Cancer risk after hospital discharge diagnosis of benign ovarian cysts and endometriosis

Data from The Swedish Hospital Discharge Registry (The National Board of Health and Welfare) were used to identify a cohort of women born in Sweden before 1970. All were discharged from hospital during the period 1969–96 with the diagnoses of ovarian cyst, functional ovarian cyst, or endo- metriosis. The unique personal identification number, assigned to each resident in Sweden, was used. The Swedish Hospital Discharge Registry started in 1969. At that time it covered 60% of the Swedish population, in 1978 75 %, and in 1983 85 %. Since 1987, all patients discharged from hospitals in Sweden are included. The register contains information about surgical procedures and up to eight dis- charge diagnoses, coded according to the International Classification of Diseases (ICD-8 1969–86 and ICD-9 1987–96). The diagnostic codes used in the present study were ovarian cyst, including benign or unknown ovarian tumor (220.0–9 benign tumor in ovary; 620.2 unspecified cyst in ovary; 235.99, 236.2 unknown ovarian tumor), func- tional ovarian cyst/Corpus luteum (615.2 and 620.0–1), and endometriosis (625.3 and 617.0–9). To each case three controls, with the same date of birth, were matched using the Swedish Population Register, Statistics Sweden, which includes all persons living in Sweden. Codes from The Swedish Hospital Discharge Registry for surgical procedures and dates of operations were linked to the cases and con- trols. Data from the National Swedish Cancer Register, founded in 1958, matched all incident cancers diagnosed among cases and controls up to December 31, 1997. The Cancer Registry has coded malignant neoplasms according to the ICD-7 classification during the entire period of the study. The Fertility Register, also kept by Statistics Sweden, contains information on all children born to Swedish women. From the Fertility Registry, the dates of birth of children born to the cases and controls were obtained. Women having changed their personal identification number were excluded ( n ¼ 122). The cases were then given a file number and the controls a corresponding control number in order to unidentify the database. This database was used in the statistical analyses. In order to avoid misclassification of benign cysts or undiagnosed malignancy at the time of hospital discharge, 1 year was allowed to elapse before any diagnosis of malignancy was accepted in the studied population. For statistical analysis the Mantel-Haenszel procedure was applied for the determination of odds ratio (OR) and the 95 % confidence intervals (CI) after stratifications as specified (13).

Risk of developing gynecologic cancer There was no change in the risk of developing ovarian cancer having had an ovarian cyst or functional ovarian cyst, while women with endo- metriosis had an increased risk for ovarian malig- nancy (Table I). It was not possible to determine precisely the ovarian status on the basis of the operation code for hysterectomy. Therefore, women who had undergone hysterectomy with or without oophorectomy were excluded, but there were only minor changes in the results after the exclusion of these women. No change in breast cancer risk was found. A decreased risk of developing endometrial cancer was found in women with ovarian cysts and endometriosis. Women with endometriosis also had a decreased risk for cervical cancer. Women with endometri- osis showed an even more pronounced risk of developing ovarian cancer more than 10 years after diagnosis (OR 1.46, 95 % CI 1.01–2.11). There was no change in total cancer risk (all types of invasive malignancies) after having had an ovarian cyst, OR 0.99 (95% CI 0.85–1.14) or a functional ovarian cyst, OR 0.90 (95 % CI 0.69– 1.19), whereas women with endometriosis had an increased overall risk for malignancy, OR 1.15 (95% CI 1.00–1.32). Risk of developing ovarian cancer related to previous ovarian surgery When analyzing women having undergone ovar- ian cyst resection and/or unilateral oophorect- omy separately, the risk for subsequent ovarian cancer was almost nine-fold higher, OR 8.8 (95% CI 5.2–14.8), as opposed to those who were not treated surgically, OR 0.48 (95 % CI 0.35–0.66). Among women who later developed ovarian can- cer ( n ¼ 367), nine women had been registered with the surgical code for bilateral oophorectomy at least 1 