Dr S Palomba, Department of Obstetrics & Gynaecology, University ‘Magna Graecia’ of Catanzaro, Via T. Campanella 182/1, 88100 Catanzaro, Italy. Email email@example.com
Please cite this paper as: Palomba S, Falbo A, Russo T, Battista L, Tolino A, Orio F, Zullo F. Uterine blood flow in pregnant patients with polycystic ovary syndrome: relationships with clinical outcomes. BJOG 2010; DOI: 10.1111/j.1471-0528.2010.02525.x.
Objective To study the impedance to blood flow through the uterine artery in pregnant women with polycystic ovary syndrome (PCOS), and to evaluate its predictive value for adverse pregnancy and perinatal outcomes in this population.
Design Prospective case-control study.
Setting Academic Departments of Obstetrics and Gynaecology in Italy.
Population Seventy-three pregnant women with ovulatory PCOS (PCOS group) and 73 age- and body mass index-matched healthy pregnant controls (control group).
Methods Serial Doppler velocimetry measurements of the uterine artery.
Main outcome measures Blood flow impedance indices and pregnancy/perinatal outcomes.
Results A significantly (P < 0.05) higher rate of subjects with abnormal velocimetry findings was observed in the PCOS group than in the control group. In the PCOS group, the pulsatility index (PI) at first (P = 0.042) and mid-second (P = 0.039) trimesters of pregnancy, and bilateral notch at first (P = 0.025) and mid-second (P = 0.007) trimesters of pregnancy, were the strongest independent predictors of adverse outcomes. Conversely, in the control group, PI at the first trimester of pregnancy was a predictor of adverse outcomes only when combined with bilateral notch (P = 0.042), whereas at mid-second trimester of pregnancy PI (P = 0.033) and bilateral notch (P = 0.048) were independent predictors of adverse outcomes.
Conclusions Uterine artery Doppler indices are more commonly altered in pregnant patients with PCOS than in controls, showing a high predictive value for abnormal pregnancy/perinatal outcomes.
Polycystic ovary syndrome (PCOS) is one of the most common endocrine disorders in women of childbearing age.1,2 The impact that the syndrome has on human reproduction does not only include anovulatory infertility, but also detrimental effects on oocyte3–6 and embryo7 quality, endometrial receptivity,8,9 and pregnancy development.10
Retrospective11–15 and prospective16–18 noncontrolled data designated PCOS as a risk factor for increased incidence of complications throughout pregnancy. A recent meta-analysis10 confirmed an increased risk for gestational diabetes mellitus (DM), pre-eclampsia (PE), pregnancy-induced hypertension (PIH), and preterm birth. In addition, babies born from mothers with PCOS seem to have a significantly higher risk of admission to neonatal intensive care units and perinatal mortality, unrelated to multiple births.10
Pregnancy complications seem to be characterised by a failure of trophoblastic invasion into the musculoelastic coat of the spiral arteries, resulting in incomplete vascular transformation and persisting increased impedance of the uterine arteries.19 In fact, an inverse relationship between vascular impedance at uterine arteries and percentage of trophoblastic vessels has been demonstrated, and incomplete spiral arteries invasion has been clearly associated with PE and fetal growth restriction (FGR).20
An increase in blood flow resistance in the uterine arteries and a reduction in subendometrial and endometrial vascularity have been recently demonstrated in young nonpregnant women with PCOS.8,9 Although no clear data are available in the literature regarding the potential pathogenic mechanism, failure of trophoblastic invasion of the musculoelastic coat of uterine spiral arteries and abnormal impedance in the uterine arteries might be suggested in women affected by PCOS in order to explain the higher pregnancy complication rates observed.21–23
Based on these considerations, the goals of the current prospective, case-controlled study were firstly to study the impedance to blood flow through the uterine artery during the first and mid-second trimester of pregnancy in women with PCOS, and secondly to correlate ultrasonographic indices of impedance to blood flow through the uterine artery with their pregnancy/perinatal outcomes.
The procedures used in the present study were in accordance with the Helsinki Declaration on human experimentation guidelines. The study was approved by the Institutional Review Board. The purpose of the protocol was carefully explained to all women before entering the study, and their written consent was obtained.
