- Email: [email protected]
Residual and Recurrent Disease After Laser Conization for Cervical Intraepithelial Neoplasia
Residual and Recurrent Disease After Laser Conization for Cervical Intraepithelial Neoplasia F. E. SKJELDESTAD, MD, PhD, B. HAGEN, MD, PhD, A. K. LIE, MD, AND C. ISAKSEN, MD Objective: To assess the risk of residual and recurrent disease after carbon dioxide laser conization treatment for high-grade lesions of the cervix uteri, and thus to refine intervals for follow-up. Methods: The study population comprised all women treated for cervical intraepithelial neoplasia grade II–III over a 10-year period (1983–1992). All women who resided within the county for the entire follow-up period were followed for residual or recurrent disease, verified histologically until the closure date of January 31, 1996. The cumulative incidence of recurrent disease was assessed by survival analyses, and logistic regression was used to predict clinical features at initial treatment that were associated with the risk of residual or recurrent disease. Results: Only 19 of 1081 women (1.8%) were lost to follow-up. Residual disease was diagnosed in 20 (1.9%) study participants. The cumulative incidence of recurrent disease was extremely low, increasing nearly linearly with an annual incidence of three per 1000 woman-years observed. Involved resection margins were associated significantly with both residual disease (crude odds ratio [OR] 18.1; 95% confidence interval [CI] 5.2, 64.0) and recurrent disease (adjusted OR 3.0; 95% CI 1.2, 7.5) when compared with disease-free resection margins as reference. Conclusion: We recommend a differential follow-up interval depending upon the histologic evaluation of cone margins. If there is no residual disease, women who have free resection margins should return at a 3-year interval for follow-up. Women who have disease extended to the cone margins are recommended Papanicolaou smears at annual intervals through the fourth postoperative year before returning at a 3-year interval as practiced in the general screening program. (Obstet Gynecol 1997;90:428 –33. © 1997 by The American College of Obstetricians and Gynecologists.)
Organized screening with the Papanicolaou smear for cervical cancer was introduced on January 1, 1995, in Norway.1 Women aged 25– 69 years are invited by From the Departments of Obstetrics and Gynecology and Pathology, Medical Faculty, Norwegian University of Science and Technology, Trondheim, Norway.
428 0029-7844/97/$17.00 PII S0029-7844(97)00276-7
individual letters sent at 3-year intervals from the Mass Screening Unit located at the Norwegian Cancer Registry to participate in the national screening program against cervical cancer. Before the introduction of a nationwide screening program, opportunistic screening had for many years captured a great majority of cases, especially in younger women.2– 4 Depending upon attendance rates, smear sensitivity, and treatment efficacy for precancerous lesions, estimates from mathematical modeling of organized screening against cervical cancer with Papanicolaou smears show that under optimal conditions, the incidence of cervical cancer can be reduced by 70 –90%.5,6 Guidelines for the screening interval differ among the Scandinavian countries,7 as do recommendations for the follow-up interval after treatment of precancerous lesions of the cervix uteri.8 –13 Risk assessment data for residual and recurrent disease among women who have been treated for precancerous lesions are important for deciding upon the interval for Papanicolaou smear follow-up. Furthermore, knowledge of risk assessment is important for both patients and their physicians for improving compliance with existing follow-up guidelines. The aim of our study was to assess the risks for residual and recurrent disease by studying a cohort of women treated by laser conization for high-grade lesions of the cervix uteri (cervical intraepithelial neoplasia [CIN] II–III) and thereby provide data to refine the Papanicolaou smear follow-up interval after treatment.
Materials and Methods In 1982, the carbon dioxide laser replaced the cold-knife method of conization in women with precancerous lesions of the cervix at the Department of Gynecology, University Hospital of Trondheim, Norway. From January 1, 1983 to December 31, 1992, 1349 laser conizations were performed among 1320 women. Over the entire study period, laser conization was performed mainly using local anesthesia (96%) and on an outpa-
Obstetrics & Gynecology
tient basis (98%) by a few staff members. Our operative technique followed that described by Dorsey and Diggs,14 using a colposcope coupled to the laser. A focused 25–30-W beam was used for excision, and a cone-shaped specimen was produced by simultaneous traction with a hook. After removal of the cone, the surgical bed was vaporized by a defocused beam. Additional hemostatic sutures were rarely applied. Endocervical curettage (ECC) at the end of the surgery was not practiced. During the entire study period, the national guidelines for abnormal smears advised observation for CIN I, whereas for women with persistent CIN I, CIN II–III, or suspected malignant cells, the recommendation was referral to a gynecologic unit for colposcopically directed biopsies and ECC. Further treatment was always based upon a histologic diagnosis. In each case, we identified the most severe histopathologic diagnosis in the preoperative biopsies, ECC, or the cone specimen. We excluded from further analyses 55 women who had invasive disease and 32 women with diagnostic conization because of repeated discrepancy between the Papanicolaou smears and punch biopsies and/or ECCs. Eligible for the follow-up analyses were 1233 women treated initially for CIN II or CIN III. However, when defining the population-based cohort, we excluded 152 women who were not permanent residents of the county at initial treatment or who moved from the county during the follow-up period. A total of 1081 women were finally included in the study: 82 treated initially for CIN II and 999 for CIN III. The first follow-up visit took place at our clinic 3– 4 months postoperatively. Papanicolaou smear was the only screening method used to identify women who were eligible for histologic verification of residual and recurrent disease. If the first smear was normal, the woman was discharged to her general practitioner for further follow-up.8,9 Women with abnormal smears were referred for colposcopic biopsies and ECC. Women were defined as being free of residual disease when the first two smears taken within 6 –12 months postoperatively were normal.13 Women who had histologic findings extending to the margins of the cone were followed at equal intervals as women who had diseasefree margins. Our department of pathology is the only one in the county examining all cytologic and histologic specimens collected in our geographic area. Data before 1986 were obtained from individual cytology and histology records. Since the introduction of a computerized registry on January 1, 1986, we have gathered follow-up information on women treated for CIN II–III by linking the records to the registry at the department of pathology. In this way, we also identified women who had
VOL. 90, NO. 3, SEPTEMBER 1997
surgery for other gynecologic disorders, as well as those who had died and their plausible cause of death. All these data and other relevant clinical data were transferred to a standardized case-report form used for the clinical management of our study. We assessed the risk of recurrent disease by survival analysis among women who were free from residual disease. Time at risk was estimated as the time from conization to the most recent normal Papanicolaou smear. Women with abnormal smears but with normal histologic follow-up findings continued in the study as recurrence-free cases. However, women with a histologic confirmation of the abnormal smear were identified as having recurrent disease at the time of the actual smear. Women who had hysterectomy for reasons other than CIN were followed until the time of the hysterectomy, provided that histologic findings from the cervix uteri were normal. Women who died during follow-up were followed until their most recent normal smear unless an autopsy showed normal cervical histology, in which case they were followed to the date of their death. All analyses were done using SPSS version 4.1 (Statistical Package for Social Sciences; SPSS Inc., Chicago, IL). When appropriate, x2 statistics, logistic regression (forward stepwise approach), and survival analyses were applied. Differences between group means of continuous variables were examined by the MannWhitney test. P , .05 was considered statistically significant.
