- Email: [email protected]
The frequency of significant pathology in women attending a general gynaecological service for postcoital bleeding
British Journal of Obstetrics and Gynaecology January 2001, Vol. 108, pp. 103±106
The frequency of signi®cant pathology in women attending a general gynaecological service for postcoital bleeding Adam N. Rosenthal a, Theo Panoskaltsis a, Trudy Smith a, W.P. Soutter b,* Objectives To document the frequency of pathology in women who complain of postcoital bleeding. To determine whether negative cervical cytology excludes serious pathology in women with postcoital bleeding. To determine whether postcoital bleeding increases the risk of serious pathology in women with an abnormal smear. Design A retrospective study. Setting A university teaching hospital. Population 314 women with postcoital bleeding seen in the gynaecology service from ®rst January 1988 to 31 December 1994. Methods Women were identi®ed from the computerised records of the colposcopy service and copies of correspondence, which was routinely retained on computer. The latter was searched for the text strings coital and intercourse. Main outcome measure Histopathological diagnosis. Results Twelve women (4%) had invasive cancer: 10 were cervical or vaginal cancers and two endometrial cancers. Eight of the 10 cervical or vaginal cancers were clinically apparent. Four women of these 10 had had a normal smear before being referred for further investigation of postcoital bleeding. Two of these cancers were visible only with the aid of the colposcope. Thus, 0.6% of women attending a gynaecology service with postcoital bleeding, a normal looking cervix and a normal smear had invasive cancer of the cervix. Cervical intraepithelial neoplasia were found in 54 women (17.%) and 15 women (5%) had cervical polyps. Nineteen of the 63 women (30%) with signi®cant pathology had a normal or in¯ammatory cervical smear. No explanation for the postcoital bleeding was found in 155 women (49 %). Conclusions Although invasive cancer is rare in women with postcoital bleeding, it is much commoner than in the general population. It seems likely that cervical intraepithelial neoplasia is also associated with postcoital bleeding, perhaps because the fragile cervical epithelium becomes detached during intercourse. Postcoital bleeding should continue to be regarded as an indication of high risk for invasive cervical cancer and for cervical intraepithelial neoplasia. Prompt referral to a colposcopy clinic is indicated, but most women with postcoital bleeding will have no serious abnormality.
INTRODUCTION Postcoital bleeding is regarded as a cardinal symptom of cervical cancer. In two case series from the United States 1,2 postcoital bleeding occurred as a presenting symptom in 6% and 10% of 81 and 231 women with cervical cancer, respectively, while in a smaller case series from the United Kingdom 3, all women with cervical cancer under the age of 65 who presented with sympa
Departments of Obstetrics and Gynaecology, Imperial College School of Medicine, Hammersmith Hospital, London, UK b Division of Paediatrics, Obstetrics and Gynaecology, Imperial College School of Medicine, Hammersmith Hospital, London, UK * Correspondence: Mr W. P. Soutter, Institute of Obstetrics and Gynaecology, Hammersmith Hospital, Du Cane Road, London W12 0HS, UK. q RCOG 2001 British Journal of Obstetrics and Gynaecology PII: S 0306-545 6(00)00008-5
toms had postcoital bleeding. However, postcoital bleeding is a common gynaecological symptom with many causes and very few women with postcoital bleeding have invasive cervical cancer. To our knowledge, only one study has investigated the frequency of pathological changes in the cervix in association with postcoital bleeding. This study was conducted in India where the incidence of cervical cancer is very much higher than in European countries, and so extrapolation of these results to the United Kingdom may not be appropriate. The objective of this study was to determine the prevalence of pathological changes in the cervix in women referred with postcoital bleeding to a hospital in the United Kingdom. We also wanted to determine whether negative cervical cytology excludes serious pathology in women complaining of postcoital bleeding and whether postcoital bleeding increases the risk of serious pathological changes in the cervix in women with an abnormal smear. It was hoped that such information might enable us to develop recommendations for treatment. www.bjog-elsevier.com
104 A.N. ROSENTHAL ET AL.
METHODS
Table 1. The main reasons for referral in women with postcoital bleeding.