year before the cancer diagnosis (cases n ¼ 2 and controls n ¼ 7). This may be because of misclassification (using code for bilateral instead of the code for unilateral oophorectomy) or the 396 C. Borgfeldt and E. Andolf # Acta Obstet Gynecol Scand 83 (2004) fact that the ovarian cancer had started as a peritoneal cancer. Risk of developing ovarian cancer in relation to age of previous ovarian cyst The youngest age group (10–29 years old) with an ovarian cyst or a functional cyst, treated or not treated, showed an increased risk of developing ovarian cancer (Table II). A decreased risk for ovarian cancer was observed in women aged older than 50 years discharged from hospitals with the diagnosis of ovarian cyst. Young women with endometriosis showed an even higher increased risk of developing ovarian can- cer compared with the total age group. Risk of developing ovarian cancer in relation to number of children There was almost no difference in risk figures according to the number of children born after the diagnosis, whereas the number of children before diagnosis of ovarian cyst, functional cyst or endometriosis was of importance (Table III). The ovarian cancer risk was significantly increased in nulliparous with functional cysts or endometriosis. In women with an ovarian cyst diagnosis, the cancer risk also decreased with parity. The overall tendency was decreasing ovar- ian cancer risk with increasing number of chil- dren born in all three studied groups. Mean age at ovarian cancer diagnosis The mean age for ovarian cancer was signifi- cantly lower in the women discharged from hos- pital for ovarian cyst (47.6 years, SEM 1.27; n ¼ 80) vs. the control group (54.1 years, SEM 0.56; n ¼ 287) ( p < 0.001); for the women dis- charged from hospital for functional cyst (40.7 years, SEM 1.70; n ¼ 31) vs. the control group (49.1 years, SEM 1.12; n ¼ 72) ( p < 0.02); and also for the women discharged from hospital with endometriosis (49.0 years, SEM 0.91; n ¼ 90) vs. the control group (51.6 years, SEM 0.60; n ¼ 183) ( p < 0.001). Even when the youngest women below 30, respectively, 35 years of age with ovarian cancer were excluded in the calcula- tions, the mean age for ovarian cancer was sig- nificantly lower in the women treated for ovarian cysts, functional cysts or endometriosis as compared with their respective control group (p < 0.05 for all six comparisons). Table I. Odds ratio for gynecologic cancers in women with hospital discharge diagnoses, ovarian cyst, functional cyst or endometriosis Ovarian cyst Functional cyst Endometriosis Cancer type OR 95 % CI Case n Control n OR 95 % CI Case n Control n OR 95 % CI Case n Control n Ovary 0.86 0.67–1.10 78 280 1.24 0.81–1.89 31 72 1.34 1.03–1.75* 81 181 Breast 1.07 0.98–1.18 586 1680 1.00 0.85–1.18 189 569 1.10 0.98–1.23 427 71165 Endometrium 0.66 0.51–0.86* 66 308 0.66 0.39–1.12 17 77 0.58 0.42–0.81* 39 211 Cervix 0.78 0.58–1.04 55 217 1.30 0.90–1.90 39 90 0.57 0.37–0.90* 23 120 No cancer 40406 124542 17461 52539 26969 81073 *95% confidence intervals excluded 1.0 Only cancer diagnoses more than 1 year since the primary hospital dischargediagnoses are included. Women with surgery as bilateral oophorectomy and hysterectomy with or without bilateral oophorectomy are excluded. Stratification is performed for women’s age and the number of children born by the women before and after the primary hospital discharge diagnosis. Table II. Odds ratio for ovarian cancer diagnosis related to woman’s age more than 1 year since the woman’s first hospital discharge with diagnoses of ovarian cyst, functional cyst, and endometriosis Ovarian cyst Functional cyst Endometriosis Age (years) OR CI 95 % OR CI 95 % OR CI 95 % 10–29 2.23 1.29–3.86* 1.76 1.50–2.00* 3.52 1.56–7.95* 30–49 0.83 0.60–1.16 0.86 0.46–1.52 1.26 0.50–3.16 50þ 0.44 0.25–0.77* – – 0.98 0.42–2.31 *95% confidence intervals excluded 1.0 –: strata missing cases or controls. Women with surgery as bilateral oophorectomy and hysterectomy with or without bilateral oophorectomy are excluded. Stratification is performed for the number ofc h i l d r e nb o r nb yt h ew o m e nb e f o r ea n da f t e rthe primary hospital discharge diagnosis. Ovarian cysts and risk of ovarian cancer 397 # Acta Obstet Gynecol Scand 83 (2004)

