A considerable proportion of patients diagnosed with a first primary cancer develop a second primary, particularly those with cancers with a good prognosis. 1,2 Second primaries may represent the effect of treatment for the first cancer, they may be independent events, or they may be influenced by the same environmental and genetic factors that caused the first cancer. 3 There are limited data on second primary cancers in women after diagnosis of cervical cancer 2 and on first primary cancers after diagnosis of in situ cervical cancer. 4,5 Human papilloma virus (HPV), particularly of types 16 and 18, is the main cause of cervical cancer, the third most common cancer in women worldwide. 6 The relative risks (RRs) in studies in which HPV DNA has been demonstrated are >20, and the attributable risk has been estimated at 89% worldwide. 6,7 Additionally, HPV has been implicated in other anogenital cancers, including those of the anus, vulva, and penis and, less convincingly, those of the head, neck, and esophagus. 6,8 Host susceptibility may also be a factor in cervical cancer, because susceptibility is associated with certain human leukocyte antigen haplotypes. 9 Immunosuppressed patients are at an increased risk of cervical cancer and other malignancies, such as squamous cell carcinoma of the skin and lip, and of Hodgkin’s and non-Hodgkin’s lymphoma. 10,11 Some of the sites involved are targets for HPV, whereas for skin and lymphoid tissues HPV involvement has not been demonstrated. 6,12
Here we analyze the occurrence of second primary cancers in women who had been diagnosed with invasive cervical cancer as compared with age- and period-standardized incidence rates for all women. To differentiate the effects of treatment, we carried out similar analyses for first cancers diagnosed after in situ cervical cancer. As the in situ cervical cancer has the same relation to HPV as the invasive cancer 6 but is treated without radiotherapy, we attempt to assess the potential role of HPV at all of the common cancer sites. A total of 135,386 subjects and 9,426 second cancers were included, making this the largest follow-up study published and the only one following in situ and invasive cancers in the same population.
Subjects and Methods
The Swedish Family-Cancer Database includes all persons born in Sweden after 1934 with their biological parents, totaling more than 9 million individuals; an updating has been carried out since our previous studies. 13,14 Data on cancers, including in situ cervical cancers, were retrieved from the nationwide Swedish Cancer Registry from the years 1958–1996. We used four-digit diagnostic code according to the International Classification of Diseases, 7th revision (ICD-7). We pooled the following ICD-7 codes: upper aerodigestive tract cancer codes 161 (larynx) and 140–148 (lip, mouth, and pharynx), except for code 142 (salivary glands); “lymphoma” codes 200 (non-Hodgkin’s lymphoma), 201 (Hodgkin’s disease), and 202 (reticulosis); and “leukemia” codes 204–207 (leukemias), 208 (polycytemia vera), and 209 (myelofibrosis). We separated rectal cancer (ICD-7 code 154) into anal (squamous cell carcinoma, 154.1) and mucosal rectum cancer (154.0). Basal cell carcinoma of the skin is not registered in the Cancer Registry.
We calculated standardized incidence ratios (SIRs) for women diagnosed with invasive or in situ cervical cancer by dividing the observed numbers of second events by the expected ones, calculated as person-years at risk from the age-specific (5-year age groups), period-specific (5-year periods), and sex-specific incidence rates of all women. 15 Ninety-five per cent confidence intervals were calculated assuming a Poisson distribution. 15
In this study, we followed the development of second primary events until the end of 1996 after diagnosis of 117,830 women with in situ and 17,556 women with invasive cervical cancers. SIRs of second primaries (first cancer after in situ cervical cancer and second primary after invasive cervical cancer) at different sites after in situ and invasive cervical cancer are shown in Table 1. Several sites showed increased SIRs after both types of second primaries: upper aerodigestive tract, anus, pancreas, lung, other female genital, and urinary bladder. SIRs were >3 for cancers in the known HPV-related sites of the anus and female genitals, but even urinary bladder cancer had an SIR of 4.15 after invasive cervical cancer. Second primaries of all sites were increased among the cervical cancer patients.