We performed a retrospective study of women referred to the colposcopy clinic with postcoital bleeding between 1 January 1988 to 31 December 1994. Identi®cation of these women was taken from computer records, and details were extracted from the case notes. Not all women with postcoital bleeding would have been referred to the colposcopy clinic; some women with clinically obvious cervical or vaginal cancer did not require a colposcopic examination, and some women with minor postcoital bleeding and normal cervical smear were not referred for a colposcopic examination. Women who had postcoital bleeding but did not attend the colposcopy clinic were identi®ed by searching the correspondence ®les of all women attending the two general gynaecology clinics during the study period. All letters and discharge summaries are routinely stored in a ®le on computer, and these were searched for the text strings coital and intercourse. The ®les identi®ed were then inspected to identify women about whom it was written that they had postcoital bleeding or `bleeding after intercourse'. The clinical records of these women were scrutinised to determine the investigations performed and the ®nal diagnosis. For all these women, the reasons for referral, the results of investigation and the ®nal diagnosis were recorded on a form and entered into a computerised database. An adequate length of follow up was ensured by not including women after the end of 1994. In a second matched control study of the effect of postcoital bleeding on the prevalence of pathology in women with abnormal cytology, a control group of 73 women without postcoital bleeding was identi®ed who had been investigated in the colposcopy clinic between 1 January 88 and 31 December 94 because of an abnormal smear. They were matched for age (to within ®ve years) and grade of cervical smear with the women postcoital bleeding who had an abnormal smear. The study and control groups were compared overall using the x 2 test. The proportions of women with invasive cervical or vaginal cancer were compared using Fisher's exact test.
Main reason for referral
n
(%)
Postcoital bleeding Abnormal smear Menstrual abnormality Pain Postmenopausal bleeding Others Total
179 73 43 12 3 4 314
57 23 14 4 1 1 100
A cervical smear was taken from all but seven women, none of whom was subsequently found to have signi®cant pathology. Four of the seven women underwent uterine curettage, one became pregnant, and the symptoms disappeared in the other two women. In all, 222 women were examined colposcopically. Ninety-two women with postcoital bleeding had a normal cervical smear and were seen in a general gynaecology clinic, 11 had cervical polyps and one had a frankly invasive cervical cancer. Diagnostic biopsies were taken from 160 women: 111 were punch biopsies; 35 were endometrial biopsies; and 14 were cone biopsies of one sort or another. Nine cancers of the cervix, one of the vagina and two endometrial cancers were diagnosed (Table 2). Seven of the cervical or vaginal tumours were clinically obvious lesions and four of them had been referred for investigation of postcoital bleeding, having had a previously normal smear. Two cancers were visible only colposcopically, one microinvasive and the other located entirely within the endocervical canal. Thus, invasive cancer of the cervix was found in 0.6% of women attending hospital with postcoital bleeding, a normal looking cervix and a normal smear. Eight of the 37 women with cervical intraepithelial neoplasia Stages II-III, and seven of the 17 with cervical intraepithelial neoplasia Stage I were referred because of postcoital bleeding only, the smear having been reported as negative or in¯ammatory. In all, 19 of the 63 (30%) women with signi®cant cervical pathology had a negative or in¯ammatory cervical smear. Table 2. Diagnoses in women with postcoital bleeding.
RESULTS
Diagnosis
In all, 314 women with postcoital bleeding were identi®ed: 189 from the colposcopy records and 125 from correspondence. Their mean age was 34.4 years (SD 9.4, range 18-64). The main indications for referral are shown in Table 1. Most women had more than one reason for being referred. General practitioners referred 257 of the women, the remaining 57 women being referred either from other sources. Four of the 10 women with cervical or vaginal cancer were referred from other hospitals.
