|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2000 American Society for Clinical Oncology Risk of Second Malignancy After Hodgkin’s Disease in a Collaborative British Cohort: The Relation to Age at TreatmentFrom the Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine; the Department of Medical Statistics and Evaluation, Imperial College School of Medicine; the British National Lymphoma Investigation, University College Hospital; and the Imperial Cancer Research Fund Medical Oncology Unit, St Bartholomew’s Hospital, London; the Lymphoma Unit and the Academic Unit of Radiotherapy and Oncology, The Institute of Cancer Research and Royal Marsden Hospital, Sutton, Surrey; and the Department of Clinical Oncology, Weston Park Hospital, Sheffield, United Kingdom. Address reprint requests to A.J. Swerdlow, DM, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, Keppel St, London WC1E 7HT, United Kingdom; email a.swerdlow{at}lshtm.ac.uk
PURPOSE: To assess long-term site-specific risks of second malignancy after Hodgkin’s disease in relation to age at treatment and other factors. PATIENTS AND METHODS: A cohort of 5,519 British patients with Hodgkin’s disease treated during 1963 through 1993 was assembled and followed-up for second malignancy and mortality. Follow-up was 97% complete. RESULTS: Three hundred twenty-two second malignancies occurred. Relative risks of gastrointestinal, lung, breast, and bone and soft tissue cancers, and of leukemia, increased significantly with younger age at first treatment. Absolute excess risks and cumulative risks of solid cancers and leukemia, however, were greater at older ages than at younger ages. Gastrointestinal cancer risk was greatest after mixed-modality treatment (relative risk [RR] = 3.3; 95% confidence interval [CI], 2.1 to 4.8); lung cancer risks were significantly increased after chemotherapy (RR = 3.3; 95% CI, 2.4 to 4.7), mixed-modality treatment (RR = 4.3; 95% CI, 2.9 to 6.2), and radiotherapy (RR = 2.9; 95% CI, 1.9 to 4.1); breast cancer risk was increased only after radiotherapy without chemotherapy (RR = 2.5; 95% CI, 1.4 to 4.0); and leukemia risk was significantly increased after chemotherapy (RR = 31.6; 95% CI, 19.7 to 47.6) and mixed-modality treatment (RR = 38.1; 95% CI, 24.6 to 55.9). These risks were generally greater after treatment at younger ages: for patients treated at ages younger than 25 years, there were RRs of 18.7 (95% CI, 5.8 to 43.5) for gastrointestinal cancer after mixed-modality treatment, 14.4 (95% CI, 5.7 to 29.3) for breast cancer after radiotherapy, and 85.2 (95% CI, 45.3 to 145.7) for leukemia after chemotherapy (with or without radiotherapy). CONCLUSION: Age at treatment has a major effect on risk of second malignancy after Hodgkin’s disease. Although absolute excess risks are greater for older patients, RRs of several important malignancies are much greater for patients who are treated when young. The increased risk of gastrointestinal cancers may relate particularly to mixed-modality treatment, and that of lung cancer to chemotherapy as well as radiotherapy; there are also well-known increased risks of breast cancer from radiotherapy and leukemia from chemotherapy. The roles of specific chemotherapeutic agents in the etiology of solid cancers after Hodgkin’s disease require detailed investigation.
IT HAS BEEN 30 YEARS since the introduction of intensive radiotherapy and chemotherapy transformed the prognosis of patients with Hodgkin’s disease. Because most of the patients are young when treated and many now survive for several decades, the long-term side effects of treatment are becoming increasingly important. Both radiotherapy and alkylating chemotherapy are themselves carcinogenic, and hence second malignancy after treatment is important to characterize and to try to prevent. Alkylating chemotherapy has been shown to cause a large relative risk of leukemia within a few years after treatment.1-5 Solid cancer risks increase more gradually, but in the long-term, these cancers constitute the great majority of second malignancies,1,3,4,6-9 and the extent of these risks and their etiology are incompletely understood. With longer follow-up, previously unrecognized risks are becoming apparent. It has recently been shown that after radiotherapy, the risk of breast cancer is increased after 10 to 15 years of follow-up, but that this risk is critically dependent on age at treatment: relative risks have been greatly increased in women who are treated when they are younger than approximately 25 years (especially those who are treated in childhood) but have been increased little or not at all in women who are treated when they are older than 25 years.9-13 Recent evidence has also suggested that risks of certain other cancers may be greater after childhood11,12 than after adult1,4,7-9,14-17 treatment. There is little evidence, however, on the effect of age within adults on second cancer risk1,8,9,18-21 and none on whether long-term age-related effects similar to those for breast cancer occur for other major sites of second malignancy. The present analyses explored the relationship of risk of second malignancy to age at treatment, type of treatment, and time since first treatment in a large cohort of British patients with Hodgkin’s disease treated over the past 35 years.
