Aust N Z J Obstet Gynaecol 2002; 42: 1: 111

Letters

Y

ear 2000

Y

ear 2000

Incontinence SurgeonsPO Box 438

Edgecliff NSW 202Australia

1 Dietz HP, Clarke B. Is the irritable bladder associated with ante-rior compartment relaxation? A critical look at the ‘integral the-ory of pelvic floor dysfunction’. Aust N Z J Obstet Gynaecol. 2001Aug; 41 (3): 317–319.

2 Petros PE, Ulmsten UI. An integal theory of female urinary incon-tinence. Experimental and clinical considerations. Acta ObstetGynecol Scand. 1990; S153, 69: 7–31.

3 Petros PE. New ambulatory surgical methods using an anatomi-cal classification of urinary dysfunction improve stress, urge, andabnormal emptying. Int Urogyn J Pelvic Floor Dysfunct. 1997; 8(5): 270–278.

To the Editor;

Regarding the recent contribution from Dietz andClarke1 on the ‘integral theory of pelvic floor dysfunc-tion’, I wish to raise a number of matters regardingtheir approach to the examination of this theory.

It is important to understand that the ‘integral theory’looks at the whole pelvic floor as an inter-related sys-tem. As well, the classification system defines anatom-ical defects in three zones of the vagina, anterior, mid-dle and posterior.2

In their paper, Dietz and Clarke have made statisticalcorrelations for instability symptoms against parame-ters which are used as confirmatory for stress inconti-nence only. From this unsatisfactory basis, they havemade various statistical correlations. Nowhere, how-ever, is there a matched control group with no symp-toms of instability. Moreover, consideration of the pos-terior compartment has been omitted. This compart-ment is a major cause of frequency, urgency and noc-turia.

In attempting to prove or disprove the ‘integral theory’hypothesis, the work of Tunn et al3 is worthy of note.Using magnetic resonance imaging in two groups ofpatients, one with prolapse, the other with urinaryincontinence, these authors found no difference inmuscle morphometry between the control and experi-mental groups. They did confirm connective tissuelaxity in the hammock for incontinence patients. Theyconcluded: “we can thus confirm the hypothesis byPetros and Ulmsten2 that an intact musculo-vaginal

To the Editor

I commend Dietz and Clarke1 for attempting to evalu-ate the ‘integral theory of female urinary inconti-nence. Unfortunately the authors have failed to fullyunderstand the scientific basis of the theory, and thishas resulted in some incorrect assumptions and, as aresult, some erroneous findings.

The ‘integral theory’ first published by Petros andUlmsten2 states that stress, urge and abnormal empty-ing mainly derive for different reasons from laxity inthe vagina or its supporting ligaments as a result ofaltered connective tissue. It is critical that the pelvisbe assessed as a functional unit and in order to aid theclinician in this task Petros has devised a pictorialalgorithm3 which assesses symptoms and signs in thethree compartments of the pelvis so that they can becorrelated with a clinical diagnosis.

Dietz and Clarke have taken a number of anteriorcompartment ultrasound measurements and extrapo-lated them as an index of all three compartments ofthe pelvis. No assessment has been made of laxity inthe posterior compartment, which I have found to bemost important in the causation of urgency, fre-quency, nocturia and pelvic pain. No control patientsare studied for comparison.

A theory can only be tested effectively for truth or fal-sity by testing its predictions. Dietz and Clarke havepresumed to invalidate the ‘integral theory’ withouttesting any outcomes. An increasing number of sur-geons are now evaluating the ‘integral theory’ andtesting the operations that are derived from it. In datanow submitted for publication I have treated a groupof 90 patients with posterior compartment ligamen-tous laxity with a surgical repair (Infracoccygealsacropexy or posterior intravaginal slingplasty).When reviewed after 12 months a significant improve-ment in symptoms was noted (urgency 80%, nocturia81%, pelvic pain 79%).

Only by applying rigorous standards to the assess-ment of the ‘integral theory’ and other new technolo-gies can the advancement of our science proceed.

