Advertisement

Optimizing the definition of intrauterine growth restriction: the multicenter prospective PORTO Study

      Objective

      The objective of the Prospective Observational Trial to Optimize Pediatric Health in Intrauterine Growth Restriction (IUGR) (PORTO Study), a national prospective observational multicenter study, was to evaluate which sonographic findings were associated with perinatal morbidity and mortality in pregnancies affected by growth restriction, originally defined as estimated fetal weight (EFW) <10th centile.

      Study Design

      Over 1100 consecutive ultrasound-dated singleton pregnancies with EFW <10th centile were recruited from January 2010 through June 2012. A range of IUGR definitions were used, including EFW or abdominal circumference <10th, <5th, or <3rd centiles, with or without oligohydramnios and with or without abnormal umbilical arterial Doppler (pulsatility index >95th centile, absent or reversed end-diastolic flow). Adverse perinatal outcome, defined as a composite outcome of intraventricular hemorrhage, periventricular leukomalacia, hypoxic ischemic encephalopathy, necrotizing enterocolitis, bronchopulmonary dysplasia, sepsis, and death was documented for all cases.

      Results

      Of 1116 fetuses, 312 (28%) were admitted to neonatal intensive care unit and 58 (5.2%) were affected by adverse perinatal outcome including 8 mortalities (0.7%). The presence of abnormal umbilical Doppler was significantly associated with adverse outcome, irrespective of EFW or abdominal circumference measurement. The only sonographic weight-related definition consistently associated with adverse outcome was EFW <3rd centile (P = .0131); all mortalities had EFW <3rd centile. Presence of oligohydramnios was clinically important when combined with EFW <3rd centile (P = .0066).

      Conclusion

      Abnormal umbilical artery Doppler and EFW <3rd centile were strongly and most consistently associated with adverse perinatal outcome. Our data call into question the current definitions of IUGR used. Future studies may address whether using stricter IUGR cutoffs comparing various definitions and management strategies has implications on resource allocation and pregnancy outcome.

      Key words

      Inadequate fetal growth affects up to 10% of all pregnancies. While the majority of such pregnancies will have a physiologically normal fetus that is simply small for gestational age (SGA), the ability to differentiate such a fetus from the pathologically growth-restricted fetus is limited. Pregnancies affected by intrauterine growth restriction (IUGR) pose a major public health problem and are associated with increased neonatal morbidity and mortality.
      • Badawi N.
      • Kurinczuk J.J.
      • Keogh J.M.
      • et al.
      Antepartum risk factors for newborn encephalopathy: the Western Australian case-control study.
      In addition, through fetal programming, adverse intrauterine environment may increase disease risk in adulthood leading to hypertension, diabetes, coronary heart disease, and stroke.
      • Barker D.J.
      Adult consequences of fetal growth restriction.
      The most commonly adopted definition to describe a fetus that has not reached its target weight based on sonographic estimated fetal weight (EFW) for a certain gestation is an abdominal circumference (AC) or EFW measurement <10th centile. The American Congress of Obstetricians and Gynecologists (ACOG)
      American College of Obstetricians and Gynecologists
      Intrauterine growth restriction.
      and Royal College of Obstetricians and Gynecologists (RCOG)
      Royal College of Obstetricians and Gynecologists
      The investigation and management of the small-for-gestational-age fetus (guideline no. 31).
      agree that at this cutoff the risk of perinatal morbidity and mortality increases. ACOG considers amniotic fluid an “important diagnostic and prognostic parameter in fetuses with IUGR,” whereas the RCOG notes that amniotic fluid assessment has “minimal value in diagnosing” growth restriction. Both guidelines agree that umbilical artery (UA) Doppler is not a reliable screening technique for IUGR, but is a useful assessment tool once IUGR is diagnosed. The Society of Obstetricians and Gynecologists of Canada
      • Lausman A.
      • McCarthy F.P.
      • Walker M.
      • Kingdom J.
      Screening, diagnosis, and management of intrauterine growth restriction.
      uses an EFW <10th centile for diagnosis of SGA and suggests that UA and uterine artery Doppler studies in combination with ultrasound of the placental morphology is useful to establish a more refined diagnosis of IUGR.
      The use of customized fetal growth charts
      • Gardosi J.
      • Chang A.
      • Kalyan B.
      • Sahota D.
      • Symonds E.M.
      Customized antenatal growth charts.
      has been proposed for more appropriate identification of fetal growth restriction taking into account anthropomorphic variables of the mother and fetus. However, these are not universally available or used in routine clinical practice in Ireland. The Hadlock formula
      • Hadlock F.P.
      • Harrist R.B.
      • Sharman R.S.
      • Deter R.L.
      • Park S.K.
      Estimation of fetal weight with the use of head, body, and femur measurements–a prospective study.
      is the most widely accepted method of estimating fetal weight using a composite sonographic measurement of fetal head, abdomen, and femur.
      Ireland has the highest birth rate in Europe with >72,000 singleton births per year.
      Perinatal Statistics Report, Economic and Social Research Institute (ESRI) on behalf of the Department of Health and the Health Service Executive Ireland, June 2010.
      The 7 academic obstetric centers that participated in the Prospective Observational Trial to Optimize Pediatric Health in IUGR (PORTO Study) deliver approximately 70% of these infants. In 2010, 5.1% of singletons were born with low birthweight, defined as <2500 g at >37 weeks' gestation, and the overall perinatal mortality rate corrected for fetal anomalies was 6.5 per 1000 births.
      Perinatal Statistics Report, Economic and Social Research Institute (ESRI) on behalf of the Department of Health and the Health Service Executive Ireland, June 2010.
      The objective of this study was to evaluate which sonographic parameters are associated with perinatal morbidity and mortality in pregnancies affected by IUGR, defined as EFW <10th centile, to establish more explicit criteria for diagnosis of IUGR. Our data link various sonographic findings with respective pregnancy outcomes. It challenges the reader to rethink the cutoffs used for IUGR definition. Further studies are needed to address optimal management and intervention strategies comparing these various IUGR definitions.