The British National Lymphoma Investigation (BNLI), Royal Marsden Hospital, and St Bartholomew’s Hospital each maintain databases on the treatment and follow-up of patients with Hodgkin’s disease who have been treated since 1970 or earlier. We extracted from these databases information on all recorded patients except those who were foreign residents and then updated the treatment and follow-up data as necessary. The BNLI database has recorded clinical trials patients treated since the foundation of the BNLI in 1970, and we extracted data on all new patients between then and April 1993, with follow-up to April 30, 1993. For The Royal Marsden Hospital, the database started in 1963, and data were extracted on all patients first treated from then to March 1991 in one part of the hospital, and to February 1989 in another part, with follow-up to the end of June 1991. For St Bartholomew’s Hospital, data were available on patients first treated from 1967 to April 1993, and follow-up was to the end of 1993. For each of these data sets, checks on completeness of follow-up for second malignancy have been conducted. For the BNLI3 and The Royal Marsden Hospital,20 these checks revealed few second cancers that were not already known from the databases, and for St Bartholomew’s Hospital, checks against the Thames Cancer Registry for cases incident from 1985 onwards (the Registry did not contain information on cancers occurring in the relevant region before then) found no cancers that were not already in the hospital’s own database. The sites of second cancer were coded to the International Classification of Diseases (ICD)22 revisions that were in force in England and Wales at the time of occurrence of the cancer: ICD7 for 1963 through 1967, ICD8 for 1968 through 1978, and ICD9 for 1979 onward. We bridge-coded the data to the ICD9 categories listed in Table 1.
For each patient in the cohort, person-years at risk of second cancer were calculated by 5-year age group, sex, and calendar year, from the date of first treatment to the date of end of follow-up, or of second cancer incidence, death, or loss to follow-up, if earlier. For analyses of time since first treatment and type of treatment, subjects were allocated at each point in their follow-up to the analytic category applicable at that time. Type of treatment was categorized into chemotherapy, radiotherapy, and mixed-modality treatment; it was not practical to obtain more detailed therapeutic information for the overall cohort, and this will be addressed elsewhere by nested case-control studies. Expected numbers of cancers incident in the cohort were calculated for each cancer site by multiplying age-, sex-, and calendar year-specific person-years at risk in the cohort by the corresponding cancer registration rates in the general population of England and Wales. Because these national registration statistics were not available in computer readable form before 1971 or after 1989, we used 1971 data to give expectations from 1963 to 1971 and 1989 data to give expectations from 1989 to 1993. The ratio of observed to expected numbers of cancers, the standardized incidence ratio (SIR; referred to in the text as relative risk [RR]) was then calculated, with likelihood-based 95% confidence intervals (CI) from Poisson models.23 All significance levels cited are two-sided. Adjustment of age at treatment risks for time since treatment, and vice versa, was also conducted using Poisson regression. Absolute excess risks of second cancer were calculated by subtracting the expected from the observed number of cases and dividing by the person-years at risk. Cumulative (actuarial) probabilities of second cancer were calculated by the Kaplan-Meier method.24
The cohort consisted of 5,519 patients: 3,772 from the BNLI, 1,039 from The Royal Marsden Hospital, and 708 from St Bartholomew’s Hospital (Table 2). Sixty-two percent (3,434) were male and 38% (2,085) were female. One hundred fifty-eight were younger than 15 years at first treatment, 3,990 were 15 to 44 years of age, and 1,371 were 45 years or older. Twenty-seven percent of cases were treated solely with radiotherapy, slightly more were treated solely with chemotherapy, and more than 40% received both modalities. During follow-up, 344 patients developed a second cancer, 1,601 others died, 90 emigrated, 93 (1.7%) were lost to follow-up in other ways, and the remaining 3,391 survived without second cancer to the end of the study period. Only three of the second cancers occurred during the first 6 months after treatment, and thus there is unlikely to have been appreciable, if any, inclusion of cancers incident but undetected before Hodgkin’s disease treatment and detected because of it. Six patients suffered a third cancer: three gastrointestinal cancers, one bone cancer, one breast cancer, and one non-Hodgkin’s lymphoma (NHL); these tumors were not included in the analyses. In total, follow-up was for 46,990 person-years, an average of 8.5 years per cohort member.