Between February 2003 and April 2008, 97 primigravid women who were suffering from PCOS were consecutively screened at two academic Departments of Obstetrics and Gynaecology (in Italy) in order to study pregnancy and neonatal outcomes in a PCOS population. Of these, 73 subjects with ovulatory PCOS (PCOS group) were enrolled in the study protocol. Women with non-ovulatory PCOS were excluded from the present study, whereas they were included in other protocols reported elsewhere.24 Ovulatory PCOS was diagnosed before pregnancy according to the presence of polycystic ovaries and clinical/biochemical hyperandrogenism without chronic oligo-anovulation.25 The diagnosis of PCOS was also confirmed at study entry.
In addition, 73 age- and body mass index (BMI)-matched healthy primigravidas, seen in our department for pre-conceptual counselling, were enrolled and considered as controls (control group). The matching procedure was one-to-one, and women were defined as age- and BMI-matched when the differences between them were less than 2 years and 1 kg/m2 for age and BMI, respectively. The same group was considered as a control group in another study.24
The health of women in the control group was determined by their medical history, a physical and pelvic examination, and complete blood chemistry. All patients in the control group had regular menstrual cycles (26–32 days in length), no signs of clinical hyperandrogenism, normal ranges of serum androgen levels, and no polycystic ovary morphology in transvaginal ultrasonography.
For cases and controls, the following exclusion criteria were considered: age >35 years; obesity (BMI > 30 kg/m2); multiple pregnancy; gestational age of greater than 7 weeks, as assessed by the crown–rump length (CRL) measurement; premalignancies or malignancies; medical conditions or other concurrent medical illnesses; cigarette smoking; drug/alcohol use; organic pelvic disease; uterine malformations; previous pelvic surgery; women noncompliant with our study protocol; and current or previous (within the last 6 months) use of any hormonal and/or antidiabetic drugs. Previous infertility treatments were considered as elective exclusion criteria in order to avoid any bias resulting from previous drug use or from sperm abnormalities. We also excluded women who intended to start a diet or a specific program of physical activity.
At study entry and throughout the study period, all women received folic acid (0.4 mg daily), and were instructed to follow their usual diet and physical activity.
Each woman was monitored throughout pregnancy with clinical, biochemical, and ultrasonographic assessments, as detailed below.
Clinical and ultrasonographic assessments were performed at study entry (up to 7 weeks of gestation), every 2 weeks for three times (at 8, 10 and 12 weeks of gestation), and every 4 weeks until delivery.24 Gestational age was calculated from the last menstrual period, and was confirmed by first-trimester ultrasound (CRL measurement).24
Clinical evaluation consisted of obstetric examination, Papanicolaou smear test (at study entry alone), Ferriman–Gallwey score,26 anthropometric measurements (including height, weight, BMI, and waist-to-hip ratio, WHR), and heart rate and blood pressure assessments.
During the same visit, a semi-quantitative questionnaire to evaluate physical activity, and job and daily activities was completed by each woman.27 Socio-economic, work, and educational status, ethnicity, and associated medical conditions were carefully assessed for each woman.24 Finally, women were asked to complete a questionnaire on family history of PE, DM, or complicated pregnancies.24
Biochemical assessment was performed for each woman. Specifically, liver and renal function, complete blood count, and serum glucose levels were assessed monthly, whereas a complete hormonal assay was evaluated at study entry. A complete urine assay was also performed monthly.
At study entry, glucose and insulin concentrations were measured at basal and after a 2-hour oral glucose tolerance test (OGTT). The glucose and insulin response to OGTT was analysed by calculating the area under the curve (AUC), and the AUCglucose/AUCinsulin ratio was calculated for each woman.28
The homeostasis model of assessment (HOMA) [fasting glucose (mmol/l) × fasting insulin (μU/ml)/22.5],29 the fasting glucose-to-insulin ratio (GIR) (mg/10−4 U),30 and the free androgen index (FAI) [total testosterone (nmol/l)/sex hormone-binding globulin (SHBG) × 100] were also calculated for each woman.31
During the study, all subjects underwent ultrasonographic evaluations. All exams were performed using an Aplio (Toshiba Medical System, Rome, Italy) equipped with 7.5-MHz transvaginal and 5-MHz transabdominal curvilinear probes. Ultrasonography was performed transvaginally until 12 weeks of gestation, and, thereafter, was performed transabdominally.