Results Over the 10-year study period, the average number of laser conizations per year increased from 96 during 1983–1985 to 106 during 1986 –1988, and to 119 during 1989 –1992. The mean age of the women treated for CIN II–III decreased significantly over the study period, from 35.7 years in the first period to 34.3 years in the last period (Table 1) (Mann-Whitney; z 5 22.03, P , .05). Over the study period, the proportion of nulliparous women and women with only one child also increased significantly (Table 1) (x2 trend 34.4, P , .001). Even though the highest CIN grade reported showed significantly more cases of CIN II during 1989 –1992, still more than 90% of the cases in this period were classified as CIN III disease (Table 2) (x2 trend 21.0, P , .001). The number of cases with free resection margins decreased significantly over the study years (Table 2) (x2 trend 18.8, P , .001), whereas few cases were reported as inconclusive with uncertain resection margins because of thermal trauma (Table 2). Twenty-eight women were excluded from analyses of residual disease because they never attended their first
Skjeldestad et al
Disease After Treatment of CIN
429
Table 1. Characteristics of the Study Population Characteristic Age (y) #29 30 – 40 $40 Mean Range Parity 0 1 $2 Unknown Previous treatment for CIN None Cryotherapy Laser vaporization
1983–1985 (n 5 287)
1986 –1988 (n 5 319)
1989 –1992 (n 5 475)
79 (27.5%) 129 (45.0%) 79 (27.5%) 35.7 21–76
89 (27.9%) 157 (49.2%) 73 (22.9%) 35.1 18 – 86
173 (36.4%) 190 (40.0%) 112 (23.6%) 34.3 19 –73
30 (10.5%) 50 (17.4%) 202 (70.4%) 5 (1.7%)
45 (14.1%) 70 (21.9%) 186 (58.3%) 18 (5.6%)
101 (21.3%) 132 (27.8%) 238 (50.1%) 4 (0.8%)
267 (93.0%) 16 (5.6%) 4 (1.4%)
315 (98.7%) 3 (1.0%) 1 (0.3%)
461 (97.1%) 5 (1.1%) 9 (1.9%)
CIN 5 cervical intraepithelial neoplasia.
scheduled visit (n 5 5), only completed their first visit with a normal smear (n 5 14), or underwent elective hysterectomy during the first 6 postoperative months for reasons other than CIN (n 5 9). Thus, 97.4% (1053 of 1081) of the initial cohort was eligible for analyses of residual disease. Over the study period, the number of cases with residual disease was low (20 of 1081, 1.9%). None of the 80 women with CIN II had residual disease. Seventeen of the 245 women (6.9%) with CIN III and involved resection margins had residual disease, compared with three of 728 women (0.4%) with CIN III and free resection margins (x2 trend 39.4, P , .001; crude odds ratio [OR] 18.1; 95% confidence interval [CI] 5.2, 64.0). However, three cases of invasive cancer that were not diagnosed during primary histologic examination were found among women with residual disease. These cancers were all diagnosed in tissue from ECC specimens and were not present in tissue from ectocervical punch biopsies. Table 3 lists the method of treatment by the most severe histologic finding among the women with residual disease. Table 2. Histologic Findings Finding CIN II CIN III Resection margins Free Not free Upper Lower Inconclusive
1983–1985 (n 5 287)
1986 –1988 (n 5 319)
1989 –1992 (n 5 475)
17 (5.9%) 270 (94.1%)
21 (6.6%) 298 (93.4%)
44 (9.3%) 431 (90.7%)
240 (83.6%)
237 (74.3%)
335 (70.5%)
37* (12.9%) 11* (3.8%) 0
47 (14.7%) 23 (7.2%) 12 (3.8%)
119 (25.1%) 11 (2.3%) 10 (2.1%)
CIN 5 cervical intraepithelial neoplasia. * Includes one woman in whom both endocervical and ectocervical margins were not free.
