Advertisement

A customized standard of large size for gestational age to predict intrapartum morbidity

Published:March 07, 2011DOI:https://doi.org/10.1016/j.ajog.2011.02.068

      Objective

      The purpose of this study was to determine whether a customized standard of large-for-gestational age (LGA) identifies pregnancies with increased perinatal risk.

      Study Design

      We evaluated 7510 estimates of fetal weight to generate a fetal growth curve. Next, we analyzed the gestational age at delivery, physiologic and pathological variables from 5072 pregnancies to predict birthweight, and calculated a customized ideal birthweight and cutoff for LGA. In a separate analysis of 32,271 pregnancies, rates of macrosomia-related adverse outcomes were compared in pregnancies that had been identified as LGA by a customized standard (LGAcust) and those pregnancies that had been identified as LGA or macrosomic by conventional standards.

      Results

      LGAcust pregnancies carried increased risk of shoulder dystocia, third- or fourth-degree laceration, and cephalopelvic disproportion. LGAcust pregnancies that did not meet conventional criteria for LGA/macrosomia were at increased risk of all measured outcomes.

      Conclusion

      A customized standard of LGA identifies a previously unrecognized population that is at increased risk of perinatal morbidity.

      Key words

      With increasing birthweight, the risk of dysfunctional labor, shoulder dystocia, birth trauma, and genital tract injury increases. Published evidence consistently demonstrates an increased risk of perinatal and neonatal morbidity with delivery of a macrosomic or large-for-gestational-age (LGA) infant.
      • Spellacy W.N.
      • Miller S.
      • Winegar A.
      • Peterson P.Q.
      Macrosomia: maternal characteristics and infant complications.
      • Boulet S.L.
      • Alexander G.R.
      • Salihu H.M.
      • Pass M.
      Macrosomic births in the United States: determinants, outcomes, and proposed grades of risk.
      • Esakoff T.F.
      • Cheng Y.W.
      • Sparks T.N.
      • Caughey A.B.
      The association between birthweight 4000 g or greater and perinatal outcomes in patients with and without gestational diabetes mellitus.
      • Gregory K.D.
      • Henry O.A.
      • Ramicone E.
      • Chan L.S.
      • Platt L.D.
      Maternal and infant complications in high and normal weight infants by method of delivery.
      • Kolderup L.B.
      • Laros Jr, R.K.
      • Musci T.J.
      Incidence of persistent birth injury in macrosomic infants: association with mode of delivery.
      • Stotland N.E.
      • Caughey A.B.
      • Breed E.M.
      • Escobar G.J.
      Risk factors and obstetric complications associated with macrosomia.
      • Walle T.
      • Hartikainen-Sorri A.L.
      Obstetric shoulder injury: associated risk factors, prediction and prognosis.
      Nonetheless, a universally accepted definition of macrosomia and a clinical standard of care for treatment of the macrosomic fetus remain elusive.
      For Editors' Commentary, see Table of Contents
      See related editorial, page 457
      Increased rates of obstetric complications have been reported with delivery of an LGA infant
      • Walle T.
      • Hartikainen-Sorri A.L.
      Obstetric shoulder injury: associated risk factors, prediction and prognosis.
      and with birthweights that exceed 4000,
      • Boulet S.L.
      • Alexander G.R.
      • Salihu H.M.
      • Pass M.
      Macrosomic births in the United States: determinants, outcomes, and proposed grades of risk.
      • Esakoff T.F.
      • Cheng Y.W.
      • Sparks T.N.
      • Caughey A.B.
      The association between birthweight 4000 g or greater and perinatal outcomes in patients with and without gestational diabetes mellitus.
      • Gregory K.D.
      • Henry O.A.
      • Ramicone E.
      • Chan L.S.
      • Platt L.D.
      Maternal and infant complications in high and normal weight infants by method of delivery.
      • Kolderup L.B.
      • Laros Jr, R.K.
      • Musci T.J.
      Incidence of persistent birth injury in macrosomic infants: association with mode of delivery.
      • Stotland N.E.
      • Caughey A.B.
      • Breed E.M.
      • Escobar G.J.
      Risk factors and obstetric complications associated with macrosomia.
      4500,
      • Spellacy W.N.
      • Miller S.
      • Winegar A.
      • Peterson P.Q.
      Macrosomia: maternal characteristics and infant complications.
      • Boulet S.L.
      • Alexander G.R.
      • Salihu H.M.
      • Pass M.
      Macrosomic births in the United States: determinants, outcomes, and proposed grades of risk.
      • Stotland N.E.
      • Caughey A.B.
      • Breed E.M.
      • Escobar G.J.
      Risk factors and obstetric complications associated with macrosomia.
      or 5000 g.
      • Boulet S.L.
      • Alexander G.R.
      • Salihu H.M.
      • Pass M.
      Macrosomic births in the United States: determinants, outcomes, and proposed grades of risk.
      • Stotland N.E.
      • Caughey A.B.
      • Breed E.M.
      • Escobar G.J.
      Risk factors and obstetric complications associated with macrosomia.
      Conway and Langer
      • Conway D.L.
      • Langer O.
      Elective delivery of infants with macrosomia in diabetic women: reduced shoulder dystocia versus increased cesarean deliveries.
      have proposed divergent management schemes for diabetic pregnancies that are based on estimated fetal weight (EFW) at 37 weeks' gestation, with expectant management of pregnancies with appropriate fetal growth, prophylactic cesarean delivery of infants with an EFW of >4250 g, and induction of labor if the EFW is >90th percentile but <4250 g. National practice guidelines state that 4500 g is the appropriate threshold for the designation of fetal macrosomia but suggest that prophylactic cesarean delivery should be reserved for pregnancies with an EFW that exceeds 5000 g.
      American College of Obstetricians and Gynecologists
      Fetal macrosomia: ACOG practice bulletin no. 22.
      These disparate criteria reflect the lack of a universally accepted standard of excessive fetal growth.
      Customized growth curves that are adjusted to reflect individual factors such as maternal height, weight, race, parity, and fetal sex, have been purported to improve the ability to identify abnormal fetal growth
      • Gardosi J.
      • Francis A.
      Controlled trial of fundal height measurement plotted on customised antenatal growth charts.
      and pregnancies that are at increased risk of adverse outcomes related to growth restriction.
      • Clausson B.
      • Gardosi J.
      • Francis A.
      • Cnattingius S.
      Perinatal outcome in SGA births defined by customised versus population-based birthweight standards.
      • Gardosi J.
      • Francis A.
      Adverse pregnancy outcome and association with small for gestational age birthweight by customized and population-based percentiles.
      • Gardosi J.
      • Clausson B.
      • Francis A.
      The value of customised centiles in assessing perinatal mortality risk associated with parity and maternal size.
      • Figueras F.
      • Figueras J.
      • Meler E.
      • et al.
      Customised birthweight standards accurately predict perinatal morbidity.
      The British Royal College of Obstetricians and Gynaecologists has recognized customized growth curves as the preferred clinical standard.
      Royal College of Obstetricians and Gynaecologists
      The investigation and management of the small-for-gestational-age fetus: guideline no. 31.
      The usefulness of customized growth standards in the identification of pregnancies that are at increased risk of morbid outcomes associated with delivery of a large infant has not been investigated thoroughly. The purpose of this retrospective study was to determine whether the use of a customized birthweight standard could improve the ability of caregivers to identify pregnancies at increased risk of macrosomia-related morbidity.