In this case–control study we found no association in the total material between the hospital discharge diagnosis of an ovarian or functional cyst and later development of ovarian cancer. However, there was a significantly increased risk for ovarian can- cer in young women discharged from hospital for an ovarian or functional ovarian cyst, especially if they were childless. Endometriosis was associated with an overall increased risk of developing cancer (all types of invasive malignancies included). Young age and nulliparity increased the risk. Women with unilateral oophorectomy or ovarian cyst resection showed an increased ovarian cancer risk later as compared with those with ovarian cysts in whom no surgery was performed. In add- ition, the mean ages at diagnosis of ovarian cancer were significantly lower in the cases as compared with the control groups. There are several limitations to these data. Pri- marily, one may speculate over the generalization of results on patients hospitalized for ovarian cysts/benign tumors, functional cysts, or endometri- osis. Women hospitalized for an adnexal lesion probably had persisting and/or symptomatic adnexal lesions. Also, there may by an increased risk for dysplasia in a long-standing cystic lesion especially if epithelial cells increase in number by mitosis as the cyst grows. Further, in women with- out histopathological verification, functional cysts might have been misclassified as ovarian cysts and vice versa. When the reliability of the registers was tested, the Swedish Hospital Discharge Registry was found to have a misclassification rate of 7 % (14). The reliability of the surgical codes was found to be good (15). One percent of the codes are missing and 5 % are erroneous. As for the National Swedish Cancer Registry, approximately 98% of the diagnoses are being morphologically verified (3). The missing cancer diagnoses are very few when compared with the National Death Certificate Register in Sweden (3). A qual- ity study of the Fertility Register has also been performed lately showing good accuracy (personal communication, Statistics Sweden). The discre- pancy between the Fertility Register and the Swedish Population Register was 5–8 % in the 1930s and 1940s, 2–4 % in the 1950s and 1960s, and less than 1 % since 1967. Nevertheless, the missing codes and the errors in the registers should be proportional in the case and control groups, minimizing the importance of the errors. The observation that nulliparity increases the risk of ovarian cancer is in accordance with earl- ier reports, where multiparas have a 40–60 % risk reduction (16–19). Infertile women, unsuccess- fully treated with hormones have a higher risk of developing ovarian cancer, especially after long- term use (20,21). However, women successfully treated did not have an increased risk (20). The possible reason for this increased risk was ovar- ian dysfunction and not the hormonal treatment. In addition, another interesting hypothesis has been proposed to explain the reduced risk of ovarian cancer in parous women: namely that pregnancy hormones or other immunological changes during pregnancy may clear the ovaries from cells that have undergone malignant trans- formation (18). As for endometriosis, our results are in accord- ance with an earlier report based on discharge diagnoses where cases were compared with the standardized incidence ratio of ovarian cancer but parity was not considered (9). Other reports have also suggested a histological transformation of benign endometriosis to early epithelial ovarian cancer (22,23). There are several theories concerning the development of endometriosis. One theory is that endometriosis arises from metaplasia of the coelomic epithelium, another that endometrial cells regurgitate through the fallopian tubes at menstruation and imp- lant on pelvic structures. Deficiency in the Table III. Odds ratio for ovarian cancer diagnosis from the first hospital discharge diagnoses of ovarian cyst, functional cyst, or endometriosis related to the number of children given birth to before and after diagnosis Number of children Ovarian cyst Functional ovarian cyst Endometriosis before diagnosis after diagnosis OR CI 95 % OR CI 95 % OR CI 95 % 0 0–16 1.50 0.97–2.31 2.28 1.18–4.37* 1.89 1.19–3.01* 1–3 0–16 0.70 0.51–0.95* 1.21 0.88–1.66 1.21 0.88–1.66 4–5 0–16 0.21 0.03–1.35 1.01 0.19–5.29 1.27 0.24–6.63 0 0 1.43 0.87–2.32 2.47 1.15–5.30* 1.87 1.15–3.06* 1–3 0 0.68 0.49–0.95* 1.04 0.55–1.94 1.16 0.84–1.61 4–5 0 0.21 0.03–1.05 1.01 0.19–5.29 1.27 0.24–6.63 *95% confidence intervals excluded 1.0. Only cancer diagnoses more than 1 year since the primary hospital discharge diagnoses are included. Gynecologic surgical procedures are not considered. 