430 Skjeldestad et al
Disease After Treatment of CIN
Table 3. Indications and Treatment of Residual and Recurrent Disease Most severe histologic finding in follow-up tissue Treatment Residual disease Hysterectomy Cryotherapy Laser vaporization Laser conization Recurrent disease Hysterectomy Cryotherapy Laser vaporization Laser conization
Number
CIN I
CIN II
CIN III
Invasive cancer
7 1 3 9
0 0 0 0
0 0 1 1
6 1 2 6
1* 0 0 2†
6 0 3 10
0
0
5
1*
0 0
2 2
1 8
0 0
CIN 5 cervical intraepithelial neoplasia. * Patients had surgery by the Wertheim-Meigs operation. † Histology from the second cone was “free resection margins”; therefore no further treatment.
After the exclusion of 20 women with residual disease, 1033 women remained for analyses of the risk of recurrent disease. During the follow-up period, 23 women underwent hysterectomy unrelated to CIN, and seven died from causes not related to genital cancer. Nineteen of 1033 women were diagnosed with a recurrent disease verified histologically. Among these, one woman was diagnosed outside the screening program, because of symptoms (irregular bleeding and signs of blood in the discharge), with cervical squamous cell carcinoma (stage IB) 35 months after initial treatment. This woman had free cone margins and had had four postoperative normal Papanicolaou smears, the most recent one 9 months before the onset of symptoms. The most severe histologic findings and the methods of treatment at recurrence are listed in Table 3. In a logistic regression model, resection margins were the only predictor of recurrence among variables such as age, previous children, previous treatment of CIN before the index treatment, and CIN grade (Table 4). Women with involved resection margins were three times more likely to appear with a recurrence (adjusted OR 3.0; 95% CI 1.2, 7.5) than were women with free cone margins (Table 4). The incidence of recurrent disease increased nearly linearly during years 1– 6 of observation (Figure 1). The overall cumulative incidence of recurrence at 3 years of observation was seven cases per 1000 woman-years, and at 6 completed years of observation, it was 22 per 1000 woman-years. Women who had disease involving the cone margins had a significantly increased risk of recurrence compared with women who had free margins (Wilcoxon Gehan statistics; x2 7.5, P , .01) (Figure 1).
Obstetrics & Gynecology
Table 4. Risk of Recurrent Disease
Variable* Age (y) #34 $35 Parity 0 –1 $2 Unknown Previous treatment for CIN Yes No Time period 1983–1985 1986 –1988 1989 –1992 Most severe histologic finding CIN II CIN III Resection margins Free Not-free or inconclusive
Discussion
No. of No. of women in women with each category recurrence Adjusted (n 5 19) OR (n 5 1033)
95% CI
582 451
10 9
1.3 1.0
0.5, 3.6 Reference
408 599 26
10 9
1.5 1.0
0.6, 3.8 Reference Not analyzed
36 997
2 17
3.3 1.0
0.7, 14.7 Reference
269 310 454
5 9 5
1.0 1.3 0.5
Reference 0.4, 4.1 0.1, 1.7
80 953
2 17
1.0 0.7
Reference 0.2, 3.0
788 245
10 9
1.0 3.0
Reference 1.2, 7.5
OR 5 odds ratio; CI 5 confidence interval; CIN 5 cervical intraepithelial neoplasia. * The variables entered the model the way they are categorized in the table.
Separate analyses starting at 48 months after treatment did not find any difference in the cumulative incidence by resection margins.
This study confirms that laser conization is highly efficient for treating CIN II–III. The incidence of residual and recurrent disease is very low and is in agreement with other studies evaluating the treatment efficacy of laser conization6,7,15 or cold-knife conization.16,17 Invasive cancer might be underdiagnosed during initial treatment of high-grade lesions of the cervix uteri.15–17 Therefore, the most important issue in the follow-up of these patients is to rule out persistent disease. This point is underscored by the finding of three microinvasive cancers among 20 women with residual disease in our study.15 Both residual and recurrent disease were highly associated with disease extending to the resection margins. Despite these associations, barely 10% of the women with involved resection margins returned with either residual disease or recurrent disease,11 reflecting that resection margins are most often free even if the margins of the cone are determined to be involved. Others have reported a higher prevalence of residual disease among women with incomplete excisions.15,17 The most likely explanation for the varying prevalence of residual disease is varying depth of the conization. If the cone is excised too low, some affected endocervical tissue might easily remain at the site. Therefore, the need for first follow-up at a specialized unit is clearly documented by the finding of three microinvasive cancers in our study. We released the patients to their general practitioner for further follow-up if the first Papanicolaou smear was normal.8,9,15 We have not found continued colpo-
Figure 1. Cumulative incidence of recurrence by histologic findings in the resection margin.