      Materials and Methods

      All data were collected from the Magee Obstetric Medical and Infant database. The Magee Obstetric Medical and Infant database, established in 1995, routinely collects comprehensive maternal, fetal, and neonatal outcomes from electronic and medical record data on all women who deliver at Magee-Womens Hospital, University of Pittsburgh Medical Center, Pittsburgh, PA. The database is surveyed periodically at random to maintain its accuracy by direct comparison with patient charts. Personal identifying information in the database was eliminated to ensure confidentiality.

      Determination of intrauterine growth standard

      A subset of subjects included in the Magee Obstetric Medical and Infant database had ultrasound data available for analysis. We performed a retrospective analysis of 7510 EFWs that were obtained during clinically indicated sonograms in pregnancies that resulted in delivery of a live singleton at term at Magee-Womens Hospital between March 2004 and May 2007. All pregnancies that were analyzed to generate an intrauterine growth standard were dated by either first-trimester or mid-trimester ultrasound scanning that corroborated dating based on the last menstrual period. EFWs obtained during the sonogram that established or confirmed pregnancy dating were not used for purposes of analysis. Thus, all subjects whose data were analyzed had a minimum of 2 sonograms that were performed during their pregnancy, and all EFWs analyzed were obtained from ultrasound scans performed on subjects with previously dated pregnancies. We used the formula of Hadlock et al
      • 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.
      based on abdominal circumference and head circumference to calculate 97.8% of the estimates of fetal weight. If a head circumference could not be obtained, EFW was calculated with Hadlock's formula based on abdominal circumference and femur length. We excluded subjects with diabetes mellitus (gestational or pregestational), chronic hypertension, gestational hypertension, preeclampsia, or congenital anomalies. The intent of excluding these pregnancies and those pregnancies that resulted in preterm delivery was to develop a standard of fetal growth in the nonpathologic pregnancy. The demographic characteristics of the population evaluated are listed in Table 1. With the use of polynomial regression, the gestational age (GA) at the time of ultrasound scan was used to predict the log-transformation of EFW.
      TABLE 1Characteristics of the subjects who underwent ultrasound scanning that were used to generate intrauterine growth curve (n = 7510)
      CharacteristicMeasurement
      Maternal age at ultrasound scanning, y
      Data are given as mean ± SD.
      29.8 ± 6.2
      Gestational age at ultrasound scanning, wk
      Data are given as mean ± SD.
      32.2 ± 4.5
      Body mass index, kg/m2
      Data are given as mean ± SD.
      25.4 ± 6.6
      Parity, n (%)
       03312 (44.1)
       12565 (34.2)
       21044 (13.9)
       3377 (5.0)
       ≥4212 (2.8)
      Male fetus, n (%)3719 (49.6)
      Race, n (%)
       White5654 (75.3)
       African American1487 (19.8)
       Other369 (4.9)
      Self-reported tobacco use, n (%)1140 (15.2)
      Self-reported cocaine use, n (%)48 (0.6)
      Self-reported marijuana use, n (%)71 (1.0)
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.
      a Data are given as mean ± SD.

