- Email: [email protected]
Risk of stillbirth after 37 weeks in pregnancies complicated by small-for-gestational-age fetuses
Research
www. AJOG.org
OBSTETRICS
Risk of stillbirth after 37 weeks in pregnancies complicated by small-for-gestational-age fetuses Amanda S. Trudell, DO; Alison G. Cahill, MD, MSCI; Methodius G. Tuuli, MD, MPH; George A. Macones, MD, MSCE; Anthony O. Odibo, MD, MSCE OBJECTIVE: The evidence for delivering small-for-gestational-age
(SGA) fetuses at 37 weeks remains conflicting. We examined the risk of stillbirth per week of gestation beyond 37 weeks for pregnancies complicated by SGA. STUDY DESIGN: Singleton pregnancies undergoing routine second tri-
mester ultrasound from 1990-2009 were examined retrospectively. The risk of stillbirth per 10,000 ongoing SGA pregnancies with 95% confidence intervals (CIs) was calculated for each week of gestation ⱖ37 weeks. Using a life-table analysis with correction for censoring, conditional risks of stillbirth, cumulative risks of stillbirth per 10,000 ongoing SGA pregnancies and relative risks (RRs) were calculated with 95% CIs for each week of gestation. RESULTS: Among 57,195 pregnancies meeting inclusion criteria the
background risk of stillbirth was 56/10,000 (95% CI, 42.3⫺72.7) with stillbirth risk for SGA pregnancies of 251/10,000 (95% CI,
221.2⫺284.5). The risk of stillbirth after the 37th week was greater compared with pregnancies delivered in the 37th week (47/10,000, 95% CI, 34.6⫺62.5 vs 21/10,000, 95% CI, 13.0⫺32.1; RR, 2.2; 95% CI, 1.3⫺3.7). The cumulative risk of stillbirth rose from 28/ 10,000 ongoing SGA pregnancies at 37 weeks to 77/10,000 at 39 weeks (RR, 2.75; 95% CI, 1.79⫺4.2). Among pregnancies complicated by SGA ⬍5% the cumulative risk of stillbirth at 38 weeks was significantly greater than the risk at 37 weeks (RR, 2.3; 95% CI, 1.4⫺3.8). CONCLUSION: There is a significantly increased risk of stillbirth in preg-
nancies complicated by SGA delivered after the 37th week. Given these findings, we advocate a policy of delivery of SGA pregnancies 37-38 weeks. Key words: SGA, small for gestational age, stillbirth
Cite this article as: Trudell AS, Cahill AG, Tuuli MG, et al. Risk of stillbirth after 37 weeks in pregnancies complicated by small-for-gestational-age fetuses. Am J Obstet Gynecol 2013;208:376.e1-7.
N
early 3 million stillbirths occur worldwide each year.1 In high-income countries stillbirth rates have remained constant in recent decades with significant global variations in stillbirth rates among affluent nations, suggesting further reduction in stillbirth rates is possible.2 Fetal growth restriction has been established as a major cause of stillbirth.3 In a metaanalysis published in From the Department of Obstetrics and Gynecology, Washington University School of Medicine, St. Louis, MO. Received Nov. 9, 2012; revised Jan. 27, 2013; accepted Feb. 18, 2013. The authors report no conflict of interest. Presented in an oral format at the 33nd annual meeting of the Society of Maternal Fetal Medicine, San Francisco, CA, Feb. 11-16, 2013. The racing flag logo above indicates that this article was rushed to press for the benefit of the scientific community. Reprints not available from the authors. 0002-9378/$36.00 © 2013 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2013.02.030
376.e1
2011, the risk of stillbirth in high-income countries was demonstrated to be 4times higher in fetuses measuring smallfor-gestational-age (SGA) compared with non-SGA fetuses and SGA was noted to have the greatest populationattributable risk compared to other risks of stillbirth.4 The delivery timing of SGA pregnancies weighs competing risks. Early delivery avoids stillbirth but increases risks of respiratory distress syndrome, hypoglycemia, and neonatal sepsis while continued gestation risks stillbirth. Recent clinical trials with long-term follow-up have sought to clarify optimal timing of delivery for SGA fetuses. A recent randomized controlled trial evaluated expectant management versus induction of labor of SGA fetuses at term,5 but because of the rarity of stillbirth, the trial was not powered to study this important outcome. Given the impact of stillbirth both globally and nationally, defining the risk of stillbirth after 37 weeks is fundamental to answering the question of optimal timing of delivery for pregnancies com-
American Journal of Obstetrics & Gynecology MAY 2013
plicated by SGA, and ultimately, an important step in stillbirth prevention. The specific aim of our study was to estimate the risk of stillbirth for each week of gestation beyond 37 weeks in pregnancies complicated by SGA.
M ATERIALS AND M ETHODS We conducted a retrospective cohort study within our prospectively collected perinatal database at Washington University School of Medicine. Before initiation of the study, we obtained approval from the institutional review board. Sociodemographic, obstetric, medical and pregnancy follow-up data were obtained via self-report questionnaires at the initiation of prenatal care, at the time of ultrasound and from the hospital medical record on delivery. Detailed patient information was entered into the database by trained research nurses. Patients and/or primary obstetricians were contacted via telephone to obtain complete follow-up information if necessary, or in the rare event when the patient delivered at an outside facility. The study included
Obstetrics
www.AJOG.org
FIGURE 1
Flow diagram detailing exclusion criteria
This flow diagram details the study population and the exclusion criteria used to identify singleton pregnancies without prenatally diagnosed chromosomal disorders or fetal anomalies. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
all singleton pregnancies seen for anatomy survey between 16 and 23 weeks from January 1990 to December 2009. Approximately 97% of our study population are seen during this period. We excluded pregnancies complicated by prenatally diagnosed major congenital anomalies and aneuploidy and those without documented birthweight (Figure 1). Ultrasounds were performed by certified, dedicated obstetric sonographers and final ultrasound interpretation and diagnosis made after review by a fellowship trained maternal fetal medicine specialist. Gestational age was assigned by last menstrual period (LMP) and confirmed with ultrasound. Gestational age was assigned by ultrasound when the LMP was noted to be discrepant by ⱖ7 days in the first trimester or ⱖ10 days in the second trimester. If LMP was unknown, gestational age was assigned by ultrasound. The primary outcome was stillbirth, defined as intrauterine fetal death at or beyond 20 weeks’ gestation. SGA was defined as birthweight ⬍10th percentile
based on the Alexander growth standard.6 The incidence of stillbirth per week of gestation was calculated as the number of SGA stillbirths per 10,000 ongoing SGA pregnancies at the start of that gestational week. The choice of ongoing pregnancies as the denominator is a logical approach because compared with the traditional stillbirth risk assessment that uses the number of deliveries occurring in a given week as the denominator, ongoing pregnancies are the pregnancies that continue to be at risk of stillbirth. Thus, this is the contemporary method of calculating stillbirth risk. 7-9 Although the number of stillbirths/ 10,000 ongoing pregnancies will assess the risk of stillbirth per week of gestation, it does not assess the cumulative risk of stillbirth. The cumulative risk of stillbirth becomes more clinically relevant than the conditional risk of stillbirth when one is trying to decide between a recommendation for delivery or expectant management. Furthermore, the risk of stillbirth/10,000 ongoing pregnancies includes in the denominator the pregnancies that might deliver live births
Research
during the proposed week. Therefore, including these pregnancies in the denominator will tend to under estimate the risk of stillbirth. This effect of censoring, while not significant in the conditional probability estimate of stillbirth, is magnified and may become clinically significant when one calculates cumulative probability. Therefore, to calculate a cumulative risk of stillbirth beyond 37 weeks and to account for censoring, we further analyzed our data using the method of life-table analysis as proposed by Smith in 2001.7 For the life-table analysis, the estimation of stillbirth risk was calculated as follows: if Pn is the number of ongoing pregnancies at gestational week n, Bn is the number of births during gestational week n and Sn is the number of stillbirths during gestational week n then the conditional probability of stillbirths PSn at gestational week n: PSn ⫽ Sn/[Pn – (0.5 ⫻ Bn)], and the cumulative probability of stillbirths CSn is calculated as the product of the conditional probabilities of survival, where survival is 1 ⫺ Probability of death, at gestational week n: CSn ⫽ 1 ⫺ [(1 ⫺ PS37) ⫻ (1 ⫺ PS38) . . . ⫻ (1 ⫺ PSn)]. Descriptive statistics were used to provide an overview of the study population and estimate the baseline risk of stillbirth. Maternal demographic, medical and obstetric information were described among SGA pregnancies ⬍37 weeks, SGA pregnancies ⱖ37 weeks, and non-SGA pregnancies. To provide a more detailed overview of the SGA population, we also compared maternal demographics and obstetric information between SGA stillbirths and SGA live births. Descriptive statistics were performed using 2 for categorical variables and student t test for continuous variables. The risk of stillbirth with 95% confidence interval (CI) as calculated for each week of gestation 37-37 6/7, 38-38 6/7, 39-39 6/7, and ⱖ40 weeks. The relative risk (RR) of stillbirth with 95% CI was calculated for SGA pregnancies delivered after the 37th week compared with the risk of stillbirth in pregnancies delivered in the 37th week. Conditional risks and cumulative risks of stillbirth per week of gestation with 95% CIs were
