (14.8%) | 59 (20.3%) | 28 (17.8%) | .12 |
| Body mass index, kg/m 2 a | 36.8 ± 6.0 | 38.4 ± 7.0 | 40.6 ± 7.5 | < .01 |
| Obesity category | < .01 | |||
| Category 1 (30.0–34.9 kg/m 2 ), n | 265 (43.6%) | 115 (34.2%) | 56 (25.5%) | |
| Category 2 (35.0–39.9 kg/m 2 ), n | 192 (31.6%) | 99 (29.5%) | 51 (23.2%) | |
| Category 3 (≥40.0 kg/m 2 ), n | 151 (24.8%) | 122 (36.3%) | 113 (51.4%) | |
| Gestational age at delivery, wk a | 38.8 ± 2.7 | 38.7 ± 2.3 | 39.1 ± 1.6 | .15 |
| Birthweight, g a | 3213 ± 687 | 3178 ± 704 | 3353 ± 553 | < .01 |
SGA was diagnosed in 59 pregnancies (5.1%). Of these, only 7 SGA pregnancies (0.6%) were diagnosed at <32 weeks’ gestation; 20 SGA pregnancies (1.7%) were diagnosed between 32 weeks and 35 weeks 6 days’ gestation, and 38 SGA pregnancies (3.3%) were diagnosed at ≥36 weeks’ gestation ( Table 2 ). The sensitivity of ultrasound scanning for the diagnosis of SGA increased as gestational age progressed, and the number of sonograms that needed to be performed to detect 1 true case of SGA decreased from 159 (95% CI, 69–490) at 24 weeks to 31 weeks 6 days’ gestation to 25 (95% CI, 17–39) at ≥36 weeks’ gestation.
| Diagnosis | Ultrasound scans performed, n | Cases diagnosed on ultrasound scan, n a | No. needed to screen (95% CI) b | Sensitivity, % | Specificity, % | Positive predictive value, % | Negative predictive value, % |
|---|---|---|---|---|---|---|---|
| Small for gestational age (n = 59) | |||||||
| 24-31.6 wk | 796 | 7 | 159 (69–490) | 6.1 | 99.7 | 71.4 | 90.3 |
| 32-35.6 wk | 600 | 20 | 43 (25–78) | 22.6 | 98.7 | 66.7 | 91.7 |
| ≥36 wk | 606 | 38 | 25 (17–39) | 32.0 | 97.0 | 58.5 | 91.0 |
| Large for gestational age (n = 38) | |||||||
| 24-31.6 wk | 796 | 1 | — c | — c | 99.9 | — c | 89.8 |
| 32-35.6 wk | 600 | 3 | 200 (69–969) | 4.6 | 100 | 100 | 89.5 |
| ≥36 wk | 606 | 35 | 22 (15–32) | 38.6 | 98.5 | 77.1 | 92.5 |
| Macrosomia (n = 23) | |||||||
| 24-31.6 wk | 796 | 0 | — c | — c | — c | — c | — c |
| 32-35.6 wk | 600 | 0 | — c | — c | — c | — c | — c |
| ≥36 wk | 606 | 23 | 29 (19–46) | 27.4 | 99.4 | 87.0 | 90.9 |
| Oligohydramnios (n = 19) | |||||||
| 24-31.6 wk | 796 | 7 | 113 (55–282) | N/A | N/A | N/A | N/A |
| 32-35.6 wk | 600 | 1 | 600 (108–23,696) | N/A | N/A | N/A | N/A |
| ≥36 wk | 606 | 11 | 55 (31–109) | N/A | N/A | N/A | N/A |
| Polyhydramnios (n = 25) | |||||||
| 24-31.6 wk | 796 | 8 | 100 (50–230) | N/A | N/A | N/A | N/A |
| 32-35.6 wk | 600 | 10 | 60 (32–125) | N/A | N/A | N/A | N/A |
| ≥36 wk | 606 | 10 | 61 (33–126) | N/A | N/A | N/A | N/A |
| Any growth or fluid abnormality d | |||||||
| 24-31.6 wk | 796 | 21 | 38 (25–61) | N/A | N/A | N/A | N/A |
| 32-35.6 wk | 600 | 33 | 18 (13–25) | N/A | N/A | N/A | N/A |
| ≥36 wk | 606 | 87 | 7 (6–8) | N/A | N/A | N/A | N/A |
a Number of cases diagnosed may total more than the total number of cases because each subject underwent multiple ultrasound scans, presented as n
b For growth abnormalities (small for gestational age, large for gestational age, macrosomia), the number needed to screen was calculated with birthweight as the true positive; for amniotic fluid abnormalities, the number needed to screen was determined by the number of amniotic fluid abnormalities detected, because these cannot be confirmed at birth
c Not calculated because there were no cases
d Small for gestational age, large for gestational age, oligohydramnios, or polyhydramnios.
LGA was diagnosed in 38 pregnancies (3.3%). The only case of LGA that was diagnosed at <32 weeks’ gestation was a false positive (ie, was not LGA at birth). At 32 weeks to 35 weeks 6 days’ gestation, 3 cases of LGA were diagnosed, and all were LGA at birth; at this gestational age, 200 (95% CI, 69–969) sonograms would need to be performed to diagnose 1 case of LGA. At ≥36 weeks’ gestation, 35 cases of LGA were diagnosed by sonogram, which reduced the NNS to 22 (95% CI, 15–32). Macrosomia was not diagnosed in any pregnancy at <36 weeks’ gestation; at ≥36 weeks’ gestation, 29 (95% CI, 19–46) sonograms needed to be performed to diagnose 1 case. As with SGA, sensitivity for the detection of LGA increased at later gestational ages.