Cervical cancer 9 3 Vaginal cancer 1 0.3 Endometrial cancer 2 0.6 CIN II-III 37 12 CIN I 17 5 Human papilloma virus 18 6 In¯ammatory changes or Metaplasia 49 16 Cervical polyp 15 5 Dysfunctional uterine bleeding 11 3 No abnormality 155 49 Total 314 100
n
(%)
No. with normal or in¯ammatory smear 4 0 0 8 7 6 29 13 11 112 190
q RCOG 2001 Br J Obstet Gynaecol 108, pp. 103±106
FREQUENCY OF PATHOLOGY IN POSTCOITAL BLEEDING 105 Table 3. Diagnoses in women with postcoital bleeding and an abnormal smear and age-matched controls with abnormal smear only. Values are given as n (%). x 2 test P 0.08. Population Abnormal smear and PCB Abnormal smear only a
Invasion
CIN 2/3
CIN 1
No. abnormality detected
Total n
4 (5 a) 1 (1 a)
26 (36) 40 (55)
11 (15) 8 (11)
33 (44) 24 (33)
73 73
Fisher's exact test P 0.36.
Both of the endometrial cancers were diagnosed by curettage but smears from both showed abnormal glandular cells. One of these women complained of postmenopausal bleeding. Neither the severity nor the duration of the postcoital bleeding was a reliable indication of invasive cancer. A total of 73 women were referred because of both postcoital bleeding and an abnormal smear. They were matched as described in the methods with a control group who had been referred because of an abnormal smear but did not complain of postcoital bleeding. There were more women with invasive cancer and fewer with cervical intraepithelial neoplasia Stages II-III in the postcoital bleeding group but the differences were not statistically signi®cant (Table 3). DISCUSSION In this well-screened population, invasive cancer was found in only 3.8% of women who complained of postcoital bleeding and who were referred to the hospital. Women with invasive cervical cancer who have symptoms at the time of referral have more advanced disease 4, so it was not surprising that all but two of the 10 invasive cervical or vaginal cancers in this series were clinically apparent. Because some of these women were tertiary referrals known to have cervical cancer, 3.8% may be an overestimate of the true prevalence in women with postcoital bleeding. However, the background incidence of cervical cancer in England and Wales is only 0.02% so the risk of invasive disease in women with postcoital bleeding does seem to be increased dramatically. Comparison of the women with postcoital bleeding and abnormal cervical smear with the matched controls without postcoital bleeding is hampered by the small numbers. There was a tendency towards an increase in the rate of invasive cancer in women with postcoital bleeding, supporting orthodox teaching. Four of the 10 women with invasive cervical or vaginal cancer had normal smears at the time of referral, and so gynaecologists should be aware that a normal smear must not be regarded as reassuring in a woman with postcoital bleeding. We therefore concur with the recommendations of a Working Group of the Royal Australian Colleges of q RCOG 2001 Br J Obstet Gynaecol 108, pp. 103±106
General Practice and of Obstetrics and Gynaecology and the legal profession 5which recommended that general practitioners refer women for colposcopy if they have: 1. persistent postcoital bleeding; 2. postcoital bleeding associated with a single smear suggestive of cervical intraepithelial neoplasia stage 1 or worse; or 3. postcoital bleeding associated with repeated smears with minor atypia or wart virus changes. In a recent survey 6, 89% of general practitioners and 86% of family planning doctors said they would repeat the cervical smear in a women with postcoital bleeding and a normal smear 18 previously, suggesting that doctors in primary care recognise that postcoital bleeding is a potentially serious symptom. Cervical polyps have long been associated with postcoital bleeding but cervical intraepithelial neoplasia, metaplasia and non-speci®c in¯ammation have not. Cervical intraepithelial neoplasia is usually regarded as an asymptomatic condition, but in our study 15 of the 54 women (28%) with cervical intraepithelial neoplasia were referred solely because of postcoital bleeding, their smear having been normal or in¯ammatory. Cervical intraepithelial neoplasia and metaplastic epithelium are both thin and friable and readily become detached from the cervix. A population-based study of women with postmenopausal bleeding found that 2.0% had cervical intraepithelial neoplasia 7, and in another study, 5% of women with cervical ectropion reported postcoital bleeding 8. It is possible that postcoital bleeding in these women is due to the friability of cervical intraepithelial neoplasia. This hypothesis should be tested in a further study. In conclusion, these data show that women with postcoital bleeding appear to have a much greater risk of invasive cancer than the general population. A normal cervical smear in women with postcoital bleeding does not rule out the possibility of cervical intraepithelial neoplasia or invasive cancer. Prompt referral to a colposcopy clinic is indicated, but women with postcoital bleeding should be reassured that colposcopy is being performed only as a precaution, as in the vast majority of instances there will be no serious abnormality.
