Chapter 6

CONCLUSIONS, IMPLICATIONS and RECOMMENDATIONS

6.1 Introduction

This chapter assesses the original thesis synopsis and discusses why changes were needed to the final objectives of the study. It then summarises the findings of all the studies and discusses the implications of the individual studies and suggests future research.

6.2 Original Synopsis

The original synopsis for this thesis asked the following questions:

  1. Are the existing fetal weight charts / percentile measures relevant to babies born today? Should the weight for calling macrosomia in babies be adjusted upwards?
  1. Can the increase in failure to progress in labour, intervention in labour and post partum haemorrhage be due to the increase in average birth-weight?
  1. Can serial ultrasounds on patients with larger than expected fundal height help to predict the heavier babies, and assist in management of labour, recommendation for elective Caesarean section or intervention?
  1. Can the ASUM ultrasonic fetal measurement charts, which are based on the mean Australian population, be used to confidently evaluate fetal size & growth in the Chinese Australian population?
  1. Is there a difference in fetal growth patterns between babies of Chinese women living in Australia and Caucasians?
  1. How does gestational diabetes affect fetal growth in:
  2. Chinese
  3. Caucasians

6.2.1Weight Chart Percentiles and Macrosomia

Question one of the synopsis queried whether the existing fetal weight charts 5th and 95th percentile limits, which were established in the nineteen eighties, are relevant to babies born today. It was also suggested that the weight for defining macrosomia should be increased upwards. Although the number of macrosomic births, or those greater than 4000 grams, is increasing (Graph 6/1) the average Caucasian birth weight remained static between 1992 and 2000 (3313g). In the same period Chinese birth weights shifted upwards significantly although the average Chinese birth weight increased by less than 100 grams (3248g > 3334g). The women involved in this study were from the Antenatal Clinics of two NorthShore hospitals with a high percentage of middle class clients and these results only relate to this population. Without data from other regions with varying socio-economic populations it would be inappropriate to change fetal weight percentile limits. The same could be said for increasing the birth weight for defining macrosomia. The results of the sub-study on fetal macrosomia and birth complications concluded that it was recommended that the weight for defining macrosomia in Chinese women in our population be dropped to 3500 grams.

Graph 6/1Caucasian/Chinese Birth Weight Comparison

Graph 6/1 shows the shift in birth weights between the two time periods. The number of births in the lower weight groups decreased whilst the number above 3500 grams increased.

6.2.2Intervention and Post Partum Haemorrhage

The synopsis also thought that the increase in failure to progress in labour, intervention in labour and post partum haemorrhage could be due to the increase in the average birth weight. As the average birth weights for the various populations have hardly changed in a decade this theory could be that the increase in complications is due to the increase in birth weights, not average birth weight. Graph 6/1 shows the shift in birth weight groupings of both Caucasian and Chinese populations with the majority of Caucasian births now between 3500 and 4000 grams compared with 3000 to 3500 grams in 1992. Most Chinese babies however are in the 3000 to 3500 gram group. The sub-study of Chapter Four highlighted the increased rate of post partum haemorrhage but could not specify why this increase has occurred. By mid 2005 PPH and intervention rates at the two hospitals involved in the study have remained relatively static even though macrosomia rates for Caucasians reached 15.6%.

6.2.3Predicting Fetal Size with Ultrasound

Questions three, four and five asked if serial ultrasounds on patients with larger than expected fundal height could help predict the heavier babies, and assist in management of labour and whether the existing ASUM fetal measurement charts were accurate for both Caucasian and Chinese populations.

Seventy-five women involved in this study were scanned between four and ten times each in the third trimester. Only three women thought to be large for dates were actually above the 95th percentile on the growth charts. Although the predicted fetal weights did vary from the actual birth weights in a small percentage of cases, the overall majority of predicted weights were within the plus / minus range of the weight formula used. There were more scans for questioning small for dates, mainly in the Chinese population. The ASUM fetal measurement charts were found to accurately evaluate size in both Caucasian and Chinese populations, except for the abdominal circumference, which showed a difference in the last five weeks of gestation, with a term Chinese fetus being three weeks less than the mean Australian population.

6.2.4Is Fetal Growth Affected by GDM?

The final question was how gestational diabetes mellitus affected fetal growth in both the Chinese and Caucasian populations. This study recruited women with GDM, all of whom controlled glucose levels by diet, and found no significant difference in fetal size/growth. Fetal growth in women with GDM requiring insulin therapy may show variations from the mean not demonstrated in our study population. Although the Chinese fetuses were slightly smaller than the Caucasian fetuses the growth patterns and growth per day were the same between the two groups and the general population based on growth charts.

