Cliniciii and .

Applied

Mean Maternal Second-Trimester Hemoglobin Concentration and Outcome of Pregnancy: A Population-Based Study

Viilumc 14 NumlitT i .liinuan 2008 \'-)-2H ZOOS Sage J'ublic;.ti<)nK 10.11 77/1 [)760296073{)474K h ttp; //cat h. sagt-pi 11>. i-11 m ImsU'd al hup;//cjnlinc.siigt'pub.tiini

Georg-Friedrich von Tempelhoff, MD, FCATH, Lothar Heilmann, MD, Lothar Rudig, MD, Kunhard Pollow, PhD, MD, Gerhard Hommcl, PhD, and Jurgen Koscielny, MD
Summary: Both anemia and the lack of physiological maternal plasma volume expansion during the second trimester are associated with higher maternal morhidity and poor tetal outcome. Mean hemoglobin levels hetween the 14th and 30th gestational weeks were calculated in 4985 consecutive pregnant women and were correlated with outcome data of pregnancy. It was found that 9.4% of participants (n = 3959) had normal pregnancy outcome. Mean maternal hemoglobin levels were significantly lower in women with a normal pregnaney (11.96 0.94 g/dL) compared with women who had adverse outcome events (preeelampsia, n - 4 2 3 , 12.5 1.0 g/dL, P < ,0001: early birth, n = 464, 12.2 1.01 g/dL, P < .0001; low birth weight newborn, n 473, 12.2 l.IO g/dL, P< .0001; intrauterine growth retardation, n= 250, 12.2 1.0 g/dL, P < .0001). The risk for any adverse outcome event was lowest with a mean hemoglohin hetween 11.0 and 12.0 g/dL (odds ratio, 0.625; 95% confidence interval, 0.43-0.89) and highest hetween 13.0 and 15.0 g/dL (odds ratio, 2.24; 95% confidence inter\'al. 1.54-3.31). In this population-hased study from a community in Western Germany, impaired plasma volume expansion was an independent risk factor for the development of an adverse outcome of pregnancy. Keywords: plasma volume; hemoglobin, preeclampsia; low birth weight; preterm delivery; intrauterine growth retardation; hemodilution

ver the past years, only anemia has heen recognized as a risk factor for poor fetal outcome, whereas high hemoglohin concentrations have heen considered optimal for pregnancy development. Meanwhile, it is well estahlished that hetween the second trimester and the middle of the third trimester of pregnancy, plasma volume expands hy 25% to 80% of prepregnancy volumes.' The increase in red cell mass is restricted to about 30%,

Address correspondence to Georg-Friedrich von Tempelhoff, MD, Department of Obstetrics and Gynecology, GP-Ruesselsheim, Augiist-Hebel Strasse 59. 65428 Ruesselsheim. Germany; e-mail: G-F.von.Tempelhoff@gmx.de Department of Obstetrics and Gynecology; Health and Care Center Ruesselsheim. Ruesselsheim, Germany (GFVT. LH, LR); Institute for Fxpcrimcntal Endocrinolog); University of Mainz, Mainz. Germany (KP); Institute for Medical Biometrics, Epidemiology and Informatic; University of Mainz, Mainz, Germany (GH); Institute for Transfusion Medicine. Charite Humboldt-University, Berlin. Germany (JK).

which results in a decrease of hematocrit of ahout 3% to 5% hetween gestational weeks 20 and 30^ without suhstantial accompanying changes in mean corpuscular volume and hemoglohin content of the red blood cells.-^ In large European studies, the majority of pregnant women with normal maternal and fetal outcomes had a range of hemoglobin concentrations during the first and second trimester distinctly lower than that of nonpregnant women. A prospective trial from the northwest Thames region of London found the lowest perinatal mortality among 222 614 women with their first singleton pregnancies, who had the lowest recorded maternal hemoglobin concentration at the first antenatal checkbetween 9.0 and lLOg/dL.-* In another large British population-hased study that included 153 602 pregnant women, babies born to women who had hemoglobin values between 8.5
19