106 A.N. ROSENTHAL ET AL.
Acknowledgements The authors would like to thank Ms J. Thomas for assisting with the statistical analysis. References 1. Pretorius R, Semrad N, Watring W, Fotheringham N. Presentation of cervical cancer. Gynecol Oncol 1991;42:48±53. 2. Pardanini NS, Tischler LP, Brown WH, de Feo E. Carcinoma of cervix. NY State J Med 1975;75:1018±1021. 3. Slater DN. Multifactorial audit of invasive cervical cancer: key lessons for the national screening programme. J Clin Path 1995;48:405±407. 4. de Souza NM, Soutter WP, McIndoe AG, Gilderdale DJ, Krausz T. Stage I cervical cancer: tumor volume by magnetic resonance imaging
5.
6. 7. 8.
of screen-detected versus symptomatic lesions. J Natl Cancer Instit 1997;89:1314±1315. Fraser IS, Petrucco OM. Management of intermenstrual and postcoital bleeding, and an appreciation of the issues arising out of the recent case of O'Shea versus Sullivan and Macquarie pathology. Aust NZ J Obstet Gynecol 1996;36:67±73. Woodman CBJ, Richardson J, Spence M. Why do we continue to take unnecessary smears? Br J Gen Prac 1997;47:645±656. Gredmark T, Kvint S, Havel G, Mattsson L-A. Histopathological ®ndings in women with postmenopausal bleeding. Br J Obstet Gynaecol 1995;102:133±136. Goldacre MJ, Loudon N, Watt B, Grant G, Loudon JDO, McPherson K, Vessey MP. Epidemiology and clinical signi®cance of cervical erosion in women attending a family planning clinic. BMJ 1978;1:748±750.
Accepted 30 August 2000
q RCOG 2001 Br J Obstet Gynaecol 108, pp. 103±106
The frequency of signi®cant pathology in women attending a general gynaecological service for postcoital bleeding Adam N. Rosenthal a, Theo Panoskaltsis a, Trudy Smith a, W.P. Soutter b,* Objectives To document the frequency of pathology in women who complain of postcoital bleeding. To determine whether negative cervical cytology excludes serious pathology in women with postcoital bleeding. To determine whether postcoital bleeding increases the risk of serious pathology in women with an abnormal smear. Design A retrospective study. Setting A university teaching hospital. Population 314 women with postcoital bleeding seen in the gynaecology service from ®rst January 1988 to 31 December 1994. Methods Women were identi®ed from the computerised records of the colposcopy service and copies of correspondence, which was routinely retained on computer. The latter was searched for the text strings coital and intercourse. Main outcome measure Histopathological diagnosis. Results Twelve women (4%) had invasive cancer: 10 were cervical or vaginal cancers and two endometrial cancers. Eight of the 10 cervical or vaginal cancers were clinically apparent. Four women of these 10 had had a normal smear before being referred for further investigation of postcoital bleeding. Two of these cancers were visible only with the aid of the colposcope. Thus, 0.6% of women attending a gynaecology service with postcoital bleeding, a normal looking cervix and a normal smear had invasive cancer of the cervix. Cervical intraepithelial neoplasia were found in 54 women (17.%) and 15 women (5%) had cervical polyps. Nineteen of the 63 women (30%) with signi®cant pathology had a normal or in¯ammatory cervical smear. No explanation for the postcoital bleeding was found in 155 women (49 %). Conclusions Although invasive cancer is rare in women with postcoital bleeding, it is much commoner than in the general population. It seems likely that cervical intraepithelial neoplasia is also associated with postcoital bleeding, perhaps because the fragile cervical epithelium becomes detached during intercourse. Postcoital bleeding should continue to be regarded as an indication of high risk for invasive cervical cancer and for cervical intraepithelial neoplasia. Prompt referral to a colposcopy clinic is indicated, but most women with postcoital bleeding will have no serious abnormality.