398 C. Borgfeldt and E. Andolf # Acta Obstet Gynecol Scand 83 (2004) immunological system has also been considered to contribute to the development of endometrio- sis. Local inflammatory response at the endome- triotic implantations initiates proteolytic systems which are involved in carcinogenesis (24). Chronic inflammation and/or a deficient immu- nological response to endometriosis may contri- bute to the increased risk of developing ovarian cancer in patients suffering from endometriosis (25). In the present study, women with ovarian cyst resection or unilateral oophorectomy for benign causes had a highly increased risk of later devel- oping ovarian cancer. This may be because of premalignant dysplasia in the contra-lateral ovary or misclassified borderline tumors which may have had unrecognized peritoneal implants. Histo-pathological changes have been found in macroscopically normal ovaries in women oper- ated on because of a strong family history (7,26). In ovarian cancer stage Ia, microscopic changes are found in up to 7% of the contra-lateral ovary (27), which may indicate that ovarian cancer is a multifocal disease. This emphasizes the import- ance of a thorough examination of the whole abdominal cavity, both at laparoscopic and open surgery, for ovarian cysts and also peri- toneal washing and biopsies of the peritoneum. Whether benign macroscopic changes, except endometriosis, precede ovarian cancer is unclear. Our finding may indicate that this is the case. Several studies have been performed to verify whether an inclusion cyst in the other ovary is more common in ovarian cancer patients, but studies are inconclusive (5,7,26,28,29). Another explanation may be that the surgical trauma itself starts the process toward dysplasia and malig- nancy, as the removal of a cyst traumatizes the ovarian-surface epithelium. The healing process entails increased cell division activity and inflam- mation with activation of proteolytic enzymes similar to those in cancer invasion and metastases (24,25). In addition, inflammatory cytokines activ- ating nitric oxide have shown to cause DNA damage and inhibit DNA repair proteins (30). Thus after ovarian surgery, as well as in endo- metriosis, an inflammatory response may be the key to carcinogenesis. Cases were significantly younger when receiv- ing the diagnosis of ovarian cancer as compared with women in the control groups. In families with a hereditary risk for ovarian cancer, the median age at diagnosis is significantly lower than the median age of women without a history of familiar ovarian cancer (31). However, only approximately 5–10 % of all ovarian cancer is considered to be caused by inherited mutations, which is why the differences in mean age at ovar- ian cancer diagnosis can not be fully explained by hereditary factors. In the present study women with endometriosis had a decreased risk of developing endometrial and cervical cancer. Treatment of endometriosis often includes synthetic progestin, gonadotropin- releasing hormone analogs, or combined oral contraceptives known to reduce the risk of endo- metrial hyperplasia and cancer (32,33). The reduced risk of cervical cancer may be the result of intensified screening and treatment of pre- malignant lesions of the cervix in the group of women with endometriosis, as they may have contacted their gynecologist more frequently. In conclusion, this hospital registry study indi- cates an increased risk for ovarian cancer in women suffering from endometriosis and young women treated for ovarian cysts, especially if the cyst is removed. This latter association encourages expectant monitoring in young women with asymptomatic cysts if malignancy is not suspected. The tentative biological explan- ations such as the inflammatory response after surgery or endometriosis-mediating carcinogen- esis needs to be proven. Acknowledgment The statistical assistance of Professor Bengt Ka¨lle´n is grate- fully acknowledged. We also wish to acknowledge Mats Talba¨ck at the National Board of Health and Welfare and A˚ ke Jalo at Statistics Sweden for helping us to match the data registers. Financial support was from the Sigrid Simonssons and Agni Olssons Foundation.