The RR of second malignancy overall, excepting nonmelanoma skin cancer, was 2.9, with significantly increased risks of cancers of the tongue, mouth and pharynx, stomach, small intestine, colon, liver, lung, pleura, bone, nervous system and thyroid, and NHL and leukemia (Table 1). In terms of absolute excess risk, the greatest contributions were from lung cancer, NHL, and leukemia. Nonmelanoma skin cancers are excluded from further analysis in this study because registration of them is believed to be seriously incomplete. Increased risk of breast cancer was confined to patients who had received solely radiotherapy, and increased risk of leukemia was confined to patients who had received chemotherapy (with or without radiotherapy), whereas risks of NHL and lung cancer were increased fairly similarly, and highly significantly, in all treatment groups (Table 3). Risks of cancers of the stomach and colon were significantly increased only in patients who had received mixed-modality treatment, and cancers of the gastrointestinal tract overall were most greatly increased in this therapeutic group. Cancer of the small intestine, cancer of the bone, and melanoma occurred solely in patients treated with radiotherapy (with or without chemotherapy), and pleural and thyroid cancer risks were significantly increased only in the radiotherapy-treated groups, whereas risks of cancers of the tongue, mouth and pharynx, liver, and soft tissue were significantly increased only after mixed-modality treatment (not shown). Risk of nervous system tumors was increased in each treatment group but significantly in none.
Table 4 lists risks of the main types of second cancer in relation to sex and age at first treatment. Lung cancer is separated from other solid second cancers because of its frequency. The RR of lung cancer was borderline significantly greater in female than in male patients (P = .049), but there was not a significant difference in RR between the sexes for other cancers or for second malignancies overall. Absolute excess risks were greater for male than female patients for NHL and for second malignancy overall but were not markedly different between the sexes for other sites.
RRs of each major malignancy except NHL diminished sharply with increasing age at first treatment (Table 4). Absolute excess risks of lung cancer and NHL, however, increased greatly with age, leukemia risk increased modestly with age, and nonlung solid cancer risks were greater for children than adults but varied little by adult age. As a consequence, absolute excess risks of second malignancy overall were greater in children than young adults but increased with older age among adults. Adjustment of the RRs by age at treatment for the potentially confounding effect of time since first treatment (not shown) did not materially alter the results: the SIRs for leukemia and NHL were negligibly altered, and for solid cancers, there was a slight alteration upward in the SIRs for older compared with younger age groups, but persons 55 years of age and older still had the lowest SIR. Examining the relationship between age and solid malignancy risks in more detail by site (Table 5), breast cancer risk was only increased in women who were treated when they were younger than 25 years. For gastrointestinal tract cancer, RRs increased significantly with younger age at first treatment, but absolute excess risks (not shown) increased with older age, from 2.6 per 10,000 person-years at ages younger than 25 years to 21.4 per 10,000 person-years at ages 55 years and older. Based on small numbers, bone and soft tissue cancers and thyroid cancer only occurred in patients who were treated when they were younger than 45 years, but RRs were greatly increased for those treated at these ages. The RR of thyroid cancer was 16.6 (95% CI, 2.0 to 59.8) for patients younger than 25 years who received any radiotherapy and 9.6 (95% CI, 1.2 to 34.5) for patients 25 to 44 years of age who received this treatment (not shown).
The 20-year cumulative risk of second malignancy overall was 14.7%, of lung cancer 3.6%, other solid cancers 7.8%, NHL 2.3%, and leukemia, excluding chronic lymphocytic leukemia, 1.7% (Table 6). Actuarial risks for each of these categories were much greater for patients who were first treated at 45 years of age and older than for those who were treated when younger. Lung cancer and other solid malignancy risks were particularly great in the older patients: 14.8% and 19.0%, respectively. The 20-year actuarial risk of breast cancer was 4.2% (95% CI, 2.0% to 8.6%) for women who were first treated when they were younger than 25 years (not shown) and 2.9% (95% CI, 1.2% to 6.9%) for women who were treated at older ages.