      Materials and Methods

      The PORTO Study is a multicenter prospective study conducted at 7 academic obstetric centers in Ireland. For the purpose of the study, IUGR was defined as EFW <10th centile based on sonographic measurements of fetal biparietal diameter, head circumference, AC, and femur length (Hadlock-4). The diagnosis was made by conventional population–based growth standards
      • Hadlock F.P.
      • Harrist R.B.
      • Sharman R.S.
      • Deter R.L.
      • Park S.K.
      Estimation of fetal weight with the use of head, body, and femur measurements–a prospective study.
      as customized centile charts were not available at the time of this analysis. From January 2010 through June 2012, the PORTO Study recruited 1200 consecutive ultrasound-dated singleton pregnancies. Inclusion criteria were a gestational age (GA) between 24 0/7 and 36 6/7 weeks' and an EFW ≥500 g. Fetuses with major structural and/or chromosomal abnormalities were excluded retrospectively from the final analysis. Institutional review board approval was obtained at each participating site, and all study participants gave written informed consent.
      Referral to the study occurred if small fetal size was suspected due to clinical evaluation in the antenatal setting. A PORTO Study research sonographer then confirmed that EFW was <10th centile, and performed a detailed sonogram of the fetal anatomy and uterine artery Doppler at enrollment. Baseline demographic data were recorded. All eligible pregnancies underwent serial sonographic evaluation of fetal weight at intervals of every 2 weeks until birth. All normally formed fetuses underwent evaluation of amniotic fluid volume; biophysical profile scoring; multivessel Doppler of UA, middle cerebral artery, ductus venosus, umbilical vein, and aortic isthmus; and myocardial performance index at every subsequent contact with the research sonographers. All prenatal and ultrasound data were contemporaneously transferred to an ultrasound software system (Viewpoint; MDI Viewpoint, Jacksonville, FL) and uploaded onto a live World Wide Web–based central consolidated database.
      In cases of absent end-diastolic flow (AEDF) or reversed end-diastolic flow in the UA, the patient was admitted to hospital and a daily computerized short-term variation cardiotocograph was carried out. Corticosteroids for fetal lung maturation were administered between 24-36 weeks' gestation if delivery was thought to be likely within 1 week. Delivery was prompted by fetal (eg, nonreassuring fetal testing) or maternal (eg, deterioration of preeclampsia, antepartum hemorrhage) reasons, however the ultimate management decisions relating to timing and mode of delivery were at the discretion of the lead clinician managing each case and were not prespecified by the study design. There was however general agreement among clinicians in Ireland to deliver AEDF cases by 34 weeks' gestation. Tertiary-level neonatal care facilities were available in all 7 sites.
      Pediatric outcomes for infants not requiring neonatal intensive care were recorded by the research sonographers and uploaded onto the database. Infants requiring neonatal intensive care unit (NICU) admission had their outcomes recorded by neonatology medical or nursing staff. Adverse perinatal outcome was defined as a composite outcome of intraventricular hemorrhage, periventricular leukomalacia, hypoxic ischemic encephalopathy, necrotizing enterocolitis, bronchopulmonary dysplasia, sepsis, and death. Given that all study sites were members of the Vermont Oxford Network,
      • Horbar J.D.
      The Vermont Oxford Network: evidence-based quality improvement for neonatology.
      definitions for intraventricular hemorrhage, periventricular leukomalacia, hypoxic ischemic encephalopathy, necrotizing enterocolitis, bronchopulmonary dysplasia, and sepsis were standardized across all centers derived from the Vermont Oxford Network manual. The outcomes were analyzed based on various IUGR definitions including EFW or AC <10th, <5th, or <3rd centiles, with or without associated oligohydramnios (defined as amniotic fluid volume of <8 cm or deepest vertical pocket <2 cm) and with or without abnormal UA Doppler (defined as pulsatility index >95th centile, AEDF or reversed end-diastolic flow).
      • Archaya G.
      • Wilsgaard T.
      • Berntsen G.K.
      • Maltau J.M.
      • Kiserud T.
      Reference ranges for serial measurements of umbilical artery Doppler indices in the second half of pregnancy.
      A sample size calculation was performed. Depending upon the baseline rates of outcomes considered, the study had sufficient power (80%) with a sample size of 1100 to show a risk reduction in 2 equally exposed groups (550 vs 550) of 8% for an outcome with a 30% overall baseline rate in our study population, and 2% for an outcome with a 5% overall baseline rate. For a subanalysis with a 2:1 study exposure partition of 1100 pregnancies (733 vs 367 pregnancies) the study would also determine an 8% and 4% reduction in baseline outcome rates of 30% and 5%, respectively.