      Determination of customized birthweight standards

      Following methods established by Gardosi et al,
      • Gardosi J.
      • Francis A.
      Adverse pregnancy outcome and association with small for gestational age birthweight by customized and population-based percentiles.
      • Gardosi J.
      • Francis A.
      A customized standard to assess fetal growth in a US population.
      • Gardosi J.
      • Chang A.
      • Kalyan B.
      • Sahota D.
      • Symonds E.M.
      Customised antenatal growth charts.
      multivariable linear regression was used to generate a predictive model of birthweight, using demographic and clinical variables as covariates. Five thousand seventy-two women who delivered live, full-term singleton infants without congenital anomalies at Magee-Womens Hospital between March 2004 and May 2007 were evaluated. The demographic characteristics of this population are listed in Table 2. Information was extracted from the same database that was used to generate a fetal growth curve; as a consequence, there is overlap between the populations that were used for these analyses. While subjects with diabetes mellitus, hypertension, or preeclampsia were not excluded from birthweight analysis, subjects with missing data for any of the variables analyzed were excluded. Physiologic variables (maternal height, weight, parity, race, fetal sex, and GA at delivery) and pathologic factors (smoking history, illicit drug use, diabetes mellitus, and hypertension/preeclampsia) were included. Covariates were selected by backward elimination with a significance level of .05.
      TABLE 2Characteristics of the population that was studied to generate the birthweight model (n = 5072)
      CharacteristicMeasurement
      Maternal age, y
      Data are given as mean ± SD.
      29.9 ± 6.1
      Gestational age at delivery, wk
      Data are given as mean ± SD.
      39.1 ± 1.2
      Body mass index, kg/m2
      Data are given as mean ± SD.
      25.6 ± 6.4
      Parity, n (%)
       02438 (48.1)
       11655 (32.6)
       2658 (13.0)
       3215 (4.2)
       ≥4106 (2.1)
      Male fetus, n (%)2559 (50.5)
      Race, n (%)
       White3946 (77.8)
       African American864 (17.0)
       Native American12 (0.2)
       Hispanic14 (0.3)
       Asian121 (2.4)
       Other103 (2.0)
      Self-reported tobacco use, n (%)698 (13.8)
      Self-reported cocaine use, n (%)26 (0.5)
      Self-reported marijuana use, n (%)43 (0.9)
      Pregnancy-related hypertension, n (%)
       Gestational hypertension300 (5.9)
       Mild preeclampsia225 (4.4)
       Severe preeclampsia22 (0.4)
       Eclampsia2 (0.04)
       Superimposed preeclampsia23 (0.5)
      Diabetes mellitus, n (%)
       Class A1/A2375 (7.4)
       Class B41 (0.8)
       Class C14 (0.3)
       Class D6 (0.1)
       Class R4 (0.1)
       Class F1 (0.02)
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.
      a Data are given as mean ± SD.
      After a predictive model of birthweight had been generated, an ideal birthweight for 40 weeks' GA was calculated, based exclusively on nonpathologic variables (maternal height, weight, parity, race, and fetal sex) and a standardized GA of 40 weeks. With the calculation of an ideal birthweight at term, a cutoff value for 10th and 90th percentile was generated. These values were based on the standard deviation/mean birthweight (coefficient of variance) multiplied by the Z score for 10th and 90th percentile, –1.28 and 1.28, respectively. The mean birthweight for the study population was 3417 g, with an SD of 485 g, which yielded a coefficient of variance of 0.14. The resulting formula was used for the calculation of 10th and 90th percentile: ideal birthweight at term ± (ideal birthweight at term × 1.28 × 0.14).
      Once ideal birthweight at term and the corresponding 10th and 90th percentile were calculated, the ideal birthweight and corresponding cutoff for 10th and 90th percentile could be extrapolated for any GA, using the intrauterine growth curve previously generated. With this method, a customized ideal birthweight and a customized standard for suboptimal or excessive growth could be generated for any pregnancy at any GA.