20

Clinical mid Applied Thrombosis/Hemostasis / Vol. 14, No. 1, January 2008

and 9.5 g/dL at approximately week 28'' had the highest mean birth weight. These results are of great interest hecause such values are considered to represent anemia according to the World Health Organization.*" Such failure of plasma expansion between 20 and 25 weeks may result in or favor the development of adverse outcome events during pregnancy. Inappropriate plasma volume expansion has heen associated with the development of growth-retarded infants,'" delivery of low hirth weight newhorns,^'^ and placenta! infarction that may he caused by reduced uteroplacental perfusion attributable to increased biood viscosity.'^ Finally, failure of plasma volume expansion during the second trimester has been held responsible for intrauterine fetal demise in women with chronic hypertension.'"* In this population-based study from the middle region of western Germany, we retrospectively analyzed prenatal maternal records and evaluated an association between mean maternal hemoglobin concentrations during the second trimester and the outcome of pregnancy.

malformations or chromosomal abnormalities), the presence and tyjie of preeclampsia (hemolysis, elevated liver enzymes, low platelet syndrome), abruptio placentae, fetal death and perinatal mortality, venous pH in the umbilical cord, and 1-minute, 5-minite, and 10-minute Apgar scores. To stratify normal and abnormal, we defined the terms similarly for all subjects: Low birtb weight: <25OO g Early birth: <37 weeks of gestation Intrauterine growth retardation (IUGR): <5th percentile Preeclampsia: blood pressure (BP) 140/90 and proteinuria >300 mg/24 hours
I

Data Collection
The obstetrics department is linked to tbe perinatal registration of Hessen. This is a central institution that administers all data on deliveries each year for statistical evaluation. Participating hospitals record information on pregnancy, delivery, and fetal outcome of the women who deliver in their department by means ofa standardized and computerized protocol. Additionally, in a supplementar)' protocol we also record clinica! and laboratory results obtained at the time of admission for delivery, which include duplex and ultrasound, blood pressure, urinary status, cardiotocography, hematology, and blood rheology of the fetus and the mother during delivery and at the time of discharge. The merge of both protocols enables a combined evaluation of clinical findings and blood test results of more than 200 variables. Data were computerized using a program produced by the authors and converted for statistical evaluation.

Patients And Methods Pregnant Women

From January 1990 to the end of December 1996, 498 S consecutive women with singleton pregnancy wbo delivered in our department were eligible to be included in this retrospective investigation. To assess the maternal data, we photocopied the maternity records (Mutterpass). These records include information on previous pregnancies and deliveries and the current pregnancy (eg, results of routine physical examination and ultrasound fetometry documented by the external obstetrician), in addition to systolic and diastolic blood pressures, hemoglobin concentrations, and urine and immunologic/hematologic results. Moreover, the personal, medical, and social histories with particular emphasis on risk factors of pregnancy and delivery such as obesity, smoking, comorbidity, and drug use are documented. To determine mean hemoglobin concentrations during the second trimester, values between the 14''' and 30'^ weeks of gestation were adopted and calculated from the maternity records. The following maternal and tetal data were assessed: geographic origin, birth weight, week of gestation at the time of delivery, small-for-date newborns (in the absence of congenital

Statistical Analysis
Descriptive analysis included mean value standard deviation, median, interquartile range, and 95% confidence interval (CI). Differences between groups were assessed with Wilcoxon's test for unpaired samples and tested against zero. Fisher's exact test (two tailed) and risk ratios were used to compare incidences in the two groups. Correlation coefficients according to Spearman were calculated. Two-sided P values of less than .05 were considered to represent statistical significance. Variables were tested univariately by logistic regression analysis for possible

Maternal Second-Trimester Hemoglobin-Concentration /Von Temjielhoff ct ul

21

Table 1.

Characteristics of Women With Normal Pregnancy Outcome and Patients With Preeclampsia, Eariy Birth (Before Week 37), Low Birth Weight (<25OO g), and Intrauterine Growth Retardation (IUGR)
Normal Outcome Preeclampsia (28.0} 28.5 5.5^ (26.2) 27.3 6.0^ Eariy Birth (<.?7 wk) (n = 464) (28.0) 28.3 S.8^ (23.2) 24.2 4.9 Fetal Birth Weight <2500 g (n = 473) (27.0) 27.7 5.6 (22.7)23.8 + 4.6 IUGR (n = 250)^ (27.0) 27.0 =i.8 (22.6) 23.7 4.5

Age. y (median) (27.0) 27.1 =5.2 mean SD I!MI. kg/mMmedian) (23.2) 24.1 4 . 3 mean SD Weight before pregnaney, (63.0) 65.0 12.0 kg (median) mean SD Weight at deliver)', kg (75) 77.1 + i2.2 (median) mean SD Change in weight, kg (12.0) 12.2 5.1 (median) mean + SD Smoker (n)% (505) 12.8 NOTE : BMI = body mass index a. 5 missing values. h. 1 missing value. e. P < .001 vs normal outeome.