      Use of statistics

      Prior to statistical analysis, all ultrasound and outcome data were screened for anomalous records or potential outliers and followed up by research sonographers for resolution. Statistical comparisons were performed using the χ2 test of association. Fisher exact test was used in cases where a table cell contained <5 observations. Demographics and baseline characteristics were presented using summary statistics (Table 1). Three centile cutoffs (3rd, 5th, and 10th) for EFW and AC, in addition to the presence of oligohydramnios or abnormal Doppler, were considered a priori as predictors of adverse perinatal outcomes, NICU admission, and perinatal mortality (TABLE 2, Table 3, Table 4). Predictors were considered over the course of a pregnancy, eg, EFW <3rd centile at any time after recruitment. Using a nominal 5% level of statistical significance, comparisons were made using a Bonferroni-corrected significance level of 0.3% (17 comparisons of predictors). The adjusted P values in TABLE 2, Table 3, Table 4 were derived from a multivariate analysis adjusting for GA at delivery. SAS version 9.2 (SAS Institute, Cary, NC) was used for data management and statistical analysis.
      TABLE 1Maternal demographics and fetal characteristics (n = 1116)
      CharacteristicValue
      Age, y30 ± 6
      White European ethnicity907 (83%)
      Spontaneous conception1100 (99%)
      Maternal height, cm162 ± 12
      Maternal weight at booking, kg64 ± 13
      BMI, kg/m224.1 ± 4.7
      Smokers261 (23%)
      Hypertensive disease/preeclampsia134 (12%)
      GA at enrollment, wk30.1 ± 3.9
      GA at delivery, wk37.8 ± 3.0
      Weight at delivery, g2495 ± 671
      NICU admission312 (28%)
      Apgar score <7513 (1%)
      Stillbirths4 (1:280)
      Neonatal deaths4 (1:280)
      Continuous variables are summarized with mean ± SD and categorical variables with n (%).
      BMI, body mass index; GA, gestational age; NICU, neonatal intensive care unit.
      Unterscheider. Optimizing the definition of IUGR. Am J Obstet Gynecol 2013.
      TABLE 2Predictors of adverse perinatal outcome
      Predictorn = 1116Adverse outcome (n = 58)Normal outcome (n = 1058)P valueAdjusted P value
      Derived from multivariate analysis adjusting for gestational age at delivery;
      EFW <3rd82651 (6.2%)775 (93.8%).0131.3409
      EFW <5th104755 (5.3%)992 (94.7%).7428.0894
      AC <3rd90952 (5.7%)857 (94.3%).0988.9027
      AC <5th99853 (5.3%)945 (94.7%).6194.2589
      AC <10th108056 (5.2%)1024 (94.8%).9215.1921
      EFW <3rd + oligohydramnios12713 (10.2%)114 (89.8%).0066.4623
      EFW <5th + oligohydramnios18913 (6.9%)176 (93.1%).2533.7184
      EFW <10th + oligohydramnios23115 (6.5%)216 (93.5%).3189.4362
      AC <3rd + oligohydramnios18014 (7.8%)166 (92.2%).0885.2251
      AC <5th + oligohydramnios20514 (6.8%)191 (93.2%).2439.3263
      AC <10th + oligohydramnios23114 (6.1%)217 (93.9%).5067.4508
      EFW <3rd + abnormal UA25142 (16.7%)209 (83.3%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0163
      EFW <5th + abnormal UA34645 (13.0%)301 (87.0%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0220
      EFW <10th + abnormal UA41347 (11.4%)366 (88.6%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0136
      AC <3rd + abnormal UA31542 (13.3%)273 (86.7%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0046
      AC <5th + abnormal UA35343 (12.2%)310 (87.8%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0108
      AC <10th + abnormal UA40046 (11.5%)354 (88.5%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0031
      Abnormal UA refers to absent or reversed end-diastolic flow in UA or pulsatility index >95th centile. Oligohydramnios refers to amniotic fluid index <8 cm or amniotic fluid deepest pool <2 cm. Adverse perinatal outcome was defined as adverse composite outcome of intraventricular hemorrhage, periventricular leukomalacia, hypoxic ischemic encephalopathy, necrotizing enterocolitis, bronchopulmonary dysplasia, sepsis, and death.
      AC, abdominal circumference; EFW, estimated fetal weight; UA, umbilical artery Doppler.
      Unterscheider. Optimizing the definition of IUGR. Am J Obstet Gynecol 2013.
      a Derived from multivariate analysis adjusting for gestational age at delivery;
      b Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      Table 3Predictors of admission to neonatal intensive care unit