      Morbidity in customized vs population-based classification of LGA

      After generating a model of customized fetal growth, we compared morbid outcomes relating to macrosomia in pregnancies that were classified as LGA using customized vs conventional population-based birthweight standards. A separate population of 32,271 pregnancies that delivered at Magee-Womens Hospital from 2003-2008, distinct from that used to generate a fetal growth curve or customized birthweight standard, was used for this analysis. The demographic composition of this population is listed in Table 3. This population was meant to represent a general obstetric population and included all women who delivered live-born infants with available records that were not included in previous analyses. We did not exclude subjects with multiple gestations, congenital anomalies, or preterm delivery. We classified pregnancies as LGAcust if the birthweight was greater than the customized cutoff for the 90th percentile. Pregnancies were classified as LGApop if the recorded birthweight was higher than the 90th percentile for GA according to the national standards published by Alexander et al.
      • Alexander G.R.
      • Himes J.H.
      • Kaufman R.B.
      • Mor J.
      • Kogan M.
      A United States national reference for fetal growth.
      Multivariate logistic regression was used to compare the risk of shoulder dystocia, cesarean section delivery for cephalopelvic disproportion (CPD), and third- or fourth-degree laceration in pregnancies that were classified as LGApop and LGAcust and in pregnancies with birthweights in excess of 4000 or 4500 g. Covariates were selected by backward elimination with a significance level of .05.
      TABLE 3Characteristics of the population with perinatal outcome evaluation (n = 32,271)
      CharacteristicMeasurement
      Maternal age, y
      Data are given as mean ± SD.
      29.4 ± 6.1
      Gestational age at delivery, wk
      Data are given as mean ± SD.
      38.5 ± 2.29
      Pregravid body mass index, kg/m2
      Data are given as mean ± SD.
      25.1 ± 5.8
      Cesarean delivery, n (%)8554 (26.5)
      Parity, n (%)
       014,632 (45.3)
       110,557 (32.7)
       24590 (14.2)
       31564 (4.9)
       ≥4928 (2.9)
      Male fetus, n (%)16,636 (51.6)
      Race, n (%)
       White24,781 (76.8)
       African American5763 (17.9)
       Native American194 (0.6)
       Hispanic108 (0.3)
       Asian509 (1.6)
       Other635 (2.0)
      Self-reported tobacco use, n (%)4682 (14.5)
      Self-reported cocaine use, n (%)295 (0.9)
      Self-reported marijuana use, n (%)429 (1.3)
      Pregnancy-related hypertension, n (%)
       Gestational hypertension1718 (5.3)
       Mild preeclampsia1325 (4.1)
       Severe preeclampsia446 (1.4)
       Eclampsia20 (0.06)
       Superimposed preeclampsia179 (0.6)
      Diabetes mellitus, n (%)
       Class A1/A21368 (4.2)
       Class B140 (0.4)
       Class C54 (0.2)
       Class D27 (0.08)
       Class R18 (0.06)
       Class F4 (0.01)
      Multiple gestation, n (%)761 (2.36)
      Congenital anomaly, n (%)1130 (3.50)
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.
      a Data are given as mean ± SD.
      All analyses were performed with the use of Stata software (versions 10 and 11; Stata Corporation, College Station, TX).