Bruce FarnsworthSecretary

Australasian Association of Ambulatory Vaginal and

112 LETTERS

unit plays a decisive role in preserving continence”.As well, at a clinical level, ambulatory vaginal andincontinence surgeons test the ‘integral theory’ on aregular basis in their day to day practice. Every time asurgeon uses the relevant diagnostic algorithm,he/she challenges the theory itself for truth or falsity.It is therefore puzzling that Dietz and Clarke have cho-sen to advise against surgery for patients with insta-bility, which is in marked contrast to the usual prac-tice of many ambulatory vaginal and incontinencesurgeons.

My own initial observations were published in 19944

and I reviewed my first 184 cases in 1997.5 Since then Ihave used the diagnostic algorithm regularly and haveperformed more than 650 operations. I can confirmthat a cure rate of 80% is the norm in patients withinstability whose ligaments have been reconstructed.The comments I made in 1994, still in my early years oflearning, hold true today: “they (the surgical tech-niques) promise a new era of virtually pain-free cureof prolapse and incontinence without catheters, and areturn to normal activities within days”. This has nowbeen confirmed world-wide by at least 200,000 opera-tions based on the ‘integral theory’.

Peter RichardsonConsultant Gynaecologist

24 Hyatt CentrePerth WA 6004

Australia

1 Dietz HP, Clarke B. Is the irritable bladder associated with ante-rior compartment relaxation? A critical look at the ‘integral the-ory of pelvic floor dysfunction’. Aust N Z J Obstet Gynaecol. 2001Aug; 41 (3): 317–319.

2 Petros PE, Ulmsten UI. An integral theory and its method for thediagnosis and management of female urinary incontinence.Scand J Urol Nephrol Suppl. 1993; S153: 1–93.

3 Tunn R, Fischer W, Hamm B, Kuchinke J. Static magnetic reso-nance imaging of the pelvic floor muscle morphology in womenwith urinary incontinence and pelvic prolapse. Neurol Urodyn.1998; 17 (6): 579–589.

4 Richardson PA. Ambulatory surgery for urinary incontinenceand vaginal prolapse. Med J Aust. 1994 Dec 17; 161 (8): 511.

5 Richardson PA. The intravaginal slingplasty operation, a per-sonal experience of 184 cases. Aust N Z J Obstet Gynaecol. 1997; 37:4: 485.

Authors’ reply,

Many thanks for giving us the opportunity to respondto two letters you have received regarding our recentpaper.1

Both letters make virtually the same points in criti-cising this paper, and none of these points seem tohave any bearing on the issues at hand. The ‘integraltheory’ postulates an association between anteriorvaginal wall relaxation and symptoms of the overac-tive bladder – anybody who has read any of the numer-ous papers published on this subject by Dr Papa Petroswould have to agree. The ‘integral theory contains anumber of other assertions, some of which would be

very difficult to test using scientific methodology. Wetested one hypothesis, ie whether anterior vaginalwall relaxation is associated with symptoms and signsof bladder over-activity, and in our population theopposite was true. The hypothesis is therefore provento be incorrect.

It is entirely irrelevant in testing of this hypothesis asto whether posterior compartment relaxation has anyinfluence on such symptoms. That is a differenthypothesis, which we have in fact tested recently on adifferent population; a manuscript is in preparation.Results will not influence the findings alreadyreported. The same goes for other assertions con-tained in the ‘integral theory’.

It is also quite irrelevant to the testing of our hypoth-esis whether any particular surgeon has performedhundreds of procedures while adhering to the tenetsof the ‘integral theory’, and whether that surgeon hascured anybody with symptoms of bladder irritabilityin this way. If, indeed, surgeons are routinely perform-ing surgical procedures for the cure of irritabilitysymptoms (on the basis of personal case series andwithout so much as a single external audit, let aloneintervention trial or RCT) then we can see major prob-lems ahead. Clearly, the concepts of evidence-basedmedicine need to be more broadly embraced.

HP DietzRoyal Hospital for Women

SydneyNew South Wales

B ClarkeRoyal Women’s Hospital

Brisbane Queensland

1 Dietz HP, Clark B. Is the irritable bladder associated with ante-rior compartment relaxation? A critical look at the ‘integral the-ory of pelvic floor dysfunction’. Aust N Z J Obstet Gynaecol. 2001Aug; 41 (3): 317–319.