      Predictorn = 1116NICU admission (n = 312)No admission (n = 804)P valueAdjusted P value
      Derived from multivariate analysis adjusting for gestational age at delivery;
      EFW <3rd826267 (32.3%)559 (67.7%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .1038
      EFW <5th1047300 (28.7%)747 (71.3%).0435.4764
      AC <3rd909275 (30.3%)634 (69.7%).0003
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .1890
      AC <5th998294 (29.5%)704 (70.5%).0011
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .1043
      AC <10th1080306 (28.3%)774 (71.7%).1249.9728
      EFW <3rd + oligohydramnios12766 (52.0%)61 (48.0%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0008
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      EFW <5th + oligohydramnios18973 (38.6%)116 (61.4%).0003
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0463
      EFW <10th + oligohydramnios23178 (33.8%)153 (66.2%).0272.2409
      AC <3rd + oligohydramnios18069 (38.3%)111 (61.7%).0007
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0165
      AC <5th + oligohydramnios20573 (35.6%)132 (64.4%).0069.0469
      AC <10th + oligohydramnios23175 (32.5%)156 (67.5%).0863.2534
      EFW <3rd + abnormal UA251147 (58.6%)104 (41.4%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      < .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      EFW <5th + abnormal UA346158 (45.7%)188 (54.3%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0974
      EFW <10th + abnormal UA413169 (40.9%)244 (59.1%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .3945
      AC <3rd + abnormal UA315149 (47.3%)166 (52.7%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0205
      AC <5th + abnormal UA353159 (45.0%)194 (55.0%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .0168
      AC <10th + abnormal UA400166 (41.5%)234 (58.5%)< .0001
      Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      .1294
      Abnormal UA refers to absent or reversed end-diastolic flow in UA or pulsatility index >95th centile. Oligohydramnios refers to amniotic fluid index <8 cm or amniotic fluid deepest pool <2 cm.
      AC, abdominal circumference; EFW, estimated fetal weight; NICU, neonatal intensive care unit; UA, umbilical artery Doppler.
      Unterscheider. Optimizing the definition of IUGR. Am J Obstet Gynecol 2013.
      a Derived from multivariate analysis adjusting for gestational age at delivery;
      b Statistically significant after adjusting for multiple comparisons (Bonferroni-adjusted significance level of 0.3%, P value < .003). Fisher exact test was used in cases where table cell contained <5 observations.
      Table 4Predictors of perinatal mortality
      Predictorn = 1116Perinatal mortality (n = 8)No mortality (n = 1108)P valueAdjusted P value
      Derived from multivariate analysis adjusting for gestational age at delivery.
      EFW <3rd8268 (1.0%)818 (99.0%).1209.9533
      EFW <5th10478 (0.8%)1039 (99.2%).4662.9751
      AC <3rd9098 (0.9%)901 (99.1%).3640.9684
      AC <5th9988 (0.8%)990 (99.2%).3290.9676
      AC <10th10808 (0.7%)1072 (99.3%).6043.9816
      EFW <3rd + oligohydramnios1273 (2.4%)124 (97.6%).0525.2249
      EFW <5th + oligohydramnios1893 (1.6%)186 (98.4%).1392.2612
      EFW <10th + oligohydramnios2314 (1.7%)227 (98.3%).0622.0660
      AC <3rd + oligohydramnios1803 (1.7%)177 (98.3%).1242.2168
      AC <5th + oligohydramnios2053 (1.5%)202 (98.5%).1677.2367
      AC <10th + oligohydramnios2313 (1.3%)228 (98.7%).3735.2579
      EFW <3rd + abnormal UA2516 (2.4%)245 (97.6%).0023.6429
      EFW <5th + abnormal UA3466 (1.7%)340 (98.3%).0123.9855
      EFW <10th + abnormal UA4136 (1.5%)407 (98.5%).0576.9180
      AC <3rd + abnormal UA3156 (1.9%)309 (98.1%).0079.4246
      AC <5th + abnormal UA3536 (1.7%)347 (98.3%).0146.5619
      AC <10th + abnormal Doppler4006 (1.5%)394 (98.5%).0282.7112
      Abnormal UA refers to absent or reversed end-diastolic flow in UA or pulsatility index >95th centile. Oligohydramnios refers to amniotic fluid index <8 cm or amniotic fluid deepest pool <2 cm.
      AC, abdominal circumference; EFW, estimated fetal weight; UA, umbilical artery Doppler.
      Unterscheider. Optimizing the definition of IUGR. Am J Obstet Gynecol 2013.
      a Derived from multivariate analysis adjusting for gestational age at delivery.