      Results

      Our final equation predicted EFW based on GA: ln(EFW) = 0.719 + (0.341 × GA) – (0.00383 × GA2), where GA is the GA in weeks at the time of delivery. A graphic depiction of our curve, along with that of Hadlock et al,
      • Hadlock F.P.
      • Harrist R.B.
      • Martinez-Poyer J.
      In utero analysis of fetal growth: a sonographic weight standard.
      which was extrapolated from an expected weight of 3417 g at 40 weeks' GA, is shown in Figure 1. Previous studies of customized growth standards have incorporated the intrauterine growth curve of Hadlock et al,
      • Hadlock F.P.
      • Harrist R.B.
      • Martinez-Poyer J.
      In utero analysis of fetal growth: a sonographic weight standard.
      which uses GA and its square as the only predictive variables for EFW. For purposes of consistency, we used an intrauterine growth curve with GA and its square as the only predictors of EFW. However, we found that GA had a significant interaction with African American race, fetal sex, body mass index, and parity in the prediction of EFW (data not shown).
      Figure thumbnail gr1
      FIGURE 1Comparison of Hadlock and Magee-Womens Hospital fetal growth curves
      Both growth curves are extrapolated from an estimated fetal weight of 3417 g at 40 weeks' gestational age.
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.
      The results of our birthweight prediction model are shown in Table 4. Fetal sex, maternal height, weight, parity, and African American and Asian race were significant physiologic predictors of birthweight. Cigarette, cocaine, or marijuana use, superimposed preeclampsia, and class A, B, C, or R diabetes mellitus were pathologic predictors of birthweight that reached statistical significance. The precision of our estimates of effect of specific variables is limited by small cell numbers of specific demographic groups, which are reflective of the predominantly white and African American populations at our center.
      TABLE 4Coefficients from multiple regression birthweight prediction model (coefficient of model = 3655 g; root MSE = 419.9; R2 = 0.25)
      VariableCoefficientSEP value
      Gestational age (from 40 wk)
       Linear term105.638.19< .001
       Quadratic term−13.483.54< .001
      Female sex (male referent)−161.2211.83< .001
      Maternal height (from 164 cm)5.660.94< .001
      Maternal weight (from 69.5 kg)
       Linear term6.780.47< .001
       Quadratic term−0.120.22< .001
       Cubic term0.000720.00027.009
      Parity (nulliparous referent)
       Para 182.5613.76< .001
       Para 2102.9513.85< .001
       Para 3110.0930.39< .001
       Para≥498.8942.31.019
      Race (white referent)
       African American−166.9516.24< .001
       Native American79.08121.64.516
       Hispanic−42.70112.55.704
       Asian−90.9839.61.022
       Other−58.3742.31.167
      Cigarettes/d−17.501.62< .001
      Cocaine use−306.9084.01< .001
      Marijuana use−146.9065.15.024
      Pregnancy-related hypertension
       Gestational hypertension−35.8525.37.158
       Mild preeclampsia16.4029.39.577
       Severe preeclampsia−95.0690.08.291
       Eclampsia319.79297.99.283
       Superimposed preeclampsia−330.6088.50< .001
      Diabetes mellitus
       Class A1/A270.3323.22.002
       Class B329.3266.77< .001
       Class C383.80112.80.001
       Class D91.74172.04.594
       Class R657.15210.93.002
       Class F560.08420.94.183
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.
      After the generation of a fetal growth curve and predictive model of birthweight, a separate study population of 32,271 pregnancies was evaluated, and appropriate subjects were classified as LGApop and LGAcust: 9.6% of the pregnancies were classified LGApop; 6.3% of the pregnancies were classified LGAcust; 9.3% of the subjects delivered infants with birthweights of >4000 g (Macro4000), and 1.2% of the subjects delivered infants with birthweights of >4500 g (Macro4500). Subjects who delivered infants who were LGApop, LGAcust, Macro4000, or Macro4500 were at an increased risk of third- or fourth-degree laceration, shoulder dystocia, and cesarean delivery for CPD. Table 5 shows the rates of these outcomes and the corresponding adjusted and unadjusted odds ratios in pregnancies that were deemed LGA or macrosomic by the different classification methods. Women with cesarean deliveries for indications other than CPD were excluded from our analysis of cesarean delivery for CPD. Inclusion of these women did not change the significance of our findings (data not shown).
      TABLE 5Rates of macrosomia-related adverse outcomes
      OutcomeClassificationnRate/100Odds ratio (95% CI)Adjusted odds ratio
      For third- or fourth-degree laceration and shoulder dystocia, data were adjusted for gestational age, performance of episiotomy, diabetes mellitus, parity, and use of forceps; for cesarean delivery for cephalopelvic disproportion, data were adjusted for gestational age, parity, diabetes mellitus, hypertensive disease, spontaneous vs induced labor, and history of cesarean delivery;
      (95% CI)
      Third- or fourth-degree laceration
      Cesarean deliveries excluded from analysis;
      Population23,7174.99
      LGApop19199.542.19 (1.86–2.59)2.70 (2.25–3.24)
      LGAcust119210.822.45 (2.02–2.97)2.87 (2.33–3.55)
      Macro400018759.652.22 (1.88–2.62)2.49 (2.07–2.99)
      Macro450020915.793.65 (2.50–5.31)3.98 (2.59–6.11)
      Shoulder dystocia
      Cesarean deliveries excluded from analysis;
      Population23,7172.38
      LGApop191911.98.68 (7.29–10.34)7.76 (6.48–9.29)
      LGAcust119212.67.69 (6.31–9.36)7.89 (6.45–9.64)
      Macro4000187511.68.17 (6.85–9.75)7.10 (5.89–8.56)
      Macro450020922.012.5 (8.93–17.6)9.65 (6.82–13.6)
      Cesarean delivery for cephalopelvic disproportion
      Indications other than cephalopelvic disproportion were excluded from analysis.
      Population24,2032.1
      LGApop20255.43.10 (2.49–3.85)3.40 (2.70–4.29)
      LGAcust12565.42.92 (2.25–3.79)3.34 (2.52–4.41)
      Macro400019956.13.63 (2.95–4.48)3.42 (2.73–4.29)
      Macro45002225.92.95 (1.67–5.19)2.50 (1.36–4.58)
      LGAcust, by a customized standard; LGApop, by the population standard; Macro4000, infants with birthweights above 4000 g; Macro4500, infants with birthweights above 4500 g.
      CI, confidence interval; LGA, large-for-gestational age.
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.