Table 7 lists treatment-related risks of selected malignancies by age. The RR of gastrointestinal cancer after mixed-modality therapy was 18.7 for patients treated at ages younger than 25 years and diminished significantly with increasing age, whereas after chemotherapy or radiotherapy alone, there was no significant trend with age. The large risk of gastrointestinal cancers in the youngest age group after mixed-modality treatment reflected much increased risks of cancers of several sites: stomach, RR = 28.9 (95% CI, 1.6 to 127.1); colon, RR = 13.8 (95% CI, 0.8 to 60.8); and esophagus, RR = 140.5 (95% CI, 23.4 to 434.0). The RR of lung cancer was also greatly increased in patients who were treated with mixed modalities at ages younger than 25 years (RR = 38.3), and again, this diminished steeply with age. There was also a significant reduction in RR of lung cancer with age after radiotherapy but not after chemotherapy alone. The RR of breast cancer was increased only in women who were treated with radiotherapy when they were younger than 25 years; in all of these cases, the young women had received supradiaphragmatic lymph node irradiation. The risk of breast cancer in women who were treated when they were younger than 25 years was greater for patients who were treated with radiotherapy alone (RR = 14.4) than for those who were treated with mixed modalities (RR = 4.6), and no cases occurred (0.18 expected) in women who were treated solely with chemotherapy. When women who were younger than 45 years at treatment were considered, the corresponding treatment-specific risks were 3.4 (95% CI, 1.6 to 6.3), 0.7 (95% CI, 0.1 to 2.6), and 0.7 (95% CI, 0.02 to 3.7), respectively, and at ages 45 years and older, the corresponding treatment-specific risks were 1.5 (95% CI, 0.4 to 3.8), 0.8 (95% CI, 0.02 to 4.4), and 0.4 (95% CI, 0.01 to 2.5), respectively. For NHL RRs (not shown), there was no systematic variation by age and treatment. The RR of leukemia decreased significantly with age in both chemotherapy- and mixed modality–treated patients (Table 7); the RR at ages younger than 25 years in these two therapeutic groups combined was 85.2 (95% CI, 45.3 to 145.7). Based on smaller numbers (not shown), the RRs of bone and soft tissue cancers were greatly increased both at ages younger than 25 years and ages 25 to 44 years in patients who had received radiotherapy only (RRs of 8.9 [95% CI, 0.51 to 39.2] and 5.9 [95% CI, 0.34 to 25.9], respectively) or mixed modalities (RRs of 29.2 [95% CI, 7.3 to 75.7] and 13.9 [95% CI, 2.3 to 42.9], respectively), and the same was true for patients with thyroid cancer (radiotherapy, RR = 16.3 [95% CI, 0.9 to 71.7] and RR = 9.1 [95% CI, 0.5 to 39.9], respectively; mixed modalities, RR = 16.9 [95% CI, 1.0 to 74.4) and RR = 10.1 [95% CI, 0.6 to 44.5], respectively).
The relative and absolute excess risks of leukemia and NHL reached a peak 5 to 9 years after first treatment, although they continued to be substantially increased beyond this, even at  15 years after treatment (Table 8). The person-years that were in the category of 15 years were mainly (85%) at 15 to 19 years of follow-up. For lung cancer and other solid cancers, no peak was reached; RRs continued to increase throughout follow-up and absolute excess risks increased more steeply. When the RRs by time since first treatment were adjusted for age at treatment (not shown), there were no substantial changes to the above findings: there was a slight alteration downward in the RRs for long compared with short follow-up periods, such that the lung cancer RRs stabilized rather than increased beyond 5 to 9 years of follow-up and the risks for other solid cancers increased less greatly with longer follow-up than in the unadjusted analysis.
RRs of gastrointestinal cancer increased highly significantly with time since first treatment (Table 9). This reflected a highly significant trend for patients who were treated with mixed modalities, for whom the RR at 15 years of follow-up was 10.4 (95% CI, 5.2 to 18.2), but there was no significant trend for the other two treatment groups (not shown). There was also a significant trend of greater risk with longer follow-up for thyroid cancer, although this was based on small numbers (Table 9).
The temporal pattern of second cancer risks varied considerably by age (Table 10). In children, solid cancer RRs were greatly increased within 5 years after treatment and showed no trend thereafter, based on small numbers. In young adults, by contrast, solid cancer RRs started little raised but increased consistently throughout follow-up, and in older adults, RRs were lower than in young adults and did not increase with follow-up beyond 5 years. Considering the main solid cancers by site, gastrointestinal cancer RRs were moderately—and in general, not significantly—increased at ages younger than 45 years in the period of less than 15 years from first treatment and at older ages throughout follow-up, but in younger patients at 15 years after treatment, the RR was significantly increased at 7.4 (the corresponding RR for patients treated with mixed-modality therapy was 14.1 [95% CI, 6.1 to 27.2]). For each follow-up period after the initial 5 years, the RR of lung cancer was approximately twice as great in patients who were first treated when they were younger than 45 years as in those treated at an older age. Breast cancer risks were only significantly increased at 10 years of follow-up in patients who were treated when they were younger than 25 years; there was no increased risk in any follow-up period for older patients.