During the first 12 weeks of gestation, the embryo heartbeat was recorded, the CRL was carefully measured, and blood flow through the uterine arteries was assessed. Successively, the ultrasonographic evaluations consisted of measurement of fetal growth, placenta location and grade (described as grades 0, I, II, or III, according to the Grannum scheme), amniotic fluid index, and umbilical artery pulsatility index (PI), when required.
In each patient, a detailed ultrasound scan for the detection of fetal malformations was performed at 20 weeks of gestation, and screening for gestational DM was performed at 26 weeks of gestation, according to the guidelines of the American Diabetes Association.32
Pregnancy/perinatal outcomes were also evaluated. In particular, for each woman miscarriage, gestational DM32, PIH33, PE33, antepartum haemorrage24, gestational age at delivery, type of delivery (instrumental, including forceps and/or vacuum extraction, or Caesarean section), fetal growth (evaluated by serial ultrasonographic assessments and classifying the fetus as appropriate for gestational age [AGA], small for gestational age [SGA], and large for gestational age [LGA], according to reference standards), birthweight, Apgar score, fetal malformations, and intrauterine deaths were recorded.
Assessment of impedance to blood flow through the uterine arteries
Uterine artery velocimetry was assessed serially during the first (at 8, 10, and 12 weeks of gestation) and mid-second (at 20 weeks of gestation) trimesters, according to standardised procedures.
For transvaginal ultrasonographic assessments, women were placed in a lithotomy position, with an empty bladder. A sagittal section of the uterus and cervical canal was obtained, and the cervical canal was identified. The probe was then moved gently laterally until the paracervical vascular plexus was seen. Colour flow was turned on and the uterine artery was identified at the level of the cervico–corporeal junction. For transabdominal examinations, women were placed in the lithotomy position, with a full bladder, and a transabdominal transducer was placed transversely on the lower quadrants of the abdomen. A pulsed Doppler gate was placed over the vessel to identify the apparent crossing of uterine and external iliac vessels to obtain flow velocity waveforms from the ascending branch of the uterine artery, before it branches into the arcuate arteries. Angle correction was then applied, ensuring that the angle was <60°, and the signal was updated until at least three similar consecutive flow velocity waveforms of good quality were obtained.
During every ultrasound examination, the presence or absence of an early diastolic notch, defined as a definitive upward change in velocity after the initial deceleration slope of the primary wave, was noted. The PI and resistance index (RI) were also automatically generated for each side, and the mean of the two measurements was calculated for each woman.
Cases and controls were defined as having abnormal uterine artery blood flow in the presence of a unilateral or bilateral notch and/or a PI value higher than the 95th percentile of the reference range for our population at that gestational age.
All scans were performed by the same experienced operator (TR), and therefore only the intra-operator variability for each measurement was calculated. In particular, the intra-operator coefficient of variation (CV) was calculated comparing a set of measurements in 20 women and repeating them in the same conditions after 10 minutes. The intra-operator CVs were less than 10% and 0% for PI values and notch detection, respectively.
Correlation between impedance in the uterine arteries and pregnancy/perinatal outcomes
Pregnancy/perinatal outcomes were categorised in normal and pathological cases. A normal outcome was defined as delivery at term of an AGA fetus (ranged between the 10th and 95th percentile) without gestational DM, PIH, PE, and/or antepartum haemorrhage.
As detailed in the section ‘Statistical analysis’, logistic regression was used for the analysis of the multivariate relations of adverse pregnancy/perinatal outcomes by one or more parameters of blood flow waveform in the uterine arteries, assessed during the first (12 weeks of gestation) and mid-second (20 weeks of gestation) trimester of pregnancy.
Categorical variables were compared by using Pearson’s chi-squared test; Fisher’s exact test was used for the frequency tables when more than 20% of the expected values were less than five.