      a For third- or fourth-degree laceration and shoulder dystocia, data were adjusted for gestational age, performance of episiotomy, diabetes mellitus, parity, and use of forceps; for cesarean delivery for cephalopelvic disproportion, data were adjusted for gestational age, parity, diabetes mellitus, hypertensive disease, spontaneous vs induced labor, and history of cesarean delivery;
      b Cesarean deliveries excluded from analysis;
      c Indications other than cephalopelvic disproportion were excluded from analysis.
      As shown in Figure 2, there is some degree of overlap between the group of subjects who were identified by the customized standard of LGA and those who were identified by other conventional methods. To quantify the risk of perinatal morbidity in subjects who would be overlooked with the exclusive use of a customized or particular conventional standard, we evaluated subjects who were identified exclusively with either the customized standard or one of the conventional alternatives. When we used both a customized and population-based standard for LGA, women whose newborns were identified by the population standard only (LGApoponly) tended to deliver larger infants at a later GA in comparison with women whose newborns were identified by the customized standard only (LGAcustonly). Figure 3 shows the distribution of GA and birthweight for LGAcustonly and LGApoponly pregnancies in relation to the 90th birthweight percentile on the Alexander curve. LGAcustonly newborns were smaller and delivered at earlier GAs in comparison to LGApoponly newborns and newborns identified exclusively by birthweight criteria of macrosomia.
      Figure thumbnail gr2
      FIGURE 2Classification of LGA by a customized standard (LGAcust) vs conventional criteria of excessive fetal growth (total population = 32,271)
      A, LGAcust vs LGA by the population standard (LGApop); B, LGAcust vs infants with birthweights above 4000 g (Macro4000); C, LGAcust vs infants with birthweights above 4500 g (Macro4500).
      LGA, large-for-gestational age.
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.
      Figure thumbnail gr3
      FIGURE 3Distribution of birthweight and gestational age in LGApoponly and LGAcustonly cohorts
      Distribution of gestational age and birthweight in LGA by the population standard only (LGApop only; red dots) and LGA by a customized standard only (LGAcust only; green dots). The line represents Alexander 90%
      LGA, large-for-gestational age.
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.
      The adjusted odds ratios of adverse outcomes in the LGAcustonly, LGApoponly, and LGAboth cohorts are shown in Figure 4. These comparisons were repeated for the customized standard of LGA in combination with birthweight criteria of macrosomia. Women whose neonates were identified as LGAcustonly, regardless of the alternative standard for LGA/macrosomia, carried significantly increased risk of third- or fourth-degree laceration, shoulder dystocia, and cesarean delivery for CPD. Exclusion of multiple gestations and women who delivered infants with congenital anomalies did not change the significance of our findings (data not shown).
      Figure thumbnail gr4a
      FIGURE 4Adverse outcomes for the pregnancies that were identified exclusively as LGA/macrosomic by either customized or conventional criteria
      A, Shoulder dystociaa,b; LGAcust, by a customized standard; LGApop, by the population standard; Macro4000, infants with birthweights above 4000 g; Macro4500, infants with birthweights above 4500 g; B, third- or fourth-degree lacerationa,b; LGAcust, by a customized standard; LGApop, by the population standard; Macro4000, infants with birthweights above 4000 g; Macro4500, infants with birthweights above 4500 g; C, cesarean delivery for cephalopelvic disproportiona,c; LGAcust, by a customized standard; LGApop, by the population standard; Macro4000,infants with birthweights above 4000 g; Macro4500, infants with birthweights above 4500 g.
      LGA, large-for-gestational age.
      aThird- or fourth-degree laceration and shoulder dystocia analyses adjusted for gestational age, performance of episiotomy, diabetes mellitus, parity, and the use of forceps; cesarean delivery for cephalopelvic disproportion analysis adjusted for gestational age, parity, diabetes mellitus, hypertensive disease, spontaneous vs induced labor, and history of previous cesarean delivery; bCesarean deliveries excluded from analysis; cCesarean deliveries for indication other than cephalopelvic disproportion were excluded from analysis; there were no cases of cesarean delivery for cephalopelvic disproportion in the cohort with birthweights above 4500 g only (Macro4500 only).
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.
      Figure thumbnail gr4b
      FIGURE 4Adverse outcomes for the pregnancies that were identified exclusively as LGA/macrosomic by either customized or conventional criteria
      A, Shoulder dystociaa,b; LGAcust, by a customized standard; LGApop, by the population standard; Macro4000, infants with birthweights above 4000 g; Macro4500, infants with birthweights above 4500 g; B, third- or fourth-degree lacerationa,b; LGAcust, by a customized standard; LGApop, by the population standard; Macro4000, infants with birthweights above 4000 g; Macro4500, infants with birthweights above 4500 g; C, cesarean delivery for cephalopelvic disproportiona,c; LGAcust, by a customized standard; LGApop, by the population standard; Macro4000,infants with birthweights above 4000 g; Macro4500, infants with birthweights above 4500 g.
      LGA, large-for-gestational age.
      aThird- or fourth-degree laceration and shoulder dystocia analyses adjusted for gestational age, performance of episiotomy, diabetes mellitus, parity, and the use of forceps; cesarean delivery for cephalopelvic disproportion analysis adjusted for gestational age, parity, diabetes mellitus, hypertensive disease, spontaneous vs induced labor, and history of previous cesarean delivery; bCesarean deliveries excluded from analysis; cCesarean deliveries for indication other than cephalopelvic disproportion were excluded from analysis; there were no cases of cesarean delivery for cephalopelvic disproportion in the cohort with birthweights above 4500 g only (Macro4500 only).
      Larkin. Customized standard of LGA. Am J Obstet Gynecol 2011.