      Results

      Of 1200 recruited pregnancies with EFW <10th centile, 32 (2.7%) were excluded due to chromosomal and/or structural abnormalities, 13 (1%) withdrew their consent, 13 (1%) delivered outside Ireland, and 26 (2.2%) were lost to follow-up. This resulted in 1116 patients completing the study protocol.
      The mean maternal age was 30 years with 83% of mothers being of white European descent. This is consistent with the demographic profile of the overall obstetric population attending for antenatal care in Ireland
      Perinatal Statistics Report, Economic and Social Research Institute (ESRI) on behalf of the Department of Health and the Health Service Executive Ireland, June 2010.
      reflecting an unselected group of recruited pregnancies. The mean GA at enrollment to the study was 30.1 weeks and the mean GA at delivery was 37.8 weeks. Table 1 outlines maternal demographics and fetal characteristics.
      One in 20 infants was affected by the composite adverse perinatal outcome and 1 in 4 infants required admission to the NICU reflecting a truly at-risk population. Table 2 describes the predictors of adverse perinatal outcome. The majority of fetuses (72%, n = 800) with an EFW <10th centile had a normal perinatal outcome, meaning they neither required admission to NICU nor were they affected by morbidity or mortality. Infants with EFW <3rd centile had rates of composite adverse outcome of 6.2% (51/826) while the cohort with EFW in the 3rd-10th centiles had a rate of 2% (5/254). Mean infant birthweight was 2495 g. In all, 58 infants (5.2%) had an adverse perinatal outcome. Whenever abnormal UA Doppler was found there was a significant increased risk of adverse perinatal outcome. The strong statistical significance for abnormal Doppler velocimetries in Table 2 arises from the considerable reduction in false-positive results associated with abnormal Doppler. EFW or AC cutoffs <5th or <10th centiles were not significantly linked to adverse perinatal outcome. An EFW <3rd centile alone was found to be statistically significant (at the nominal 5% significance level). However, in those who did not have an adverse outcome (n = 1058), 775 (73.3%) had an EFW <3rd centile at any stage during the study period as compared to 209 (19.8%) who had an EFW <3rd centile plus an abnormal UA Doppler.
      Irrespective of the EFW or the AC measurement, the strongest and most significant association with adverse perinatal outcome in the overall population was found when an abnormal UA Doppler was present. Even after multivariate analysis adjusting for GA at delivery expressed by respective P values in Table 2, abnormal Doppler remained significantly associated with adverse perinatal outcome. The only sonographic weight-related definition consistently associated with adverse perinatal outcome was an EFW <3rd centile (P = .0131), although this did not reach the Bonferroni-corrected significance level of 0.3%. The presence of oligohydramnios was only important when combined with an EFW <3rd centile (P = .0066).
      As outlined in Table 3, 28% (n = 312) of the cohort required admission to the NICU with a median length of stay of 13 days. Table 3 demonstrates better predictive ability of fetal biometry markers (EFW <3rd centile, AC <3rd or <5th centiles) for NICU admission in the absence of abnormal Doppler. However as with predictors of adverse perinatal outcome, whenever abnormal Doppler was present, the risk of NICU admission was significantly increased irrespective of the biometry cutoff used.
      There were 4 stillbirths and 4 neonatal deaths in this cohort of nonanomalous infants corresponding to a perinatal mortality rate of 7.2 per 1000 births. As demonstrated in Table 4, all 8 perinatal deaths occurred in the setting of an EFW <3rd centile. The lack of statistical significance associated with any of the cutoffs used in Table 4 reflects the small number of deaths in the overall cohort.

      Comment

      The appropriate management of IUGR pregnancies relies on accurate identification and diagnosis. The optimal definition of growth restriction, in particular the differentiation between physiological (SGA) and pathological (IUGR) small fetal size, is one of the most common, controversial, and complex problems in modern obstetrics.