VOL. 90, NO. 3, SEPTEMBER 1997
Skjeldestad et al
Disease After Treatment of CIN
431
scopic examinations or blind biopsies necessary in the surveillance of residual or recurrent disease.8,9,15 This position is supported by the extremely low incidence of recurrence.8 –11,15 During the study period, the national guidelines for Papanicolaou smear follow-up were four visits during the first year, two visits during the second year, and annual visits subsequently.8 –11 We have not practiced these guidelines strictly, but have advocated a follow-up based on two visits during the first 12 months and annually thereafter. These guidelines were introduced in Norway during 1993.18 Based on the linear shape of the cumulative incidence curve (Figure 1) of recurrent disease among women with free resection margins at an annual incidence of two per 1000 women, we conclude that these women can return, after residual disease has been ruled out, at a 3-year interval as practiced in the national screening program. The slope of the cumulative incidence curve of recurrent disease among women who have involved resection margins is steep between 12 and 48 months after treatment; thereafter, the curve runs parallel to that of women with free cone margins. After 48 months of observation, very few recurrences are diagnosed, even among women with disease extending to the cone margins. In our opinion, the cumulative incidence of recurrence among women with disease involving the cone margins qualifies them for annual visits through the fourth year after treatment and thereafter release to the general 3-year screening interval. Overall risk assessment is important for deciding on and ensuring compliance with the interval for followup. It should be reassuring for patients, gynecologists, and general practitioners alike to know that among women with free resection margins, the risk of recurrence is as low or even lower than the risk of ever being diagnosed with a histologically confirmed CIN II–III lesion in the general screening program.3,4 Furthermore, knowledge about the natural progression of CIN II–III to invasive disease is important for interval decisions. This process is modeled to take 14 or more years.5,6,19 –23 Therefore, among women in whom representative smears are collected, there is time enough within our differential approach to capture recurrent disease. So far, we have done surveillance on the assumption that these women are at high risk of developing cancer, implying that our therapy might be incomplete. Our results, however, show that the longterm efficacy of therapy is very high and that we can take advantage of the information in the pathology report on resection margins to develop a differential follow-up regimen.
432 Skjeldestad et al
Disease After Treatment of CIN
References 1. Bjørge T, Thoresen SØ, Gunbjørud AB. Masseundersøkelsen mot livmorhalskreft. Nå starter invitasjonene til kvinner på landsbasis [Norwegian]. Tidsskr Nor Lægeforen 1995;115:815– 6. 2. Haugen OA. Forbruk av vaginalcytologiske prøver i Norge [Norwegian]. Tidsskr Nor Lægeforen 1984;104:952– 6. 3. Forsmo S, Buhaug H, Stalsberg H. Use of Papanicolaou smears in a population without a mass screening program. Acta Obstet Gynecol Scand 1994;73:824 – 8. 4. Bjørge T, Gunbjørud AB, Langmark F, Skare GB, Thoresen SØ. Cervical mass screening in Norway—510,000 smears a year. Cancer Detect Prev 1994;6:463–70. 5. Forsmo S, Buhaug H, Skjeldestad FE, Haugen OA. Treatment of preinvasive conditions during opportunistic screening and its effectiveness on the cervical cancer incidence in one Norwegian county. Int J Cancer 1997;71:4 – 8. 6. IARC Working Group on Evaluation of Cervical Cancer Screening Programmes. Screening for squamous cervical cancer: Duration of low risk after negative results of cervical cytology and its implication for screening policies. BMJ 1986;293:659 – 64. 7. Sigurdsson K. Effect of organized screening on the risk of cervical cancer. Evaluation of screening activity in Iceland, 1964 –91. Int J Cancer 1993;54:563–70. 8. Scheistrøen M. Laserbehandling av premaligne sykdommer i cervix uteri [Norwegian]. Tidsskr Nor Lægeforen 1992;112:1145–7. 9. Forsmo S, Øian P, Stalsberg H. Cervikal intraepitelial neoplasi. Persistens og residiv etter laserbehandling [Norwegian]. Tidsskr Nor Lægeforen 1992;112:1148 –51. 10. Tabor A, Berget A. Cold-knife and laser conization for cervical intraepithelial neoplasia. Obstet Gynecol 1990;76:633–5. 11. Andersen ES, Pedersen B, Nielsen K. Laser conization: The results of treatment of cervical intraepithelial neoplasia. Gynecol Oncol 1994;54:201– 4. 12. Sigurdsson K. Quality assurance in cervical cancer screening: The Icelandic experience 1964 –1993. Eur J Cancer 1995;31A:728 –34. 13. Rylander E, Isberg A, Joelson I. Laser vaporization of cervical intraepithelial neoplasia. A five-year follow-up. Acta Obstet Gynecol Scand Suppl 1984;125:33– 6. 14. Dorsey JH, Diggs ES. Microsurgical conization of the cervix by carbon dioxide laser. Obstet Gynecol 1979;54:565–70. 15. Lopez A, Morgan P, Murdoch J, Piura B, Monaghan JM. The case of conservative management of “incomplete excision” of CIN after laser conization. Gynecol Oncol 1993;49:247–9. 16. Kolstad P, Klem V. Long-term follow-up of 1121 cases of carcinoma in situ. Obstet Gynecol 1976;48:125–9. 17. Vedel P, Jakobsen H, Kryger-Baggesen N, Rank F, Bostofte E. Five-year follow up of patients with cervical intraepithelial neoplasia in the cone margins after conization. Eur J Obstet Gynecol Reprod Biol 1993;50:71– 6. 18. Nøstdahl W. Cytologiske kontroller etter behandling av premaligne lidelser i cervix uteri [Norwegian]. Tidsskr Nor Lægeforen 1993;113:1884 –5. 19. Gustafsson L, Adami H-O. Natural history of cervical neoplasia: Consistent results obtained by an identification technique. Br J Cancer 1989;60:132– 41. 20. Van Oortmarssen GJ, Habbema JDF. Duration of preclinical cervical cancer and reduction in incidence of invasive negative Pap smears. Int J Epidemiol 1995;24:300 –7. 21. Albert A. Estimated cervical cancer disease state incidence and transition rates. J Natl Cancer Inst 1981;67:571– 6. 22. Prorok PC. Mathematical models and natural history in cervical cancer screening. In: Hakama M, Miller AB, Day NE, eds. Screening for cancer of the uterine cervix. Lyon: International Union Against Cancer, 1986:185–96.