      Comment

      Our analysis demonstrates that subjects who were identified as LGAcust are at increased risk of morbid perinatal outcomes. Twenty to 80% of these pregnancies would not be identified by conventional standards of excessive fetal growth. As highlighted in Figure 3, both the mean birthweight and GA at delivery of the LGAcustonly cohort were significantly lower than that of the LGApoponly cohort. Because the risk of macrosomia-related morbidity increases with advancing GA and birthweight, these differences create a potential source of bias towards reduced risk in the LGAcustonly group. Despite this trend, LGAcustonly subjects were at increased risk of every adverse outcome that was evaluated.
      Genital tract injury, shoulder dystocia, and CPD stem from an imbalance between maternal and fetal factors that prevent smooth fetal passage through the pelvis and birth canal. It follows that adjustment for variation in maternal characteristics would identify perinatal risk in pregnancies that would be overlooked with the use of conventional methods that do not account for maternal factors. Our findings reinforce the importance of maternal characteristics in the determination of perinatal risk.
      The observed effect of both physiologic and nonphysiologic factors on birthweight is similar to those published in previous studies of US populations.
      • Gardosi J.
      • Francis A.
      A customized standard to assess fetal growth in a US population.
      • Odibo A.O.
      • Francis A.
      • Cahill A.G.
      • Macones G.A.
      • Crane J.P.
      • Gardosi J.
      Association between pregnancy complications and small-for-gestational-age birth weight defined by customized fetal growth standard versus a population-based standard.
      The ideal birthweight at term (3655 g) of a male infant who is born to a white mother of standard height and weight is slightly larger than that reported by Gardosi and Francis
      • Gardosi J.
      • Francis A.
      A customized standard to assess fetal growth in a US population.
      (3519 g) and Odibo et al
      • Odibo A.O.
      • Francis A.
      • Cahill A.G.
      • Macones G.A.
      • Crane J.P.
      • Gardosi J.
      Association between pregnancy complications and small-for-gestational-age birth weight defined by customized fetal growth standard versus a population-based standard.
      (3592 g). One potential contributor to this increased ideal birthweight is our lower degree of precision in recorded GA. While Gardosi and Francis
      • Gardosi J.
      • Francis A.
      Controlled trial of fundal height measurement plotted on customised antenatal growth charts.
      and Odibo et al
      • Odibo A.O.
      • Francis A.
      • Cahill A.G.
      • Macones G.A.
      • Crane J.P.
      • Gardosi J.
      Association between pregnancy complications and small-for-gestational-age birth weight defined by customized fetal growth standard versus a population-based standard.
      have reported GA in days, our data list GA in weeks. As a consequence, the coefficient for ideal birthweight applies to women who deliver between 40 and 41 weeks of completed gestation and is likely larger than a coefficient that would apply to women who deliver at precisely 280 days of gestation.
      Despite multiple differences in study populations, our fetal growth curve was remarkably similar to that of Hadlock et al.
      • Hadlock F.P.
      • Harrist R.B.
      • Martinez-Poyer J.
      In utero analysis of fetal growth: a sonographic weight standard.
      The significant interaction between race, parity, body mass index, fetal sex, and GA in the prediction of EFW suggests that a customized intrauterine growth curve may further enhance the accurate identification of the abnormally grown fetus or neonate. Future investigations that will test this hypothesis are warranted.
      Our study offers a number of methodologic strengths. In contrast to previous studies of customized growth standards,
      • Clausson B.
      • Gardosi J.
      • Francis A.
      • Cnattingius S.
      Perinatal outcome in SGA births defined by customised versus population-based birthweight standards.
      • Gardosi J.
      • Francis A.
      Adverse pregnancy outcome and association with small for gestational age birthweight by customized and population-based percentiles.
      • Figueras F.
      • Figueras J.
      • Meler E.
      • et al.
      Customised birthweight standards accurately predict perinatal morbidity.
      • McCowan L.M.
      • Harding J.E.
      • Stewart A.W.
      Customized birthweight centiles predict SGA pregnancies with perinatal morbidity.
      our analyses of morbid outcomes were performed on a population that was distinct from the study population used to generate our predictive model of birthweight, avoiding statistical pitfalls associated with models that are not validated in independent datasets. Also distinguishing our findings from previous studies of customized growth, we performed adjusted analyses to account for potential confounders, which included differences in GA at delivery.
      As with all retrospective observational studies, our findings are limited by possible bias. Pregnant women at our institution do not undergo ultrasound scans routinely beyond the midtrimester anatomic evaluation. All ultrasound data were collected for clinical indications outside of a research protocol, which could have introduced the potential for selection bias towards a population with greater risk of morbid outcomes than a general obstetric population. The exclusion of women with diabetes mellitus and women with preeclampsia/hypertension was intended to eliminate a significant proportion of “higher risk” pregnancies from the analyses performed to generate an intrauterine growth trajectory. Clinical decisions regarding mode and timing of delivery were likely influenced by estimates of fetal weight, introducing further potential for bias. Furthermore, the inaccuracy of ultrasound scanning adds an element of uncertainty to prenatal decision-making that is not accounted for in a retrospective study of birthweight data.
      Although previous investigations have shown limited benefit of obstetric interventions based on the suspicion of macrosomia, the design of these studies limits their relevance to our findings. In a randomized controlled trial, Gonen et al
      • Gonen O.
      • Rosen D.J.
      • Dolfin Z.
      • Tepper R.
      • Markov S.
      • Fejgin M.D.
      Induction of labor versus expectant management in macrosomia: a randomized study.
      showed no reduction in rates of cesarean delivery or neonatal morbidity with induction of labor at 39 weeks' gestation for pregnancies with an EFW of 4000-4500 g. However, maternal factors were not considered in inclusion, randomization, or outcome analysis. Levine et al
      • Levine A.B.
      • Lockwood C.J.
      • Brown B.
      • Lapinski R.
      • Berkowitz R.L.
      Sonographic diagnosis of the large for gestational age fetus at term: does it make a difference?.
      reported that antenatal diagnosis of macrosomia led to increased rates of elective cesarean delivery and increased frequency of the diagnosis of labor abnormalities in comparison with pregnancies with similar birthweights that did not carry an antenatal diagnosis. However, there was no uniform management or intervention introduced with the antenatal diagnosis of macrosomia, precluding any potential benefit that could have resulted from the diagnosis. Furthermore, induction of labor and maternal characteristics were not considered in the analysis.
      Our findings demonstrate a novel technique that identifies a previously unrecognized population at increased risk of adverse perinatal outcomes. These findings warrant further investigation into the utility of introducing a customized standard of LGA into clinical practice. Unlike the study of customized standards of small-for-gestational age by Gardosi and Francis,
      • Gardosi J.
      • Francis A.
      Adverse pregnancy outcome and association with small for gestational age birthweight by customized and population-based percentiles.
      the LGApoponly cohort in the present study carries an elevated risk of perinatal morbidity, and the LGAboth cohort consistently carries the highest risk of morbid outcomes. It is possible that an intervention such as scheduled induction or elective cesarean delivery would be of benefit only to pregnancies that are classified as LGA by both population-based and customized methods. Randomized trials of obstetric interventions that are allocated on the basis of the presence or absence of LGAcust, either alone or in addition to conventional standards, may hold promise for benefit. Irrespective of future studies, use of a customized standard of LGA identifies a previously unrecognized population at increased risk of perinatal morbidity and may improve the ability to gauge and anticipate clinical risk.