(71,0) 74.3 17.3^ (62.0) 64.2 13.6

(60.0) 62.7 12.7" (59.0) 62.6 13.

(85.0) 87.3 17.1'-' (72.0) 74.3 J3.8'- (71.0) 72.8+ 13.2'-' (70.0) 73.1 14. (13.0) I3.26.1^ (64) 15.T (10.0) 10.3+4.7'^ (86) 18.5^ (10.0) 1O.34.7'" (112)23.7'^ (10.0) 10.6 4.8' (68)27.3^

prediction of adverse outcome events. Only variables with P < .05 were selected as possibly predictive of iidverse pregnancy outcome events and were entered in ii stepwise logistic regression analysis (multivariate analysis). Odds ratios (ORs) together with the 95% CIs were calculated. Receiver operating characteristic (ROC) curves were huilt to describe the observed sensitivity and specificity of ranked mean second trimester hemoglobin concentrations for the prediction of adverse outeome events. The area under the resulting ROC curve (AUC) was then calculated. Statistical analyses were conducted in coHaboration with the Institute for Medicai Biometrics, Epidemiology and Informatic of the University of Mainz. The statistical analysis was performed using the SAS 9.1 program package (SAS Institute, Berkley, Calif).

Results Patients Studied

During the observation time, 5035 consecutive Caucasian women delivered in the department, of wbom SO nonsingieton pregnant women were excluded from calculation. We found that 79.4% {n = 3,959) had normal pregnancy outcome, 8.4% (n = 423) developed preeclampsia, 9.5% (n = 473) delivered a low birth weight newborn. 9.3% (n = 464) had early birth, and 5.0% (n = 250) were diagnosed with IUGR. In patients with preeclampsia, the prevalence

of classical risk factors was significantly higher such as older age, higher body mass index (BMI) and increased weight gain during pregnancy, and status as a smoker compared with women who had normal pregnancy outcome. By contrast, women with a diagnosis of IUGR and women who delivered a low birth weight baby had significantly lower weight before pregnancy and significantly less weight gain during pregnancy, whereas the relative number of smokers was higher compared with that among women with normal pregnancy outcome (Table I). The number of patients receiving iron supplement during pregnancy was not significantly different among women with normal pregnancy (n = 334; 8.5%), preeelampsia (n = 24; 5.7%; P = .05), low hirth weight habies {n = 35; 7.4%; P = .48), early birth (n = 32; 6.9%; P = .28), and IUGR (n = 24; 9.6%; P = .29). Women with a complicated pregnancy had significantly lower mean gestational age at the time of delivery than those with normal pregnancy. The number of female newhorns was significantly higher among those with low birth weight babies and a diagnosis of IUGR compared with those who had normal pregnancy outcome. Table 2 compares fetal outeome results in women with normal versus complicated pregnancies. Maternal Hemoglobin Concentration D u r i n g the S e c o n d Trimester During the second trimester, 34 895 hemoglohin estimations were included in the calculations. The

22

Clinical and Applied Titrombosis/Hemostasis / \/o\. 14, No. 1, Januiiry 2008

Table 2. Outcome Data of Women Witb Normal Pregnancy and Patients With Preeclampsia, Early Birth (Before wk 37), Low Birth Weigbt (<25OO g), and Intrauterine Growth Retardation (IUGR) Normal Outcome (n - 3959)^' Week of delivery (median) mean + SD Birth weight, g (median) mean SD Male (n) % rt-male (n) % Umbilical arter)' pH (umbilical eord) (median) mean + SD Apgar 5 (median) mean SD Apgar 10 (median) mean SD a. One missing value. h. P < .001 vs normal outcome. (39) 39 1 (3400) 3433 405 (1875) 47.4 (2080) 52.6 (7.32) 7.32 0.06 (10) 9.7 0.9 (10) 9.9 0.7 Preeclampsia (n = 423) (38) 38 3" (3160) 3046 742^ (204) 48.2 (219) 51.8 (7.31) 7.30 0.08" (10) 9.2 1.5" (10) 9.6+ 1.4" Early Birth (<37 wk) (n = 464) (35) 34 2" (2430) 2380 675" (235) 50.6 (229) 49.4 (7.32) 7.31 0.09 (9) 8.4 2.3'' (10) 8.9 2.3" Fetal Birth Weight <250O g (n = 473) (36) 35 3" (2300) 2149 464" (194)41.0^ (279) 59.0 (7.32) 7.30 0.08 (9) 9.0 2.2" (10) 9.0 2.2" IUGR (n^250) (38) 38 2" (2500) 2423 500" (102)41.0" (147) 59.{r (7.32) 7.31 0.08 (10) 9.2 1.7'' (10) 9.5 1.7"

