Skip Navigation
Skip to contents

Cancer Res Treat : Cancer Research and Treatment

Search
Advanced Search
OPEN ACCESS
mobile search button mobile menu button

Articles

Page Path
HOME > J Korean Cancer Assoc > Ahead-of print articles > Article
Original Article IARC Group 1 Pharmaceuticals and Associated Cancer Risks: A Nationwide Population-Based Cohort Study in Korea
Woojin Lim1,2,3orcid, Na Rae Lee4, Ho Gyun Shin4, Su-Yeon Yu5, Sue K. Park1,2,6orcid

DOI: https://doi.org/10.4143/crt.2024.1201
Published online: April 24, 2025

1Department of Preventive Medicine, Seoul National University College of Medicine, Seoul, Korea

2Cancer Research Institute, Seoul National University, Seoul, Korea

3Department of Biomedical Sciences, Seoul National University Graduate School, Seoul, Korea

4National Evidence-based Healthcare Collaborating Agency (NECA), Seoul, Korea

5College of Pharmacy, Kangwon National University, Chuncheon, Korea

6Integrated Major in Innovative Medical Science, Seoul National University Graduate School, Seoul, Korea

Correspondence: Sue K. Park, Department of Preventive Medicine, Seoul National University College of Medicine, 103 Daehak-ro, Jongno-gu, Seoul 03080, Korea
Tel: 82-2-740-8338 E-mail: suepark@snu.ac.kr
• Received: December 12, 2024   • Accepted: April 23, 2025