The normal distribution of continuous variables data was evaluated with the Kolmogrov–Smirnov test. Thus, our data were expressed as medians and interquartile ranges (IQRs), with minimum and maximum values, and the differences between groups were analysed using the Mann–Whitney U-test. The Friedman test was used for repeated measures analysis with Dunn’s test as the post hoc analysis.
Logistic regression was used for the analysis of multivariate relations of one dependent variable by one or more factors. The model included the presence/absence of adverse pregnancy/perinatal outcomes, consisting of miscarriage, gestational DM, PIH, PE, antepartum haemorrage, FGR, instrumental and Caesarean delivery, delivery before 37 weeks of gestation, 5-minute Apgar score <6, fetal malformations, and intrauterine deaths, as dependent variables. In particular, we evaluated each adverse pregnancy/perinatal outcome per woman by considering each woman as a unit coded with 1 or 0, respectively, to indicate whether they had or had not, respectively, an adverse outcome during the observation period.
Backward stepwise elimination was used for the multivariate logistic analysis of prediction of any adverse event for independent variables consisting of PI values at uterine arteries at the first (12 weeks of gestation) and mid-second (20 weeks of gestation) trimester, and for the presence of a unilateral or bilateral notch at the same assessments.
As uterine artery PI was not normally distributed by the Kolmogorov–Smirnov test, PI values were log-transformed before entering them into the logistic model.
A P < 0.05 was used as the cutoff level for the elimination of nonsignificant predictors from the model.
Sensitivity, specificity, and positive predictive value (PPV), negative predictive value (NPV), and relative risk (RR), with 95% confidence interval (CI), were calculated for each significant predictor. Sensitivity was defined as the percentage of women with adverse outcomes who had a positive test result. Specificity was defined as the percentage of women with a normal pregnancy and perinatal outcome that had a negative test result. PPV and NPV were defined as the percentage of women with positive and negative test results, respectively, who were correctly diagnosed.
Statistical significance was set at P < 0.05. A statistical trend was arbitrarily established for P values of between 0.05 and 0.07.
We used spss 14.0.1 (SPSS Inc., Chicago, IL, USA) for all statistical analyses. Sensitivity, specificity, PPV, NPV, and RR were calculated with StatsDirect 2.4.3.
The drop-out rate was similar in the two groups (three and four women in the PCOS and control groups, respectively). These women were excluded because they missed their first follow-up visit and their clinical data were unavailable for the final analysis.
The main clinical and biochemical data at baseline from PCOS and control groups are shown in Tables 1 and 2. In particular, two groups were significantly different in WHR and Ferriman–Gallwey scores. Similarly, significant differences were observed between groups in testosterone, androstenedione, dehydroepiandrosterone sulphate, SHBG, and fasting insulin levels. FAI, GIR, HOMA, AUCglucose, AUCinsulin, and the AUCglucose/AUCinsulin ratio was calculated for each woman, and were also significantly different between groups. No other significant difference in any clinical and biochemical data was detected between PCOS and control groups.
Table 1. Clinical data for cases (PCOS group) and controls (control group) at baseline
PCOS group (n = 70)
Control group (n = 69)
Data are expressed as medians (IQR; range). BMI, body mass index; DBP, diastolic blood pressure; HR, heart rate; SBP, systolic blood pressure; WHR, waist-to-hip ratio.
29.5 (5; 20–33)
30 (7; 19.0–34.0)
24.1 (3.5; 18.0–29.4)
24.0 (3.4; 17.8–29.4)
0.81 (0.17; 0.64–0.93)
0.77 (0.07; 0.61–0.85)
11 (3; 8–14)
3 (2; 0–7)
84 (15; 68–110)
84 (16; 64–105)
122 (20; 96–133)
124 (18; 96–133)
72 (9; 62–83)
72 (9.5; 60–82)
Table 2. Main hormonal and metabolic data for cases (PCOS group) and controls (control group) at baseline
PCOS group (n = 70)
Control group (n = 69)
Data are expressed as medians (IQR; range). The biochemical assays are reported in metric units. DHEAS, dehydroepiandrosterone sulphate.