number of estimations per woman was not different in women with normal pregnancy and patients with preeclampsia (mean SD [median]: 7.5 2.7 [7] vs 7.6 3.0 [8]; P = 0.82), but patients who delivered a low birth weight baby (6.6 2.9 [6]; P < .0001). had early birth (6.5 2 . 9 [6j; P < .0001), or had IUGR (6.8 2.5 [6]; P < .0001) had hemoglohin estimations less often. Hemoglobin was lower in pregnant women with normal outcome (mean + SD: 1 1.9 + 0.9 g/dL) compared with women who had adverse outcome events (mean 12.1-12.5 g/dL; P < .0001; Fig. 1). Distribution of mean hemoglobin values at this time followed a bell-shaped pattern in all pregnancies and was disposed to the righl in patients with compHcated outcome. Ninety percent of women with normal outcome had values below 13.0 g/dL, whereas 33.0% of patients with subsequent preeclampsia, 24% with IUGR, and 20% who delivered a low birth weight baby or had early birth had values above 13,0 g/dL. According to the distribution categories of the hemoglobin used in F'igure 2, relative risk for the development of complicated pregnancy was lowest, with a mean second trimester hemoglobin value between 11.0 g/dL and 12.0 g/dL (OR,,,^, 0.625; 95% Cl, 0.43-0.89) and was highest in the range between 13.0 g/dL and 15.0 g/dL (OR^,^^, 2.24; 95% GI, 1.54-2.31). Low and high ORs were statistically significant for each adverse outcome event (Fisher's exact test, P < .001) and most remarkable in preeelampsia (OR|,,^, 0.43, 95% Cl, 0.35-0.53; Oa,,K- 3.41, 95% Cl, 2.84-4.09). Altogether, the rate of complicated and normal pregnancies was not significantly different within other ranges of the

hemoglobin (OR, -1.0; P > .05). Less than 5% of patients with preeclampsia had hemoglobin values below 11.0 g/dL and 1.5% had values below 10.0 g/dL compared with 18.0% and 3.0% of the women with normal pregnancies, respectively. ROC curves were plotted for the prediction of mean second trimester hemoglobin concentrations and the development of adverse outcome events. AUC was highest in preeclampsia (0.683; 95% GI, 0.656-0.7 ] 1), and as shown in Figure 3, a mean second trimester value of 12.1 g/dL had optimal sensitivity (70%) and specificity (41%). The AUG for other adverse outcome events was between 0.58 (low birth weight) and 0.60 (IUGR). Second trimester hemoglobin concentration was not correlated with maternal age (r = 0.04; P = .003), maternal BMI (r = 0.08; P < .0001), gain of maternal weight during pregnancy (r = 0.04; P = .004), gestational age at delivery (r = 0.06: P = .007), and newborn birth weight (r = -0.07; P < .0001). The number of women with severe anemia (hemoglobin <7.0 g/dL) was small (n = 4), but none of the women with a mean hemoglobin concentration below 9.0 g/dL experienced adverse outcome events during pregnancy.

Logistic Regression for the Development of Adverse Outcome Events During Pregnancy
Odds ratios and 95% GIs were calculated for women with complicated and normal pregnancies including variables such as maternal age, BMI, weight before pregnaney, weight gain during pregnancy, smoking

Maternal Second-Trimester Hemoglobin-Concentration / Von Tenifelhojf et al

23

hemotobln [g/dL]

13,5

ROC-curve "Preeclampsia

p < 0.0001 vs normal pregnancy 25-75 % Inter quartlle < means medians

Figure 1. Mean maternal hemoglobin concentrations during second trimester in women with normal and complicated pregnancies. IUGR, intrauterine growth retardation; SSW, gestalional week.
1 -Specificity Cut-off: 12.05 g/dL Sensitivity: 70.0% Specificity: 59.0 Cut-off: 12.75 g/dL Sensitivity: 42.0 % Specificity; 83.0 Cut-off: 13.05 g/dL Sensitivity: 34.0 % Specificity: 91.0

