820 Correspondence

reached such firm conclusions on the potential of CA 125 for early diagnosis. Little data exist on this subjcct and her statement that CA 125 has no place in screening asymptomatic women for epithelial ovarian carcinoma is premature, if not wrong, for the following reasons: 1. Normal uppcr limit. The upper limit of 35 Uiml used in most studies of monitoring estab- lished ovarian cancer was arbitrarily defined by Bast et al. (1983) on the basis of serum levels in male and female blood donors of median age 34 years. The normal range of serum levels for postmenopausal women has not yet been estab- lished; 35 Uiml may be inappropriate for a screening programme which aims to achieve high sensitivity in postmenopausal women at risk. Preliminary results from the first 1010 apparently healthy postmenopausal women (mean age 54.2 years. range 45-75 years) studied at The London Hospital indicate that the 90th, 95th and 98th centiles for serum CA 125 in these women are 23.2 Uiml, 26.3 Uiml and 30.85 U/mi, respectively. 2. Specificity. It is not necessary for a serum test for early ovarian cancer to be 100% specific. Ultrasound scanning is an extremely good method of imaging the postmenopausal ovary (Campbell et al. 1982) and a suitable second line screening test (cf. cervical cytology and colpo- scopy). It may be appropriate to sacrifice a degree of specificity in order to increase the sen- sitivity for early stage disease. Furthermore many of the non-malignant causes of an elevated CA 125 level do not occur in postmenopausal women who would be the target of a screening progranime (e.g. endometriosis, pregnancy. pelvic inflammatory disease). 3. Sensitivity. Dr Lambert emphasized the need for a critical tumour mass to cause elevated serum CA 125 levels, referring in particular to 11 patients with persistent disease nodules less than 1 cm diameter at second look surgery who had CA 125 levels less than 35 Uiml (Canney el al. 1984). As clinically diagnosed stage I ovarian malignancies are rarely less than 5 cm diameter at presentation, critical mass isunlikely to be the cause of low serum levels of CA 125 in these women. Preoperative serum CA 125 levels in stage I and TI disease have not been reported in a large series of cases or in relation to an appropri- ate upper limit of normal for screening postmenopausal women.

CA 125 may not be the ultimate answer to the quest for early diagnosis of ovarian cancer but it

would be foolish to dismiss this possibility before the relevant information is available.

lan Jacobs Research Fdlow

David Oram Consultant Gynaecologist

Department of Obstetrics urid Gynascology The Loridon Hospital (Whitechapelj

London E l 1 B B

References

Bast, R. C., Klug, T. L., St John. E. et al. (1983) A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer. N Eng! .T Med 309, 883-887.

Campbell, S . , Goesscns, L., Goswamy, R. & White- head. M. (1982) Real time ultrasonography for determination of ovarian morphology and volume: possible early screening test for ovarian cancer? Lancet i, 425.

Canney, P. A , , Moore. M. , Wilkinson, P. M. &James. K. D. (1YS4) Ovarian cancer antigen CA 125: a prospective clinical asscssnient of its role as a tumour marker. Br J Cancer 50, 765-769.

Lambert. J. (1987) Commentary. The value o f 125 serum assay in the management of ovarian cancer. Br J Obster Gynuecol94, 193-195.

Authors’ reply

Dear Sir. Effective screening for occult ovarian cancer requires tests that are both sensitive and specific.

Data so far suggest that CA 125 is elevated in most patients with ovarian cancer. As a single test, however, CA 125 is not sufficiently specific to permit cost effective screening of an appar- ently healthy population. Drs Jacobs and Oram report on their preliminary results from 1010 apparently healthy postmenopausal women, indicating that in order to achieve maximal sen- sitivity, the upper normal limit for serum CA 125 should bc set at 23 unitsiml. This i s probably a very valid suggestion? but more data arc still required. Although ultrasound scanning is an extremely good method of imaging the postmenopausal ovary, its place as a screening test has not been fully established, nor whether it offers any significant advantages over vaginal examination.