MAY 2013 American Journal of Obstetrics & Gynecology
376.e2
Research
Obstetrics
calculated using life-table analysis with correction for censoring as stated above. Risk ratios for stillbirth with 95% CIs were calculated for each week of gestation relative to the risk of stillbirth at 37 weeks using the cumulative risks for stillbirth. To evaluate the cumulative risk of stillbirth among neonates with a lower threshold of SGA, the risk of SGA for neonates with birthweight ⬍5% was also determined by life-table analysis. To estimate if changes in clinical management or technology over the period of the study affected stillbirth risks, we calculated and compared risks of stillbirth among SGA fetuses ⱖ37 weeks’ gestation between 2 time periods determined by days from initiation of enrollment of 50% of the study population, approximately 1990-2000 and 2001-2009. The risk of neonatal morbidity in association with early term delivery was evaluated in a subanalysis for all neonates admitted to the neonatal intensive care unit or special care nurseries after delivery. Postnatally diagnosed aneuploidy and congenital anomalies were excluded. Adverse neonatal outcomes measured included intensive care unit admission, respiratory distress syndrome, meconium aspiration syndrome, and length of stay ⬎5 days. The risk of adverse outcomes at 38, 39, and ⱖ40 weeks relative to the risk at 37 weeks was calculated with 95% CI. Statistical analysis was performed with STATA 12 (StataCorp, College Station, TX).
www.AJOG.org
TABLE 1
Demographic characteristics of SGA and non-SGA pregnancies
Characteristic Maternal age, y (SD)
SGA delivered SGA delivered <37 wks <37 wks Non-SGA n ⴝ 884 (1.5%) n ⴝ 3333 (5.8%) n ⴝ 52,978 (92.6%) 29.3 ⫾ 6.6
29.0 ⫾ 7.0
30.4 ⫾ 6.2
..............................................................................................................................................................................................................................................
AMA
226 (25.6)
877 (26.3)
15,779 (29.8)
..............................................................................................................................................................................................................................................
Race
.....................................................................................................................................................................................................................................
Black
378 (42.8)
1253 (37.6)
11,370 (21.5)
White
384 (43.4)
1536 (46.1)
33,397 (63.0)
Other
122 (13.8)
544 (16.3)
8211 (15.5)
Nulliparous
450 (50.9)
1644 (49.3)
19,981 (37.7)
86 (9.7)
93 (2.8)
1248 (2.4)
327 (37.3)
340 (10.3)
3919 (7.5)
..................................................................................................................................................................................................................................... ..................................................................................................................................................................................................................................... .............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
Chronic hypertension
..............................................................................................................................................................................................................................................
Preeclampsia
..............................................................................................................................................................................................................................................
Pregestational diabetes
37 (4.2)
22 (0.7)
1051 (2.0)
Gestational diabetes
40 (4.6)
139 (4.2)
2841 (5.4)
Gestational age at delivery, wk
33.4 ⫾ 4.1
.............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
39.1 ⫾ 1.1
38.8 ⫾ 2.3
..............................................................................................................................................................................................................................................
Birthweight, g
1601 ⫾ 618
2659 ⫾ 238
3397 ⫾ 730
Stillbirth, n (%)
86 (9.7)
20 (0.6)
217 (0.4)
.............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
AMA, advanced maternal age; SD, standard deviation; SGA, small-for-gestational-age. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
39.1 ⫾ 1.1 weeks. Table 2 compares maternal demographic characteristics of SGA stillbirths delivered ⱖ37 weeks and SGA live births delivered ⱖ37 weeks. There was no significant difference in maternal age, race, parity, diabetes, or
hypertensive disorders, including preeclampsia, among women who delivered SGA stillbirths after 37 weeks and those who delivered SGA live births after 37 weeks. As expected, SGA stillbirths were significantly smaller at delivery than
TABLE 2
Demographic characteristics of SGA pregnancies delivered >37 weeks Characteristic
SGA stillbirth >37 (n ⴝ 20)
Maternal age (SD)
29.9 ⫾ 6.4
SGA live birth >37 (n ⴝ 3313) 29.0 ⫾ 7.0
P value .59
..............................................................................................................................................................................................................................................
R ESULTS
AMA, n (%)
Among 57,195 pregnancies meeting inclusion criteria, the background risk of stillbirth was 56/10,000 (95% CI, 42.3⫺72.7). SGA complicated 4217 (7.4%) pregnancies, of which 3333 (5.8%) delivered ⱖ37 weeks. Table 1 demonstrates relevant demographic information of the study cohort including the SGA pregnancies delivered ⬍37 weeks, the SGA pregnancies delivered ⱖ37 weeks and the non-SGA population. The primary concern of this investigation was the SGA group delivered ⱖ37 weeks of which the average gestational age at the time of delivery was
Race
376.e3
5 (25)
872 (26)
.89
.............................................................................................................................................................................................................................................. .....................................................................................................................................................................................................................................
Black, n (%)
8 (40)
1245 (38)
.83
White, n (%)
9 (45)
1527(46)
.92
Other, n (%)
3 (15)
541 (16)
.87
Nulliparous, n (%)
8 (40)
1636 (49)
.40
Chronic hypertension, n (%)
1 (5)
92 (2.8)
.55
Preeclampsia, n (%)
1 (5)
339 (10)
.43
Pregestational diabetes, n (%)
0 (0)
22 (0.6)
.71
Gestational diabetes, n (%)
2 (10)
..................................................................................................................................................................................................................................... ..................................................................................................................................................................................................................................... .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
137 (4)
.20
..............................................................................................................................................................................................................................................
Gestational age, wk (SD)
38.8 ⫾ 1.3
39.1 ⫾ 1.1
.27
..............................................................................................................................................................................................................................................