6.3 Final Objectives of the Thesis

The original synopsis was redesigned to be more relevant to the needs of the individual studies. Instead of only one objective for each of the sub-studies there were varying numbers depending on the complexity of the study. There were four major objectives:

  1. The first objective was to describe fetal biometry graph formation and then to construct ultrasonic fetal biometry graphs based on an Australian population for the crown rump length, head circumference and abdominal circumference. The new graphs are then compared with the graphs in the literature that are relevant to this new work.
  1. The second objective aimed to examine fetal measurement discrepancies. Initially in the first trimester CRL measuring accuracy was evaluated, along with the BPD, head and abdominal circumference and long bone length searching for the stage at which dates could be conflicting, leading to problems with timing for invasive procedures such as CVS, amniocentesis and maternal blood analysis. In the third trimester inter and intra sonographer fetal measurement reproducibility was evaluated in the final 6 weeks of pregnancy to detect any common pitfalls of measuring that could affect the report on fetal size. The hypothesis being that sonographers entrusted with the measuring of the fetus in the third trimester could do so in a competent and reproducible manner. It is also hoped to detect any broad based problems of measuring that could be addressed to improve outcomes.
  1. The third objective was to compare birth weights and rates of fetal macrosomia (birth weight > 4000 grams) and birth complications in both Chinese and Caucasian women for two time periods: 1992 and 1999/2000 and within each of these time periods, to compare obstetric complication and intervention rates for these two groups of women. This was to be done by accessing and downloading statistics from the data base for all births of Caucasian and Chinese women at the RoyalNorthShore and HornsbyHospital during the time periods in question.
  1. The fourth and final objective was to compare fetal size in the thirdtrimester of pregnancy in bothChinese and Caucasian women with gestationaldiabetes mellitus and to determine whether the mean fetal size in Chinese women with GDM varied from either Caucasian pregnancies affected by GDM or normal Chinese pregnancies. This involved the serial measurement of the fetuses of three groups of women in the third trimester of pregnancy – Caucasians affected by GDM and Chinese both with and without GDM. The hypothesis was that the normal Chinese fetus was smaller than the normal Caucasian fetus and Caucasian and Chinese fetuses affected by GDM, in the last five weeks of gestation. It was also aimed to determine if any ultrasonically measured fetal parameter predicts excessive birth weight of Chinese or Caucasian babies and to calculate the fetal growth per day in grams in Caucasian and Chinese pregnancies.

6.3.1Objective One: Fetal Biometry and Australian Fetal Measurement Charts

Fetal biometry is relevant to every obstetric ultrasound and yet over 90% of all articles in the literature relates to antenatal problems that affect only 2% of all pregnancies (Dudley and Chapman 2002). During the nineteen eighties there was a plethora of work on fetal measurements, with new charts and graphs appearing in almost every issue of every ultrasound journal in the western world. These charts laid the foundation for future work but there was reluctance by many researchers to revisit and update this important facet of obstetric ultrasound. Hadlock, Deter and Jeanty were amongst the very few who upgraded their existing work when new data, new equipment and new techniques became available. One of the problems with many of the charts is that they were formulated using biased populations such as middle class white women and, as in the case of the 1975 crown rump length chart by Robinson et al, on crude equipment with manual manipulation of data collection and measuring. The callipers in use on the early generation of real time ultrasound systems could only measure to an accuracy of, at best, 5mm compared with today’s systems that can measure to 0.01mm. Ellipse and trace facilities were still in their development and as a result the data used to generate the abdominal and head circumference charts was collected from polaroid images of the fetal parameter and utilized a hand held map measurer to trace the circumference.

It is surprising that in over thirty-five years of diagnostic ultrasound in Australia the only fetal measurement charts that were based on an Australian population were, until 2001, the BPD chart of de Crespigny et al and the OFD and BPD charts of the RoyalHospital for Women in Sydney.

As discussed in the literature review the way the fetal measurement charts are formulated is paramount to their accuracy in the clinical sphere. Chapter Two described a correct method for chart creation and showed how the new graphs were formulated. There were 500 CRL measurements and 1,855 abdominal circumference measurements performed on fetuses from 5 weeks 2 days to 41 weeks gestation. The head circumference utilised data obtained from 2,600 BPD and OFD measurements collected for the 1999 study. Using the suggestions of amongst others, Jeanty (1991) and Altman and Chitty (1993), the selection of participants, collection of data and application of mathematical modelling to fetal parameter charts are demonstrated with the end resultant new graphs (Westerway et al 2000) shown in Tables 2/2 and 2/3.

ASUM accepted the Westerway fetal measurement charts of the CRL, BPD, OFD, HC, AC and femur and humerus lengths in 2001. The new Australian Charts were criticised by de Crespigny et al (2002) due to failure to take into account the length of menstrual cycle and the fact that the data was collected by a number of sonographers from many ultrasound practices. In reality more than two thirds of the women included in the study had cycle length recorded but this was deliberately omitted from the methodology in the original article (Westerway 2000) on the suggestion of a number of mentors, as it was not deemed to be necessary due to the statistical averaging used. It is of interest to note that fetal biometry charts for the BPD, HC, AC and femur produced by another team of Australian researchers (Schulter, Pritchard and Gill, 2005), attempted to address the perceived problems of the ASUM 2001 charts, but were heavily criticised (Mongelli and Benzie 2005) for being a biased population with no demographic characteristics discussed.