[percent]

Figure 3. Receiver operating characteristic (ROC) curve for mean second trimester hemoglobin concentration and a diagnosis ot preeclampsia at tbe time of delivery, Sensitivity and specificity calculations using 3 different cutoff values are shown.

ig/dL]

' normal pregnancy lUGH

^ B CD

pre-eclampsia early birth< 37 w

^M

birih weight < 2500g

Figure 2. {distribution of the mean maternal hemoglobin concentration during second trimester in women with normal (line) and complicated pregnancies (beams). IUGR, Intrauterine growth retardation.

habit, presence of proteinuria (>300 mg/24 h), increased diastolic or systolic maternal blood pressure (BP^^,^^^,.^. >140 mm Hg and/or BP.,.^^^^,,,,^, >90 mtn Hg), and mean maternal hemoglobin concentration during the second trimester. Variables were treated as continuous. All tested variables except smoking and the number of consumed cigarettes per day were significant risk factors for tbe occurrence of preeclampsia (Fig. 4A). When entered in a multivariate analysis, mean maternal second trimester hemoglobin (OR, 1.48; 95% CI, 1.291-1.699; P < .0001),

mean second trimester systolic BP (OR, 5.854; 95% CI, 4.285-7.996; P < .0001), mean second trimester diastolic BP (OR, 7.672; 95% CI, 5.608-10.496; P < .000i), maternal age (OR, 1.033; 95% CI, 1.0081.058; P = .009), and proteinuria during second trimester (OR, 1372; 95% CI, 1.058-1.780; P = .01) were statistically significant independent prognostic markers for the development of preeclampsia. in tbe univariate analysis, increasing mean maternal hemoglobin concentration during the second trimester, pathological BP at that time, status as a smoker, the number of cigarettes per day, maternal age, and a lower maternal weight before pregnancy were statistically significant risk factors for delivery of a low birth weight newborn (Fig. 4B). In tbe multivariate calculations, mean maternal second trimester hemoglobin (OR, 1.241; 95% CI, 1.166-1.381; P < .0001), mean second trimester systolic BP (OR, 2.2; 95% CI, 1.557-3.107; P < .0001), mean second trimester diastolic BP (OR, 3.451; 95% CI, 2.4374.888; P < .0001), weight before pregnancy (OR, 0.966;

24

Clinical and Applied Thrombosis/Hemostasis / \o\. 14, No. 1, January 2008

Hemoglobin II trim. Weight before pregn. Proteinuria II trim. RRsys>140lltrlm. RR dias > 90 II trim. Smoking Number of cigVcJay BMI (Kg/m2} Chi - Squfire iower risk higher risk

B
Hemoglobin il trim. Weight before pregn. Proteinuria il trim. RR sys> 140 il trim. RR dias > 90 il trim. Smoking Number of cig./day Age BMi {Kg/m2) Chi - Square *: p<0.0()01 lower risk higher risk

Maternal Second-Trimester Hemoglobin-Concentration / Von Tempelhoff et al

25

e
Hemoglobin II trim. Weight before pregn. Proteinuria II trim. RR sys > 140 II trim. RRdias>90 II trim. Smoking Number of cig./day Age BMI (Kg/m2)
Chi Square : p<0.0001

lower risk

higher risk

Hemoglobin II trim. Weight before pregn. Proteinuria II trim. RR sys > 140 H trim. RR dias > 90 II trim. Smoking Number of cig./day Age BMI (Kg/m2)
Ch) - Square p<0.0001

lower risk

higher risk

I igure 4. Logistic regression analysis for the (A) development of preeclampsia, (B) debvery of a low birth weight newborn, (C) delivery ol a newborn with intrtiutcrine growth retardation, and (D) early birth including maternal age. body mass index (BMI). weight before pregnancy, weight gain during pregnancy, smoking habit, presence of proteinuria (>3()0 mg/24 hours), increased diastolic and systolic maternal blood pressures (>I40 and/or >90 mm Hg, respectively), and mean maternal bemoglobin concentration dtiring tbe second trimester (11 trim.). Variables were treated as continuous (box plots, 9S% confidence interval; lines, odds ratios). RR, blood pressure.