AMA, advanced maternal age; SD, standard deviation; SGA, small-for-gestational-age. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
American Journal of Obstetrics & Gynecology MAY 2013
Obstetrics
www.AJOG.org
TABLE 3
Risk of stillbirth per week of gestation Ongoing SGA pregnancies
SGA stillbirths (n ⴝ 20)
Stillbirth risk/10,000 ongoing SGA pregnancies (95% CI)
37
3333
7
21 (13.0–32.1)
38
2776
3
11 (5.5–19.7)
39
1953
5
26 (17.0–38.1)
ⱖ40
832
5
60 (45.8–77.2)
GA, wk
.............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
CI, confidence interval; GA, gestational age; SGA, small-for-gestational-age. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
SGA live births (2370 ⫾ 314 vs 2596 ⫾ 237, P ⬍ .01). Among SGA pregnancies, 106 (2.5%), were complicated by stillbirth with a stillbirth risk of 251/10,000 (95% CI, 221.2⫺283.5). Table 3 demonstrates the risk of stillbirth per week of gestation. In the 37th week, the risk of stillbirth was 21/10,000 ongoing SGA pregnancies (95% CI, 13.0⫺32.1) compared with 47/ 10,000 (95% CI, 34.6⫺62.5) after the 37th week (RR, 2.2; 95% CI, 1.3⫺3.7). Using life-table analysis with correction for censoring, the cumulative risk of stillbirth was calculated for each week of gestation (Table 4). The risk of stillbirth in the 37th week was noted to be 28/ 10,000 (95% CI, 18.6⫺40.0). As expected, the cumulative risk of stillbirth rose with each advancing week of gestation to 41/10,000 (95% CI, 29.4⫺55.6) in the 38th week, 77/10,000 (95% CI, 60.8⫺96.1) in the 39th week and 194/ 10,000 (95% CI, 168.9⫺222.4) at ⱖ40 weeks. Figure 2 graphically illustrates a rise in cumulative risk of stillbirth for
each advancing week of gestation. The risk ratio for stillbirth for each week of gestation relative to 37 weeks is also demonstrated in Table 4. The cumulative risk of stillbirth at each advancing week of gestation relative to the cumulative risk of stillbirth at 37 weeks was significantly higher at 39 weeks (RR, 2.75; 95% CI, 1.8⫺4.2). At ⱖ40 weeks the risk of stillbirth rose nearly 7-fold compared to the risk of stillbirth at 37 weeks (95% CI, 4.7⫺10.3). The life-table analysis for SGA ⬍5% is located in Table 5. Of the 3333 SGA pregnancies delivered ⱖ37 weeks, 1422 (42.7%) were SGA ⬍5% with a total of 12 stillbirths for an overall stillbirth rate of 84/10,000. At this lower threshold of SGA the cumulative risk of stillbirth at 38 weeks was significantly greater compared with the risk of stillbirth at 37 weeks (RR, 2.3; 95% CI, 1.4⫺3.8). Adverse neonatal outcomes are demonstrated in Table 6. The risk of NICU admission, respiratory distress syndrome, and meconium aspiration syn-
Research
drome at 38 weeks was found to be lower relative to the risk at 37 weeks (RR, 0.73; 95% CI, 0.6⫺0.9; RR, 0.67; 95% CI, 0.5⫺0.9; RR, 0.71; 95% CI, 0.6⫺0.9), respectively. Length of stay ⬎5 days was not significantly different at 38 compared with 37 weeks (RR, 0.77; 95% CI, 0.6⫺1.0). At ⱖ40 weeks, the risk of meconium aspiration syndrome is greater than the risk at 37 weeks (RR, 2.0; 95% CI, 1.1⫺3.5). Sensitivity analysis showed no evidence that advancing technology and changes in clinical management over the 19-year study period affected the risk of stillbirth among SGA fetuses. Risks of SGA stillbirths after 37 weeks were not significantly different in the 2 halves of the study period (54/10,000 vs 67/10,000 P ⫽ .62, for 1990-2000 and 2001-2009, respectively).
C OMMENT Our study demonstrates a 2-fold risk of stillbirth after the 37th week of gestation compared with pregnancies delivered in the 37th week. The cumulative risk of stillbirth for each week of gestation beyond 37 weeks increased. At 39 weeks and ⱖ40 weeks, the stillbirth risk was nearly 3-fold and 7-fold. The relative risk of stillbirth at 38 weeks reached statistical significance when the SGA threshold was set at the 5th percentile. Although, the risk of stillbirth in SGA pregnancies has been evaluated,10 our investigation is unique in estimating the cumulative risk of stillbirth in SGA pregnancies stratified by gestational age. This provides impor-
TABLE 4
Life-table analysis SGA <10%: risk of stillbirth/10,000 ongoing SGA pregnancies per week of gestation beyond 37 weeks Conditional probabilitya of stillbirth/10,000 (95 % CI)
Cumulative probabilitya of stillbirth/10,000 (95% CI)
550
28 (18.6–40.4)
28 (18.6–40.0)
Referent
Live SGA births
Risk ratio (95% CI)
Ongoing SGA pregnancies
SGA stillbirths
37
3333
7
38
2776
3
820
13 (6.9–22.2)
41 (29.4–55.6)
1.46 (0.9–2.4)
39
1953
5
1116
36 (25.2–49.8)
77 (60.8–96.1)
2.75 (1.8–4.2)
ⱖ40
832
5
827
120 (99.6–143.3)
194 (168.9–222.0)
6.90 (4.7–10.3)
GA, wk
................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ a
Corrected for censoring.
Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
MAY 2013 American Journal of Obstetrics & Gynecology
376.e4
Research
Obstetrics
www.AJOG.org
FIGURE 2
Cumulative probability of SGA stillbirths over time
This graph displays the rise in the stillbirth risk for the SGA fetus as pregnancy progresses beyond 37 weeks. The risk of stillbirth is reported as SGA stillbirths/10,000 ongoing SGA pregnancies (y-axis). SGA, small-for-gestational-age. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
tant data for the clinical management of SGA pregnancies. The significance of the study findings in contemporary perspective is best appreciated when the risk estimates for stillbirth are compared with the risk of adverse neonatal outcomes resulting from early term delivery. For example, the highest odds ratios reported by Tita et al11 for adverse neonatal outcomes in
elective early term deliveries was 4.2, the odds ratio for neonatal respiratory distress syndrome at 37 weeks (95% CI, 1.9⫺3.3).11 Our investigation of neonatal outcomes among the growth restricted cohort demonstrates the risks of respiratory distress is reduced in the later term period compared with 37 weeks; however, these risks are relatively low at 127/10,000 at 37 weeks and this did not
translate into a significant difference in the length of hospital stay between 37 and 38 weeks. Importantly, although it is difficult to measure in conventional mathematic terms, the gravity of RDS is minor compared with a single stillbirth. On the other hand, the risk of meconium aspiration syndrome in this population increased significantly after 40 weeks. The timing of delivery for pregnancies complicated by SGA continues to be an important controversial clinical question. In 2010, the induction versus expectant monitoring for intrauterine growth restriction at term or DIGITAT trial5 demonstrated no increased maternal or fetal risk in the induction group compared with the expectant monitoring group, and the authors concluded that either induction or expectant monitoring were acceptable management strategies for presumed growth restriction at term. The follow-up data to DIGITAT was recently published demonstrating no difference in neurodevelopmental outcomes among children that were induced vs those that were expectantly managed.12 The information gained from this well-designed randomized controlled trial is important; however, because of time and cost constraints innate to any prospective trial, DIGITAT was not powered to examine the risk of stillbirth.5 In 2012, Rosenstien et al9 estimated the risk of stillbirth and infant mortality stratified by gestational age using California vital statistics to assess the risk of expectant management at term from 1997-2006. By using state vital statistics,
TABLE 5
Life-table analysis SGA <5%: risk of stillbirth/10,000 ongoing SGA pregnancies per week of gestation beyond 37 weeks Conditional probabilitya of stillbirth/10,000
Cumulative probabilitya of stillbirth/10,000
Ongoing SGA pregnancies <5%
SGA stillbirths <5%
Live SGA births
37
1422
3
238
23
23
38
1181
3
383
30
53
39
795
3
454
53
106
4.6 (2.9-7.2)
ⱖ40
338
3
335
176
280
12.1 (8.0-19.0)
GA, wk
Risk ratio (95% CI) Referent
................................................................................................................................................................................................................................................................................................................................................................................
2.3 (1.4-3.8)
................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................
CI, confidence interval; GA, gestational age; SGA, small-for-gestational-age. a
Corrected for censoring.
Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
376.e5
American Journal of Obstetrics & Gynecology MAY 2013
Obstetrics
www.AJOG.org
Research
TABLE 6
Neonatal outcomes for SGA neonates born during each week of gestation beyond 37 weeksa
GA, wk
NICU admit/10,000 live SGA births
RR (95% CI)
RDSb/10,000 live SGA birsth 127
RR (95% CI)
MASc/10,000 live SGA births
RR (95% CI)
LOSd >5 days/10,000 live SGA births
RR (95% CI)
37
200
Ref
Ref
18
Ref
223
Ref
38
146
0.73 (0.6–0.9)
85
0.67 (0.5–0.9)
12
0.71 (0.6–0.9)
159
0.77 (0.6–1.0)
39
81
0.40 (0.3–0.5)
18
0.14 (0.1–0.2)
18
1.0 (0.5–1.9)
45
0.2 (0.1–0.3)
ⱖ40
85
0.4 (0.3–0.6)
12
0.67 (0.5–0.9)
36
2.0 (1.1–3.5)
121
0.5 (0.4–0.7)
................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................
CI, confidence interval; GA, gestational age; LOS, length of stay; MAS, meconium aspiration syndrome; NICU, neonatal intensive care unit; RDS, respiratory distress syndrome; Ref, reference; RR, relative risk; SGA, small-for-gestational-age. a
SGA defined as birthweight ⬍10%, excludes postnatally diagnosed aneuploidy and or major anomalies; b Respiratory distress syndrome; c Meconium aspiration syndrome; d Length of stay.
Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
the authors reported on over 3 million nonanomalous pregnancies including 3999 stillbirths. The stillbirth rate per week of gestation was added to the infant mortality rate in the subsequent week to estimate the risk of expectant management per week of gestation. The authors found the risk of expectant management outweighed the risk of delivery for weeks 39, 40, and 41.9 This is consistent with the publication by Smith7 which used life-table analysis. Most recently, the risk of stillbirth in SGA pregnancies was estimated per week of gestation by Pilliod et al10 using the same administrative data used by Rosenstein and colleagues.9 The authors compared the risk of stillbirth between growth centiles ⬍3rd, ⬍5th, and ⬍10th. Consistent with previous studies, the authors showed the risk of stillbirth in SGA pregnancies advances with advancing gestational age. The risk of stillbirth was also found to be substantially greater among the smaller growth centiles 3rd and 5th compared with the 10th. Given these results, the authors suggested using different SGA thresholds for clinical decision making. In contrast to our study, the authors did not address the cumulative risk of stillbirth with advancing gestational age. In addition, the finding of increased risk of stillbirth with decreasing SGA centiles is important as it demonstrates a dose response effect with worsening SGA. We also demonstrated this dose response relationship in our secondary analysis examining SGA ⬍5%; however, because of limited sam-
ple size, we could not further stratify the cumulative risk of stillbirth by lower centiles of SGA. Our study offers several strengths. The large, validated database containing individual patient level data was a significant advantage for our analysis. Vital statistics data, as used by previous studies,9,10 have the limitations of relying on birth certificate information and missing or incomplete documentation of patient-level data. In comparison, our database is maintained by trained research nurses; it is relatively complete and contains detailed maternal, obstetric and follow-up information. In addition, given that stillbirth is a rare outcome, our study was able to evaluate a large stillbirth cohort that gave us the ability to stratify the risk of stillbirth by gestational age and assess the cumulative risk of stillbirth as gestational age progresses. Further, we used rigorous approaches in our investigation of stillbirth risk. Our initial method of calculating risk of stillbirth was performed to estimate the simplest conditional risk of stillbirth per week of gestation. We used the contemporary, now widely accepted method of calculating stillbirth risk as stillbirths per ongoing pregnancies. In addition, we also performed a life-table analysis,7 in which we were able to account for censoring and calculate cumulative risk of stillbirth. To further strengthen our conclusion, we were able to examine the risk of stillbirth for SGA ⬍5% and demonstrate a significant increase in the cumulative risk of
stillbirth at 38 weeks compared with 37 weeks. Our study is not without limitations, including the use of birthweight to determine SGA status. This is of particular concern when investigating stillbirths as the timing of the stillbirth is usually unknown and birthweight may not correlate with gestational age at delivery, which may, in part, explain the significant difference in birthweight between SGA stillbirths and live births demonstrated in Table 2. This limitation is partly mitigated using the censoring approach described above. In addition, the use of estimated fetal weight is also complicated by the known inaccuracies of ultrasound13 and timing of ultrasound examination that is often remote from delivery. Finally, the use of birthweight allows us to examine the direct relationship between small fetuses and stillbirth whereas the estimated fetal weight examines the accuracy of ultrasound to predict fetal weight and is a different question altogether. It is only after we have tested the direct association between stillbirth and SGA as we have in this study that improvement of our technical ability to predict SGA by ultrasound becomes relevant. Our study was also limited by our inability to report on Doppler velocimetry assessment in association with fetal growth restriction given that many of the subjects did not receive sonographic assessment of fetal weight after the second trimester. Another important consideration for a study of this nature that spans a long period is whether changes in clinical
MAY 2013 American Journal of Obstetrics & Gynecology
376.e6
Research
Obstetrics
management affected the risk of stillbirth. However, sensitivity analysis demonstrated no difference in stillbirth risk for SGA pregnancies ⬎37 weeks in 1990-2000 compared with 2001-2009. The study was, however, underpowered to compare the risk of stillbirth per week for each gestational age after 37 weeks for these 2 periods. In conclusion, our data demonstrates a significant increase in the cumulative probability of stillbirth in SGA pregnancies after 37 weeks. Optimal timing of delivery for SGA pregnancies requires a balance of the risks of both delivery and of continued gestation. Without a large randomized controlled trial powered to examine the risk of stillbirth, given our findings, we advocate a policy of delivery of SGA pregnancies at 37-38 weeks as the strategy that minimizes the risk of stillbirth. f
376.e7
www.AJOG.org REFERENCES 1. Cousens S, Blencowe H, Stanton C, et al. National, regional, and worldwide estimates of stillbirth rates in 2009 with trends since 1995: a systematic analysis. Lancet 2011;377:1319-30. 2. Flenady V, Middleton P, Smith GC, et al. Stillbirths: the way forward in high-income countries. Lancet 2011;377:1703-17. 3. Mullan Z, Horton R. Bringing stillbirths out of the shadows. Lancet 2011;377:1291-2. 4. Flenady V, Koopmans L, Middleton P, et al. Major risk factors for stillbirth in high-income countries: a systematic review and meta-analysis. Lancet 2011;377:1331-40. 5. Boers KE, Vijgen SM, Bijlenga D, et al. Induction versus expectant monitoring for intrauterine growth restriction at term: randomised equivalence trial (DIGITAT). BMJ 2010;341:c7087. 6. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol 1996;87:163-8. 7. Smith GC. Life-table analysis of the risk of perinatal death at term and post term in singleton pregnancies. Am J Obstet Gynecol 2001;184: 489-96.
American Journal of Obstetrics & Gynecology MAY 2013
8. Smith GC. Estimating risks of perinatal death. Am J Obstet Gynecol 2005;192:17-22. 9. Rosenstein MG, Cheng YW, Snowden JM, Nicholson JM, Caughey AB. Risk of stillbirth and infant death stratified by gestational age. Obstet Gynecol 2012;120:76-82. 10. Pilliod RA, Cheng YW, Snowden JM, Doss AE, Caughey AB. The risk of intrauterine fetal death in the small-for-gestational-age fetus. Am J Obstet Gynecol 2012;207:039. 11. Tita AT, Landon MB, Spong CY, et al. Timing of elective repeat cesarean delivery at term and neonatal outcomes. N Engl J Med 2009; 360:111-20. 12. Van Wyk L, Boers KE, Van Der Post JA, et al. Effects on (neuro)developmental and behavioral outcome at 2 years of age of induced labor compared with expectant management in intrauterine growth-restricted infants: long-term outcomes of the DIGITAT trial. Am J Obstet Gynecol 2012;206:22. 13. Dudley NJ. A systematic review of the ultrasound estimation of fetal weight. Ultrasound Obstet Gynecol 2005;25:80-9.
