Objectives . To assess the relationship between the amniotic fluid volume (AFV) in low risk pregnancy and the perinatal outcome, using either AFI or SDP, and to evaluate the effects of different fetal positions and Attitudes on those measurements
Methods . A prospective study was performed, in which a sample of 3000 low-risk pregnant women were studied using routine ultrasound, including fetal biometry and measurement of AFI, and SDP. Data were analysed using multiple linear regression, and constructing a curve for both the AFI, and SDP measurements, according to gestational age, the fetal positions and attitudes, in addition to the assessment of the final perinatal outcome.
Results . The 50th percentile remained practically constant at approximately 150 mm between the 20th and 33rd week, after which there was a decline in volume, which became evident after the 38th week. At the 40th week, the 10th percentile was around 62 mm and the 2.5th percentile around 33 mm. Among the group with intact membranes, no significant differences in perinatal outcome could be seen in relationship to the AFI and SDP, although a 50% increase in emergency operations for fetal distress was seen in women with oligohydramnios. Fetal position had significantly affected the AFI, which was remarkably lower in breech pregnancies, but without similar effect on SDP. There was no significant difference for either SDP (P = 0.8) or AFI (P = 0.3) between fetuses lying on the right or the left side of the maternal abdomen.
Conclusions . The percentiles incidence of amniotic fluid measurements in low-risk pregnant women showed significant decrease with gestational age, especially after the 33rd week pregnancy. Fetal position and laterality had affected significantly the AFI, but not the SDP.
Key words: Amniotic fluid index, low-risk pregnancy, obstetric ultrasonography
Abbreviations: AFI: amniotic fluid index, AFV: amniotic fluid volume, GA: gestational age, p: percentile
The importance of variations in volume of amniotic fluid to fetal well-being has been particularly well-established, and are closely correlated to an increase in perinatal mortality and morbidity rates (21, 2), although some doubts have recently been raised (3). Fetal well-being is an important question that can, however, remain unanswered in many situations, but progress in diagnostic techniques has resulted in better perinatal outcomes, and has also contributed to understanding the complex physiological and pathological interaction between fetus and mother (4, 5).
AFI and SDP are the sonographic parameters most commonly used to estimate amniotic fluid volume. Both use a two-dimensional measurement to estimate a three- dimensional parameter and are therefore subject to error. Amniotic fluid index (AFI), a semiquantitative ultrasound measure used to denote the volume of amniotic fluid, was first described in 1987 by Phelan et al. (6, 7).
Since AFI involves measurements in four quadrants and SDP only measures the deepest pocket, it is possible that fetal position would affect these two indices differently. The relative accuracy of SDP and AFI is still controversial. Using invasive methods, some studies have shown these methods to be comparable, while others have shown that one index might be better than the other. However, none of these studies took into account the potential effect of fetal position on the amniotic fluid volume indices (8).
Many studies have shown an increased risk of intrapartal fetal distress in parturient women with oligohydramnios, as identified by ultrasound examination. The exact pathophysiologic mechanism of olighydramnios has not been defined, but one likely explanation is an increased risk of umbilical cord compression during uterine contractions (7,9). However, doubts remain concerning normal values of AFI for each gestational age.
The reference curves established some years ago are still in use in current obstetrical practice, but there is a need for new data, using a reliable reference low-risk pregnant women sample, to establish the limits of AFI that would indicate perinatal risk (9). Some existing curves (10 – 13), were based on relatively small sample sizes, and normal AFI for each gestational age was not yet definitely established.
The purpose of this study was to estimate the curve for the amniotic fluid volume in low risk pregnancy, using a set of obstetric sonograms of women between the 20th and 42nd week, using two established parameters, the AFI and SDP, and to assess the effects of those measurements on the final perinatal outcome, in addition to studying the effects of different fetal positions and attitudes on those measured parameters.
Material and methods
A prospective study was carried out to estimate and evaluate the reference curve of AFI values in low-risk pregnant women, and to follow its effects on the final perinatal outcome. The study was performed at the Feto-maternal Unit, Department of Obstetrics and Gynecology at El-Minya University, Egypt. Inclusion criteria were: gestational age clearly established by last menstrual period, and confirmed by early ultrasound examination performed in first trimester of pregnancy; and gestational age between 20 and 42 weeks.
Women excluded were those with pregnancy-induced hypertension, diabetes mellitus, chronic hypertension, gestational diabetes, fetal macrosomia, ruptured membranes, placental senescence, twin pregnancy, fetal growth restriction, fetal abnormalities, fetal death, fetal isoimmunisation, or other conditions, such as metabolic disorders, kidney and heart disorders, and hypo- and hyperthyroidism. Amniotic fluid volume was measured using a 3.5 MHz linear transducer linked to a ALOKA SS 280. A scanner using the 4-quadrant technique for the assessment of AFV, described by Phelan et al.(6, 7, 9), with a modification proposed by Jeng et al.(11).