1. The National Board Health Welfare Centre for Epidemiology. Causes of death 1997: Official Statistics of Sweden – Health and Diseases: Stockholm, Sweden: 2000; 3: 1–224. 2. Pecorelli S, Benedet J, Beller U, Creasman W, Heintz A, Pettersson F. FIGO Annual Report on the Results of Treat- ment in Gynaecological Cancer. FIGO Annual Report on the

of Treatment in Gynaecological Cancer. Oxford, England: Isis Medical Media Ltd, 2001; 6: 1–184. 3. The National Board Health Welfare Centre for Epidemiology. Cancer incidence in Sweden 1998: Official Statistics of Sweden – Health and Diseases: Stockholm, Sweden: 2000; 4: 1–155. 4. Hartge P, Whittemore AS, Itnyre J, McGowan L, Cramer D. Rates and risks of ovarian cancer in subgroups of white women in the United States. The Collaborative Ovarian Cancer Group. Obstet Gynecol 1994; 84: 760–4. 5. Westhoff C, Clark CJ. Benign ovarian cysts in England and Wales and in the United States. Br J Obstet Gynaecol 1992; 99: 329–32. 6. Crayford TJ, Campbell S, Bourne TH, Rawson HJ, Collins WP. Benign ovarian cysts and ovarian cancer: a cohort study with implications for screening. Lancet 2000; 355: 1060–3. Ovarian cysts and risk of ovarian cancer 399 # Acta Obstet Gynecol Scand 83 (2004) 7 .S a l a z a rH ,G o d w i nA K ,D a l yM B ,L a u bP B ,H o g a nW M , Rosenblum N et al. Microscopic benign and invasive malignant neoplasms and a cancer-prone phenotype in prophylactic oophorectomies. J Natl Cancer Inst 1996; 88: 1810–20. 8. Bourne TH, Whitehead MI, Campbell S, Royston P, Bhan V, Collins WP. Ultrasound screening for familial ovarian cancer. Gynecol Oncol 1991; 43: 92–7. 9. Brinton LA, Gridley G, Persson I, Baron J, Bergqvist A. Cancer risk after a hospital discharge diagnosis of endometriosis. Am J Obstet Gynecol 1997; 176: 572–9. 10. Plaxe SC, Deligdisch L, Dottino PR, Cohen CJ. Ovarian intra- epithelial neoplasia demonstrated in patients with stage I ovar- ian carcinoma. Gynecol Oncol 1990; 38: 367–72. 11. Puls LE, Powell DE, DePriest PD, Gallion HH, Hunter JE, Kryscio RJ, van Nagell JR Jr Transition from benign to malig- nant epithelium in mucinous and serous ovarian cystadenocar- cinoma. Gynecol Oncol 1992; 47: 53–7. 12. Deligdisch L, Gil J. Characterization of ovarian dysplasia by interactive morphometry. Cancer 1989; 63: 748–55. 13. Miettinen OS. Simple interval-estimation of risk ratio [abstract]. Am J Epidemiol 1974; 100: 515–6. 14. Nilsson AC, Spetz CL, Carsjo K, Nightingale R, Smedby B. Slutenvardsregistrets tillforlitlighet. Diagnosuppgifterna battre an sitt rykte. Lakartidningen 1994; 91: 598. 15. Falkeborn M, Persson I, Naessen T, Kressner U. Validity of information on gynecological operations in the Swedish in-patient registry. Scand J Soc Med 1995; 23: 220–4. 16. Risch HA, Marrett LD, Howe GR. Parity, contraception, infertility, and the risk of epithelial ovarian cancer. Am J Epi- demiol 1994; 140: 585–97. 17. Hankinson SE, Colditz GA, Hunter DJ, Willett WC, Stampfer MJ, Rosner B, Hennekens CH, Speizer FE. A prospective study of reproductive factors and risk of epithelial ovarian cancer. Cancer 1995; 76: 284–90. 