www. AJOG.org
OBSTETRICS
Risk of stillbirth after 37 weeks in pregnancies complicated by small-for-gestational-age fetuses Amanda S. Trudell, DO; Alison G. Cahill, MD, MSCI; Methodius G. Tuuli, MD, MPH; George A. Macones, MD, MSCE; Anthony O. Odibo, MD, MSCE OBJECTIVE: The evidence for delivering small-for-gestational-age
(SGA) fetuses at 37 weeks remains conflicting. We examined the risk of stillbirth per week of gestation beyond 37 weeks for pregnancies complicated by SGA. STUDY DESIGN: Singleton pregnancies undergoing routine second tri-
mester ultrasound from 1990-2009 were examined retrospectively. The risk of stillbirth per 10,000 ongoing SGA pregnancies with 95% confidence intervals (CIs) was calculated for each week of gestation ⱖ37 weeks. Using a life-table analysis with correction for censoring, conditional risks of stillbirth, cumulative risks of stillbirth per 10,000 ongoing SGA pregnancies and relative risks (RRs) were calculated with 95% CIs for each week of gestation. RESULTS: Among 57,195 pregnancies meeting inclusion criteria the
background risk of stillbirth was 56/10,000 (95% CI, 42.3⫺72.7) with stillbirth risk for SGA pregnancies of 251/10,000 (95% CI,
221.2⫺284.5). The risk of stillbirth after the 37th week was greater compared with pregnancies delivered in the 37th week (47/10,000, 95% CI, 34.6⫺62.5 vs 21/10,000, 95% CI, 13.0⫺32.1; RR, 2.2; 95% CI, 1.3⫺3.7). The cumulative risk of stillbirth rose from 28/ 10,000 ongoing SGA pregnancies at 37 weeks to 77/10,000 at 39 weeks (RR, 2.75; 95% CI, 1.79⫺4.2). Among pregnancies complicated by SGA ⬍5% the cumulative risk of stillbirth at 38 weeks was significantly greater than the risk at 37 weeks (RR, 2.3; 95% CI, 1.4⫺3.8). CONCLUSION: There is a significantly increased risk of stillbirth in preg-
nancies complicated by SGA delivered after the 37th week. Given these findings, we advocate a policy of delivery of SGA pregnancies 37-38 weeks. Key words: SGA, small for gestational age, stillbirth
Cite this article as: Trudell AS, Cahill AG, Tuuli MG, et al. Risk of stillbirth after 37 weeks in pregnancies complicated by small-for-gestational-age fetuses. Am J Obstet Gynecol 2013;208:376.e1-7.
N
early 3 million stillbirths occur worldwide each year.1 In high-income countries stillbirth rates have remained constant in recent decades with significant global variations in stillbirth rates among affluent nations, suggesting further reduction in stillbirth rates is possible.2 Fetal growth restriction has been established as a major cause of stillbirth.3 In a metaanalysis published in From the Department of Obstetrics and Gynecology, Washington University School of Medicine, St. Louis, MO. Received Nov. 9, 2012; revised Jan. 27, 2013; accepted Feb. 18, 2013. The authors report no conflict of interest. Presented in an oral format at the 33nd annual meeting of the Society of Maternal Fetal Medicine, San Francisco, CA, Feb. 11-16, 2013. The racing flag logo above indicates that this article was rushed to press for the benefit of the scientific community. Reprints not available from the authors. 0002-9378/$36.00 © 2013 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2013.02.030
376.e1
2011, the risk of stillbirth in high-income countries was demonstrated to be 4times higher in fetuses measuring smallfor-gestational-age (SGA) compared with non-SGA fetuses and SGA was noted to have the greatest populationattributable risk compared to other risks of stillbirth.4 The delivery timing of SGA pregnancies weighs competing risks. Early delivery avoids stillbirth but increases risks of respiratory distress syndrome, hypoglycemia, and neonatal sepsis while continued gestation risks stillbirth. Recent clinical trials with long-term follow-up have sought to clarify optimal timing of delivery for SGA fetuses. A recent randomized controlled trial evaluated expectant management versus induction of labor of SGA fetuses at term,5 but because of the rarity of stillbirth, the trial was not powered to study this important outcome. Given the impact of stillbirth both globally and nationally, defining the risk of stillbirth after 37 weeks is fundamental to answering the question of optimal timing of delivery for pregnancies com-
American Journal of Obstetrics & Gynecology MAY 2013
plicated by SGA, and ultimately, an important step in stillbirth prevention. The specific aim of our study was to estimate the risk of stillbirth for each week of gestation beyond 37 weeks in pregnancies complicated by SGA.
M ATERIALS AND M ETHODS We conducted a retrospective cohort study within our prospectively collected perinatal database at Washington University School of Medicine. Before initiation of the study, we obtained approval from the institutional review board. Sociodemographic, obstetric, medical and pregnancy follow-up data were obtained via self-report questionnaires at the initiation of prenatal care, at the time of ultrasound and from the hospital medical record on delivery. Detailed patient information was entered into the database by trained research nurses. Patients and/or primary obstetricians were contacted via telephone to obtain complete follow-up information if necessary, or in the rare event when the patient delivered at an outside facility. The study included
Obstetrics
www.AJOG.org
FIGURE 1
Flow diagram detailing exclusion criteria
This flow diagram details the study population and the exclusion criteria used to identify singleton pregnancies without prenatally diagnosed chromosomal disorders or fetal anomalies. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
all singleton pregnancies seen for anatomy survey between 16 and 23 weeks from January 1990 to December 2009. Approximately 97% of our study population are seen during this period. We excluded pregnancies complicated by prenatally diagnosed major congenital anomalies and aneuploidy and those without documented birthweight (Figure 1). Ultrasounds were performed by certified, dedicated obstetric sonographers and final ultrasound interpretation and diagnosis made after review by a fellowship trained maternal fetal medicine specialist. Gestational age was assigned by last menstrual period (LMP) and confirmed with ultrasound. Gestational age was assigned by ultrasound when the LMP was noted to be discrepant by ⱖ7 days in the first trimester or ⱖ10 days in the second trimester. If LMP was unknown, gestational age was assigned by ultrasound. The primary outcome was stillbirth, defined as intrauterine fetal death at or beyond 20 weeks’ gestation. SGA was defined as birthweight ⬍10th percentile
based on the Alexander growth standard.6 The incidence of stillbirth per week of gestation was calculated as the number of SGA stillbirths per 10,000 ongoing SGA pregnancies at the start of that gestational week. The choice of ongoing pregnancies as the denominator is a logical approach because compared with the traditional stillbirth risk assessment that uses the number of deliveries occurring in a given week as the denominator, ongoing pregnancies are the pregnancies that continue to be at risk of stillbirth. Thus, this is the contemporary method of calculating stillbirth risk. 7-9 Although the number of stillbirths/ 10,000 ongoing pregnancies will assess the risk of stillbirth per week of gestation, it does not assess the cumulative risk of stillbirth. The cumulative risk of stillbirth becomes more clinically relevant than the conditional risk of stillbirth when one is trying to decide between a recommendation for delivery or expectant management. Furthermore, the risk of stillbirth/10,000 ongoing pregnancies includes in the denominator the pregnancies that might deliver live births
Research
during the proposed week. Therefore, including these pregnancies in the denominator will tend to under estimate the risk of stillbirth. This effect of censoring, while not significant in the conditional probability estimate of stillbirth, is magnified and may become clinically significant when one calculates cumulative probability. Therefore, to calculate a cumulative risk of stillbirth beyond 37 weeks and to account for censoring, we further analyzed our data using the method of life-table analysis as proposed by Smith in 2001.7 For the life-table analysis, the estimation of stillbirth risk was calculated as follows: if Pn is the number of ongoing pregnancies at gestational week n, Bn is the number of births during gestational week n and Sn is the number of stillbirths during gestational week n then the conditional probability of stillbirths PSn at gestational week n: PSn ⫽ Sn/[Pn – (0.5 ⫻ Bn)], and the cumulative probability of stillbirths CSn is calculated as the product of the conditional probabilities of survival, where survival is 1 ⫺ Probability of death, at gestational week n: CSn ⫽ 1 ⫺ [(1 ⫺ PS37) ⫻ (1 ⫺ PS38) . . . ⫻ (1 ⫺ PSn)]. Descriptive statistics were used to provide an overview of the study population and estimate the baseline risk of stillbirth. Maternal demographic, medical and obstetric information were described among SGA pregnancies ⬍37 weeks, SGA pregnancies ⱖ37 weeks, and non-SGA pregnancies. To provide a more detailed overview of the SGA population, we also compared maternal demographics and obstetric information between SGA stillbirths and SGA live births. Descriptive statistics were performed using 2 for categorical variables and student t test for continuous variables. The risk of stillbirth with 95% confidence interval (CI) as calculated for each week of gestation 37-37 6/7, 38-38 6/7, 39-39 6/7, and ⱖ40 weeks. The relative risk (RR) of stillbirth with 95% CI was calculated for SGA pregnancies delivered after the 37th week compared with the risk of stillbirth in pregnancies delivered in the 37th week. Conditional risks and cumulative risks of stillbirth per week of gestation with 95% CIs were