Amniotic fluid abnormalities were diagnosed in 44 pregnancies (3.8%): oligohydramnios in 19 pregnancies (1.6%) and polyhydramnios in 25 pregnancies (2.1%). Of these 19 cases, 7 were diagnosed at <32 weeks’ gestation (NNS, 100; 95% CI, 50–230), and 1 was diagnosed at 32 weeks to 35 weeks 6 days’ gestation (NNS, 600; 95% CI, 108–23,696). Oligohydramnios was diagnosed in 11 women at ≥36 weeks’ gestation (NNS, 55; 95% CI, 31–109). Of the 25 cases who were diagnosed with polyhydramnios, 8 cases were detected at <32 weeks’ gestation (NNS, 100; 95% CI, 50–230); 10 cases were detected at 32 weeks to 35 weeks 6 days’ gestation (NNS, 60; 95% CI, 32–125), and 10 cases were detected at ≥36 weeks’ gestation (NNS, 61; 95% CI, 33–126).
Considering the presence of any fluid or growth abnormality, the number of sonograms needed to be performed to diagnose 1 case declined from 38 (95% CI, 25–61) at <32 weeks’ gestation to 18 (95% CI, 13–25) from 32 weeks to 35 weeks 6 days’ gestation, with a further decline to 7 (95% CI, 6–8) at ≥36 weeks’ gestation.
A similar pattern was seen when the women were examined by BMI category ( Table 3 ). As gestational age progressed, the number of cases that were detected increased.
| Category | Ultrasound scans, n | Small for gestational age, n | Large for gestational age, n | Macrosomia, n | Oligohydramnios, n | Polyhydramnios, n |
|---|---|---|---|---|---|---|
| 1: BMI, 30-34.9 kg/m 2 (n = 436) | ||||||
| 24-31.6 wk | 250 | 6 | 0 | 0 | 1 | 2 |
| 32-35.6 wk | 201 | 13 | 1 | 0 | 1 | 4 |
| ≥36 wk | 230 | 20 | 17 | 9 | 6 | 4 |
| 2: BMI, 35.0-39.9 kg/m 2 (n = 342) | ||||||
| 24-31.6 wk | 226 | 0 | 0 | 0 | 5 | 3 |
| 32-35.6 wk | 164 | 2 | 1 | 0 | 0 | 2 |
| ≥36 wk | 159 | 10 | 9 | 8 | 0 | 1 |
| 3: BMI, ≥40.0 kg/m 2 (n = 386) | ||||||
| 24-31.6 wk | 320 | 1 | 1 | 0 | 1 | 3 |
| 32-35.6 wk | 235 | 5 | 1 | 0 | 0 | 4 |
| ≥36 wk | 217 | 8 | 9 | 6 | 5 | 5 |
The ongoing risk of abnormal fluid or growth after a normal ultrasound scan at each gestational age was calculated. Of the 689 women with a fluid and growth ultrasound scan from 24 weeks to 31 weeks 6 days’ gestation, no abnormality was detected at that time for 668 women. Of the 668 women with a normal ultrasound scan from 24 weeks to 31 weeks 6 days’ gestation, 439 women had a repeat ultrasound scan at ≥32 weeks’ gestation, and 57 of those results (13.0%) were abnormal. Of the 147 women who had a normal ultrasound scan at both 24 weeks to 31 weeks 6 days’ gestation and 32 weeks to 35 weeks 6 days’ gestation, 21 women (14.3%) went on to experience a fluid or growth abnormality at their next ultrasound scan. Of the 243 women whose first normal ultrasound scan for fluid and growth occurred at 32 weeks to 35 weeks 6 days’ gestation, 66 women had a repeat ultrasound scan at ≥36 weeks’ gestation, and 7 of these women (10.6%) went on to experience a fluid or growth abnormality.
The charts of women who were diagnosed with an abnormal ultrasound finding were then reviewed to assess what actions were taken based on the abnormal ultrasound finding (ie, repeat ultrasound scan, antenatal testing, or delivery). We also assessed if the finding may have been detected on clinical examination (ie, fundal height measurement) or if other clinical assessment would have prompted the ultrasound scan. Table 4 presents the abnormal ultrasound findings at <32 weeks’ gestation. Of the 8 pregnancies diagnosed with SGA (with or without oligohydramnios), only 2 cases (case 3, case 4) would not have been detected based on patient complaints (vaginal bleeding, decreased fetal movement), blood pressure findings, or fundal height measurements (either greater or less than expected based on gestational age). These pregnancies both underwent repeat ultrasound scans and were grown appropriately on follow-up examination. All cases of oligohydramnios would have been detected based on patient complaints. Of note, several women were referred for suspicion of ruptured membranes or fetal anomaly on outside hospital scans; both fetal anomalies and preterm rupture of membranes were exclusion criteria for this study and should not have been included in the analysis. However, to avoid ascertainment biases, we retained them in the analysis. All cases of polyhydramnios that were diagnosed at <32 weeks’ gestation were resolved on repeat ultrasound scans; of the 4 cases with fundal height measurements, all were at least 4 cm greater than expected by gestational age.