For NHL, the pattern of risks over time did not vary materially by age (Table 10). For leukemia, however, the peak in risk at 5 to 9 years after treatment was present only in patients who were treated at ages younger than 45 years (Table 10). This was a consequence of variations in the time course of risks by age in patients who were treated with chemotherapy (with or without radiotherapy) for whom the RR of leukemia at 5 to 9 years after treatment was 94.5 (95% CI, 52.9 to 156.0) at ages younger than 45 years, compared with 19.4 (95% CI, 5.3 to 49.7) at older ages (not shown).
The study showed that long-term risks of second malignancy after Hodgkin’s disease are highly dependent on age at treatment. The RR of leukemia increased significantly with younger age at first exposure, in accord with some8,18 but not all1 previous results, which were based on much smaller numbers. Absolute excess risks, however, were greatest for patients treated at older ages. The time course of the leukemia risk varied by age. As in numerous previous studies,1,4,7-9 all-age RRs of leukemia reached a peak in the early years after first treatment, but on examination of these risks by age in the present data, it became apparent that this peak was solely a function of risk in patients who were treated when they were younger than 45 years; for older patients, no peak occurred. There seem to be no previous published analyses with which to compare this finding, which was a consequence of different time courses for different age groups in patients who received chemotherapy, which was the main cause of leukemia in Hodgkin’s disease patients in our study as in previous data.1,3-5 The pattern does show some resemblance, however, to the age-specific patterns of RR of leukemia after radiation exposure in atomic bomb survivors: RRs in the early years were much greater for persons who were exposed at younger than at older ages, and the decline in risk with time was slight or nonexistent for those who were exposed when they were older than 40 years.25 NHL RRs also showed a peak at 5 to 9 years after first treatment, as in one previous study,9 although in general, there has been no consistent pattern.1,4,6-8,11 There was no age or treatment relationship, however, of the NHL RRs. Although the large RRs of leukemia and NHL in the early years after treatment of Hodgkin’s disease were striking, the results of longer follow-up emphasized the predominance of solid malignancies, and particularly of lung cancer, which accounted for one quarter of the excess risk of second cancer in the cohort. RRs of solid cancers, notably lung, gastrointestinal, and breast cancers, were greater for patients who were younger at first treatment, but absolute excess risks of lung cancer were greatest for patients who were treated at older ages, and for other solid cancers, absolute excess risks showed no consistent trend with age. The absolute excess measure is the more important one for clinical consideration of hazard. The rates of second cancer in patients who were treated when they were young are, in absolute terms, low while they are at young ages, when background cancer rates are relatively low, but the feature of concern about the large RRs in these subjects is the possibility that these increased RRs might persist as the patients grow older and background rates increase, in which case they would translate into large absolute excesses. Analysis of the time course of risks by age (Table 8) suggested that this may be so. In children, risks were greatly increased even in the first 5 years after treatment and persisted at a similar level thereafter, following the same pattern as was observed in the only previous similar analysis.11 There do not seem to have been previous analyses of the time course of risks by age within adults, but our data showed rising RRs for younger and not older adults. With few person-years of observation beyond 20 years of follow-up in the study, however, the absolute excess risks in patients who were treated as young adults, although increasing, remained relatively low; it will be important to determine risks in these subjects over the next decade. The large cumulative risks of second solid malignancy in patients who were treated at older ages need to be interpreted with caution. In part, the cumulative risks are high because background cancer rates are high in older people in general: the cumulative measure of risk, unlike the relative and absolute excess risks, makes no adjustment for expectations from background population risks at these ages. Furthermore, the 20-year cumulative risks are a notional construct that is conditional on patients surviving for 20 years (or until cancer incidence). In practice, however, many older patients will die before 20 years have elapsed from causes other than second malignancy. The largest component of the solid cancer risk was from lung cancer, for which RRs in all time periods beyond 5 years were much greater in patients who were treated when they were young than those who were treated at older ages. There seem to be no data from other cohorts of Hodgkin’s disease patients with which to compare these findings. In several cohorts of people who were exposed to radiation for other reasons, RRs have not been related to age at exposure.26,27 The increased risk of lung cancer after radiotherapy for Hodgkin’s disease in our cohort is a usual finding,4,8,28 but the risk after chemotherapy has been inconsistent in previous studies4,5,8,14,28,29; our data add highly significant results, based on substantial numbers, to suggest that chemotherapy may indeed be a risk factor. The large increased risk of breast cancer in long-term follow-up of women with Hodgkin’s disease who were treated with radiotherapy at ages younger than 25 years and the lack of an increased risk at older ages of treatment are in agreement with findings from previous studies.4,9,10,13,30 The major increase in risk occurred a little sooner among patients in our study, at 10 to 14 years of follow-up, than in previous studies, in which it occurred at 15 to 19 years.4,10 It is difficult to compare the extent of risk in young women in