2.1 (1.4; 1.1–4.1)
0.9 (0.3; 0.7–3.0)
4.3 (4.1; 1.4–7.3)
1.7 (0.9; 0.9–3.8)
Dehydroepiandrosterone sulphate (ng/ml)
2689.9 (362.1; 2397.6–2998.9)
1712.4 (164.2; 1589.7–2198.0)
18.0 (5.0; 14.0–27.0)
42.0 (5.7; 17.0–56.0)
13.4 (7.4; 5.0–23.3)
4.2 (2.3; 2.4–6.3)
Fasting glucose (mg/dl)
82.5 (9.8; 65–98)
77 (18; 60–93)
Fasting insulin (μU/ml)
13.9 (3.7; 10.8–18.6)
5.9 (1.3; 3.6–9.7)
GIR (mg/10−4 units)
4.6 (1.1; 2.8–5.6)
4.8 (0.8; 4.0–5.6)
16.0 (7.8; 11.5–24.5)
14.0 (4.9; 10.9–17.4)
AUCglucose (mg/dl/120 min)
1033 (101.5; 879–1321)
1011 (79; 879–1199)
AUCinsulin (μU/ml/120 min)
8766 (3233; 4988–11909)
4592 (3744; 1988–8766)
0.1 (0.1; 0.1–0.3)
0.2 (0.2; 0.1–0.6)
A total of 22 out of 70, and eight out of 69 from the PCOS and control group, respectively, had adverse pregnancy/perinatal outcomes. The cumulative rate of women with adverse outcomes was significantly higher in the PCOS group than in the control group (31.4 versus 11.6%; P = 0.005).
Miscarriage, PIH, PE, gestational DM, and antepartum haemorrhage were significantly more frequent in the PCOS group than in the control group (Table 3). A significant difference (P < 0.001) between groups was detected in SGA, LGA, and AGA distribution. In this regard, birthweight was significantly lower in women with PCOS than in controls (Table 3; P = 0.006). A trend (P = 0.059) towards a higher instrumental and Caesarean delivery rate in the PCOS group in comparison with the control group was detected. In particular, in the PCOS and control groups, the Caesarean section rates were 31.4 (22/70) and 23.2% (16/69), and use of forceps and/or vacuum extraction were 7.1 (5/70) and 1.4% (1/69), respectively (Table 3).
Table 3. Main pregnancy and perinatal outcomes in cases (PCOS group) and controls (control group)
PCOS group (n = 70)
Control group (n = 69)
DM, diabetes mellitus; FGR, fetal growth restriction; PE, pre-eclampsia; PIH, pregnancy-induced hypertension. The total percentage is more than 100 for the coexistence of two or more conditions.
Miscarriage, n (%)
PIH, n (%)
PE, n (%)
Gestational DM, n (%)
Antepartum haemorrhagia, n (%)
Gestational age at delivery (weeks)
39 (2 IQR; 32–41 range)
39 (2 IQR; 37–40 range)
Preterm delivery, n (%)
Type of delivery
3100 (530 IQR; 1800–4000 range)
3300 (700 IQR; 2120–4100 range)
NGA, n (%)
LGA, n (%)
SGA, n (%)
FGR, n (%)
10 (1 IQR; 4–10 range)
10 (0.5 IQR; 5–10 range)
Five-minute Apgar score
Fetal malformations, n (%)
No differences between groups were observed in gestational age at delivery, preterm delivery, FGR, Apgar score, and fetal malformations. In one case (PCOS group; 1/70, 1.4%), an placental abruption was observed, whereas no intrauterine and/or neonatal deaths were observed in either group.
Finally, no differences between groups were recorded in placenta location and grade distribution, amniotic fluid index, and mean umbilical artery PI (data not shown).
Impedance in the uterine arteries
There were no differences observed between PI and RI values, thus only PI values are shown.
The PI values were significantly higher in the PCOS group than in the control group at each follow-up visit (P < 0.001). In the PCOS group, a significant reduction in PI was observed at 20 weeks of gestation, in comparison with the previous follow-up visits. On the other hand, the control group already presented a significant (P < 0.001) reduction in PI at 12 weeks of gestation in comparison with previous assessments, and a further significant (P < 0.001) reduction at 20 weeks of gestation was observed (Figure 2).