A total of 3000 women between the 20th and 42nd week of pregnancy were evaluated in this study, between August 2008 and December 2010. In order to avoid any possible bias due to repetition of examinations in women with some undetected problem, an independent sample was chosen. Therefore, only the first ultrasonographic examination of each woman was included in the study, and different sample populations were used for each gestational age, in a cross-sectional design. A formal consent had been taken form the women included in the study, after full explanation and counseling, and approval of the regional ethical committee.
The uterus was imaginarily divided into right and left halves along the linea nigra on the surface of the maternal abdomen. Using the mid-point between the fundus uteri and the pubic symphysis, the uterus was also divided into upper and lower halves. With the transducer head perpendicular to the ground, the largest amniotic fluid pocket in each quadrant was identified. The vertical diameter of this largest pocket of each one of the four quadrants was then measured. The AFI was defined as the sum of the measurements of each quadrant in millimeters. All examinations were performed by only one professional in order to avoid inter-observer variability. The intra-observer variability of the measures performed with this technique was estimated to be high (correlation coefficient 0.92) (14-18).
Both AFI and SDP were measured at the same time during the examination. SDP was obtained by measuring the depth of the single deepest vertical amniotic fluid pocket that was clear of umbilical cord or fetal parts (19,20). AFI was calculated as the sum of the depths of the deepest pockets from each of the four quadrants of the uterus. The position of the fetal trunk was characterized by three parameters. Initially, the ultrasound probe was placed transversely on the maternal abdominal wall, with the midpoint of the probe over the sagittal midline of the maternal abdomen, at the level of the fetal abdominal circumference (Figure 1).
A vertical line (Line Y) was drawn downwards from the center of the ultrasound probe. A horizontal line (Line X) was drawn across the maximum diameter of the fetal abdominal circumference. Line X was thus divided by Line Y into a shorter part (S) and a longer part (L). The first parameter to be determined was the position of the fetal trunk. This was assigned as either fetal trunk left or fetal trunk right depending on whether L was on the left or the right side, respectively, of the maternal abdomen (21-24).
Next, we determined by how much the fetal trunk lay to one side of the uterus, by calculating the laterality score, defined as S/(S + L). This score ranged from 0 to 0.5; a score of exactly 0.5 meant that the fetal trunk was on the sagittal midline of the maternal abdomen, and a score of 0 meant that the fetal trunk was to the side and did not cross over Line Y. The use of the laterality score has not been reported previously.
Finally, we determined the orientation of the ventral part of the fetal abdomen: a line (Line Z) was drawn from the fetal hepatic vein to the fetal spine, and the angle (A) between Lines Z and Y was determined. Fetuses were classified into one of three groups: ventral anterior (A = 300.1 – 360- or0-60-), ventral lateral (A = 60.1 – 120- or 240.1 – 300-) and ventral posterior (A= 120.1 – 240-). Figure I.
The study population was categorized into different groups according to the fetal position, and AFI and SDP in the different groups were compared. Pearson’s correlation coefficient between laterality score and AFI was considered the primary outcome measure. For an r of 0.25, a minimum of 62 cases was needed at a Type I error of 0.05. Based on the curve of Jeng et al. (25), and adopting a mean AFI measurement of 140 mm at 40 weeks, and a standard deviation of 48 mm, a sample size of at least 120 measurements for each week of pregnancy was estimated, assuming an α error of 0.05 and a maximum difference of 10 mm between population and sample measurements.
The AFI was correlated to perinatal outcome based on the Apgar score, umbilical cord blood pH, birthweight, frequency of cesarean section for fetal distress, operative delivery for fetal distress, including both cesarean section, vaginal forceps, and ventous extractions, and referral to the neonatal intensive care unit (NICU). Fisher’s exact test was used for statistical evaluation. P < 0.05 was considered statistically significant. The computer program ‘nQuary Advisor Release 3’ (Statistical Solutions Ltd, Cork, Ireland) was used to calculate the sample size needed in order to obtain significance levels at p < 0.05 and 0.01 with 90% confidence intervals (CI).
Data were analysed using multiple linear regression, and by constructing a curve of the 2.5th, 10th, 50th, 90th and 97.5th percentiles of the amniotic fluid measurements according to gestational age. All statistical analyses were performed using the Statistical Package for Social Sciences for Windows version 10.0 (SPSS Inc, Chicago, IL, USA). Student’s t- test, Pearson’s correlation coefficient, linear regression and ANOVA were used as appropriate. A P-value of < 0.05 was considered statistically significant.