To the Editor,

Short-term neonatal outcomes of growth restrictedinfants by their mode of delivery

There is no evidence regarding the best mode of deliv-ery of the intrauterine growth restricted (IUGR) fetus.There are no studies comparing caesarean section(CS) with vaginal delivery (VD). It is not known howmuch additional hypoxaemia labour and VD mightcause and effect outcome of the already compromisedfetus.

We performed a retrospective study with the aim tocompare the immediate/short-term outcomes of IUGRinfants (< 10th centile) with birthweight < 1500 gdepending on the mode of delivery. All IUGR infantsborn at the Royal Women’s Hospital, Brisbane, from

113

1996 to 1998 were included in the study. Infants born byemergency caesarean section (CS) for fetal distress,antepartum haemorrhage, chorioamnionitis or cordcomplications were excluded. Outcome measuresincluded: Apgar scores, need for intubation for resus-citation at birth, metabolic acidosis, ventilationrequirements, mortality to discharge and chronic lungdisease (O2 requirement at 36 weeks corrected).

There were 153 neonates born with the above men-tioned criteria. Thirteen (8%) were delivered vagi-nally and 140 (92%) by CS. The IUGR was known ante-natally in 7/13 (54%) of the VD group and in 103/140(74%) of the VD group (p = 0.19. Fisher’s exact test).

We conclude, therefore, that most IUGR infants arecurrently born by caesarean section. Those infantsdelivered vaginally are bigger and more mature. Anyvalid conclusions are difficult due to small numbers ofinfants born by VD and multiple confounders. Fromthese data, the VD infants are not worse off than theCS infants. However, we cannot say that infants whowere delivered by CS would have been better or worseoff if delivered by VD.

A larger study is needed, in increasing order of desir-ability; a retrospective cohort, a prospective cohort ora RCT.

Dr Zsuzsoka Kecskes FRACPDr Janet Berrington MRCPDr Mark W Davies FRACPPerinatal Research Centre

Royal Women’s HospitalButterfield Street

HerstonQueensland 4029 Australia

SVD CS

Mean (standard deviation)

Gestational age (weeks) 32.9 (3.3) p = 0.3* 30.6 (32.6)

Birthweight (grams) 1244 (310) p = 0.02* 1002 (332)

Number (%)

Antenatal steroids 7 (54) p = 0.20† 101 (72)

Intubated for resuscitation 5 (38) p = 1.0† 55 (39)

Base excess (1st 12 hours) < -5 1 (8) p = 0.30† 31 (22)

Died 0 (0) p = 0.36† 16 (11)

In O2 at 36 weeks 1 (8) p = 1.0† 15 (11)

Median (IQR)

1 minute Apgar score 6 (2.5–8.5) p = 0.44‡ 7 (5–8)

5 minute Apgar score 9 (7–9) p = 0.19‡ 9 (8–9)

Total time IPPV hours 0 (0–0) p = 0.04‡ 18 (0–212)

Total time O2 hours 2 (0–22) p = 0.05‡ 45 (2–329)

Max FiO2 (1st 12 hours) 0.27 (0.21–0.3) p = 0.01‡ 0.40 (0.27–0.6)

* = Student’s t-test; † = Fisher’s exact test; ‡ = Mann-Whitney U test

To the Editor,

We write with regard to the Australian Society ofUltrasound in Medicine (ASUM) Policy Statement onnormal ultrasonic fetal measurements.1

The Campbell Westerway ultrasound fetal measure-ment charts2 were discussed in detail at a meeting ofthe Victorian Obstetric Sonologists (VOS), a groupcomprised predominantly of obstetricians who prac-tice ultrasound. This group has a number of concernsregarding these charts.