A significantly higher percentage of women in the PCOS group had abnormal PIs in comparison with women in the control group at each time point (Table 4). A significant difference (P < 0.001) between groups was detected in women without a bilateral notch at each assessment. At 20 weeks of gestation, the presence of a unilateral or bilateral notch was significantly (P < 0.001) higher in the PCOS group than in the control group (Table 4).
Table 4. Serial ultrasonographic data of uterine artery velocimetry in cases (PCOS group) and controls (control group)
PCOS group (n = 70)
Control group (n = 69)
*P < 0.001 versus control group; **P < 0.05 versus control group; ***P < 0.001 versus previous assessments; ****P < 0.05 versus previous assessments.
Abnormal PI, n (%)
No notch, n (%)
Unilateral notch, n (%)
Bilateral notch, n (%)
The proportion of women in the control group with unilateral or bilateral notches changed throughout gestation, whereas a significantly lower percentage of subjects without a notch was detected in the PCOS group at each assessment. In particular, the proportion of controls with unilateral or bilateral notches was significantly (P < 0.05) lower at 20 weeks of gestation in comparison with all previous ultrasonographic examinations (Table 4).
Correlation with pregnancy/perinatal outcomes
Using the backward elimination procedure, PI values at uterine arteries in the first trimester of gestation and presence of unilateral or bilateral notch at the same assessments were selected as significant variables to be included in the final model.
The final step of the multivariate analysis showed that, in the PCOS group, PI in the first (OR: 2.98, 95% CI: 1.12–5.39; P = 0.042) and mid-second (OR: 3.2, 95% CI: 1.41–7.69; P = 0.039) trimesters of pregnancy, and presence of a bilateral notch in the first (OR: 2.21, 95% CI: 1.63–6.98; P = 0.025) and mid-second (OR: 6.64, 95%: CI 4.87–13.32; P = 0.007) trimesters of pregnancy, were the strongest independent predictors for adverse outcomes.
On the other hand, in the control group, PI in the first trimester of pregnancy was a predictor of adverse outcomes only when combined with the presence of a bilateral notch (OR: 1.83, 95% CI: 1.39–3.17; P = 0.042), whereas PI in the mid-second trimester of pregnancy (OR: 2.31, 95% CI: 1.87–4.34; P = 0.033) and bilateral notch (OR: 5.22, 95% CI; 1.99–9.23; P = 0.048) were independent predictors of adverse outcomes.
Test characteristics for predicting adverse pregnancy/perinatal outcomes are detailed in Table 5.
Table 5. Test characteristics for predicting adverse pregnancy and perinatal outcomes in cases (PCOS group) and controls (control group)
To our knowledge, this is the first clinical study evaluating the impedance to blood flow through the uterine artery in pregnant women with PCOS. Specifically, we studied the changes in uterine blood flow throughout the first and mid-second trimester of pregnancy in a well-selected population of pregnant women with PCOS. In fact, in order to avoid possible biases related to the syndrome, such as fertility/hormonal treatments or obesity, we carefully selected only women with ovulatory PCOS, who were compared with age- and BMI-matched healthy controls, and excluded obese women.
Our previous preliminary data8,9 showed an alteration of impedance in the uterine arteries in anovulatory women with PCOS, suggesting functional and possible morphologic endometrial derangements that could per se impair preimplantation and early implantation phases.
In the current study, we demonstrate that in women with PCOS the alterations in the impedance to blood flow through the uterine artery are sustained during both the first and the mid-second trimester of pregnancy. In fact, PI was significantly higher in women affected by PCOS than in healthy controls at each follow-up visit. In addition, the percentages of subjects with abnormal PI remained significantly higher in the PCOS group than in the control group at each follow-up visit. Similarly, the rate of women without a notch was significantly lower in PCOS women in comparison with healthy controls at each follow-up assessment.
A recent experimental study on a small population of women who underwent pregnancy termination for nonmedical reasons,20 showed a significant relationship between decidual vessels with endovascular trophoblast invasion and resistance indices in the uterine artery, assessed by Doppler velocimetry. In particular, in women with high-resistance uterine artery blood flow at 10–14 weeks of gestation, the proportion of decidual vessels with endovascular trophoblast invasion was significantly lower when compared with women with low resistance.20 These data suggest that in women with a low-resistance uterine artery flow pattern in the early stages of pregnancy there is more extensive trophoblastic invasion of the decidual vessels compared with women with a high-resistance pattern.