18. Adami HO, Hsieh CC, Lambe M, Trichopoulos D, Leon D, Persson I, Ekbom A, Janson PO. Parity, age at first childbirth, and risk of ovarian cancer. Lancet 1994; 344: 1250–4. 19. Whiteman DC, Murphy MF, Cook LS, Cramer DW, Hartge P, Marchbanks PA, Nasca PC, Ness RB, Purdie DM, Risch HA. Multiple births and risk of epithelial ovarian cancer. J Natl Cancer Inst 2000; 92: 1172–7. 20. Whittemore AS, Harris R, Itnyre J. Characteristics relating to ovarian cancer risk: collaborative analysis of 12 US case-control studies. II. Invasive epithelial ovarian cancers in white women. Collaborative Ovarian Cancer Group. Am J Epidemiol 1992; 136: 1184–203. 21. Rossing MA, Daling JR, Weiss NS, Moore DE, Self SG. Ovarian tumors in a cohort of infertile women. N Engl J Med 1994; 331: 771–6. 22. Sainz de la Cuesta R, Eichhorn JH, Rice LW, Fuller AF, JrNikrui, Goff BA. Histologic transformation of benign endo- metriosis to early epithelial ovarian cancer. Gynecol Oncol 1996; 60: 238–44. 23. Heaps JM, Nieberg RK, Berek JS. Malignant neoplasms arising in endometriosis. Obstet Gynecol 1990; 75: 1023–8. 24. Andreasen PA, Egelund R, Petersen HH. The plasminogen activation system in tumor growth, invasion, and metastasis. Cell Mol Life Sci 2000; 57: 25–40. 25. Balkwill F, Mantovani A. Inflammation and cancer: back to Virchow? Lancet 2001; 357: 539–45. 26. Sherman ME, Lee JS, Burks RT, Struewing JP, Kurman RJ, Hartge P. Histopathologic features of ovaries at increased risk for carcinoma. A case-control analysis. Int J Gynecol Pathol 1999; 18: 151–7. 27. Williams TJ, Dockerty MB. Status of the contralateral ovary in encapsulated low grade malignant tumors of the ovary. Surg Gynecol Obstet 1976; 143: 763–6. 28. Werness BA, Afify AM, Bielat KL, Eltabbakh GH, Piver MS, Paterson JM. Altered surface and cyst epithelium of ovaries removed prophylactically from women with a family history of ovarian cancer. Hum Pathol 1999; 30: 151–7. 29. Werness BA, Afify AM, Eltabbakh GH, Huelsman K, Piver MS, Paterson JM. p53, c-erbB, and Ki-67 expression in ovaries removed prophylactically from women with a family history of ovarian cancer. Int J Gynecol Pathol 1999; 18: 338–43. 30. Jaiswal M, LaRusso NF, Burgart LJ, Gores GJ. Inflammatory cytokines induce DNA damage and inhibit DNA repair in cholangiocarcinoma cells by a nitric oxide-dependent mechan- ism. Cancer Res 2000; 60: 184–90. 33. Ford D, Easton DF, Peto J. Estimates of the gene frequency of BRCA1 and its contribution to breast and ovarian cancer incidence. Am J Hum Genet 1995; 57: 1457–62. 32. Hankinson SE, Colditz GA, Hunter DJ, Spencer TL, Rosner B, Stampfer MJ. A quantitative assessment of oral contraceptive use and risk of ovarian cancer. Obstet Gynecol 1992; 80: 708–14. 33. Weiderpass E, Adami HO, Baron JA, Magnusson C, Bergstrom R, Lindgren A et al. Risk of endometrial cancer following estrogen replacement with and without progestins. J Natl Cancer Inst 1999; 91: 1131–7. Address for correspondence: Christer Borgfeldt Department of Obstetrics and Gynecology University Hospital S-221 85 Lund Sweden e-mail: [email protected] 400 C. Borgfeldt and E. Andolf # Acta Obstet Gynecol Scand 83 (2004)