INTRODUCTION Postcoital bleeding is regarded as a cardinal symptom of cervical cancer. In two case series from the United States 1,2 postcoital bleeding occurred as a presenting symptom in 6% and 10% of 81 and 231 women with cervical cancer, respectively, while in a smaller case series from the United Kingdom 3, all women with cervical cancer under the age of 65 who presented with sympa
Departments of Obstetrics and Gynaecology, Imperial College School of Medicine, Hammersmith Hospital, London, UK b Division of Paediatrics, Obstetrics and Gynaecology, Imperial College School of Medicine, Hammersmith Hospital, London, UK * Correspondence: Mr W. P. Soutter, Institute of Obstetrics and Gynaecology, Hammersmith Hospital, Du Cane Road, London W12 0HS, UK. q RCOG 2001 British Journal of Obstetrics and Gynaecology PII: S 0306-545 6(00)00008-5
toms had postcoital bleeding. However, postcoital bleeding is a common gynaecological symptom with many causes and very few women with postcoital bleeding have invasive cervical cancer. To our knowledge, only one study has investigated the frequency of pathological changes in the cervix in association with postcoital bleeding. This study was conducted in India where the incidence of cervical cancer is very much higher than in European countries, and so extrapolation of these results to the United Kingdom may not be appropriate. The objective of this study was to determine the prevalence of pathological changes in the cervix in women referred with postcoital bleeding to a hospital in the United Kingdom. We also wanted to determine whether negative cervical cytology excludes serious pathology in women complaining of postcoital bleeding and whether postcoital bleeding increases the risk of serious pathological changes in the cervix in women with an abnormal smear. It was hoped that such information might enable us to develop recommendations for treatment. www.bjog-elsevier.com
104 A.N. ROSENTHAL ET AL.
METHODS
Table 1. The main reasons for referral in women with postcoital bleeding.
We performed a retrospective study of women referred to the colposcopy clinic with postcoital bleeding between 1 January 1988 to 31 December 1994. Identi®cation of these women was taken from computer records, and details were extracted from the case notes. Not all women with postcoital bleeding would have been referred to the colposcopy clinic; some women with clinically obvious cervical or vaginal cancer did not require a colposcopic examination, and some women with minor postcoital bleeding and normal cervical smear were not referred for a colposcopic examination. Women who had postcoital bleeding but did not attend the colposcopy clinic were identi®ed by searching the correspondence ®les of all women attending the two general gynaecology clinics during the study period. All letters and discharge summaries are routinely stored in a ®le on computer, and these were searched for the text strings coital and intercourse. The ®les identi®ed were then inspected to identify women about whom it was written that they had postcoital bleeding or `bleeding after intercourse'. The clinical records of these women were scrutinised to determine the investigations performed and the ®nal diagnosis. For all these women, the reasons for referral, the results of investigation and the ®nal diagnosis were recorded on a form and entered into a computerised database. An adequate length of follow up was ensured by not including women after the end of 1994. In a second matched control study of the effect of postcoital bleeding on the prevalence of pathology in women with abnormal cytology, a control group of 73 women without postcoital bleeding was identi®ed who had been investigated in the colposcopy clinic between 1 January 88 and 31 December 94 because of an abnormal smear. They were matched for age (to within ®ve years) and grade of cervical smear with the women postcoital bleeding who had an abnormal smear. The study and control groups were compared overall using the x 2 test. The proportions of women with invasive cervical or vaginal cancer were compared using Fisher's exact test.