26

Clinical and Applied Thrombosis/Hemostasis / Vol. 14, No. 1, January 2008

95% Cl, 0.951 -0.984; P < .0001), maternal age (OR, 1.044; 95% Cl, 1.005-1.024; P = .005). and smoking (OR, 1.748; 95% Cl, 1.177-2.597; P = .006) were statistically significant independent prognostic markers for the delivery of low birth weight children. The risk for delivery of a growth-retarded newborn was significantly associated with increasing mean maternal hemoglohin concentration and pathological BP during the second trimester, smoking, the numher of cigarettes per day, and a lower weight hefore pregnancy (Fig. 4C). Mean maternal second trimester hemoglobin (OR, 1.352; 95% Cl, 1.175-1.556; P < .0001). weight hefore pregnancy (OR, 0.967; 95% Cl, 0.955-0.979; P < .0001). mean second trimester diastolic BP (OR. 2.790; 95% Cl, 1.769-4.402; P < .0001), smoking (OR. 2.216; 95% Cl, 1381-3.558; P = .001), and mean second trimester systolic BP (OR, 2.138; 95% Cl, 1.366-3.346; P = .003) were statistically significant independent prognostic markers for delivery of a growth-retarded child. Increasing mean hemoglohin concentration, a pathological BP during the second trimester, smoking, and maternal age were significant risk factors for early hirth according to the univariate analysis (Fig. 4D). An older maternal age (OR, 1.039; 95% Cl, 1.020-1.058; P < .0001), mean second trimester diastolic BP (OR, 2.729; 95% Cl, 1.929-3.862; P < .0001). mean second trimester hemoglohin (OR, 1.210; 95% Cl, 1.087-1.347; P = .0005), mean second trimester systolic BP (OR, 1.831; 95% Cl, 1.297-2.585; P = .0006). and the numher of cigarettes per day (OR, 1.032; 95% Cl, 1.004-1.361; P = .03) were independent prognostic markers.

Discussion
In a number of large and well-designed trials, a high hemoglohin concentration early during pregnancy was followed hy a higher rate of adverse outcome events, such as stillhirth,'^'"' early hirth,"'' delivery of a growth-retarded infant,'**'^ delivery of a low-weight newhorn,"*'^"'" or preeclampsia.'^'^ Rasmussen and Olan'' found an independent inverse correlation hetween second trimester hemoglobin values and fetal hirth weight as well as a strong positive correlation hetween first to second trimester decreases in hemoglohin and fetal hirth weight. In a retrospective trial of 54 382 singleton pregnancies, perinatal mortality and the incidence of low hirth weight newhorns as well as early hirth deliveries

were strikingly higher in women whose hemoglohin concentrations at 13 to 19 weeks of gestation were greater than 13.2 g/dL compared with those who had a value between 10.4 and li.2 g/dL. Forty-two percent of primiparae women with hemoglohin values greater than 14.5 g/dL had hypertensive disorders, a number that was significantly higher than the 7% incidence found in women who had mean second trimester hemoglobin values less than 10.5 g/dL (P< .001).' Based on more than 34 800 hemoglobin estimations from 4985 consecutive women with singleton pregnancies, women with adverse outcome events (namely preeclampsia, IUGR. early birth, or delivery of a low weight newhorn) had significantly higher mean hemoglohin concentrations during the second trimester compared with women whose pregnancy remained uneventful. The risk of development of each adverse outcome event increased with a rising mean hemoglohin concentration. Moreover, mean hemoglohin value during the second trimester was an independent prognostic marker for subsequent diagnosis of all 4 adverse outcome events. At this time, the highest mean hemoglobin concentrations were calculated in patients with suhsequent preeclampsia. Mean hemoglohin concentrations were independent of maternal age. BMI, or smoking hahit of women. Suhgroup analysis revealed no correlation hetween high hemoglohin and pathological hlood pressure, proteinuria, or gestational age at term in a complicated pregnancy. The hnk hetween a high hemoglohin concentration early during pregnancy and the development of a complicated pregnancy is unclear. A high hemoglohin concentration during the second trimester may he a symptom of ineffective plasma volume expansion'^ or the result of increased erythropoiesis attrihutahle to a compensatory mechanism caused hy impaired oxygenation in the placenta''^ and/or the release of placental factors that promote erythropoiesis. such as activin A."" This may have a profound effect on the maintenance of the uteroplacental circulation.^' because concentrations are positively correlated unth perivillous fihrin deposition and placenta! infarction. Moreover, an early increase in progesterone concentration, as well as a high progesterone/estradiol ratio with less expansion in plasma volume from weeks 14 to 17 until term, may he of concern in the development of preeclampsia.^ Women who received iron supplementation had no higher rate of adverse outcome events in our