our cohort with that in young women in other studies4,9-12,30 because of differences in age groupings, duration of follow-up, completeness of follow-up in certain instances,31 and treatment patterns. The somewhat reduced risks of breast cancer in women who were treated with mixed modalities or solely chemotherapy at premenopausal ages would be compatible with an effect of chemotherapy in inducing premature menopause and hence reduced risk, but the reductions were not significant and reduced risks were also present after chemotherapy at older ages (although again, nonsignificantly), so conclusions must be made with caution. The increasing risk of gastrointestinal cancers with longer follow-up has been observed previously.4,21 Radiation exposure has been shown to cause gastrointestinal malignancies in other irradiated groups,26,27 but there are very limited data available on treatment-related risks after Hodgkin’s disease. Our findings concur with those in a study from Stanford, CA,21 in finding a borderline significantly increased risk after radiotherapy but a larger and highly significant risk after combined-modality treatment, which warrants further investigation. The risk after chemotherapy alone was not significantly or greatly increased in our data; in the Stanford study,21 there were few patients who received this treatment, and no cases of gastrointestinal cancer occurred in those patients. One previous study reported a relationship between gastrointestinal cancer risk and doxorubicin treatment,5 but only two patients with such a cancer in our study had received this drug. The larger RR of gastrointestinal cancers in patients who were treated at young rather than older ages accords with previous data,9,11,12,21 but the effect of duration of follow-up on the age-specific risks does not seem to have been examined before. Our data suggest particular long-term risk from the combination of chemotherapy and radiotherapy administered to young patients. The increased risk of thyroid cancer is to be expected, as it was demonstrated in previous studies,4-8,11,12,32,33 and this risk has been found to be greater in patients who were treated as children11,12,33 than as adults,3-8 although there do not seem to be previous investigations of risks in relation to different adult ages at treatment. Although the risk in relation to modality of treatment for Hodgkin’s disease does not seem to have been analyzed before, cases have been shown to occur mainly within radiotherapy fields,12,32 and the sensitivity of the thyroid to radiation carcinogenesis is evident from studies of other groups of exposed people.34 The significant 10-fold risk that occurred after radiotherapy when such treatment was administered to patients at ages 25 to 44 years in our data contrasts with the findings in radiation-exposed cohorts in other circumstances, where there was little raised risk for persons who were exposed when they were older than 20 years.34 The large increase in risk of thyroid cancer with greater time since first treatment accords well with findings from other radiation-exposed cohorts,33 although such an effect has been unclear6 or not present11 in previous studies of Hodgkin’s disease patients. The much increased RRs of bone and soft tissue cancers accord with findings from other cohorts,1,4-7,9,11,12 although results have varied in terms of whether risks have been greater for patients treated at young ages11,12,35,36 than at all ages in total.1,4-7,9 The relationship between bone and soft tissue cancers and radiotherapy has been found previously14 and accords with studies of radiation effects in general.35 One study of Hodgkin’s disease patients found a significant increased risk in relation to chemotherapy,5 but in our data, no cases occurred in patients who received this treatment only. The increased risk of melanoma in the cohort is in accordance with previous findings,1,4-9,11,12 and the timing of this risk in the years soon after treatment accords with one previous study1,36 but less clearly with another.6 The early risk of melanoma might, it has been suggested,37 reflect immunologic dysfunction and immunosuppressive effects of treatment in the early years. There is limited previous evidence of an association of risk after Hodgkin’s disease with radiotherapy.14 All melanoma patients in the present study had received radiotherapy, although we do not have information on whether the tumors occurred within the radiation fields. In a previous study, several patients with melanoma after Hodgkin’s disease had shown clinical evidence of radiosensitivity at the time of treatment.37 Melanoma is not a tumor, however, for which increased risk has in general been found in relation to radiation.26 Significant increased risks were also present in the cohort for cancers of the mouth and pharynx, pleura, and nervous system. For oral/pharyngeal and pleural cancers, there is some support for an increased risk, although not decisively, from previous studies.4,6-8,12,14 For nervous system tumors, data have been inconsistent in terms of whether risks were increased.6,7,11,12,14 Based on small numbers, there were significant treatment relationships between oral and pharyngeal cancer and mixed-modality treatment and between pleural cancer and radiotherapy; from the general literature on radiation carcinogenesis, one would expect relationships between nervous system cancers (and, to some extent, between oral and pharyngeal cancers) and radiotherapy.26 In conclusion, follow-up of this large cohort shows increasing RRs of solid tumors, with longer survival of patients who are treated for Hodgkin’s disease at young ages, and especially demonstrates the growing importance of lung and gastrointestinal second cancers in such patients and the possible role of chemotherapy as well as radiotherapy in the etiology of these tumors. The patients who were treated when they were young are still at ages when cancer rates are relatively low; it is important to continue follow-up of their risks of second cancer as these patients grow older, because if there were continued high RRs, then these would translate into large absolute risks of cancer.