SIRs for cancers of different sites were determined after diagnosis of in situ cervical cancer (Figure 1). Many sites showed an increased SIR within the first year of diagnosis, including those shown in Figure 1. SIRs for other female genitals and anus were highest and remained so till the end of the observation period. Also, the SIR for lung and urinary bladder cancer remained above unity. Similar data are shown in Figure 2 after invasive cervical cancer. SIRs of other female genital and bladder cancers were initially very high and then decreased to about 2 at 5–9 years after diagnosis, followed by an increase until the end of the follow-up. Lung cancer showed a different trend, whereas the curve for anal cancer was irregular because of the small number of cases.
We divided analysis of follow-up into two periods, 0–9 years and more than 9 years after diagnosis, to allow separation of the apparent treatment-related effects after invasive cervical cancer. 2 The second primaries after in situ cervical cancer were increased in the two follow-up periods for cancers of the anus, lung, cervix, and other female genitals (Table 2). Skin cancer was increased in the first period only, whereas upper aerodigestive tract, esophagus, endometrium, kidney, and urinary bladder cancer were increased in the latter period only. Second primaries after invasive cervical cancer were increased in the two follow-up periods for cancer of the anus, lung, other female genitals, and bladder (Table 3). Upper aerodigestive tract, pancreatic, and renal cancers and leukemia were increased in the first period; colon, rectum, and connective tissue cancers were increased in the latter period.
We reviewed histologies of urinary bladder cancers in all patients and of those diagnosed after cervical cancers. The main histologic classification was transitional cell carcinoma, and the small proportion of squamous cell carcinomas was equal among all bladder cancers and those diagnosed after cervical cancer (data not shown).
Evidence of HPV infection can be obtained by demonstrating viral DNA or specific antibodies in the human body. 6,8 For carcinogenesis, past infections are most relevant but difficult to document in an epidemiologic setting. Oncogenic HPV types have been reported in the affected organs of internal cancers, but the possibility of contamination during isolation has not been ruled out. 6 Here we used the nationwide Swedish Family-Cancer Database to address the possible risks of HPV in a follow-up of patients with cervical cancer. 13 In the database, cancer cases were identified through the Swedish Cancer Registry, which has recorded all invasive and in situ cancers since 1958. 16 The advantage of using both invasive and in situ cancers is the ability to differentiate the effect of treatment. Surgery and radiotherapy have been the standard methods of treatment for invasive cervical cancer, whereas for in situ cancer superficial ablative therapy, performed often with laser and on an outpatient basis, has been the standard treatment. Consistently, no evidence for radiotherapy-related effects were seen after in situ cancer: SIRs for cancers at radiosensitive sites, such as female genitals, urinary bladder, and rectum decreased rather than increased during the follow-up. 2 Almost 100% of cervical cancers were verified by cytologic/histologic examination. A total of 117,830 patients with in situ cancer and 17,556 with invasive cervical cancer were followed up. Below, we will discuss separately four aspects of the results obtained.
First, the results of follow-up after both in situ and invasive cervical cancers were consistent in showing an increase in anal and female genital cancers, both being known targets of HPV. 6,7 The SIRs at these sites were higher after invasive than in situ cancer, probably relating to the length and severity of infection; invasive cancers are diagnosed on average 10 years later than in situ cancers, and many in situ cancers regress spontaneously. 17,18 The high SIRs for genital and anal cancers immediately after diagnosis (<1 year, Figures 1 and 2) may be ascribed to an exceptionally high diagnostic intensity during the year of cervical cancer treatment, but an alternative explanation is given later. The data showed consistent increases also for cancers of the upper aerodigestive tract and esophagus, suggested targets of HPV, 6,8 but because these are also tobacco-related sites, they are discussed together with lung cancer (see next paragraph). There was a weak association only with in situ cervical cancer and squamous cell carcinoma of the skin, lending at most weak support to the suggested role of HPV infection on nonmelanomatous skin cancer. 19 The SIR for invasive cervical cancer after in situ cancer was 2.30. In a similar study on 37,000 patients with in situ cancer from Norway, no excess risk was noted, 4 whereas in a study on 2,200 patients from Switzerland the SIR was 3.4. 5