Obstetrics & Gynecology
23. Eddy DM. The frequency of cervical cancer screening. Comparison of a mathematical model with empirical data. Cancer 1987;60: 1117–22.
Received January 24, 1997. Received in revised form May 6, 1997. Accepted May 21, 1997.
Address reprint requests to:
F. E. Skjeldestad, MD, PhD Department of Obstetrics and Gynecology University Hospital of Trondheim Trondheim, 7006 Norway E-mail: [email protected]
STANDARDS
Copyright © 1997 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
FOR REPORTING TRIALS
The CONSORT listing of standards for reporting randomized trials has been adopted as policy by Obstetrics & Gynecology. Investigators who are planning, conducting, or reporting randomized trials should be thoroughly familiar with these standards. A copy can be obtained by contacting: Obstetrics & Gynecology, 1100 Glendon Avenue, Suite 1655, Los Angeles, CA 90024-3520; FAX (310) 208-2838.
VOL. 90, NO. 3, SEPTEMBER 1997
Skjeldestad et al
Disease After Treatment of CIN
433
Organized screening with the Papanicolaou smear for cervical cancer was introduced on January 1, 1995, in Norway.1 Women aged 25– 69 years are invited by From the Departments of Obstetrics and Gynecology and Pathology, Medical Faculty, Norwegian University of Science and Technology, Trondheim, Norway.
428 0029-7844/97/$17.00 PII S0029-7844(97)00276-7
individual letters sent at 3-year intervals from the Mass Screening Unit located at the Norwegian Cancer Registry to participate in the national screening program against cervical cancer. Before the introduction of a nationwide screening program, opportunistic screening had for many years captured a great majority of cases, especially in younger women.2– 4 Depending upon attendance rates, smear sensitivity, and treatment efficacy for precancerous lesions, estimates from mathematical modeling of organized screening against cervical cancer with Papanicolaou smears show that under optimal conditions, the incidence of cervical cancer can be reduced by 70 –90%.5,6 Guidelines for the screening interval differ among the Scandinavian countries,7 as do recommendations for the follow-up interval after treatment of precancerous lesions of the cervix uteri.8 –13 Risk assessment data for residual and recurrent disease among women who have been treated for precancerous lesions are important for deciding upon the interval for Papanicolaou smear follow-up. Furthermore, knowledge of risk assessment is important for both patients and their physicians for improving compliance with existing follow-up guidelines. The aim of our study was to assess the risks for residual and recurrent disease by studying a cohort of women treated by laser conization for high-grade lesions of the cervix uteri (cervical intraepithelial neoplasia [CIN] II–III) and thereby provide data to refine the Papanicolaou smear follow-up interval after treatment.
Materials and Methods In 1982, the carbon dioxide laser replaced the cold-knife method of conization in women with precancerous lesions of the cervix at the Department of Gynecology, University Hospital of Trondheim, Norway. From January 1, 1983 to December 31, 1992, 1349 laser conizations were performed among 1320 women. Over the entire study period, laser conization was performed mainly using local anesthesia (96%) and on an outpa-
Obstetrics & Gynecology
tient basis (98%) by a few staff members. Our operative technique followed that described by Dorsey and Diggs,14 using a colposcope coupled to the laser. A focused 25–30-W beam was used for excision, and a cone-shaped specimen was produced by simultaneous traction with a hook. After removal of the cone, the surgical bed was vaporized by a defocused beam. Additional hemostatic sutures were rarely applied. Endocervical curettage (ECC) at the end of the surgery was not practiced. During the entire study period, the national guidelines for abnormal smears advised observation for CIN I, whereas for women with persistent CIN I, CIN II–III, or suspected malignant cells, the recommendation was referral to a gynecologic unit for colposcopically directed biopsies and ECC. Further treatment was always based upon a histologic diagnosis. In each case, we identified the most severe histopathologic diagnosis in the preoperative biopsies, ECC, or the cone specimen. We excluded from further analyses 55 women who had invasive disease and 32 women with diagnostic conization because of repeated discrepancy between the Papanicolaou smears and punch biopsies and/or ECCs. Eligible for the follow-up analyses were 1233 women treated initially for CIN II or CIN III. However, when defining the population-based cohort, we excluded 152 women who were not permanent residents of the county at initial treatment or who moved from the county during the follow-up period. A total of 1081 women were finally included in the study: 82 treated initially for CIN II and 999 for CIN III. The first follow-up visit took place at our clinic 3– 4 months postoperatively. Papanicolaou smear was the only screening method used to identify women who were eligible for histologic verification of residual and recurrent disease. If the first smear was normal, the woman was discharged to her general practitioner for further follow-up.8,9 Women with abnormal smears were referred for colposcopic biopsies and ECC. Women were defined as being free of residual disease when the first two smears taken within 6 –12 months postoperatively were normal.13 Women who had histologic findings extending to the margins of the cone were followed at equal intervals as women who had diseasefree margins. Our department of pathology is the only one in the county examining all cytologic and histologic specimens collected in our geographic area. Data before 1986 were obtained from individual cytology and histology records. Since the introduction of a computerized registry on January 1, 1986, we have gathered follow-up information on women treated for CIN II–III by linking the records to the registry at the department of pathology. In this way, we also identified women who had
VOL. 90, NO. 3, SEPTEMBER 1997
surgery for other gynecologic disorders, as well as those who had died and their plausible cause of death. All these data and other relevant clinical data were transferred to a standardized case-report form used for the clinical management of our study. We assessed the risk of recurrent disease by survival analysis among women who were free from residual disease. Time at risk was estimated as the time from conization to the most recent normal Papanicolaou smear. Women with abnormal smears but with normal histologic follow-up findings continued in the study as recurrence-free cases. However, women with a histologic confirmation of the abnormal smear were identified as having recurrent disease at the time of the actual smear. Women who had hysterectomy for reasons other than CIN were followed until the time of the hysterectomy, provided that histologic findings from the cervix uteri were normal. Women who died during follow-up were followed until their most recent normal smear unless an autopsy showed normal cervical histology, in which case they were followed to the date of their death. All analyses were done using SPSS version 4.1 (Statistical Package for Social Sciences; SPSS Inc., Chicago, IL). When appropriate, x2 statistics, logistic regression (forward stepwise approach), and survival analyses were applied. Differences between group means of continuous variables were examined by the MannWhitney test. P , .05 was considered statistically significant.