      There is agreement internationally that an EFW <10th centile for gestation should alert clinicians to potential small fetal size. However, our data call into question whether having an EFW <10th centile on its own really matters in predicting adverse perinatal outcome. Our data suggest that all fetuses with an EFW <3rd centile, or those fetuses with a combination of EFW <10th centile and abnormal UA Doppler, are at increased risk of either adverse perinatal outcome or NICU admission when compared to those with EFW or AC <10th or <5th centiles or normal UA Doppler indices. Whether a composite morbidity rate of 2% in the group of infants with EFW in the 3rd-10th centiles is acceptable to clinicians to change their utilization of antenatal surveillance is debatable. Until such time that patients identified with various cutoffs are subjected to a randomized trial or a blinded trial the question regarding optimal IUGR definition, surveillance, and intervention cannot be answered satisfactorily. In contrast to the ACOG bulletin on IUGR, we found that amniotic fluid abnormalities per se do not significantly increase the risk for adverse outcome.
      A major strength of this study is the prospective study design. It took only 2 years to recruit 1200 pregnancies in 7 centers, all of which were subjected to a high degree of fetal surveillance using the most advanced Doppler techniques available, which were performed by a small group of trained research sonographers. While the study was multicenter in nature, inconsistencies in assessment were overcome with regular training sessions of the cohort of research sonographers by experienced maternal–fetal medicine specialists. All sonographers used the same ultrasound equipment (Voluson E8; GE Healthcare, Buckinghamshire, UK) and underwent regular quality assurance assessments. Another strength of the study is that the cohort reflects a true at-risk population with 1 in 20 infants affected by composite adverse perinatal outcome and 1 in 4 infants requiring admission to the NICU. A possible weakness of our study relates to the slightly late GA at which IUGR pregnancies were enrolled (mean GA at enrollment was 30 weeks). This might underestimate the true implications of IUGR as most later-gestation pregnancies generally have a better perinatal prognosis.
      A possible weakness of our study is that we had no ability to compare our IUGR cohort to an appropriately grown reference population. Therefore a firm conclusion regarding when it is reasonable to avoid fetal surveillance for the SGA fetus cannot be drawn from our data. Also, there may be some limitations in generalizing our results to the US population, given that the majority of our patients were white European women with a mean body mass index of 24. Additionally, there are some minor variations in clinical practice between our population and the US population, such as the administration of corticosteroids for IUGR fetuses up until 36 weeks' gestation in our population as recommended by the RCOG.
      Royal College of Obstetricians and Gynecologists
      Antenatal corticosteroids to reduce neonatal morbidity (guideline no. 7).
      We also acknowledge that our perinatal outcomes are not solely a reflection of ultrasound parameters and that respective pregnancy outcomes must always be seen in the context of mode and frequency of antenatal surveillance and interventions such as administration of steroids and timely delivery, especially given that we did not use preset criteria for delivery.
      In conclusion, the terminology currently used to describe inadequate fetal growth is inconsistent and confusing; therefore a clear definition of SGA and IUGR is needed to improve outcomes for this cohort. Our data provide evidence from a large prospective cohort of >1100 pregnancies with small fetal size that correlates various sonographic parameters with perinatal outcome. Pregnancies at increased risk of adverse outcome are those with associated abnormal UA Doppler and in particular those with EFW <3rd centile with or without accompanying oligohydramnios.