Second, there were consistent increases in second primaries at tobacco-related sites, including upper aerodigestive tract, pancreas, lung, and urinary bladder, and additionally esophagus for in situ cancer only. The data on lung cancer are consistent with previous studies on second primary cancers after invasive cervical cancer and other genital cancers. 2,20,21 Many of the above sites, however, are also assumed targets for HPV. We assessed the contribution of smoking by considering the following available RRs at these sites among female Swedish smokers of >15 cigarettes/day 22 : 7.75 for lung; 1.89 for upper aerodigestive tract, including esophagus; 3.44 for urinary bladder; and 1.64 for pancreas. The RR for lung cancer was identical for women smoking 8–15 cigarettes/day, but the RRs at the other sites were lower. Assuming that the increased SIR for lung cancer in cervical cancer patients was solely due to smoking, then the increases in SIR of oral and pancreatic cancers can be explained by smoking. For esophageal cancer the increase after in situ cervical cancer cannot be fully explained by smoking (SIR for esophagus, 1.85, and for lung, 2.17), but the strong modification of esophageal cancer by alcohol consumption invalidates conclusions about small differences in SIR. Furthermore, esophageal cancer showed no increase after invasive cervical cancer. By contrast, increased SIRs in urinary bladder cancer were higher than expected by smoking alone. For example, 0–9 years after invasive cancer, the SIR of lung cancer was 4.45, and that of bladder cancer was 2.99 (Table 3); the corresponding SIRs after in situ cancer were 1.81 and 1.52 (Table 2). It would appear that up to one-half of the increase in SIR of urinary bladder cancer remained unexplained by smoking and was possibly contributed by HPV infection or related immunosuppression. Although the relative excess was modest, the number of attributable cases could be close to the attributable number of cases of anal cancer. In a previous Swedish study of 5,325 patients with invasive cervical cancer, the SIR of urinary bladder cancer was increased during the follow-up period. 2 Similarly, the Norwegian study of women with in situ cervical cancers noted an increase in urinary bladder cancer as a second primary. 4
Third, the study provided evidence of treatment-related increases in cancer at radiosensitive sites in the vicinity of the uterine cervix, in agreement with previous studies. 2,23 The increases at these sites were observed 10 or more years after diagnosis of invasive cervical cancer and included the following sites: colon, rectum, anus, other female genitals, and urinary bladder (Table 3 and Figure 2). The increase in connective tissue cancers also appeared to be treatment related.
Fourth, the data suggest a contributing role for a depressed immune function in permitting second primary cancers to appear. The progression of cervical cancer signals faltering immunosurveillance, 6,10 which may affect other parts of the body. Thus, even if the underlying reason for the genital and anal cancers was transformation by HPV, the emergence of clinical cancer was triggered by malfunctioning cellular immunity. In urinary bladder, the similarity of the histologic distribution between primary bladder cancers unrelated to cancers of other sites and those occurring after cervical cancer did not support a direct role for HPV. Rather, depressed immunosurveillance may permit progression of preexisting transformed cells to clinical malignancy. The critical role of immune function in controlling squamous cell carcinomas and lymphomas is observed in immunosuppressed patients. 10 A more subtle depression may contribute to familial cervical cancer 18 and to second primary cancers after squamous cell carcinoma of the skin 24 and, as suggested here, after cervical cancer.
1. Sankila R, Pukkala E, Teppo L. Risk of subsequent malignant neoplasms
among 470,000 cancer patients in Finland 1953–1991. Int J Cancer 1995; 60:464–470.
2. Bergfeldt K, Einhorn S, Rosendahl I, Hall P. Increased risk of second primary malignancies in patients with gynecological cancer. Acta Oncol 1995; 34:771–777.
3. Hemminki K, Vaittinen P. Familial risks in second primary breast cancer based on the Family-Cancer Database. Eur J Cancer 1999; 35:455–458.
4. Bjorge T, Hennig EM, Skare GB, Soreide O, Thoresen SO. Second primary cancers in patients with carcinoma in situ
of the uterine cervix: the Norwegian experience, 1970–1992. Int J Cancer 1995; 62:29–33.