Results Over the 10-year study period, the average number of laser conizations per year increased from 96 during 1983–1985 to 106 during 1986 –1988, and to 119 during 1989 –1992. The mean age of the women treated for CIN II–III decreased significantly over the study period, from 35.7 years in the first period to 34.3 years in the last period (Table 1) (Mann-Whitney; z 5 22.03, P , .05). Over the study period, the proportion of nulliparous women and women with only one child also increased significantly (Table 1) (x2 trend 34.4, P , .001). Even though the highest CIN grade reported showed significantly more cases of CIN II during 1989 –1992, still more than 90% of the cases in this period were classified as CIN III disease (Table 2) (x2 trend 21.0, P , .001). The number of cases with free resection margins decreased significantly over the study years (Table 2) (x2 trend 18.8, P , .001), whereas few cases were reported as inconclusive with uncertain resection margins because of thermal trauma (Table 2). Twenty-eight women were excluded from analyses of residual disease because they never attended their first
Skjeldestad et al
Disease After Treatment of CIN
429
Table 1. Characteristics of the Study Population Characteristic Age (y) #29 30 – 40 $40 Mean Range Parity 0 1 $2 Unknown Previous treatment for CIN None Cryotherapy Laser vaporization
1983–1985 (n 5 287)
1986 –1988 (n 5 319)
1989 –1992 (n 5 475)
79 (27.5%) 129 (45.0%) 79 (27.5%) 35.7 21–76
89 (27.9%) 157 (49.2%) 73 (22.9%) 35.1 18 – 86
173 (36.4%) 190 (40.0%) 112 (23.6%) 34.3 19 –73
30 (10.5%) 50 (17.4%) 202 (70.4%) 5 (1.7%)
45 (14.1%) 70 (21.9%) 186 (58.3%) 18 (5.6%)
101 (21.3%) 132 (27.8%) 238 (50.1%) 4 (0.8%)
267 (93.0%) 16 (5.6%) 4 (1.4%)
315 (98.7%) 3 (1.0%) 1 (0.3%)
461 (97.1%) 5 (1.1%) 9 (1.9%)
CIN 5 cervical intraepithelial neoplasia.
scheduled visit (n 5 5), only completed their first visit with a normal smear (n 5 14), or underwent elective hysterectomy during the first 6 postoperative months for reasons other than CIN (n 5 9). Thus, 97.4% (1053 of 1081) of the initial cohort was eligible for analyses of residual disease. Over the study period, the number of cases with residual disease was low (20 of 1081, 1.9%). None of the 80 women with CIN II had residual disease. Seventeen of the 245 women (6.9%) with CIN III and involved resection margins had residual disease, compared with three of 728 women (0.4%) with CIN III and free resection margins (x2 trend 39.4, P , .001; crude odds ratio [OR] 18.1; 95% confidence interval [CI] 5.2, 64.0). However, three cases of invasive cancer that were not diagnosed during primary histologic examination were found among women with residual disease. These cancers were all diagnosed in tissue from ECC specimens and were not present in tissue from ectocervical punch biopsies. Table 3 lists the method of treatment by the most severe histologic finding among the women with residual disease. Table 2. Histologic Findings Finding CIN II CIN III Resection margins Free Not free Upper Lower Inconclusive
1983–1985 (n 5 287)
1986 –1988 (n 5 319)
1989 –1992 (n 5 475)
17 (5.9%) 270 (94.1%)
21 (6.6%) 298 (93.4%)
44 (9.3%) 431 (90.7%)
240 (83.6%)
237 (74.3%)
335 (70.5%)
37* (12.9%) 11* (3.8%) 0
47 (14.7%) 23 (7.2%) 12 (3.8%)
119 (25.1%) 11 (2.3%) 10 (2.1%)
CIN 5 cervical intraepithelial neoplasia. * Includes one woman in whom both endocervical and ectocervical margins were not free.
430 Skjeldestad et al
Disease After Treatment of CIN
Table 3. Indications and Treatment of Residual and Recurrent Disease Most severe histologic finding in follow-up tissue Treatment Residual disease Hysterectomy Cryotherapy Laser vaporization Laser conization Recurrent disease Hysterectomy Cryotherapy Laser vaporization Laser conization
Number
CIN I
CIN II
CIN III
Invasive cancer
7 1 3 9
0 0 0 0
0 0 1 1
6 1 2 6
1* 0 0 2†
6 0 3 10
0
0
5
1*
0 0
2 2
1 8
0 0
CIN 5 cervical intraepithelial neoplasia. * Patients had surgery by the Wertheim-Meigs operation. † Histology from the second cone was “free resection margins”; therefore no further treatment.