Supported by funding from the Medical Research Council, London, to the Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine; the Lymphoma Research Trust and Cancer Research Campaign, London, Lisa Lear Fund, Windsor, and Isle of Man Anti-Cancer Association, Isle of Man, to the British National Lymphoma Investigation; and the Imperial Cancer Research Fund, London, United Kingdom, to the Department of Medical Oncology, St Bartholomew’s Hospital. We thank the collaborators in the BNLI whose patients are included in this cohort. We also thank the regional cancer registries in England and Wales for provision of follow-up information; S. Campbell, A.J. Douglas, S. Milan, J. Nicholas, R.Z. Omar, PhD, L. Wickens, and A. Wilson for help in assembling the cohort data; and E. Middleton for secretarial help.
1. Tucker MA, Coleman CN, Cox RS, et al: Risk of second cancers after treatment for Hodgkin’s disease. N Engl J Med 318:76-81, 1988[Abstract] 2. Kaldor JM, Day NE, Clarke EA, et al: Leukemia following Hodgkin’s disease. N Engl J Med, 322:7-13, 1990[Abstract] 3. Swerdlow AJ, Douglas AJ, Vaughan Hudson G, et al: Risk of second primary cancers after Hodgkin’s disease by type of treatment: Analysis of 2846 patients in the British National Lymphoma Investigation. BMJ 304:1137-1143, 1992 4. van Leeuwen FE, Klokman WJ, Hagenbeek A, et al: Second cancer risk following Hodgkin’s disease: A 20-year follow-up study. J Clin Oncol 12:312-325, 1994[Abstract]
5.
Boivin J-F, Hutchison GB, Zauber AG, et al: Incidence of second cancers in patients treated for Hodgkin’s disease. Cancer Inst 87:732-741, 1995 6. Kaldor JM, Day NE, Band P, et al: Second malignancies following testicular cancer, ovarian cancer and Hodgkin’s disease: An international collaborative study among cancer registries. Int J Cancer 39:571-585, 1987[Medline] 7. Henri-Amar M: Second cancer after the treatment for Hodgkin’s disease: A report from the International Database on Hodgkin’s Disease. Ann Oncol 3:S117–S128, 1992 (suppl 4) 8. Abrahamsen JF, Andersen A, Hannisdal E, et al: Second malignancies after treatment of Hodgkin’s disease: The influence of treatment, follow-up time, and age. Oncol 11:255-261, 1993
9.
Mauch PM, Kalish LA, Marcus KC, et al: Second malignancies after treatment for laparotomy staged IA-IIIB Hodgkin’s disease: Long-term analysis of risk factors and outcome. Blood 87:3625-3632, 1996
10.
Hancock SL, Tucker MA, Hoppe RT: Breast cancer after treatment of Hodgkin’s disease. Cancer Inst 85:25-31, 1993
11.
Bhatia S, Robison LL, Oberlin O, et al: Breast cancer and other second neoplasms after childhood Hodgkin’s disease. J Med 334:745-751, 1996
12.
Sankila R, Garwicz S, Olsen JH, et al: Risk of subsequent malignant neoplasms among 1,641 Hodgkin’s disease patients diagnosed in childhood and adolescence: A population-based co-hort study in the five Nordic countries. J Clin Oncol 14:1442-1446, 1996
13.