5. Levi F, Randimbison L, Vecchia C, Franceschi S. Incidence of invasive cancers following carcinoma in situ
of the cervix. Br J Cancer 1996; 74:1321–1323.
6. IARC. Human Papillomaviruses. IARC Monographs on the Carcinogenic Risks to Humans, No. 64. Lyon: International Agency for Research on Cancer, 1995.
7. Pisani P, Parkin D, Munoz N, Ferlay J. Cancer and infection: estimates of the attributable fraction in 1990. Cancer Epidemiol Biomarkers Prev 1997; 6:387–400.
8. Björge T, Hakulinen T, Egeland A, Jellum E, Koskela P, Lehtinen M, Luostarinen T, Paavonen J, Sapp M, Schiller J, Thoresen S, Wang Z, Youngman L, Dillner J. A prospective, seroepidemiological study of the role of human papillomavirus in esophageal cancer in Norway. Cancer Res 1997; 57:3989–3992.
9. Sanjeevi C, Hjelmström P, Hallmans G, Wiklund F, Lenner P, Angstrom T, Dillner J, Lernmark A. Different HLA-DR-DQ haplotypes are associated with cervical intraepithelial neoplasia among human papillomavirus type-16 seropositive and seronegative Swedish women. Int J Cancer 1996; 68:409–414.
10. Birkeland S, Storm H, Lamm L, Barlow L, Blohme I, Forsberg B, Eklund B, Fjeldborg O, Friedberg M, Frodin L. Cancer risk after renal transplantation in the Nordic countries, 1964–1986. Int J Cancer 1995; 60:183–189.
11. IARC. Human Immunodeficiency Viruses and Human T-Cell Lymphotrophic Viruses. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, No. 67. Lyon: International Agency for Research on Cancer, 1996.
12. McGregor J, Proby C. The role of papillomaviruses in human non-melanoma skin cancer. Cancer Surv 1996; 26:219–236.
13. Hemminki K, Vaittinen P, Kyyrönen P. Age-specific familial risks in common cancers of the offspring. Int J Cancer 1998; 78:172–175.
14. Hemminki K, Vaittinen P, Kyyrönen P. Modification of cancer risk in offspring by parental cancer. Cancer Causes Control 1999; 10:125–129.
15. Estève J, Benhamou E, Raymond L. Descriptive Epidemiology. Statistical Methods in Cancer Research. Vol. 4. IARC Scientific Pub. No. 128. Lyon: International Agency for Research on Cancer, 1994.
16. Hemminki K, Vaittinen P. Familial risks in in situ
cancers from the Family-Cancer Database. Cancer Epidemiol Biomarkers Prev 1998; 7:865–868.
17. Holowaty P, Miller A, Rohan T, To T. Natural history of dysplasia of the uterine cervix. J Natl Cancer Inst 1999; 91:252–258.
18. Hemminki K, Dong C, Vaittinen P. Familial risks in cervix cancer: is there a hereditary component? Int J Cancer 1999; 82:775–781.
19. Levi F, Randimbison L, La Vecchia C. Nonmelanomatous skin cancer following cervical, vaginal, and vulval neoplasms
: etiologic association. J Natl Cancer Inst 1998; 90:1570–1571.
20. Macara L, Lamont D, Symonds R. Second malignancies in cervical cancer patients in the west of Scotland. Scott Med J 1998; 43:16–18.
21. Sturgeon S, Curtis R, Johnson K, Ries L, Brinton L. Second primary cancers after vulvar and vaginal cancers. Am J Obstet Gynecol 1996; 174:929–933.
22. Nordlund L, Carstensen J, Pershagen G. Cancer incidence in female smokers: a 26-year follow-up. Int J Cancer 1997; 73:625–628.
23. Boice JJ, Engholm G, Kleinerman R, Blettner M, Stovall M, Lisco H, Moloney WC, Austin DF, Bosch A, Cookfair DL. Radiation dose and second cancer risk in patients treated for cancer of the cervix. Radiat Res 1988; 116:3–55.
24. Hemminki K, Dong C. Primary cancers following squamous cell carcinoma of the skin suggest involvement of Epstein-Barr virus (Letter). Epidemiology 2000; 11:94.