After the exclusion of 20 women with residual disease, 1033 women remained for analyses of the risk of recurrent disease. During the follow-up period, 23 women underwent hysterectomy unrelated to CIN, and seven died from causes not related to genital cancer. Nineteen of 1033 women were diagnosed with a recurrent disease verified histologically. Among these, one woman was diagnosed outside the screening program, because of symptoms (irregular bleeding and signs of blood in the discharge), with cervical squamous cell carcinoma (stage IB) 35 months after initial treatment. This woman had free cone margins and had had four postoperative normal Papanicolaou smears, the most recent one 9 months before the onset of symptoms. The most severe histologic findings and the methods of treatment at recurrence are listed in Table 3. In a logistic regression model, resection margins were the only predictor of recurrence among variables such as age, previous children, previous treatment of CIN before the index treatment, and CIN grade (Table 4). Women with involved resection margins were three times more likely to appear with a recurrence (adjusted OR 3.0; 95% CI 1.2, 7.5) than were women with free cone margins (Table 4). The incidence of recurrent disease increased nearly linearly during years 1– 6 of observation (Figure 1). The overall cumulative incidence of recurrence at 3 years of observation was seven cases per 1000 woman-years, and at 6 completed years of observation, it was 22 per 1000 woman-years. Women who had disease involving the cone margins had a significantly increased risk of recurrence compared with women who had free margins (Wilcoxon Gehan statistics; x2 7.5, P , .01) (Figure 1).
Obstetrics & Gynecology
Table 4. Risk of Recurrent Disease
Variable* Age (y) #34 $35 Parity 0 –1 $2 Unknown Previous treatment for CIN Yes No Time period 1983–1985 1986 –1988 1989 –1992 Most severe histologic finding CIN II CIN III Resection margins Free Not-free or inconclusive
Discussion
No. of No. of women in women with each category recurrence Adjusted (n 5 19) OR (n 5 1033)
95% CI
582 451
10 9
1.3 1.0
0.5, 3.6 Reference
408 599 26
10 9
1.5 1.0
0.6, 3.8 Reference Not analyzed
36 997
2 17
3.3 1.0
0.7, 14.7 Reference
269 310 454
5 9 5
1.0 1.3 0.5
Reference 0.4, 4.1 0.1, 1.7
80 953
2 17
1.0 0.7
Reference 0.2, 3.0
788 245
10 9
1.0 3.0
Reference 1.2, 7.5
OR 5 odds ratio; CI 5 confidence interval; CIN 5 cervical intraepithelial neoplasia. * The variables entered the model the way they are categorized in the table.
Separate analyses starting at 48 months after treatment did not find any difference in the cumulative incidence by resection margins.
This study confirms that laser conization is highly efficient for treating CIN II–III. The incidence of residual and recurrent disease is very low and is in agreement with other studies evaluating the treatment efficacy of laser conization6,7,15 or cold-knife conization.16,17 Invasive cancer might be underdiagnosed during initial treatment of high-grade lesions of the cervix uteri.15–17 Therefore, the most important issue in the follow-up of these patients is to rule out persistent disease. This point is underscored by the finding of three microinvasive cancers among 20 women with residual disease in our study.15 Both residual and recurrent disease were highly associated with disease extending to the resection margins. Despite these associations, barely 10% of the women with involved resection margins returned with either residual disease or recurrent disease,11 reflecting that resection margins are most often free even if the margins of the cone are determined to be involved. Others have reported a higher prevalence of residual disease among women with incomplete excisions.15,17 The most likely explanation for the varying prevalence of residual disease is varying depth of the conization. If the cone is excised too low, some affected endocervical tissue might easily remain at the site. Therefore, the need for first follow-up at a specialized unit is clearly documented by the finding of three microinvasive cancers in our study. We released the patients to their general practitioner for further follow-up if the first Papanicolaou smear was normal.8,9,15 We have not found continued colpo-
Figure 1. Cumulative incidence of recurrence by histologic findings in the resection margin.
VOL. 90, NO. 3, SEPTEMBER 1997
Skjeldestad et al
Disease After Treatment of CIN
431
scopic examinations or blind biopsies necessary in the surveillance of residual or recurrent disease.8,9,15 This position is supported by the extremely low incidence of recurrence.8 –11,15 During the study period, the national guidelines for Papanicolaou smear follow-up were four visits during the first year, two visits during the second year, and annual visits subsequently.8 –11 We have not practiced these guidelines strictly, but have advocated a follow-up based on two visits during the first 12 months and annually thereafter. These guidelines were introduced in Norway during 1993.18 Based on the linear shape of the cumulative incidence curve (Figure 1) of recurrent disease among women with free resection margins at an annual incidence of two per 1000 women, we conclude that these women can return, after residual disease has been ruled out, at a 3-year interval as practiced in the national screening program. The slope of the cumulative incidence curve of recurrent disease among women who have involved resection margins is steep between 12 and 48 months after treatment; thereafter, the curve runs parallel to that of women with free cone margins. After 48 months of observation, very few recurrences are diagnosed, even among women with disease extending to the cone margins. In our opinion, the cumulative incidence of recurrence among women with disease involving the cone margins qualifies them for annual visits through the fourth year after treatment and thereafter release to the general 3-year screening interval. Overall risk assessment is important for deciding on and ensuring compliance with the interval for followup. It should be reassuring for patients, gynecologists, and general practitioners alike to know that among women with free resection margins, the risk of recurrence is as low or even lower than the risk of ever being diagnosed with a histologically confirmed CIN II–III lesion in the general screening program.3,4 Furthermore, knowledge about the natural progression of CIN II–III to invasive disease is important for interval decisions. This process is modeled to take 14 or more years.5,6,19 –23 Therefore, among women in whom representative smears are collected, there is time enough within our differential approach to capture recurrent disease. So far, we have done surveillance on the assumption that these women are at high risk of developing cancer, implying that our therapy might be incomplete. Our results, however, show that the longterm efficacy of therapy is very high and that we can take advantage of the information in the pathology report on resection margins to develop a differential follow-up regimen.