Interesting findings obtained in the current study were the different pattern of changes in the impedance in the uterine arteries between women with PCOS and healthy pregnant women observed during the early stages of pregnancy, and the changes in uterine artery impedance that were detected later in pregnancy in the PCOS group. In fact, a significant reduction in uterine artery impedance was detected at each follow-up visit in healthy controls, whereas a less significant reduction in uterine artery resistance was observed only at the mid-second trimester in women affected by PCOS. These changes appear critical for the development of the fetal–placental unit, as demonstrated previously.34 Unfortunately, in order to improve the compliance of patients with the protocol, we performed the last Doppler velocimetry of uterine arteries at 20 weeks of gestation, even though the changes of impedance in the uterine arteries are normally completed only by 24 weeks of gestation.
The rate of subjects without uterine artery notches at each assessment was significantly lower in the PCOS group than in healthy controls, whereas, in the mid-second trimester of pregnancy, the rate of unilateral and bilateral notches was significantly higher in women with PCOS than in controls. At the same time, the proportion of healthy controls with unilateral and bilateral notches changed throughout time, whereas no significant change was observed in PCOS women: in the mid-second trimester of pregnancy, the proportion of healthy controls with unilateral and bilateral notch was significantly lower in comparison with previous follow-up visits. Hemodynamic changes in the uterine circulation throughout the first half of pregnancy are likely to reflect the dramatic changes occurring during the placentation process.34 In addition, differences in the sequence of changes in impedance in the uterine arteries between normal and complicated pregnancies have been demonstrated previously.34 In this regard, some authors have hypothesised that defective trophoblastic invasion of decidual vessels in the first trimester may be associated with decreased trophoblastic conversion of the spiral artery branches to uteroplacental vessels in later pregnancy, showing a correlation between uterine artery blood flow obtained in early pregnancy and subsequent development of PE and FGR, which can result from defective placentation.35,36
In the current study, we investigated if there was any stronger correlation between ultrasonographic indices of impedance to blood flow through the uterine artery, obtained in the first and mid-second trimester of pregnancy, and pregnancy/perinatal outcomes in women with PCOS compared with healthy controls. Our results confirmed that women with PCOS had a higher rate of pregnancy and perinatal complications than healthy controls.10,24 In particular, although we excluded women who were obese, were aged 35 years or older, or had multiple pregnancies, the total incidence of adverse outcomes was significantly higher in the PCOS group than in healthy women (31.4 versus 11.6%). In fact, significant differences in miscarriage, PIH, PE, gestational DM, antepartum haemorrhage, and birthweight were observed between the PCOS and control groups.
Furthermore, for the moment, it is unknown which factor, alone or combined, could adversely influence pregnancy development in PCOS. Our previous clinical data from a heterogeneous population of women with PCOS showed that the relative risk for adverse obstetric and/or neonatal outcomes varied according to the PCOS phenotype, with a higher risk for full-blown and nonpolycystic ovary phenotypes in comparison with others, and for women with oligo-anovulatory PCOS compared with ovulatory PCOS.24 Moreover, the risk for adverse outcomes was affected significantly by ovarian dysfunction and biochemical hyperandrogenism, whereas no significant effect was detected for clinical hyperandrogenism and polycystic ovaries.24
A multivariate analysis was carried out in order to identify stronger predictors for the overall adverse pregnancy/perinatal outcomes in PCOS. In particular, we investigated if abnormal uterine artery impedance could be correlated with adverse outcome in this population. In the first trimester of pregnancy, PI combined with the presence of bilateral notches was a predictor of adverse outcomes in controls (low-risk pregnant women without PCOS), whereas in the mid-second trimester of pregnancy, PI and bilateral notch were independent predictors of adverse outcomes. On the other hand, PI and bilateral notch in the first and in the mid-second trimester of pregnancy were the strongest independent predictors for adverse outcomes in PCOS. These findings were not changed by the exclusion of outcomes not associated with uterine Doppler assessments, such as fetal malformations and gestational DM, from the logistic regression model.