The literature only sites a handful of new charts since 1990. Most of these works were population based, such as Lai and Yeo (1995), on Asians in Singapore, and each of the groups of authors made reference to the early works of the Hadlock, Deter and Jeanty. It was important therefore in this body of work to compare the new ASUM charts with the above authors. Only three of their charts were recommended for use in Australia prior to 2001. No statistically significant differences were seen from 30 weeks gestation between any of the ASUM fetal measurement charts and the relevant American charts of Hadlock, Deter and Jeanty (Chapter 2 Graphs 2/23 – 2/27) and as a result the ASUM 2001 charts could justifiably be used for the comparison of works in Chapter Five.

Displaying the rigour of chart formation was only one of the aims in this chapter. The other aims were to examine ultrasonic fetal measurement discrepancies in the first trimester and to assess inter and intra-sonographer measuring errors.

6.3.2Objective Two: Measuring Accuracy

6.3.2.1First Trimester Measuring Accuracy

This was the first time in the Australian literature that the accuracy of early ultrasound dating has been evaluated. The crown rump length study (Appendix One) was published in 2002. The state of the art ultrasound systems now in use, combined with trans vaginal ultrasound produces images with excellent fetal detail and as a result the predictive values for accuracy shown in Chapter Two has, as expected, improved on those values cited by Bovicelli et al (1981), Selbing (1982) and Pedersen (1983).

Being aware of how fetal flexion can affect the assignment of fetal age is the first step towards dating accuracy and overcoming the problems associated with incorrect measurements. Provided the operator has measured the crown rump length in the neutral position the measurements, when combined with the measurements of the BPD and head circumference after 11 weeks gestation, can be used to confidently date the early pregnancy to within 6 days of actual dates. Less than 11 weeks 3 days this predictive value +/- 4 days. The differences seen could be due to timing of ovulation.

Correct age assignment for nuchal translucency assessment, the timing for chorionic villi sampling and first trimester blood analysis is an expected outcome at first trimester imaging. This confidence in the ultrasound examination has taken time to achieve and provided ultrasound practitioners are aware of the possible errors in first trimester measurements then a high level of accuracy can be maintained.

6.3.2.2 Inter / Intra Sonographer Measurement Reproducibility

This is the largest inter-sonographer study undertaken in the literature and one of the few that has investigated not only the inter/intra observer differences but explored the reason why there was a difference. The results of the study were presented internationally and published in 2004. Using the optimum head and abdominal images and the measurements of all 26 sonographers, this study had an inter-sonographer error of 8.1% for the head circumference and 4.3% for the abdominal circumference. This result should silence the critics of the ASUM 2001 charts for not including an inter/intra-sonographer analysis (Nisbet et al 2002). The range of results for the individual fetal parameters gave a 17% or 445 gram difference in the fetal weight estimation. Using the individual sonographers mean set of variables on these same images the difference was 12% or 335 grams compared with the work of Gull et al with a difference of 6.3% using only three sonographers. The intra-sonographer error for weight assessment using all 26 sonographers was 83 grams compared with the 75 grams of the 3-way intra-observer error in the study of Chang et al (1993). In this study the majority of sonographers measured the fetal parameters in an acceptable way with the only variables of the 52 measurements not within the accepted range of 2 standard deviations being the occipito-frontal diameters and the femur at the acute angle. Five of the participating sonographers with 10 months, 2, 3, 7 and 15 years experience, either did not routinely perform third trimester scans, or had limited obstetric ultrasound exposure and this was reflected in their measurements of both the acceptable and sub-optimal images, with calliper placements less accurate than those sonographers with more obstetric exposure. Three of these sonographers were the worst in their respective group for intra-sonographer evaluation. It is of concern that these sonographers do not keep up their obstetric imaging practice and yet are “qualified” to perform such scans when necessary. This is a problem of quality control in scanning procedures.

One of the aims of this project was to detect any common pitfalls of measuring that could adversely affect the report on fetal size/growth. By far the most common problem was on ultrasound images that were not standard measuring planes due particularly to angulation. The majority of ultrasounds in the third trimester are for assessing fetal size and in at least half of these the fetus will be in a position that makes it difficult to achieve ideal imaging planes for at least one of the fetal parameters needed for measuring for inclusion in a weight formula. In this instance it may be necessary to take a measurement from an unsuitable image. In this study there was an example of a head buried low in the pelvis, which was measured reasonably well. The other most prominent example was the long bone image at an acute angle where both the experienced and inexperienced sonographers measured the diaphysis, epiphysis and/or artefacts.