The initial concern was an impression that they werenot in keeping with both clinical experience and othercharts.3–5 Campbell Westerway’s early crown rumplength (CRL) measurements were particularly worry-ing. At less than eight weeks gestation they differgreatly from previous charts that are based on aknown date of ovulation.5

The selection of cases for the study appears to be lessthan exacting. Bias may be introduced by using casesscanned for confirmation of dates to compile the CRLdata. Why were these cases being scanned to confirmdates? The LMP is often a poor indicator of gestationalage, ovulation frequently occurring beyond day 14 ofthe menstrual cycle.6 We are aware of a letter to AssocProf de Crespigny from Ms Campbell Westerway inwhich she stated that a detailed menstrual history wasnot taken from the 3800 women used to develop thecharts ‘previous cycle length, contraceptive use wasnot considered’. Previous CRL charts have used meth-ods such as a carefully selected group of women4 or aknown date of conception5 to improve the accuracy oftheir data.

The stated aim of this study was to establish ‘normal’growth curves, a process that requires the collection oflongitudinal data. However, in the methods it is statedthat the study was cross sectional from 11 weeks. Theconcluding remark that existing charts are deficientin regard to fetal age/growth assessment adds to theconfusion and implies that the new charts should beused to assess both gestational age and fetal growth.

It is unclear whether all 3800 fetuses scanned at 11–40weeks were included in the compilation of each of thecharts (CRL, FL, etc.). Given the inevitability thatthere would be difficulty in obtaining measurementsin some cases, a statement should be made as towhether such cases are excluded.

The equipment used should be specified rather thansimply providing the blanket statement ‘the practicesinvited to participate all used well maintained, state ofthe art real time ultrasound systems’.

Developing charts as a reference range requires obses-sive attention to accuracy, both in relation to patientselection (a careful menstrual history) and fetal mea-surements. A minimal number of dedicated operators,

LETTERS

114

and an assessment of inter- and intra-observer vari-ability in measurements is essential. The use of alarge number of sonographers, including supervisedtrainees, at 28 practices, with no auditing of how mea-surements were being taken is not an ideal way togather data for such a purpose.

There is a minimal description of how the measure-ments were obtained. This is particularly worryingsince so many individuals were involved in collectingdata. One would hope that ‘instructions for measure-ment’ were sent to the contributing sonographers, butthe author has not provided any such instructions.

The precise gestational age for a given measurementis unclear from the tables. We are aware from letters toboth Assoc Prof de Crespigny and Dr Nisbet thatrounding down was used; however, those who haveread Campbell Westerway’s paper and use her chartswould not be aware that this process has been used.Rounding down means that ‘18 week pregnancies’included all pregnancies from 18 weeks 0 days to 18weeks 6 days. It would be more helpful (and more accu-rate) to tabulate using set increments, such as 1 mmfor the biparietal diameter (BPD), since, at some ges-tations, the existing tables will be difficult to use clin-ically. An illustration of the problem is that the differ-ence in BPD between 26 and 27 weeks is just 1 mm inthe Campbell Westerway charts where 3 mm would bethe expected difference. One explanation might be thatdata was collected from fetuses at the upper end of the26 weeks gestational age range and the lower end ofthe 27 weeks age range. It is important to be preciseabout the gestational age for a given measurement,particularly at earlier gestations, as there is a movetoward abandoning the use of the LMP in favour ofusing ultrasound to more accurately estimate the duedate.7

The author notes ‘the head circumference was calcu-lated independently using the formula (BPD + OFD) x1.57. This gave equivalent results to the ellipse methodof measuring’. Data to support this statement arerequired. The head circumference (HC) is normallymeasured at the outer perimeter of the skull. Theauthor indicated that in these charts the BPD is mea-sured outer skull to inner skull and the occipitofrontaldiameter (OFD) is measured mid skull to mid skull.These two techniques should produce significantly dif-ferent results, the results should not be ‘equivalent’.

Finally, some aspects of the statistical analysis arequestionable. It is usual to carry out a regressionanalysis of how an outcome (ie dependent) variable,such as gestational age, varies with a predictor (ieindependent) variable, such as BPD. Only the oppositeanalysis has been performed for some of the charts.Furthermore, the validity of the reference rangedepends critically on the assumption of normalityand equal standard deviations among the outcomevariable for all values of the predictor variable. There

is a lack of comment on or demonstration of equalstandard deviations for each measurement (het-eroscedasticity) through the gestation range. Theupper and lower limits of a calculated reference rangeare potentially affected by the lack of demonstrationof normality of the residuals at each data collectionpoint (gestation). It would be very helpful to see theraw data with the overlaying modeled curves.