Copyright © 2025 by the Korean Cancer Association

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

  • 1,277 Views
  • 101 Download
Full Article

Download PDF Download PDF

  • Purpose
    The aim of this study is to summarize cancer risk among patients with clinical indications of immunosuppressive and antineoplastic drugs in Korea, which are pharmaceuticals defined as group 1 by International Agency for Research on Cancer.
  • Materials and Methods
    We conducted a nationwide population-based retrospective cohort study using the Korean National Health Insurance Service claims data from 2002 to 2018. Patients with clinical indications for group 1 pharmaceuticals from 2002 to 2017 were selected as baseline population, and followed up until 2018. Cox proportional hazards regression model was used to analyze the risk of cancer and dose-response relationship between group 1 pharmaceuticals and cancer.
  • Results
    Azathioprine use increased the risk of skin and hematologic cancer (hazard ratio [HR], 4.63; 95% confidence interval [CI], 2.91 to 7.39 and HR, 3.15; 95% CI, 2.41 to 4.13). Cyclosporine use increased the risk of skin and hematologic cancer (HR, 2.30; 95% CI, 1.79 to 2.95 and HR, 2.96; 95% CI, 2.59 to 3.40). Cyclophosphamide use increased the risk of bladder and hematologic cancer (HR, 2.69; 95% CI, 1.92 to 3.78 and HR, 3.83; 95% CI, 3.20 to 4.59). Chlorambucil use increased the risk of hematologic cancer (HR, 3.51; 95% CI, 2.53 to 4.87) and melphalan use increased the risk of hematologic cancer (HR, 16.31; 95% CI, 13.41 to 19.85). Methoxsalen use increased the risk of skin cancer (HR, 2.32; 95% CI, 1.36 to 3.95).
  • Conclusion
    Group 1 pharmaceuticals were associated with increased risk of cancer. The results are expected to help establish alternative clinical strategies and policies for patients with clinical indications of group 1 pharmaceuticals, by continuous risk analysis and discussions on the surveillance systems.
  • Key words: Group 1 pharmaceuticals, Immunosuppressive drugs, Antineoplastic agents, Skin neoplasms, Hematologic neoplasms
The International Agency for Research on Cancer (IARC), a subsidiary of the World Health Organization (WHO), has been conducting and coordinating research into cancer causes. In 1970, the IARC review committee proposed that its expert groups should determine cancer-causing hazards by qualitatively assessing evidence from both human and animal studies. This led to the creation of the IARC Monographs Program, aimed at identifying carcinogens and evaluating environmental factors contributing to cancer in humans. The program classifies substances, mixtures, and exposures into four groups: group 1 (carcinogenic to humans), group 2A (probably carcinogenic to humans), group 2B (possibly carcinogenic to humans), and group 3 (not classifiable regarding their carcinogenicity to humans). Since 1975, the carcinogenic potential of numerous pharmaceuticals has been examined, with 24 pharmaceuticals categorized as group 1 by 2021 [1].
Group 1 pharmaceuticals, often include antineoplastic or immunosuppressive drugs used solely, or in combination therapies. Particularly, group 1 immunosuppressive drugs are commonly prescribed as primary treatments for patients with various autoimmune diseases or to those who have received or to be receiving solid organ transplants. Group 1 antineoplastic drugs are usually prescribed as a combination therapy along with radiotherapy and surgery for patients with solid cancers. However, scarcity of alternative treatments, lack of novel therapeutic options, and classification of some group 1 pharmaceuticals into essential medicines by the WHO compel patients to continue using these drugs globally.
The IARC working group have been thoroughly evaluating the carcinogenicity of these drugs to date, indicating a link between their use and an increased risk of cancer. The working group have been also reviewing observational studies on the associations between drug use and cancer risk, presenting the summary of each study (e.g. study design, number of participants, type of exposure, outcome and effect size) in monographs [2,3]. However, since no studies have used real world data to analyze the associations between group 1 pharmaceuticals use and cancer risk in Korean population, our study aimed to quantify and assess the cancer risk among Korean patients prescribed with group 1 pharmaceuticals by using nationwide retrospective cohort. By using nationwide data of patients with various clinical indications, we sought to improve ongoing monitoring of the use of group 1 pharmaceuticals in Korea, providing evidence for making public health policies.
1. Group 1 pharmaceuticals
To estimate the causal effect of intervention, which are group 1 pharmaceuticals, target trial emulation was conducted and presented through target trial protocol of observational study (S1 Table) [4]. From each monograph on pharmaceuticals and drugs, all group 1 pharmaceuticals were selected, comprising a total of 24 pharmaceuticals [1]. Five of these, specifically diethylstilbestrol, chlornaphazine, phenacetine, a phenacetine-containing mixture, and semustine (methyl-CCNU), have been either banned globally or are under further investigation for their carcinogenic potential [1]. Additionally, four hormone-related drugs—combined estrogen-progestogen menopausal therapy, combined estrogen-progestogen oral contraceptives, tamoxifen, and postmenopausal estrogen therapy—were not considered, as their carcinogenic risks were already specified in monographs on hormonal factors [1]. Furthermore, the study aimed to focus solely on the risks caused by synthetic pharmaceuticals, leading to the exclusion of three herbal pharmaceuticals: aristolochic acid, plants associated with aristolochic acid, and opium. In addition, regimen combining mustargen, oncovin, procarbazine, and prednisone (MOPP), a treatment protocol combining etoposide, bleomycin, and cisplatin (BEP), treosulfan and etoposide were excluded since these four pharmaceuticals had no major ingredient codes in the National Health Insurance Service (NHIS) database.
Consequently, after excluding 16 pharmaceuticals from a total of 24 pharmaceuticals, cyclosporine, azathioprine, cyclophosphamide, busulfan, methoxsalen (alongside ultraviolet radiation therapy), melphalan, chlorambucil, and thiotepa were included in this study. Among selected eight pharmaceuticals, cyclosporine and azathioprine were classified as immunosuppressive drugs, and cyclophosphamide, busulfan, melphalan, chlorambucil, and thiotepa were classified as antineoplastic drugs. Specifically, methoxsalen was classified as skin ointment.
2. Clinical indications
The clinical indications were defined with diseases identified having substantial evidence in the IARC monographs on pharmaceuticals (S2 Table) [1,5,6]. Autoimmune diseases such as rheumatoid arthritis, psoriasis, atopic dermatitis, systemic erythematosus lupus, myasthenia gravis, inflammatory bowel disease, chronic allograft rejection, graft-versus-host diseases, and solid organ transplants as clinical indications for group 1 immunosuppressive drugs were included in this study. On the other hand, malignancies such as leukemia, lymphoma, myeloma, cancer in breast, ovary, lung, thyroid and urinary bladder as indications for group 1 antineoplastic drugs were included. In addition, dermatopathy such as idiopathic vitiligo, psoriasis, and eczema as indications for group 1 skin ointments were included. The classification and coding of these cancers were based on the International Classification of Diseases, 10th revision (ICD-10).