In a recent meta-analysis,37 an increased PI with the presence of uni- or bi-lateral and bi-lateral notch in the mid-second trimester of gestation best predicted overall PE in low- and high-risk women, respectively. On the other hand, first-trimester Doppler indices showed a limited clinical role in identifying pregnancies with an increased risk of developing PE or FGR in an unselected Mediterranean population.38 Nevertheless, uterine artery Doppler screening in early pregnancy was demonstrated to be useful in the prediction of adverse pregnancy outcomes in a high-risk population.39
Higher sensitivity and lower specificity of indices of impedance in the uterine arteries in the first and mid-second trimester of pregnancy, for detecting adverse pregnancy/perinatal outcomes, were detected more frequently in women with PCOS than in healthy controls. These findings suggest that the assessment of Doppler velocity in the uterine arteries could be useful in clinical practice as an early screening test to predict adverse outcomes in high-risk populations, such as women with PCOS, in the first trimester of gestation.
To date, no pharmacologic treatment has proved to be effective in preventing pregnancy complications in women with PCOS. Although there is no indication for the antenatal use of metformin in women with PCOS, it may be time to take PCOS in pregnancy seriously, as suggested by Siassakos and Wardle.40 Even if preliminary studies seem to suggest a beneficial effect of metformin, 41 we are waiting for the results from two well-designed randomised controlled studies before drawing definitive conclusions.42,43
Women with PCOS commonly show more abnormal uterine artery Doppler indices during the early phases of pregnancy, and this might be suggested as one explanation for the increased risk for adverse pregnancy and perinatal outcome observed in these patients.
Even if our data demonstrate the high predictive value of Doppler velocimetry of uterine artery impedance in women with PCOS, and a very high sensitivity in the detection of adverse pregnancy and perinatal outcomes, its real usefulness in clinical practice is unknown, and it will be necessary to investigate if the routine examination of the uterine arteries in this specific population may improve the maternal and perinatal morbidity.
Disclosure of interests
The authors declare no potential conflicts of interest, whether of a financial or other nature.
Contribution to authorship
SP designed the study, and wrote and revised the manuscript; AF performed the statistical analysis and interpreted the data; TR enrolled the patients and performed the ultrosonographic assessments; LB enrolled the patients; AT interpreted the data; FO interpreted the data and revised the manuscript; and FZ wrote and revised the manuscript. All authors approved the final version of the manuscript.
Details of ethics approval
The procedures used during the study were in accordance with the guidelines of the Declaration of Helsinki on human experimentation and of the Good Clinical Practice (GCP) guidelines. The study protocol was approved by the Ethical Committee of the Department of Obstetrics & Gynaecology, University ‘Magna Graecia’ of Catanzaro.
No one source supported this study, and the provision of supplies or services from a commercial organisation were not required.
No financial support was provided by any pharmaceutical company to realise the present research.
1. Background: How common is polycystic ovary syndrome (PCOS)? What are the criteria for diagnosis? What is the possible pathogenetic basis for adverse pregnancy outcome in women with PCOS?
2. Methods: A previous systematic narrative review of the literature identified that most studies that found worse outcome for pregnant women with PCOS had not controlled for important confounding factors (Siassakos D, Wardle P. Polycystic ovary syndrome and pregnancy outcome: red herring or red flag? BJOG. 2007 Aug;114(8):922-32). Which are these factors, and have the authors of this study controlled for all of them? How? Discuss the choice of outcome measures and their inclusion in multivariate analyses. Describe the exclusion criteria, and possible reasons for their choice.
3. Results: Compared to matched controls women with PCOS were at higher risk of adverse outcomes; which ones? Have the authors discussed plausible reasons for this increased risk? There was a non-significant increase in the number of operative abdominal or vaginal delivery for women with PCOS compared to controls; what are the possible explanations? Was abnormal uterine artery blood flow found only in women with PCOS?
4. Implications: Will this paper change the way you counsel and follow-up antenatally pregnant women with PCOS? In view of the relatively high prevalence of PCOS, what are the implications for service provision and costs? The positive predictive value for abnormal uterine flow in the first trimester was low: less than half the women with PCOS and abnormal test results had adverse outcome. What are the implications, in view of the psychological impact of a positive test?