We feel that there are too many unanswered questionsabout the Campbell Westerway charts for them to berecognised as the Australian standard. One must becareful not to let the wish to use Australian data be areason for abandoning tried and tested charts whenthe Australian charts have not been derived with therigour expected in current practice.

Debbie NisbetHugh Robinson

Jane Halliday Lachlan de Crespigny

Victorian Obstetric UltrasonologistsC/o Womens Ultrasound Malvern

Suite 2, 147 Wattletree RoadMalvern

Victoria 3144 Australia

REFERENCES

1 Australasian Society for Ultrasound in Medicine. Statement onnormal ultrasonic fetal measurements. ASUM Bulletin. 2001; 4:28–31.

2 Campbell Westerway S. Ultrasonic fetal measurements: newAustralian standards for the new millennium. Aust N Z J ObstetGynaecol. 2000; 40: 297–302.

3 Australasian Society for Ultrasound in Medicine: Policies andStatements. November 1998; D7.

4 Robinson HP, Fleming JE. A critical evaluation of sonar ‘crown-rump length’ measurements. Br J Obstet Gynaecol. 1975 Sep; 82(9): 707–710.

5 Silva PD, Mahairas G, Schaper AM, Schauberger CW. Earlycrown rump length. A good predictor of gestational age. J ReprodMed. 1990 Jun; 35 (6): 641–644.

6 Walker EM, Lewis M, Cooper W, Marnie M, Howie PW. Occultbiochemical pregnancy: fact or fiction? Br J Obstet Gynaecol. 1988Jul; 95 (7): 659–663.

7 Gardosi J. Dating of pregnancy: time to forget the last menstrualperiod. Ultrasound Obstet Gynecol. 1997 Jun; 9 (6): 367–368.

Reply from the Author

Thank you for allowing me to respond to the con-cerns by some Victorian Obstetric Sonologists (VOS)members.

The previous chart recommendations by theAustralian Society of Ultrasound Medicine (ASUM)are for fetal age, not growth, and it is these charts thatmine replaced. The terms fetal age chart/fetal growthchart are probably incorrectly interchanged, but it is afact of life that very few ultrasound centres use differ-ent charts for age and growth reference. We must beguided by common sense that, although a biparietaldiameter (BPD) may vary by 1 mm depending onwhether an age or growth chart was used, the statisti-cal difference is not significant.

LETTERS

115

As recommended by the British Medical UltrasoundSociety (BMUS) and mentioned in my article, the dat-ing of my charts is in completed weeks.

In regards to using a cross-sectional study, Spencer etal1 made an important point ‘The narrower limits oflongitudinally derived data do not reflect cross-sec-tional population variation and so are not appropriatefor determination of population-related size limits’. Ifyou refer to the head and abdominal circumferenceworks of Deter and Hadlock2, they discuss the longitu-dinal (20 pts scanned at 2–3week intervals) vs crosssectional (252 pts) evaluation of HC and AC and foundthe predicted values were in close agreement withmean standard deviation differences of 1.97% HC and1.38% AC. Similar work and results was also done bythem with the BPD.

In retrospect, my study design should have asked forcycle length and ovulation history, but in the clinicalconditions used to collect the data, this would havebeen difficult. LMP was chosen as this was one of thecriteria used for all previous charts. I also felt thatvariances in cycle length over the large sample popu-lation are compensated for in the mean figures. Toobtain the wide range of socio economic classes andethnic groups, I had to involve experienced ultrasoundpractices from across New South Wales. This is in con-trast to all previous ASUM recommended charts(Robinson/Silva – CRL; Hadlock – HC, FL; Deter – AC;and De Crespigny – BPD) collected from restrictedgroups. eg middle class white Caucasians. The prac-tices used a variety of state of the art, well maintainedultrasound equipment operated by qualified sonogra-phers. Measuring instructions were preciselyexplained. A subsequent study of inter/intra sonogra-pher evaluation of measuring accuracy was presentedat 11th World Congress of US in O and G (ISUOG) inMelbourne and involved both qualified and unquali-fied sonographers from 20 practices, given no measur-ing instructions, with very favourable results, particu-larly with students. It must be remembered that therewas no international uniformity for BPD levels until1991 when the American College of Obstetricians andGynecologists recommended the level of the cavumseptum pellucidum and thalamic nuclei.3