The 3000 pregnant women included in the study had a mean age of 25.9 years (range 13 – 46), with low parity (45% were primigravida). The sample was basically from a low risk population because of the exclusion criteria used. The values of the 2.5th, 10th, 50th, 90th and 97.5th percentiles of the AFI, and SDP according to gestational age are shown in Figure II, III and IV show the data after being submitted to a smoothing process using quadratic polynomial adjustments.
Analysis of the 50th percentile measurements of the AFI, and SDP curve at different gestational ages revealed that these values remained practically constant, at around 150 mm, between the 20th and 33rd week of pregnancy. At this point, values began to decrease, and this decline became particularly evident after the 38th week, reaching 130 mm at the 39th week, 120 mm at the 41st week and 116 mm at the 42nd week of gestation. Table I, and II
The mean gestational age at the time of examination was 33.3 ± 2.8 weeks. The mean SDP and AFI were 5.5 (range, 2.8 – 9.3) cm and 14.5 (range, 6.7 – 29.3) cm, respectively. There were no significant differences in mean AFI measurements when these data were controlled for age, race, literacy, parity or previous caesarean scar (data already published elsewhere) (18). Measurements of the 10th percentile remained 100 mm until the 33rd week, when an accentuated decrease started, declining even more sharply after the 38th week of gestation, reaching values 80 mm and 40 mm at the 42nd week.
According to the published normal ranges, six cases had polyhydramnios (AFI = 29.3 cm at 29 weeks’ gestation, SDP = 9.1 cm at 37 weeks and SDP = 9.3 cm at 33 weeks’ gestation) and two cases had oligohydramnios (AFI = 6.7 cm at 36 weeks’ gestation). Twelve hundreds and five (42%) cases were fetal trunk right and 1663 (58%) were fetal trunk left. There were no significant differences between these groups with respect to gestational age (33.1 ± 2.4 vs. 33.4 ± 2.7 weeks, P = 0.7), mean SDP (5.4 ± 1.3 vs. 5.5 ± 1.4 cm, P = 0.3) and mean AFI (15.1 ± 5.1 vs. 14.1 ± 4.0 cm, P = 0.8).Table III
In fact, our results showed that fetal position had a significant effect on AFI but not on SDP; the more the fetus was positioned to one side of the uterus, the lower was the AFI. Both methods show good correlation between the measurements and the actual volume of amniotic fluid. The effect of laterality score on amniotic fluid volume indices was assessed by Pearson’s correlation coefficient and linear regression. It had no significant effect on SDP (r = 0.13, β = 1.1, standard error = 0.9, P = 0.23). However, it did significantly influence AFI (r = 0.31, β = Transverse section of maternal trunk level of fetal abdomen8.7, standard error = 3.0, P = 0.005).
The regression line is shown in Figure III. In other words, when the laterality score increased, the AFI increased proportionately. When the laterality score was 0.5 (fetal trunk positioned at the midline of the maternal abdomen), the AFI was, on average, 4.35 cm higher than it was when the laterality score was 0 (fetal trunk lay on the side and did not cross the midline of the maternal sagittal plane). Figure III, IV.
There were two case of high AFI (29.3 cm) in the study population. In order to exclude the possibility that the results were influenced by this single case, we repeated the analysis quadrants of the ipsilateral side, the vertical depth of these two quadrants being be much shallower compared with those on the contralateral side.
Although amniotic fluid should be displaced to the contralateral side, this may not be reflected completely in a two-dimensional measurement of the depth of the other two pockets. Therefore, it is not surprising to find that AFI measurement is lower when the fetus lies on one side of the uterus instead of centrally. The difference was statistically significant and is clinically important. When the fetus lay on one side of the uterus, the AFI was, on average, 4.35 cm lower compared with the AFI for a fetus lying centrally. On the contrary, SDP is apparently rather ‘inert’ to fetal position. Since SDP only measures the deepest pocket, it is understandable that the effect of fetal position on its measurement is less.
Based on the results of this study, SDP may be a better index for estimation of amniotic fluid volume than is AFI, because the association between SDP and laterality score remained non-significant (P = 0.4, β = 0.8, standard error = 0.9). Further analysis was also performed with linear regression to control for the effect of gestational age. These results showed that the laterality score had a significant effect on AFI (β = 9.6, standard error = 3.0, P = 0.002) that was independent of gestational age (β = −0.4, standard error = 0.2, P = 0.019).