Travis LB, Curtis RE, Boice JD Jr: Late effects of treatment for childhood Hodgkin’s disease. N Engl J Med 335:352-353, 1996 14. Boivin J-F, O’Brien K: Solid cancer risk after treatment of Hodgkin’s disease. Cancer 61:2541-2546, 1988[Medline] 15. Sont JK, van Stiphout WAHJ, Noordijk EM, et al: Increased risk of second cancers in managing Hodgkin’s disease: The 20-year Leiden experience. Ann Hematol 65:213-218, 1992[Medline]
16.
Rodriguez MA, Fuller LM, Zimmerman SO, et al: Hodgkin’s disease: Study of treatment intensities and incidences of second malignancies. Ann Oncol 4:125-131, 1993
17.
Dietrich P-Y, Henry-Amar M, Cosset J-M, et al: Second primary cancers in patients continuously disease-free from Hodgkin’s disease: A protective role for the spleen? Blood 84:1209-1215, 1994 18. van der Velden JW, van Putten WLJ, Guinee VF, et al: Subsequent development of acute non-lymphocytic leukemia in patients treated for Hodgkin’s disease. Int J Cancer 42:252-255, 1988[Medline] 19. Swerdlow AJ, Douglas AJ, Vaughan Hudson G, et al: Risk of second primary cancer after Hodgkin’s disease in patients in the British National Lymphoma Investigation: Relationships to host factors, histology and stage of Hodgkin’s disease, and splenectomy. Br J Cancer 68:1006-1011, 1993[Medline] 20. Swerdlow AJ, Barber JA, Horwich A, et al: Second malignancy in patients with Hodgkin’s disease treated at the Royal Marsden Hospital. Br J Cancer 75:116-123, 1997[Medline] 21. Birdwell SH, Hancock SL, Varghese A, et al: Gastrointestinal cancer after treatment of Hodgkin’s disease. Biol Phys 37:67-73, 1997 22. World Health Organization: Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death. Geneva, Switzerland, World Health Organization, 1978 23. Clayton D, Hills M: Statistical Models in Epidemiology. Oxford, England, Oxford University Press, 1993 24. Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 25. Preston DL, Kusumi S, Tomonaga M, et al: Cancer incidence in atomic bomb survivors: Part III. Leukemia, lymphoma and multiple myeloma, 1950-1987. Radiation Res 137:S68–S97, 1994[Medline] 26. National Research Council: Health Effects of Exposure to Low Levels of Ionizing Radiation: BEIR V. Washington DC, National Academy Press, 1990 27. Thompson DE, Mabuchi K, Ron E, et al: Cancer incidence in atomic bomb survivors: Part II. Solid tumors,1958-1987. Radiat Res 137:S17–S67, 1994
28.
van Leeuwen FE, Klokman WJ, Stovall M, et al: Roles of radiotherapy and smoking in lung cancer following Hodgkin’s disease. Natl Cancer Inst 87:1530-1537, 1995 29. Kaldor JM, Day NE, Bell J, et al: Lung cancer following Hodgkin’s disease: A case-control study. Int J Cancer 52:677-681, 1992[Medline] 30. Aisenberg AC, Finkelstein DM, Doppke KP, et al: High risk of breast carcinoma after irradiation of young women with Hodgkin’s disease. Cancer 79:1203-1210, 1997[Medline]
31.
Donaldson SS, Hancock SL: Second cancers after Hodgkin’s disease in childhood. Med, 334:792-794, 1996 32. Hancock SL, Cox RS, McDougall R: Thyroid diseases after treatment of Hodgkin’s disease. Med 325:599-605, 1991[Abstract] 33. Tucker MA, Morris Jones PH, Boice JD Jr, et al: Therapeutic radiation at a young age is linked to secondary thyroid cancer. Cancer Res 51:2885-2888, 1991 34. Ron E, Lubin JH, Shore RE, et al: Thyroid cancer after exposure to external radiation: A pooled analysis of seven studies. Radiat Res 141:259-277, 1995[Medline] 35. Tucker MA, D’Angio GJ, Boice JD Jr, et al: Bone sarcomas linked to radiotherapy and chemotherapy in children. N Engl J Med 317:588-593, 1987[Abstract] 36. Tucker MA: Solid second cancers following Hodgkin’s disease. Oncol Clin North Am 7:389-400, 1993 37. Tucker MA, Misfeldt D, Coleman CN, et al: Cutaneous malignant melanoma after Hodgkin’s disease. Ann Intern Med 102:37-41, 1985 Submitted March 17, 1999; accepted September 30, 1999. This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