Main reason for referral
n
(%)
Postcoital bleeding Abnormal smear Menstrual abnormality Pain Postmenopausal bleeding Others Total
179 73 43 12 3 4 314
57 23 14 4 1 1 100
A cervical smear was taken from all but seven women, none of whom was subsequently found to have signi®cant pathology. Four of the seven women underwent uterine curettage, one became pregnant, and the symptoms disappeared in the other two women. In all, 222 women were examined colposcopically. Ninety-two women with postcoital bleeding had a normal cervical smear and were seen in a general gynaecology clinic, 11 had cervical polyps and one had a frankly invasive cervical cancer. Diagnostic biopsies were taken from 160 women: 111 were punch biopsies; 35 were endometrial biopsies; and 14 were cone biopsies of one sort or another. Nine cancers of the cervix, one of the vagina and two endometrial cancers were diagnosed (Table 2). Seven of the cervical or vaginal tumours were clinically obvious lesions and four of them had been referred for investigation of postcoital bleeding, having had a previously normal smear. Two cancers were visible only colposcopically, one microinvasive and the other located entirely within the endocervical canal. Thus, invasive cancer of the cervix was found in 0.6% of women attending hospital with postcoital bleeding, a normal looking cervix and a normal smear. Eight of the 37 women with cervical intraepithelial neoplasia Stages II-III, and seven of the 17 with cervical intraepithelial neoplasia Stage I were referred because of postcoital bleeding only, the smear having been reported as negative or in¯ammatory. In all, 19 of the 63 (30%) women with signi®cant cervical pathology had a negative or in¯ammatory cervical smear. Table 2. Diagnoses in women with postcoital bleeding.
RESULTS
Diagnosis
In all, 314 women with postcoital bleeding were identi®ed: 189 from the colposcopy records and 125 from correspondence. Their mean age was 34.4 years (SD 9.4, range 18-64). The main indications for referral are shown in Table 1. Most women had more than one reason for being referred. General practitioners referred 257 of the women, the remaining 57 women being referred either from other sources. Four of the 10 women with cervical or vaginal cancer were referred from other hospitals.
Cervical cancer 9 3 Vaginal cancer 1 0.3 Endometrial cancer 2 0.6 CIN II-III 37 12 CIN I 17 5 Human papilloma virus 18 6 In¯ammatory changes or Metaplasia 49 16 Cervical polyp 15 5 Dysfunctional uterine bleeding 11 3 No abnormality 155 49 Total 314 100
n
(%)
No. with normal or in¯ammatory smear 4 0 0 8 7 6 29 13 11 112 190
q RCOG 2001 Br J Obstet Gynaecol 108, pp. 103±106
FREQUENCY OF PATHOLOGY IN POSTCOITAL BLEEDING 105 Table 3. Diagnoses in women with postcoital bleeding and an abnormal smear and age-matched controls with abnormal smear only. Values are given as n (%). x 2 test P 0.08. Population Abnormal smear and PCB Abnormal smear only a
Invasion
CIN 2/3
CIN 1
No. abnormality detected
Total n
4 (5 a) 1 (1 a)
26 (36) 40 (55)
11 (15) 8 (11)
33 (44) 24 (33)
73 73
Fisher's exact test P 0.36.
Both of the endometrial cancers were diagnosed by curettage but smears from both showed abnormal glandular cells. One of these women complained of postmenopausal bleeding. Neither the severity nor the duration of the postcoital bleeding was a reliable indication of invasive cancer. A total of 73 women were referred because of both postcoital bleeding and an abnormal smear. They were matched as described in the methods with a control group who had been referred because of an abnormal smear but did not complain of postcoital bleeding. There were more women with invasive cancer and fewer with cervical intraepithelial neoplasia Stages II-III in the postcoital bleeding group but the differences were not statistically signi®cant (Table 3). DISCUSSION In this well-screened population, invasive cancer was found in only 3.8% of women who complained of postcoital bleeding and who were referred to the hospital. Women with invasive cervical cancer who have symptoms at the time of referral have more advanced disease 4, so it was not surprising that all but two of the 10 invasive cervical or vaginal cancers in this series were clinically apparent. Because some of these women were tertiary referrals known to have cervical cancer, 3.8% may be an overestimate of the true prevalence in women with postcoital bleeding. However, the background incidence of cervical cancer in England and Wales is only 