The lack of specificity and sensitivity using a single marker such as CA 125 may be improved

Correspondence 821

branes are a valuable additional source of decidua vera containing portions of decidual spi- ral arteries, which may show acute atherosis. Difficulties in interpretation of pathologic lesions arise in the uteroplacental arteries them- selves. In their decidual segments the vessels havc been altered almost beyond recognition through involvement with endovascular trophoblast and it becomes very difficult to differentiate the characteristic fibrinoid material in the vessel wall from the fibrinoid necrosis of an acute vasculopathy. Furthermore: scnsesccnt cells in the physiologic changes, including trophoblast, accumulate lipid in their cytoplasm’.

We found absence of physiological changes in basal plate associated with acute atherosis in basal plate and/or parietal decidua in cases with idiopathic intrauterine growth retardation and in those with pre-eclampsia with or without intrauterine growth retardation (Althabe et al. 1985; Labarrere & Althabe 1985). Acute athe- rosis was described in idiopathic intrautcrine growth retardation, pre-eclampsia with or with- out intrauterine growth retardation. insulin- dependent diabetes, sustained chronic hyper- tension with or without superimposed pre- eclampsia, first-trimester abortions and also in some autoimmune diseases (Robertson et al. 1981; Sheppard & Bonnar 1981; Kitzmiller et al. 1981; Althabe et al. 1985; Labarrere & Althabe 1985; Lichtig et al. 1985; Labarrere et ul. 1986). Similar vascular lesions were described in rejected renal, cardiac and hepatic transplants (Dempster etal. 1964; Iless el ul. 1983; Demetris et al. 1985).

A possible immunological origin of atherosis is suggested by thc finding of immunoglobulins and complement in the walls of damaged placental vessels (Laberrere etal. 1985). A maternal immu- nological aggression against placental tissues may cause a deficit in placentation, which may be related to the matcrnal vasculopathies described elsewhere. Low birthweight may be related to placental and fetal ischaemia secondary to a high extension and severity of maternal vascular lesions in the placenta. The hypertension of preg- nacy may be a compensatory mechanism to ensure an adequate blood supply to the placcnta when the physiological changes fail to occur and then may exert a beneficial influence on the development of the fetus (Gerretsen et 01. 1981). When this compensatory mechanism is either insufficient or does not appear, the inadequate

by using serum markcrs in combination, but pro- spcctive studies are required to test this. 1 agree with Dr Jacobs and Dr Oram that a specific tumour marker for epithelial ovarian carcinoma is highly desirable, both as a screening test, and in monitoring patients who are apparcntly in clinicai remission, but I will still stick to my earlier conclubion that CA 125 on its own has no place in screening asymptomatic women tor epi- thelial ovarian carcinoma. Its value, together with other tumour markers and with other diag- nostic approaches, should be assessed in prop- erly conducted trials.

Joanna Lamhrt Consultant Radiotherapist and Oncologibt

Hammersmith Hospital Did Cane Road

London 7V12 OHS

The relation of birthweight to histological appearances in vessels of the placental bed

Dear Sir, McFadyen eral. [RrJ Ohstet Gynaecol(1986) 93, 47&481] found a significantly lowcr birthweight in the babics born to mothers with atherosis than in those with other placental bed histology. Hypertension also occurred most frequcntly in those women who had athcrosis in the placental bed. Matcrnal hypertension and a lower birth- weight in the babies were also observed in cases with absence of thc normal physiological changes of pregnancy. Interestingly, the authors pointed out the marked reduction in birthweight associated with atherosis, greater than in those whose vessels show only inadequate physiologi- cal changcs.

The authors stated that examination of the vessels attached to the chorionic membranes did not add any useful information. Robertson et al. (1986) mentioned that ‘the best material to use to identify acute atherosis is thc decidua Vera or parietalis and not the placental bed or the basal plate of the placenta. The decidua vera obviously ib devoid of migratory trophoblast and its associated effects on maternal tissues, includ- ing spiral arteries. The vasculopathy is therefore seen in the decidua Vera in its ‘pure’ form, uncomplicated by preexisting trophoblast- induced vascular changes. Thus, in btudies con- cerned mainly with vasculopathies, it is good practice to take biopsy specimens of the decidua parietalis in addition to those from the placental bed. Rolled-up strips of the placental mem-