AFI was significantly higher in cephalic fetal position, more than with breech ones. This result had been clearly apparent after 32 weeks gestation, and with less AFI with the ventral fetal trunk attitude with the breech position, than other. SDP had not show the same picture in different fetal positions, either breech or cephalic, so SDP as an AFV parameter had not been affected with the different fetal positions. Of the 3000 fetuses, 345 were ventral anterior, 1720 were ventral lateral and 803 were ventral posterior. The respective gestational ages of these groups were 33.5 ± 2.7, 32.8 ± 2.7 and 34.2 ± 2.8 weeks, the SDPs were 5.5 ± 1.4, 5.5 ± 1.3 and 5.5 ± 1.4 cm, and the AFIs were 14.5 ± 5.3, 14.4 ± 4.4 and 14.8 ± 4.4 cm. None of these was significantly different between the three groups (P = 1.0, P = 0.14 and P = 0.9, respectively).
The 3000 pregnant women were divided into two subgroups according to the status of the fetal membranes. The membranes were found to be ruptured at the time of the examination in 1400 (44%) women; 750 (25%) had oligohydramnios. The membranes were intact in 1600 (55); 350 (15%) had hydramnios. Table I shows the maternal variables of the two groups. The median interval between the ultrasound examination and delivery was 4 h (range 0-24 h) in the group with ruptured membranes and 6 h (range 0-70 h) in those with intact membranes.
In the group with ruptured membranes there was a significant difference in the frequency of operative delivery due to feta distress between the parturients with oligohydramnios and those with a normal volume of amniotic fluid [10.6% and 3.0%, respectively, p < 0.02, OR 3.86 (range 1.34-1.11)]. No significant differences were found regarding the other variables of perinatal outcome (Table II). In the group with intact membranes, there was a 50% increased risk of operative intervention due to fetal distress (OR 1.5), though not significant (CI 0.48-4.63) (Table III).
There is a variation in AFI measurements according to gestational age. Values in the current study remain relatively constant until the 33rd week of pregnancy when a progressive decrease starts, becoming particularly evident after the 38th week of gestation. The normal lower and upper limit values of the AFI commonly used up to now, which vary between 50 and 200 mm, are similar to those found in the present study up to the 40th week of pregnancy. When adopting reference values between 80 and 180 mm for every week of pregnancy (19,20), incorrect diagnosis are likely to occur.
Our findings, suggested a strong influence of fetal position on sonographic indices of amniotic fluid volume. Furthermore, we recruited women with apparently normal pregnancies and hence most likely with normal amniotic fluid volumes. Further studies should look at the relationship between fetal position and amniotic fluid volume indices in cases of oligohydramnios and polyhydramnios. Pregnant women, who are classified as having oligohydramnios by these criteria, may possibly be considered normal if a reference curve of AFI specific to gestational age were used, especially in term and post-term pregnancies.
The adopted limit values indicating an alteration in the AFV are variable. For the fetal biophysical profile, the measurement of just one pocket is adopted, varying from 1 to 3 cm, and considered the lower normal limit by some authors (21,22); however, in this case, total volume would be considered decreased if the AFI were used. In fact, a RCT comparing both techniques showed an overestimation of abnormal results with AFI in post term pregnancies, increasing the number of obstetric interventions (23).
These variations in classifying oligohydramnios reflect doubts regarding which percentiles best express the correlation between the decrease in AFV and poor fetal outcome. When the 50th percentile AFI was compared with that reported in a previous study (14), measurements were always higher in our study at all gestational ages by approximately 50 mm up to 28 weeks, and by 30 – 40 mm between 32 and 40 weeks of pregnancy. On the other hand, the current 50th percentile showed fewer variations, around 10 mm at all gestational ges, compared to the results of the indian study population (15).
The importance of a curve that includes the 10th and 90th percentiles is reflected in its greater capacity to identify abnormal cases. Therefore, if the 10th percentile is used as the lower normal limit, there would be less likelihood of missing a case of real oligohydramnios. A curve that included the 2.5th and 97.5th percentiles would diagnose fewer cases of abnormal AFI, and this could result in more cases of oligohydramnios or polyhydramnios being included within the normal range.
By adopting the 10th percentile of AFI as the diagnosis for oligohydramnios in our population, the values are higher than those found for the Chinese study up to 36 weeks, but similar around 40 weeks of pregnancy (14). When we compare the results of this study to previous published curves (10-17), similarities can be seen for the 50th percentile of AFT at all gestational ages. However, when comparing the 2.5th percentile, it is evident that the measurements in Moore and Cayle’s curve are lower up to the 35th week of pregnancy, after which they are higher than the values found in our study curve. The 97.5th percentile of the Moore and Cayle curve is slightly higher at all gestational ages except for the 41st and 42nd weeks.