Maternal Second-Trimester Hemoglobin-Concentration / Von Tempelhoffet al 27 trial. Although correction of iron deficiency in pregnancy^which is a common cause oF anemia in developing countriescan produce a significant increase in hemoglobin concentrations, it does not increase hemoglobin higher than the optimal concentration needed for oxygen delivery,^^ Therefore, iron supplementation is not likely a cofactor for higher hemoglobin concentrations in patients with adverse outcome events during pregnancy. In most developing countries, anemia is a leading cause (or symptom) in women with adverse outcome events during pregnancy. Such low hemoglobin concentrations in pregnant women often are tbe consequence of common environmental influences and conditions"^ (eg, malnutrition, comorhidity, injury, or inadequate management of bleeding complications), which themselves potentially interfere with tbe normal outcome of pregnancy. Apart from very low hemoglobin concentrations (<7.0 g/dL), a number of trials found an association between anemia (<1I.O g/dL) and adverse pregnancy outcomes, whereas other studies could not confirm such an association,-"*^^ Although our trial does not contribute to understanding the role of high hemoglobin and development of adverse outcome events during pregnancy, our results demonstrate that hemoglobin is an important determinant for the risk calculation of such complications and preeclampsia in particular, ROC plots revealed rather low sensitivity and specificity of ranked single hemoglobin values for the prediction of adverse outcome eventseven for the diagnosis of preeclampsia. However, approximately one third of patients with a complicated pregnancy compared with only 10% of normal pregnancies had mean hemoglobin concentrations greater than 13,0 g/dL during the second trimester; ihus, in this subgroup of pregnancies, high hemoglobin is at least a cofactor for an increased risk of complicated pregnancies. In contrast, the rate of normal pregnancy outcomes was highest with mean hemoglobin values during the second trimester between I 1,0 and 12,0 g/dL (40%). The frequency of complicated as well as uneventful pregnancies was not different with values below this range (apart Irom early birth rate), and therefore such hemoglobin concentrations signal optimal outcome rate in women of this trial. Although the hemoglobin increase during the third trimester and near term has also been shown to correlate with the rate of poor maternal and fetal outcome, this is of little concern with respect to the therapeutic options."~ From a clinical point of view, it seems mandatory to monitor these women carefully and more frequently, especially for the development of preeclampsia, whereas hemodilution therapy has been a successful approach for prolongation of pregnancy.^**

References
1. Pirani BBK, Campbell DM, MacGilivray I. Plasma volume expansion during normal first pregnancy, J Obstet Br Commons. 1973;80:804-808. 2. Hytten F, Blood volume changes in normal pregnancy. Clin Hematol. 1985:85:404-411. 3. Letsky E. The hematological system. In: Hytten F, Chamberlain G, eds. Clinical Physiology in Obstetrics. Oxford, UK: Blackwell Scientific; l99]:.^9-82, 4. Little MP, Brocard P, Eliott P, Steer PJ. Hemoglobinconcentration in pregnancy and perinatal morlalit\: a London-based cohort study. Am J Ohstet Gynecul. 2005:193:220-226. 5. Steer PJ, /\lam MA, Wadsworth J, Welch A, Relation between matema! haemoglobin concentration and hirth weight in differenl ethnic groups. BMj. 1995;310:489-491. 6. World Health Orgiinization Expert Committee un Nutrition in pregnancy and lactation; report of a WHO Expert Committee, World Health Organ Tech Rep Ser. I96S;302;5-19, 7. Murphy JF, O'Riordan J, Newcombc RG. Coles EC, Pearson JF, Relation of hemoglobin levels in first and second trimesters to outcome of pregnancy. La}icet. 1986;l:992-995. 8. Salas SP, Marshall G. Gutierrez BL, Rosso P. Time course of maternal plasma volume and hormonal changes in women with preeclampsia or fetal growth restriction. Hypertension. 2006;47:203-208. 9. Rasmussen S, Olan P. First- and second-trimester hemoglobin levels. Acta Obstet Gynecol Scand. 1993;72: 246-256. 10. Sagen N, Nilsen ST, Kim HC, Bergsjo P, Koller O. Maternal hemoglobin-concentration is closely related to hirth weight in normal pregnancies. Acta Obstet Gynecoi Scand. I984;63:245-248. 11. Hemminki E, Rimpela U. Iron supplementation, maternal packed cell volume, and fetal growth. Arch Dis Child. 1991:66:422-425. 12. Garn SM, Ridella SA, Petzold AS, Falkner F. Maternal hematologic levels and pregnancy outcomes, Setnin Permaiol. 198 I ;5:1 55-162. 13. Nordenvall M, Sanstedt B. Plaeental lesions and maternal hemoglobin levels: a comparative investigation. Acta Obstet Gynecol Scand. I99O;69:I27-133. 14. Sibai BM, Abdella TN, Anderson GD, MCCuhbin JH. Plasma volume determination in pregnancies complicated