      Acknowledgments

      We thank Fiona Manning, original Perinatal Ireland research manager, and the team of Perinatal Ireland research sonographers Fiona Cody, Hilda O'Keefe, Emma Doolin, Cecelia Mulcahy, Azy Khalid, Phyl Gargan, Annette Burke, Edel Varden, Wendy Ooi, and Amanda Ali.

      References

        • Badawi N.
        • Kurinczuk J.J.
        • Keogh J.M.
        • et al.
        Antepartum risk factors for newborn encephalopathy: the Western Australian case-control study.
        BMJ. 1998; 317: 1549-1553
        • Barker D.J.
        Adult consequences of fetal growth restriction.
        Clin Obstet Gynecol. 2006; 49: 270-283
        • American College of Obstetricians and Gynecologists
        Intrauterine growth restriction.
        American College of Obstetricians and Gynecologists, Washington, DC2000
        • Royal College of Obstetricians and Gynecologists
        The investigation and management of the small-for-gestational-age fetus (guideline no. 31).
        Royal College of Obstetricians and Gynecologists, London2002
        • Lausman A.
        • McCarthy F.P.
        • Walker M.
        • Kingdom J.
        Screening, diagnosis, and management of intrauterine growth restriction.
        J Obstet Gynaecol Can. 2012; 34: 17-28
        • Gardosi J.
        • Chang A.
        • Kalyan B.
        • Sahota D.
        • Symonds E.M.
        Customized antenatal growth charts.
        Lancet. 1992; 339: 283-287
        • Hadlock F.P.
        • Harrist R.B.
        • Sharman R.S.
        • Deter R.L.
        • Park S.K.
        Estimation of fetal weight with the use of head, body, and femur measurements–a prospective study.
        Am J Obstet Gynecol. 1985; 151: 333-337
      1. Perinatal Statistics Report, Economic and Social Research Institute (ESRI) on behalf of the Department of Health and the Health Service Executive Ireland, June 2010.
        (Accessed Feb. 25, 2013)
        • Horbar J.D.
        The Vermont Oxford Network: evidence-based quality improvement for neonatology.
        Pediatrics. 1999; 103: 350-359
        • Archaya G.
        • Wilsgaard T.
        • Berntsen G.K.
        • Maltau J.M.
        • Kiserud T.
        Reference ranges for serial measurements of umbilical artery Doppler indices in the second half of pregnancy.
        Am J Obstet Gynecol. 2005; 192: 937-944
        • Royal College of Obstetricians and Gynecologists
        Antenatal corticosteroids to reduce neonatal morbidity (guideline no. 7).
        Royal College of Obstetricians and Gynecologists, London2010