MAY 2013 American Journal of Obstetrics & Gynecology
376.e2
Research
Obstetrics
calculated using life-table analysis with correction for censoring as stated above. Risk ratios for stillbirth with 95% CIs were calculated for each week of gestation relative to the risk of stillbirth at 37 weeks using the cumulative risks for stillbirth. To evaluate the cumulative risk of stillbirth among neonates with a lower threshold of SGA, the risk of SGA for neonates with birthweight ⬍5% was also determined by life-table analysis. To estimate if changes in clinical management or technology over the period of the study affected stillbirth risks, we calculated and compared risks of stillbirth among SGA fetuses ⱖ37 weeks’ gestation between 2 time periods determined by days from initiation of enrollment of 50% of the study population, approximately 1990-2000 and 2001-2009. The risk of neonatal morbidity in association with early term delivery was evaluated in a subanalysis for all neonates admitted to the neonatal intensive care unit or special care nurseries after delivery. Postnatally diagnosed aneuploidy and congenital anomalies were excluded. Adverse neonatal outcomes measured included intensive care unit admission, respiratory distress syndrome, meconium aspiration syndrome, and length of stay ⬎5 days. The risk of adverse outcomes at 38, 39, and ⱖ40 weeks relative to the risk at 37 weeks was calculated with 95% CI. Statistical analysis was performed with STATA 12 (StataCorp, College Station, TX).
www.AJOG.org
TABLE 1
Demographic characteristics of SGA and non-SGA pregnancies
Characteristic Maternal age, y (SD)
SGA delivered SGA delivered <37 wks <37 wks Non-SGA n ⴝ 884 (1.5%) n ⴝ 3333 (5.8%) n ⴝ 52,978 (92.6%) 29.3 ⫾ 6.6
29.0 ⫾ 7.0
30.4 ⫾ 6.2
..............................................................................................................................................................................................................................................
AMA
226 (25.6)
877 (26.3)
15,779 (29.8)
..............................................................................................................................................................................................................................................
Race
.....................................................................................................................................................................................................................................
Black
378 (42.8)
1253 (37.6)
11,370 (21.5)
White
384 (43.4)
1536 (46.1)
33,397 (63.0)
Other
122 (13.8)
544 (16.3)
8211 (15.5)
Nulliparous
450 (50.9)
1644 (49.3)
19,981 (37.7)
86 (9.7)
93 (2.8)
1248 (2.4)
327 (37.3)
340 (10.3)
3919 (7.5)
..................................................................................................................................................................................................................................... ..................................................................................................................................................................................................................................... .............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
Chronic hypertension
..............................................................................................................................................................................................................................................
Preeclampsia
..............................................................................................................................................................................................................................................
Pregestational diabetes
37 (4.2)
22 (0.7)
1051 (2.0)
Gestational diabetes
40 (4.6)
139 (4.2)
2841 (5.4)
Gestational age at delivery, wk
33.4 ⫾ 4.1
.............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
39.1 ⫾ 1.1
38.8 ⫾ 2.3
..............................................................................................................................................................................................................................................
Birthweight, g
1601 ⫾ 618
2659 ⫾ 238
3397 ⫾ 730
Stillbirth, n (%)
86 (9.7)
20 (0.6)
217 (0.4)
.............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
AMA, advanced maternal age; SD, standard deviation; SGA, small-for-gestational-age. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
39.1 ⫾ 1.1 weeks. Table 2 compares maternal demographic characteristics of SGA stillbirths delivered ⱖ37 weeks and SGA live births delivered ⱖ37 weeks. There was no significant difference in maternal age, race, parity, diabetes, or
hypertensive disorders, including preeclampsia, among women who delivered SGA stillbirths after 37 weeks and those who delivered SGA live births after 37 weeks. As expected, SGA stillbirths were significantly smaller at delivery than
TABLE 2
Demographic characteristics of SGA pregnancies delivered >37 weeks Characteristic
SGA stillbirth >37 (n ⴝ 20)
Maternal age (SD)
29.9 ⫾ 6.4
SGA live birth >37 (n ⴝ 3313) 29.0 ⫾ 7.0
P value .59
..............................................................................................................................................................................................................................................
R ESULTS
AMA, n (%)
Among 57,195 pregnancies meeting inclusion criteria, the background risk of stillbirth was 56/10,000 (95% CI, 42.3⫺72.7). SGA complicated 4217 (7.4%) pregnancies, of which 3333 (5.8%) delivered ⱖ37 weeks. Table 1 demonstrates relevant demographic information of the study cohort including the SGA pregnancies delivered ⬍37 weeks, the SGA pregnancies delivered ⱖ37 weeks and the non-SGA population. The primary concern of this investigation was the SGA group delivered ⱖ37 weeks of which the average gestational age at the time of delivery was
Race
376.e3
5 (25)
872 (26)
.89
.............................................................................................................................................................................................................................................. .....................................................................................................................................................................................................................................
Black, n (%)
8 (40)
1245 (38)
.83
White, n (%)
9 (45)
1527(46)
.92
Other, n (%)
3 (15)
541 (16)
.87
Nulliparous, n (%)
8 (40)
1636 (49)
.40
Chronic hypertension, n (%)
1 (5)
92 (2.8)
.55
Preeclampsia, n (%)
1 (5)
339 (10)
.43
Pregestational diabetes, n (%)
0 (0)
22 (0.6)
.71
Gestational diabetes, n (%)
2 (10)
..................................................................................................................................................................................................................................... ..................................................................................................................................................................................................................................... .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
137 (4)
.20
..............................................................................................................................................................................................................................................
Gestational age, wk (SD)
38.8 ⫾ 1.3
39.1 ⫾ 1.1
.27
..............................................................................................................................................................................................................................................
AMA, advanced maternal age; SD, standard deviation; SGA, small-for-gestational-age. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
American Journal of Obstetrics & Gynecology MAY 2013
Obstetrics
www.AJOG.org
TABLE 3
Risk of stillbirth per week of gestation Ongoing SGA pregnancies
SGA stillbirths (n ⴝ 20)
Stillbirth risk/10,000 ongoing SGA pregnancies (95% CI)
37
3333
7
21 (13.0–32.1)
38
2776
3
11 (5.5–19.7)
39
1953
5
26 (17.0–38.1)
ⱖ40
832
5
60 (45.8–77.2)
GA, wk
.............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. .............................................................................................................................................................................................................................................. ..............................................................................................................................................................................................................................................