0.02% so the risk of invasive disease in women with postcoital bleeding does seem to be increased dramatically. Comparison of the women with postcoital bleeding and abnormal cervical smear with the matched controls without postcoital bleeding is hampered by the small numbers. There was a tendency towards an increase in the rate of invasive cancer in women with postcoital bleeding, supporting orthodox teaching. Four of the 10 women with invasive cervical or vaginal cancer had normal smears at the time of referral, and so gynaecologists should be aware that a normal smear must not be regarded as reassuring in a woman with postcoital bleeding. We therefore concur with the recommendations of a Working Group of the Royal Australian Colleges of q RCOG 2001 Br J Obstet Gynaecol 108, pp. 103±106
General Practice and of Obstetrics and Gynaecology and the legal profession 5which recommended that general practitioners refer women for colposcopy if they have: 1. persistent postcoital bleeding; 2. postcoital bleeding associated with a single smear suggestive of cervical intraepithelial neoplasia stage 1 or worse; or 3. postcoital bleeding associated with repeated smears with minor atypia or wart virus changes. In a recent survey 6, 89% of general practitioners and 86% of family planning doctors said they would repeat the cervical smear in a women with postcoital bleeding and a normal smear 18 previously, suggesting that doctors in primary care recognise that postcoital bleeding is a potentially serious symptom. Cervical polyps have long been associated with postcoital bleeding but cervical intraepithelial neoplasia, metaplasia and non-speci®c in¯ammation have not. Cervical intraepithelial neoplasia is usually regarded as an asymptomatic condition, but in our study 15 of the 54 women (28%) with cervical intraepithelial neoplasia were referred solely because of postcoital bleeding, their smear having been normal or in¯ammatory. Cervical intraepithelial neoplasia and metaplastic epithelium are both thin and friable and readily become detached from the cervix. A population-based study of women with postmenopausal bleeding found that 2.0% had cervical intraepithelial neoplasia 7, and in another study, 5% of women with cervical ectropion reported postcoital bleeding 8. It is possible that postcoital bleeding in these women is due to the friability of cervical intraepithelial neoplasia. This hypothesis should be tested in a further study. In conclusion, these data show that women with postcoital bleeding appear to have a much greater risk of invasive cancer than the general population. A normal cervical smear in women with postcoital bleeding does not rule out the possibility of cervical intraepithelial neoplasia or invasive cancer. Prompt referral to a colposcopy clinic is indicated, but women with postcoital bleeding should be reassured that colposcopy is being performed only as a precaution, as in the vast majority of instances there will be no serious abnormality.
106 A.N. ROSENTHAL ET AL.
Acknowledgements The authors would like to thank Ms J. Thomas for assisting with the statistical analysis. References 1. Pretorius R, Semrad N, Watring W, Fotheringham N. Presentation of cervical cancer. Gynecol Oncol 1991;42:48±53. 2. Pardanini NS, Tischler LP, Brown WH, de Feo E. Carcinoma of cervix. NY State J Med 1975;75:1018±1021. 3. Slater DN. Multifactorial audit of invasive cervical cancer: key lessons for the national screening programme. J Clin Path 1995;48:405±407. 4. de Souza NM, Soutter WP, McIndoe AG, Gilderdale DJ, Krausz T. Stage I cervical cancer: tumor volume by magnetic resonance imaging
5.
6. 7. 8.
of screen-detected versus symptomatic lesions. J Natl Cancer Instit 1997;89:1314±1315. Fraser IS, Petrucco OM. Management of intermenstrual and postcoital bleeding, and an appreciation of the issues arising out of the recent case of O'Shea versus Sullivan and Macquarie pathology. Aust NZ J Obstet Gynecol 1996;36:67±73. Woodman CBJ, Richardson J, Spence M. Why do we continue to take unnecessary smears? Br J Gen Prac 1997;47:645±656. Gredmark T, Kvint S, Havel G, Mattsson L-A. Histopathological ®ndings in women with postmenopausal bleeding. Br J Obstet Gynaecol 1995;102:133±136. Goldacre MJ, Loudon N, Watt B, Grant G, Loudon JDO, McPherson K, Vessey MP. Epidemiology and clinical signi®cance of cervical erosion in women attending a family planning clinic. BMJ 1978;1:748±750.
Accepted 30 August 2000
q RCOG 2001 Br J Obstet Gynaecol 108, pp. 103±106