The lower limit of 2 standard-deviations and the mean values of the Jeng et al. (11) curve are slightly lower in relation to the present curve at corresponding gestational ages, except from the 37th to the 42nd week, when values remain higher than those in the present curve. The definition of normal AFI cannot, in itself, guarantee good perinatal outcome. For instance, a 42-week pregnancy with an AFI of 45 mm would be considered normal, but how physiological this value is and what real risk it represents are questions that still need to be fully answered.
If the correlation between AFV and perinatal outcome can be established, this curve may have a broader clinical application in prenatal diagnosis and care. Moreover, the curve of the 2.5th, 10th, 50th, 90th and 97.5th percentiles of the AFI measurements shows a significant decrease according to gestational age, especially after the 32nd week.
This measurement could, therefore, considered a normal reference curve for the evaluation of AFI. The results of the present study suggest that oligohydramnios after rupture of the membranes in low-risk pregnancies is associated with a nearly four-fold increased risk of operative delivery due to fetal distress. An ultrasound examination of AFI could thus identify those who may need intensified fetal surveillance during labor.
The present study was performed on a selected group of women with low-risk pregnancy. As AFI is one of the parameters checked in high-risk pregnancies at our hospital, these parturients were excluded in order to make the study ‘blind’. By adding high-risk pregnancies, a much smaller sample size would be needed. The frequency of oligohydramnios in cases with intact membranes was unexpectedly high: 15% instead of 5% in the controls.
Although our pregnancies were low-risk, a few showed signs of pregnancy complications on admission to the labor ward (Table I), which might explain the higher frequency of oligohydramnios in this group. Although there was a significant correlation between operative delivery due to fetal distress and oligohydramnios in cases of ruptured membrane (Table II), sensitivity was low (11%), and false-positive and negative rates were 46% and 23%, respectively. Thus the knowledge of oligohydramnios in these low-risk pregnancies did not cause any immediate action, only more intense surveillance during labor.
In the present study there was a 50% increased risk of operative fetal delivery due to fetal distress in parturients with oligohydramnios and intact membranes. Teoh et al. studied 120 pregnancies as an admission study in early labor with intact membranes. The frequency of oligohydramnios (AFI < 5 cm) in their study was 22%, and operative delivery due to fetal distress frequency among these was 27%. Based on these data, a sample size of 100 would be sufficient. We chose, however, three times that size, as the low-risk status of their population was uncertain (9, 11).
The pathophysiology of oligohydramnios before membrane rupture is unclear. One theory is that a reduced perfusion of the placenta causes hypovolemia in the fetus, and/or an automatic redistribution of fetal blood volume to vital organs with a resultant reduced blood supply to the kidneys. This in turn could lead to reduced production of urine, and thus reduce the volume of amniotic fluid. Bar- Hava et al. studied signs of redistribution, renal blood flow, and signs of oligohydramnios, but could find no correlation. There was no change in the renal artery pulsatility index (12, 14, 19).
Oligohydramnios in labor after the rupture of membranes in a low-risk pregnancy is probably not caused by a reduced perfusion of the placenta, but is more probably caused by the loss of large amounts of amniotic fluid at the time of the rupture. One explanation for the significantly increased risk of operative delivery due to fetal distress in the group with ruptured membranes might be that there is an increased risk of the umbilical cord becoming trapped in an ad- verse position, at the time of the rupture, if a large amount of amniotic fluid is lost. Amnioinfusion may be a way to treat such cases in order to restore the volume of amniotic fluid and reduce the risk of compression of the umbilical cord, thus averting the need for operative delivery (11, 21, 23).
As a conclusion of the current study, assessment of the AFV during pregnancy using the SDP appears to be more accurate than the AFI, especially the SDP evaluation has not been affected significantly with either different fetal positions or attitudes, but still we are in need for further controlled studies to compare the accuracy of the two modes of AFV assessment. Another conclusion drawn from our study is that an ultrasound examination, including measurement of AFI as an admission test for women presenting at the labor ward with ruptured membranes after an uneventful pregnancy, could help identify those with an increased risk of intrapartum fetal distress, namely those with oligohydramnios.
Moreover Measuring AFI in low-risk pregnancies on admission to the labor ward might detect cases needing special surveillance. We are currently preparing a new ongoing study, as an extension to the current study, comparing the previous two parameters of AFV assessment in high risk pregnancies, and the preliminary results could confirm the previously mentioned results, but it is too early to get to a final conclusion.