      References

        • Spellacy W.N.
        • Miller S.
        • Winegar A.
        • Peterson P.Q.
        Macrosomia: maternal characteristics and infant complications.
        Obstet Gynecol. 1985; 66: 158-161
        • Boulet S.L.
        • Alexander G.R.
        • Salihu H.M.
        • Pass M.
        Macrosomic births in the United States: determinants, outcomes, and proposed grades of risk.
        Am J Obstet Gynecol. 2003; 188: 1372-1378
        • Esakoff T.F.
        • Cheng Y.W.
        • Sparks T.N.
        • Caughey A.B.
        The association between birthweight 4000 g or greater and perinatal outcomes in patients with and without gestational diabetes mellitus.
        Am J Obstet Gynecol. 2009; 200: 672.e1-672.e4
        • Gregory K.D.
        • Henry O.A.
        • Ramicone E.
        • Chan L.S.
        • Platt L.D.
        Maternal and infant complications in high and normal weight infants by method of delivery.
        Obstet Gynecol. 1998; 92: 507-513
        • Kolderup L.B.
        • Laros Jr, R.K.
        • Musci T.J.
        Incidence of persistent birth injury in macrosomic infants: association with mode of delivery.
        Am J Obstet Gynecol. 1997; 177: 37-41
        • Stotland N.E.
        • Caughey A.B.
        • Breed E.M.
        • Escobar G.J.
        Risk factors and obstetric complications associated with macrosomia.
        Int J Gynaecol Obstet. 2004; 87: 220-226
        • Walle T.
        • Hartikainen-Sorri A.L.
        Obstetric shoulder injury: associated risk factors, prediction and prognosis.
        Acta Obstet Gynecol Scand. 1993; 72: 450-454
        • Conway D.L.
        • Langer O.
        Elective delivery of infants with macrosomia in diabetic women: reduced shoulder dystocia versus increased cesarean deliveries.
        Am J Obstet Gynecol. 1998; 178: 922-925
        • American College of Obstetricians and Gynecologists
        Fetal macrosomia: ACOG practice bulletin no. 22.
        The College, Washington, DC2000
        • Gardosi J.
        • Francis A.
        Controlled trial of fundal height measurement plotted on customised antenatal growth charts.
        BJOG. 1999; 106: 309-317
        • Clausson B.
        • Gardosi J.
        • Francis A.
        • Cnattingius S.
        Perinatal outcome in SGA births defined by customised versus population-based birthweight standards.
        BJOG. 2001; 108: 830-834
        • Gardosi J.
        • Francis A.
        Adverse pregnancy outcome and association with small for gestational age birthweight by customized and population-based percentiles.
        Am J Obstet Gynecol. 2009; 201: 28.e1-28.e8
        • Gardosi J.
        • Clausson B.
        • Francis A.
        The value of customised centiles in assessing perinatal mortality risk associated with parity and maternal size.
        BJOG. 2009; 116: 1356-1363
        • Figueras F.
        • Figueras J.
        • Meler E.
        • et al.
        Customised birthweight standards accurately predict perinatal morbidity.
        Arch Dis Child Fetal Neonatal Ed. 2007; 92: F277-F280
        • Royal College of Obstetricians and Gynaecologists
        The investigation and management of the small-for-gestational-age fetus: guideline no. 31.
        RCOG, London2002
        • 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
        • Gardosi J.
        • Francis A.
        A customized standard to assess fetal growth in a US population.
        Am J Obstet Gynecol. 2009; 201: 25.e1-25.e7
        • Gardosi J.
        • Chang A.
        • Kalyan B.
        • Sahota D.
        • Symonds E.M.
        Customised antenatal growth charts.
        Lancet. 1992; 339: 283-287
        • Alexander G.R.
        • Himes J.H.
        • Kaufman R.B.
        • Mor J.
        • Kogan M.
        A United States national reference for fetal growth.
        Obstet Gynecol. 1996; 87: 163-168
        • Hadlock F.P.
        • Harrist R.B.
        • Martinez-Poyer J.
        In utero analysis of fetal growth: a sonographic weight standard.
        Radiology. 1991; 181: 129-133
        • Odibo A.O.
        • Francis A.
        • Cahill A.G.
        • Macones G.A.
        • Crane J.P.
        • Gardosi J.
        Association between pregnancy complications and small-for-gestational-age birth weight defined by customized fetal growth standard versus a population-based standard.
        J Matern Fetal Neonatal Med. 2011; 24: 411-417
        • McCowan L.M.
        • Harding J.E.
        • Stewart A.W.
        Customized birthweight centiles predict SGA pregnancies with perinatal morbidity.
        BJOG. 2005; 112: 1026-1033
        • Gonen O.
        • Rosen D.J.
        • Dolfin Z.
        • Tepper R.
        • Markov S.
        • Fejgin M.D.
        Induction of labor versus expectant management in macrosomia: a randomized study.
        Obstet Gynecol. 1997; 89: 913-917
        • Levine A.B.
        • Lockwood C.J.
        • Brown B.
        • Lapinski R.
        • Berkowitz R.L.
        Sonographic diagnosis of the large for gestational age fetus at term: does it make a difference?.
        Obstet Gynecol. 1992; 79: 55-58