CI, confidence interval; GA, gestational age; SGA, small-for-gestational-age. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
SGA live births (2370 ⫾ 314 vs 2596 ⫾ 237, P ⬍ .01). Among SGA pregnancies, 106 (2.5%), were complicated by stillbirth with a stillbirth risk of 251/10,000 (95% CI, 221.2⫺283.5). Table 3 demonstrates the risk of stillbirth per week of gestation. In the 37th week, the risk of stillbirth was 21/10,000 ongoing SGA pregnancies (95% CI, 13.0⫺32.1) compared with 47/ 10,000 (95% CI, 34.6⫺62.5) after the 37th week (RR, 2.2; 95% CI, 1.3⫺3.7). Using life-table analysis with correction for censoring, the cumulative risk of stillbirth was calculated for each week of gestation (Table 4). The risk of stillbirth in the 37th week was noted to be 28/ 10,000 (95% CI, 18.6⫺40.0). As expected, the cumulative risk of stillbirth rose with each advancing week of gestation to 41/10,000 (95% CI, 29.4⫺55.6) in the 38th week, 77/10,000 (95% CI, 60.8⫺96.1) in the 39th week and 194/ 10,000 (95% CI, 168.9⫺222.4) at ⱖ40 weeks. Figure 2 graphically illustrates a rise in cumulative risk of stillbirth for
each advancing week of gestation. The risk ratio for stillbirth for each week of gestation relative to 37 weeks is also demonstrated in Table 4. The cumulative risk of stillbirth at each advancing week of gestation relative to the cumulative risk of stillbirth at 37 weeks was significantly higher at 39 weeks (RR, 2.75; 95% CI, 1.8⫺4.2). At ⱖ40 weeks the risk of stillbirth rose nearly 7-fold compared to the risk of stillbirth at 37 weeks (95% CI, 4.7⫺10.3). The life-table analysis for SGA ⬍5% is located in Table 5. Of the 3333 SGA pregnancies delivered ⱖ37 weeks, 1422 (42.7%) were SGA ⬍5% with a total of 12 stillbirths for an overall stillbirth rate of 84/10,000. At this lower threshold of SGA the cumulative risk of stillbirth at 38 weeks was significantly greater compared with the risk of stillbirth at 37 weeks (RR, 2.3; 95% CI, 1.4⫺3.8). Adverse neonatal outcomes are demonstrated in Table 6. The risk of NICU admission, respiratory distress syndrome, and meconium aspiration syn-
Research
drome at 38 weeks was found to be lower relative to the risk at 37 weeks (RR, 0.73; 95% CI, 0.6⫺0.9; RR, 0.67; 95% CI, 0.5⫺0.9; RR, 0.71; 95% CI, 0.6⫺0.9), respectively. Length of stay ⬎5 days was not significantly different at 38 compared with 37 weeks (RR, 0.77; 95% CI, 0.6⫺1.0). At ⱖ40 weeks, the risk of meconium aspiration syndrome is greater than the risk at 37 weeks (RR, 2.0; 95% CI, 1.1⫺3.5). Sensitivity analysis showed no evidence that advancing technology and changes in clinical management over the 19-year study period affected the risk of stillbirth among SGA fetuses. Risks of SGA stillbirths after 37 weeks were not significantly different in the 2 halves of the study period (54/10,000 vs 67/10,000 P ⫽ .62, for 1990-2000 and 2001-2009, respectively).
C OMMENT Our study demonstrates a 2-fold risk of stillbirth after the 37th week of gestation compared with pregnancies delivered in the 37th week. The cumulative risk of stillbirth for each week of gestation beyond 37 weeks increased. At 39 weeks and ⱖ40 weeks, the stillbirth risk was nearly 3-fold and 7-fold. The relative risk of stillbirth at 38 weeks reached statistical significance when the SGA threshold was set at the 5th percentile. Although, the risk of stillbirth in SGA pregnancies has been evaluated,10 our investigation is unique in estimating the cumulative risk of stillbirth in SGA pregnancies stratified by gestational age. This provides impor-
TABLE 4
Life-table analysis SGA <10%: risk of stillbirth/10,000 ongoing SGA pregnancies per week of gestation beyond 37 weeks Conditional probabilitya of stillbirth/10,000 (95 % CI)
Cumulative probabilitya of stillbirth/10,000 (95% CI)
550
28 (18.6–40.4)
28 (18.6–40.0)
Referent
Live SGA births
Risk ratio (95% CI)
Ongoing SGA pregnancies
SGA stillbirths
37
3333
7
38
2776
3
820
13 (6.9–22.2)
41 (29.4–55.6)
1.46 (0.9–2.4)
39
1953
5
1116
36 (25.2–49.8)
77 (60.8–96.1)
2.75 (1.8–4.2)
ⱖ40
832
5
827
120 (99.6–143.3)
194 (168.9–222.0)
6.90 (4.7–10.3)
GA, wk
................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ a
Corrected for censoring.
Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
MAY 2013 American Journal of Obstetrics & Gynecology
376.e4
Research
Obstetrics
www.AJOG.org
FIGURE 2
Cumulative probability of SGA stillbirths over time
This graph displays the rise in the stillbirth risk for the SGA fetus as pregnancy progresses beyond 37 weeks. The risk of stillbirth is reported as SGA stillbirths/10,000 ongoing SGA pregnancies (y-axis). SGA, small-for-gestational-age. Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
tant data for the clinical management of SGA pregnancies. The significance of the study findings in contemporary perspective is best appreciated when the risk estimates for stillbirth are compared with the risk of adverse neonatal outcomes resulting from early term delivery. For example, the highest odds ratios reported by Tita et al11 for adverse neonatal outcomes in
elective early term deliveries was 4.2, the odds ratio for neonatal respiratory distress syndrome at 37 weeks (95% CI, 1.9⫺3.3).11 Our investigation of neonatal outcomes among the growth restricted cohort demonstrates the risks of respiratory distress is reduced in the later term period compared with 37 weeks; however, these risks are relatively low at 127/10,000 at 37 weeks and this did not
translate into a significant difference in the length of hospital stay between 37 and 38 weeks. Importantly, although it is difficult to measure in conventional mathematic terms, the gravity of RDS is minor compared with a single stillbirth. On the other hand, the risk of meconium aspiration syndrome in this population increased significantly after 40 weeks. The timing of delivery for pregnancies complicated by SGA continues to be an important controversial clinical question. In 2010, the induction versus expectant monitoring for intrauterine growth restriction at term or DIGITAT trial5 demonstrated no increased maternal or fetal risk in the induction group compared with the expectant monitoring group, and the authors concluded that either induction or expectant monitoring were acceptable management strategies for presumed growth restriction at term. The follow-up data to DIGITAT was recently published demonstrating no difference in neurodevelopmental outcomes among children that were induced vs those that were expectantly managed.12 The information gained from this well-designed randomized controlled trial is important; however, because of time and cost constraints innate to any prospective trial, DIGITAT was not powered to examine the risk of stillbirth.5 In 2012, Rosenstien et al9 estimated the risk of stillbirth and infant mortality stratified by gestational age using California vital statistics to assess the risk of expectant management at term from 1997-2006. By using state vital statistics,
TABLE 5
Life-table analysis SGA <5%: risk of stillbirth/10,000 ongoing SGA pregnancies per week of gestation beyond 37 weeks Conditional probabilitya of stillbirth/10,000
Cumulative probabilitya of stillbirth/10,000
Ongoing SGA pregnancies <5%
SGA stillbirths <5%
Live SGA births
37
1422
3
238
23
23
38
1181
3
383
30
53
39
795
3
454
53
106
4.6 (2.9-7.2)
ⱖ40
338
3
335
176
280
12.1 (8.0-19.0)
GA, wk
Risk ratio (95% CI) Referent
................................................................................................................................................................................................................................................................................................................................................................................
2.3 (1.4-3.8)
................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................
CI, confidence interval; GA, gestational age; SGA, small-for-gestational-age. a
Corrected for censoring.
Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
376.e5
American Journal of Obstetrics & Gynecology MAY 2013
Obstetrics
www.AJOG.org
Research
TABLE 6
Neonatal outcomes for SGA neonates born during each week of gestation beyond 37 weeksa
GA, wk
NICU admit/10,000 live SGA births
RR (95% CI)
RDSb/10,000 live SGA birsth 127
RR (95% CI)
MASc/10,000 live SGA births
RR (95% CI)
LOSd >5 days/10,000 live SGA births
RR (95% CI)
37
200
Ref
Ref
18
Ref
223
Ref
38
146
0.73 (0.6–0.9)
85
0.67 (0.5–0.9)
12
0.71 (0.6–0.9)
159
0.77 (0.6–1.0)
39
81
0.40 (0.3–0.5)
18
0.14 (0.1–0.2)
18
1.0 (0.5–1.9)
45
0.2 (0.1–0.3)
ⱖ40
85
0.4 (0.3–0.6)
12
0.67 (0.5–0.9)
36
2.0 (1.1–3.5)
121
0.5 (0.4–0.7)
................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................
CI, confidence interval; GA, gestational age; LOS, length of stay; MAS, meconium aspiration syndrome; NICU, neonatal intensive care unit; RDS, respiratory distress syndrome; Ref, reference; RR, relative risk; SGA, small-for-gestational-age. a
SGA defined as birthweight ⬍10%, excludes postnatally diagnosed aneuploidy and or major anomalies; b Respiratory distress syndrome; c Meconium aspiration syndrome; d Length of stay.
Trudell. SGA and Stillbirth. Am J Obstet Gynecol 2013.
the authors reported on over 3 million nonanomalous pregnancies including 3999 stillbirths. The stillbirth rate per week of gestation was added to the infant mortality rate in the subsequent week to estimate the risk of expectant management per week of gestation. The authors found the risk of expectant management outweighed the risk of delivery for weeks 39, 40, and 41.9 This is consistent with the publication by Smith7 which used life-table analysis. Most recently, the risk of stillbirth in SGA pregnancies was estimated per week of gestation by Pilliod et al10 using the same administrative data used by Rosenstein and colleagues.9 The authors compared the risk of stillbirth between growth centiles ⬍3rd, ⬍5th, and ⬍10th. Consistent with previous studies, the authors showed the risk of stillbirth in SGA pregnancies advances with advancing gestational age. The risk of stillbirth was also found to be substantially greater among the smaller growth centiles 3rd and 5th compared with the 10th. Given these results, the authors suggested using different SGA thresholds for clinical decision making. In contrast to our study, the authors did not address the cumulative risk of stillbirth with advancing gestational age. In addition, the finding of increased risk of stillbirth with decreasing SGA centiles is important as it demonstrates a dose response effect with worsening SGA. We also demonstrated this dose response relationship in our secondary analysis examining SGA ⬍5%; however, because of limited sam-
ple size, we could not further stratify the cumulative risk of stillbirth by lower centiles of SGA. Our study offers several strengths. The large, validated database containing individual patient level data was a significant advantage for our analysis. Vital statistics data, as used by previous studies,9,10 have the limitations of relying on birth certificate information and missing or incomplete documentation of patient-level data. In comparison, our database is maintained by trained research nurses; it is relatively complete and contains detailed maternal, obstetric and follow-up information. In addition, given that stillbirth is a rare outcome, our study was able to evaluate a large stillbirth cohort that gave us the ability to stratify the risk of stillbirth by gestational age and assess the cumulative risk of stillbirth as gestational age progresses. Further, we used rigorous approaches in our investigation of stillbirth risk. Our initial method of calculating risk of stillbirth was performed to estimate the simplest conditional risk of stillbirth per week of gestation. We used the contemporary, now widely accepted method of calculating stillbirth risk as stillbirths per ongoing pregnancies. In addition, we also performed a life-table analysis,7 in which we were able to account for censoring and calculate cumulative risk of stillbirth. To further strengthen our conclusion, we were able to examine the risk of stillbirth for SGA ⬍5% and demonstrate a significant increase in the cumulative risk of
stillbirth at 38 weeks compared with 37 weeks. Our study is not without limitations, including the use of birthweight to determine SGA status. This is of particular concern when investigating stillbirths as the timing of the stillbirth is usually unknown and birthweight may not correlate with gestational age at delivery, which may, in part, explain the significant difference in birthweight between SGA stillbirths and live births demonstrated in Table 2. This limitation is partly mitigated using the censoring approach described above. In addition, the use of estimated fetal weight is also complicated by the known inaccuracies of ultrasound13 and timing of ultrasound examination that is often remote from delivery. Finally, the use of birthweight allows us to examine the direct relationship between small fetuses and stillbirth whereas the estimated fetal weight examines the accuracy of ultrasound to predict fetal weight and is a different question altogether. It is only after we have tested the direct association between stillbirth and SGA as we have in this study that improvement of our technical ability to predict SGA by ultrasound becomes relevant. Our study was also limited by our inability to report on Doppler velocimetry assessment in association with fetal growth restriction given that many of the subjects did not receive sonographic assessment of fetal weight after the second trimester. Another important consideration for a study of this nature that spans a long period is whether changes in clinical
MAY 2013 American Journal of Obstetrics & Gynecology
376.e6
Research
Obstetrics
management affected the risk of stillbirth. However, sensitivity analysis demonstrated no difference in stillbirth risk for SGA pregnancies ⬎37 weeks in 1990-2000 compared with 2001-2009. The study was, however, underpowered to compare the risk of stillbirth per week for each gestational age after 37 weeks for these 2 periods. In conclusion, our data demonstrates a significant increase in the cumulative probability of stillbirth in SGA pregnancies after 37 weeks. Optimal timing of delivery for SGA pregnancies requires a balance of the risks of both delivery and of continued gestation. Without a large randomized controlled trial powered to examine the risk of stillbirth, given our findings, we advocate a policy of delivery of SGA pregnancies at 37-38 weeks as the strategy that minimizes the risk of stillbirth. f
376.e7
www.AJOG.org REFERENCES 1. Cousens S, Blencowe H, Stanton C, et al. National, regional, and worldwide estimates of stillbirth rates in 2009 with trends since 1995: a systematic analysis. Lancet 2011;377:1319-30. 2. Flenady V, Middleton P, Smith GC, et al. Stillbirths: the way forward in high-income countries. Lancet 2011;377:1703-17. 3. Mullan Z, Horton R. Bringing stillbirths out of the shadows. Lancet 2011;377:1291-2. 4. Flenady V, Koopmans L, Middleton P, et al. Major risk factors for stillbirth in high-income countries: a systematic review and meta-analysis. Lancet 2011;377:1331-40. 5. Boers KE, Vijgen SM, Bijlenga D, et al. Induction versus expectant monitoring for intrauterine growth restriction at term: randomised equivalence trial (DIGITAT). BMJ 2010;341:c7087. 6. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol 1996;87:163-8. 7. Smith GC. Life-table analysis of the risk of perinatal death at term and post term in singleton pregnancies. Am J Obstet Gynecol 2001;184: 489-96.
American Journal of Obstetrics & Gynecology MAY 2013
8. Smith GC. Estimating risks of perinatal death. Am J Obstet Gynecol 2005;192:17-22. 9. Rosenstein MG, Cheng YW, Snowden JM, Nicholson JM, Caughey AB. Risk of stillbirth and infant death stratified by gestational age. Obstet Gynecol 2012;120:76-82. 10. Pilliod RA, Cheng YW, Snowden JM, Doss AE, Caughey AB. The risk of intrauterine fetal death in the small-for-gestational-age fetus. Am J Obstet Gynecol 2012;207:039. 11. Tita AT, Landon MB, Spong CY, et al. Timing of elective repeat cesarean delivery at term and neonatal outcomes. N Engl J Med 2009; 360:111-20. 12. Van Wyk L, Boers KE, Van Der Post JA, et al. Effects on (neuro)developmental and behavioral outcome at 2 years of age of induced labor compared with expectant management in intrauterine growth-restricted infants: long-term outcomes of the DIGITAT trial. Am J Obstet Gynecol 2012;206:22. 13. Dudley NJ. A systematic review of the ultrasound estimation of fetal weight. Ultrasound Obstet Gynecol 2005;25:80-9.