With regard to CRL data, no patients with bleeding orchecking for viability were included. The resolutionand measuring callipers of circa 1997 ultrasoundmachines were certainly more accurate than 10 yearsprior when Silva collected his data and undoubtedlysince Robinson’s 1975 work where you were lucky tosee a very early fetal pole let alone measure it accu-rately. I respect Dr Robinson immensely and his con-tribution to ultrasound cannot be questioned.Comparing the figures of the combined Silva/Robinson chart in ASUM policies and statements(June 1991), the comment that CRL ‘differ(s) greatly’from previous work based on known ovulation – there

is a 2 mm (or two-day) difference from 6–7 weeks –meaning it is essentially the same.

The equation used for head circumference has beenproved by Hadlock4 and Jeanty5 to be equivocal. Iasked Dr Jeanty at ISUOG in Melbourne if he still heldthis equation to be correct and he confirmed that hedid.

All regression curves were created using the same for-mat and methods used by Hadlock and Deter to createthe charts that mine have replaced. A complete set ofregressions for all parameters were provided to theEditor, ANZJOG, but did not appear in the article dueto a communication breakdown. The standard devia-tions given were the true statistical variations seen,not a smoothed version. Complete sets of data forsmoothed, non-smoothed, differences in millimetresor days and regressions are available from ASUM.Ultrasound companies have been given all data to loadinto their systems in a form suitable for their soft-ware. Dr Rob Heard, a statistician from SydneyUniversity checked the statistical analysis by indepen-dently reproducing it using SPSS and Minitab statisti-cal packages. Apart from very minor differences dueto rounding, the results matched. He is satisfied theresults are accurate and is happy to be quoted as sup-porting the work. Two of Australasia’s most promi-nent ultrasound personnel, Dr Peter Warren from theRoyal Hospital for Women at Randwick and Dr DavidRobinson, co-author of previous ASUM BPD charts,gave important input to the study during its forma-tion. Dr Warren has read and supports this reply to theVictorian Obstetric Sonologists’ concerns.

It has taken 25 years for Australian charts to be pro-duced for our population. If some clinicians disagreewith them then maybe they can undertake anotherstudy – ensuring it incorporates all socio-economicgroups and represents a full facet of ethnic women. Inthe intervening years, however, maybe the charts putforward and accepted by ASUM in 2001 are of morevalue than data derived from the study of 20-year-oldmiddle class white Americans.

Susan Campbell Westerway MAppSc DMU AMS

REFERENCES

1 Spencer JA, Chang TC, Robson SC, Gallivan S. Fetal size andgrowth in Bangladeshi pregnancies. Ultrasound Obstet Gynecol.1995 May; 5 (5): 313–317.

2 Deter RL, Harrist RB, Hadlock FP, Carpenter RJ. Fetal head andabdominal circumferences: II. A critical re-evaluation of the rela-tionship to menstrual age. J Clin Ultrasound. 1982 Oct; 10 (8):365–372.

3 Benson CB, Doubilet PM. Sonographic prediction of gestationalage: accuracy of second- and third-trimester fetal measurements.AJR Am J Roentgenol. 1991 Dec; 157 (6): 1275–1277.

4 Hadlock FP, Kent WR, Loyd JL, Harrist RB, Deter RL, Park SK.An evaluation of two methods for measuring fetal head and bodycircumferences. J Ultrasound Med. 1982 Nov–Dec; 1 (9): 359–360.

5 Jeanty P. Fetal biometry. In: Fleischer A, Romero R, Manning F,Jeanty P, editors. Principles and Practice of Ultrasonography inObstetrics and Gynecology. 4th edition 1991. Prentice-HallInternational Editions.

LETTERS