28

Clinical and Applied Thrombosis/Hemostasis / Voi. 14, No. I, January 2008

15.

16.

17.

18.

19.

20.

21.

by chronic hypertension and intrauterine fetal demise, Obstet Gynecol. 1982:60:174-178. Stephansson O, Dickman PW, Johansson A, Cnattingius S. Maternal hemoglobin-concentration during pregnancy and risk of stillbirth. JAMA. 2000;284:2611 -2617, Lao I'V, Tam K-F, Chan LY. Third trimester iron status and pregnancy outcome in non-anaemic women; pregnancy unfavourably affected by maternal iron excess. Hum Reproditd. 2OOO;I5,1843-1848. Lu ZM, Goldenberg RL, Oliver SP, Cutter G, Blankson M, The relationship between maternal hematocrit and pregnancy outcome. Obstet Gynecol 199 I ;77:190-194. Uosso P, Donoso E, Braun S, Espinoza R, Fcrnadez C, Salas SP. Maternal hemodynamic adjustment in idiopathic fetal growth retardation. Gynecol Obstet Invest. I993;35:I62-165. Ortner A, Zech H, Humpeler E, Mairbaeurl H, May high oxygen affinity of maternal hemoglobin cause fetal growth retardation?. Oynecology. 1983;234:79-85. Troeger C, Holzgreve W, Ladewig A, Zhong XY, Hahn S. Examination of maternal plasma erythropoietin and activin A concentrations with regard to circulalorv' crythroblast levels in normal and preeclamptic pregnancies. Fetal Diagn Ther. 2OO6;21:I66-16O. Daniel SS, James LS, Stark Rl, Iropper PJ. Prevention of the normal expansion of maternal plasma volume: a model for chronic fetal hypoxaemia. J Dev Physiol. 1989;! 1:225-233.

22, Barton DP, Joy MT, Lappin TR, et al. Maternal erythropoietin in singleton pregnancies: a randomized trial on the effect of oral hematinic supplementation. Am J Obstet Gynecol. 1994;170:896-901. 23, Bodnar LM. Scanlon KS, Freedman DS, Siega Riz AM, Cogswell ME. High prevalence of postpartum anemia among low-income women in the United States. Am J Obstet Gynecol. 2001; 185:438-443, 24, Rush D. Nutrition and maternal mortality in the developing world. Am J Glin Nutr. 2000;72(suppl); 212S-240S, 25, Meis PJ. Miehielutte R, Peters TJ, et al. Factors associated with prcterm birth in Cardiff, Wales. I. Univariabie and muhivariable analysis. Am J Obstet Gynecol. 1995;172:590-596. 26, Meis PJ, Miehielutte R, Peters TJ, et ak Factors associated with preterm birth in Cardiff, Wales. 11. Indieated and spontaneous preterm hirth. Am J Obstet Gynecol. 1995:172:597-602. 27, Ganzevoort W, Rep A, Bonsel GJ, De Vries JI, Wolf H, PETRA investigators. A randomized trial of plasma volume expansion in hypertensive disorders of pregnancy: influence on the pulsatility indices of the fetal umbilical artery and middle cerebral artery. Am J Obstet Gynecol. 2005; 192:233-239. 28, von Tempelhoff G-F, Heilmann L. The effect of plasma volume e.xpansion on the utero-placental perfusion. Clin Hemorheol. 1993; 19:729-736.