Re: Ultrasound or menstrual dating
From: David J R Hutchon (DJRHutchon@Postmaster.co.uk)
Thu Jan 21 13:21:18 1999
Have you seen the paper by Smith et al in the December issue of the New
England Journal of Medicine (. N Engl J Med 1998;339:1817-22)
showing an association between low birth weight and suboptimal
first-trimester growth that has an important impact on the current
debate on ultrasound and menstrual dating. They show that women in
which the expected crown rump length is less than the measured crown
rump length by the equivalent of between 2 and 7 days, there is a risk
of low birth weight for gestational age in up to 7%. This must have an
significant impact on the suggestion that scan dating alone should be
used. These women could be targeted for growth monitoring later in the
pregnancy. If menstrual dating is ignored, the opportunity to identify
this group is lost. Menstrual dating with all its limitations, still
provides a useful measure of fetal age and I think it still needs to be
integrated with the measured ultrasound fetal size.
--
David J R Hutchon
At Thu, 07 Jan 1999, vval@earthlink.net wrote:
>
>art fougner, md wrote:
>>
>> Jim,
>>
>> when period agrees with sono, can use either date actually. when they
>> diverge, my money's still on sono. for me, due dates are not arrival
>> dates. they're more like subway schedules. and just because the train
>> is expected, that doesn't mean that's when it's pulling into the
>> station.
>>
>> in any event, what works for you works for you :)
>>
>> art
>>
>> At Tue, 05 Jan 1999, James S Smeltzer MD wrote:
>> >
>> >David,
>> >
>> >Campbell produced a study showing the superiority of sonography prior to 20
>> >weeks in prediciting the date of spontaneous live birth over known,
>> >certain, regular and reliable menses in a large population of british
>> >women. The modal duration of gestation, and the median are usually
>> >identical and roughly 278-280 days from onset of last menses. The mean can
>> >be skewed, as you have noted. The date of conception is plus or minus about
>> >5 days in normal women with regular cycles, based on last menses. Assuming
>> >that sonography and menses are within a week, it would be most reasonable
>> >to average these to predict gestation, as this is likely to be better than
>> >either alone by a ratio of 1/sqrt(2) if the variances are about equal
>> >(which they are for sonogram bvefore 18 weeks).
>> >
>> >It is my opinion that average estimate is better (and I can formally prove
>> >that it is better if the estimates are both unbiased and independent) than
>> >either component of average before 20 weeks with differences under a week.
>> >
>> >On the other hand if dates are not certain, cycles are not regular or there
>> >has been recent hormonal treatment within 3 months of last menses, then the
>> >LMP should not be averaged and sonogram alone used for dating, IMHO.
>> >
>> >[By the way I believe in counting the number of moons from last menses
>> >onset, and when I run out of fingers, the baby is due]
>> >
>> >Jim SMeltzer MD ;^}>
>> >
>> >At 04:05 PM 1/3/1999 -0600, you wrote:
>> >>Many thanks to Obgyn.net for the summaries from the Edinburgh meeting.
>> >>Reynir Geirrson and Jason Gardosi and others contend there is evidence
>> >>that, over all, scanning is a better predictor of gestation than the
>> >>last menstrual period, even when this is apparently confidently known by
>> >>the woman. They demonstrate this partly by demonstrating the
>> >>relationship between the scan in these women and established scan
>> >>charts, and partly by showing the relationship between the two methods
>> >>and the actual date of delivery. If you look carefully at the figure 5
>> >>in the recent paper by Tunon and Eik-Nes shows that for those women with
>> >>scan size within 7 days of menstrual gestation the LMP is at least as
>> >>precise as ultrasound.
>> >>
>> >>In the opinion of Ole Olsen, a Danish statistician, however the studies
>> >>are statistically flawed and although the numbers in some of the studies
>> >>are very large, as he points out study size does not compensate for
>> >>methodological faults. Olsen feels that the interval from LMP to EDC
>> >>should be readjusted to 283 days. When one thinks about it is
>> >>remarkable that the original work by Hermanni Boehaave was so accurate.
>> >>The figure was largely based on folklore although the idea of an
>> >>interval of about 40 weeks had been known for at least a thousand years.
>> >>When we talk about the mean length of pregnancy we cannot be referring
>> >>to the mean length of every pregnancy otherwise the figure would be
>> >>hopelessly skewed if every miscarriage was included. We must be talking
>> >>about the mean length of healthy pregnancy which by definition means a
>> >>healthy baby. Mittendorf in Chicago, using a large database of
>> >>menstrual data showed slight but significant differences in the length
>> >>of pregnancy in women of different parity, ethnicity etc but what we
>> >>dont know is how it translates to ultrasound dating.
>> >>
>> >>I think amongst most obstetricians the argument centres around whether
>> >>scan dates should be used exclusively or the whether the menstrual dates
>> >>should be maintained if the difference is small enough to be
>> >>statistically not significant.(e.g.about 7 days) Here in the UK the
>> >>majority of obstetricians and midwives (and a CESDI recommendation) keep
>> >>the menstrual dates if the difference is small. I would argue that this
>> >>is the logical approach but not the most practical approach and it is
>> >>far easier, with modern ultrasound machines, to simply pick a parameter
>> >>and allow the machine to generate a gestation and an estimated date of
>> >>confinement from its internal chart. The LMP will always be an
>> >>important landmark in a woman's pregnancy but its importance for
>> >>providing an EDC can be gradually diminished, as women begin to
>> >>appreciate that the scan is the easiest way to get and accurate result.
>> >>I do not think there is any great advantage in changing from "gestation"
>> >>to say fetal age as the change is likely to lead to at least as much
>> >>confusion as there is with gestation.
>> >>
>> >>An important aspect of the discussion arises from the use of menstrual
>> >>data in the generation of ultrasound scans. This leads to a normal
>> >>range in the size of the parameter for each gestation week. Clinicians
>> >>interpret the normal range in the same way as in other medical
>> >>measurements. For example the patient has a healthy haemoglobin if it
>> >>is in the normal range. No one expects it to be spot on the mean value.
>> >>
>> >>If the scan measurements are used as a landmark in pregnancy in the same
>> >>way as the LMP, and the mean interval to normal labour and a healthy
>> >>baby of a population of women with a specific scan measurement is
>> >>plotted against each of the scan measurents, a single regression line of
>> >>the range of the measurements can be generated. Since there is no
>> >>reliance on LMP for this chart, there is no need to consider how the
>> >>scan EDC relates to the menstrual EDC. This is the proposal I have put
>> >>forward at http://www.obgyn.net/us/cotm/9807/cotm_9807.htm
>> >>
>> >>It seems to me that if there is a significant difference between and the
>> >>menstrual EDC in women with certain dates and the scan EDC using
>> >>conventional charts, this simply shows that the population from which
>> >>the scan chart was derived is not the same as the population to which it
>> >>is being applied, and as such is therefore an inappropriate standard.
>> >>
>> >> Reference
>> >>Tunon K, Eik-Nes S H and Grottum P A comparison between ultrasound and a
>> >>reliable last menstrual period as predictors of the day of delivery in
>> >>15000examinations. Ultrasound in Obstetrics and Gynaecology 8 (1996)
>> >>178-185
>> >>
>>
>> --
>> art fougner, md
>> SonoScan/Genetic Sciences
>> forest hills, ny
>> evsono@pipeline.com
>
>I agree 100% Dr. Forgner. I always tell my patients that the due date
>is only an estimation, the baby will arrive whenever it chooses!
>
>Victoria A. Valentine, BA, RDMS