432 Skjeldestad et al
Disease After Treatment of CIN
References 1. Bjørge T, Thoresen SØ, Gunbjørud AB. Masseundersøkelsen mot livmorhalskreft. Nå starter invitasjonene til kvinner på landsbasis [Norwegian]. Tidsskr Nor Lægeforen 1995;115:815– 6. 2. Haugen OA. Forbruk av vaginalcytologiske prøver i Norge [Norwegian]. Tidsskr Nor Lægeforen 1984;104:952– 6. 3. Forsmo S, Buhaug H, Stalsberg H. Use of Papanicolaou smears in a population without a mass screening program. Acta Obstet Gynecol Scand 1994;73:824 – 8. 4. Bjørge T, Gunbjørud AB, Langmark F, Skare GB, Thoresen SØ. Cervical mass screening in Norway—510,000 smears a year. Cancer Detect Prev 1994;6:463–70. 5. Forsmo S, Buhaug H, Skjeldestad FE, Haugen OA. Treatment of preinvasive conditions during opportunistic screening and its effectiveness on the cervical cancer incidence in one Norwegian county. Int J Cancer 1997;71:4 – 8. 6. IARC Working Group on Evaluation of Cervical Cancer Screening Programmes. Screening for squamous cervical cancer: Duration of low risk after negative results of cervical cytology and its implication for screening policies. BMJ 1986;293:659 – 64. 7. Sigurdsson K. Effect of organized screening on the risk of cervical cancer. Evaluation of screening activity in Iceland, 1964 –91. Int J Cancer 1993;54:563–70. 8. Scheistrøen M. Laserbehandling av premaligne sykdommer i cervix uteri [Norwegian]. Tidsskr Nor Lægeforen 1992;112:1145–7. 9. Forsmo S, Øian P, Stalsberg H. Cervikal intraepitelial neoplasi. Persistens og residiv etter laserbehandling [Norwegian]. Tidsskr Nor Lægeforen 1992;112:1148 –51. 10. Tabor A, Berget A. Cold-knife and laser conization for cervical intraepithelial neoplasia. Obstet Gynecol 1990;76:633–5. 11. Andersen ES, Pedersen B, Nielsen K. Laser conization: The results of treatment of cervical intraepithelial neoplasia. Gynecol Oncol 1994;54:201– 4. 12. Sigurdsson K. Quality assurance in cervical cancer screening: The Icelandic experience 1964 –1993. Eur J Cancer 1995;31A:728 –34. 13. Rylander E, Isberg A, Joelson I. Laser vaporization of cervical intraepithelial neoplasia. A five-year follow-up. Acta Obstet Gynecol Scand Suppl 1984;125:33– 6. 14. Dorsey JH, Diggs ES. Microsurgical conization of the cervix by carbon dioxide laser. Obstet Gynecol 1979;54:565–70. 15. Lopez A, Morgan P, Murdoch J, Piura B, Monaghan JM. The case of conservative management of “incomplete excision” of CIN after laser conization. Gynecol Oncol 1993;49:247–9. 16. Kolstad P, Klem V. Long-term follow-up of 1121 cases of carcinoma in situ. Obstet Gynecol 1976;48:125–9. 17. Vedel P, Jakobsen H, Kryger-Baggesen N, Rank F, Bostofte E. Five-year follow up of patients with cervical intraepithelial neoplasia in the cone margins after conization. Eur J Obstet Gynecol Reprod Biol 1993;50:71– 6. 18. Nøstdahl W. Cytologiske kontroller etter behandling av premaligne lidelser i cervix uteri [Norwegian]. Tidsskr Nor Lægeforen 1993;113:1884 –5. 19. Gustafsson L, Adami H-O. Natural history of cervical neoplasia: Consistent results obtained by an identification technique. Br J Cancer 1989;60:132– 41. 20. Van Oortmarssen GJ, Habbema JDF. Duration of preclinical cervical cancer and reduction in incidence of invasive negative Pap smears. Int J Epidemiol 1995;24:300 –7. 21. Albert A. Estimated cervical cancer disease state incidence and transition rates. J Natl Cancer Inst 1981;67:571– 6. 22. Prorok PC. Mathematical models and natural history in cervical cancer screening. In: Hakama M, Miller AB, Day NE, eds. Screening for cancer of the uterine cervix. Lyon: International Union Against Cancer, 1986:185–96.
Obstetrics & Gynecology
23. Eddy DM. The frequency of cervical cancer screening. Comparison of a mathematical model with empirical data. Cancer 1987;60: 1117–22.
Received January 24, 1997. Received in revised form May 6, 1997. Accepted May 21, 1997.
Address reprint requests to:
F. E. Skjeldestad, MD, PhD Department of Obstetrics and Gynecology University Hospital of Trondheim Trondheim, 7006 Norway E-mail: [email protected]
STANDARDS
Copyright © 1997 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
FOR REPORTING TRIALS
The CONSORT listing of standards for reporting randomized trials has been adopted as policy by Obstetrics & Gynecology. Investigators who are planning, conducting, or reporting randomized trials should be thoroughly familiar with these standards. A copy can be obtained by contacting: Obstetrics & Gynecology, 1100 Glendon Avenue, Suite 1655, Los Angeles, CA 90024-3520; FAX (310) 208-2838.
VOL. 90, NO. 3, SEPTEMBER 1997
Skjeldestad et al
Disease After Treatment of CIN
433