3. Data source
We used the NHIS claims data from January 1, 2002 to December 31, 2018. The NHIS operates as a single, universal health insurance system that encompasses the entire population of Korea [7]. This extensive database captures a wide array of healthcare utilization data, encompassing both inpatient and outpatient services. It includes related diagnostic codes aligned with the ICD-10, alongside comprehensive records of prescriptions, medical procedures, laboratory tests, demographic details, and contributions to insurance based on income levels. In addition, the database includes differential co-payments codes. Since 2011, the differential copayment policy applies a varying coinsurance rate to prescribed medications when a patient seeks healthcare services for a mild disease [8].
4. Study subjects
Study subjects were defined as patients diagnosed with confirmed clinical indications between January 1, 2002 and December 31, 2017. Clinical indications were defined as the presence of ICD-10 code as either a main or sub-diagnosis code. Given that the clinical indications included in this study require long-term treatment for chronic conditions, the differential co-payments codes typically employed for identifying new patients were not used [9].
The start of follow-up is defined as first record of prescription (use) after diagnosis of clinical indications to the point of event, death, or the end of the study (December 31, 2018). If follow-up ended within 30 days after the initial prescription of group 1 pharmaceuticals, or if the incidence of cancer occurred before the initial prescription of group 1 pharmaceuticals, patients were excluded [10]. Patients prescribed with group 1 pharmaceuticals within 3 years prior to the diagnosis of clinical indications were also excluded in order to wash out the remaining effect of pharmaceutical. Hence, prevalent users of group 1 pharmaceuticals were excluded in this study. In addition, if data on family history or cigarette smoking were missing, patients were excluded. To minimize the exclusion of patients from the study, data on family history and cigarette smoking recorded during health check-ups within 5 years before and after the cohort enrollment were included.
5. Exposure definition
Exposure of group 1 pharmaceuticals were operationally defined in three forms: as a binary variable representing exposure (ever-use) vs. non-exposure (non-use) [11], then using repeated prescription episodes to define a lifetime cumulative dose, and finally, categorizing into low and high, and low, intermediate, and high doses based on the median and tertile of lifetime cumulative dose to assess the dose-response relationship [12]. In addition, single dose and lifetime administered duration of group 1 pharmaceuticals in each patient were identified [13,14].
The use of group 1 pharmaceuticals (exposure) was assumed to potentially lead to cancer if at least one prescription was received, thus defining exposure as having received at least one prescription of any group 1 pharmaceuticals. Patients in the overall cohort who were not defined as ‘exposed’ to a particular group 1 pharmaceuticals were classified as ‘non-exposed’ [15].
In this study, the following formula were used to estimate the lifetime cumulative dose and lifetime administered duration of group 1 pharmaceuticals in each patient.
Total dose per episode=Single dose×Number of doses administered per day×Total administered days (variables defined within 365 days) (1)
Total dose per year=Σ(Total dose per episode) (2)
Lifetime cumulative dose=Σ(Total dose per year) (3)
Lifetime administered duration=Σ(Administered days since first prescription (variables defined within 365 days)) (4)
6. Outcomes
The outcomes identified were cancer sites with sufficient evidence from the IARC monographs, as detailed in S2 Table [1,5,6]. The cancers identified by the IARC to be associated with carcinogenicity in humans due to group 1 pharmaceuticals include skin, hematologic, and urinary bladder cancer. For the purpose of this study, skin cancer encompassed all forms of melanoma, non-melanoma, and other types of skin malignancies, while hematologic cancer covered all types of hematologic malignancies, including lymphoma, leukemia, and myeloma as outcomes. The ICD-10 codes were used to determine cancer outcomes. Specifically, skin cancers were identified using ICD-10 codes C43-C44, hematologic cancers with codes C81-C96, and urinary bladder cancer with code C67. Since outcome cancers included in this study have slow onset and long latent period [16], we have excluded cases with cancer incidence within a year from the start of follow-up.
The outcome for each patient was operationally defined as the first recorded ICD-10 code of outcome cancers after the initiation of group 1 pharmaceuticals use, along with records of the differential copayment codes of cancer, V193 and V194 [17].
7. Statistical analysis
Patients with clinical indications were categorized into ever-users and non-users of each group 1 pharmaceuticals. Chi-square and t tests were used to compare the general characteristics of each group. Median and interquartile range of total follow-up years, and mean and standard deviation of age in each group were summarized. In addition, frequencies of sex, family history of cancer, risk for skin cancer, hematologic cancer, bladder cancer, and cigarette smoking in each group were summarized. Cigarette smoking was categorized into never smoking, past smoking, and currently smoking [18].
Number of cancer incidence was divided by total person-years to estimate incidence rate (IR) of cancer in each group, and Cox proportional hazards model was used to estimate hazard ratio (HR) and 95% confidence interval (CI) of HR of cancer in the ever-users.
The IR and HR of cancer in the ever-users compared to the non-users was estimated. Furthermore, lifetime cumulative dose, lifetime administered duration, and single dose among ever-users were categorized into low and high subgroups by median of the ever-users, and the IR and HR of cancer in each of subgroups compared to the non-users was estimated. The IR and HR of cancer in the high subgroup was also estimated in comparison to the low subgroup.
In addition, lifetime cumulative dose, lifetime administered duration, and single dose among ever-users were categorized into low, intermediate, and high subgroups by tertile of the ever-users, and the IR and HR of cancer in each of subgroups compared to the non-users was estimated. The IR and HR of cancer in the intermediate subgroup and the high subgroup was also estimated in comparison to the low subgroup [12].
The crude HR was estimated, followed by the estimation of HR adjusted for age, sex, family history of cancer, cigarette smoking, and past history of risk factors related to skin cancer. Past history of risk factors for skin cancer were human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS), malignant melanoma and neoplasms of the skin, and the use of other group 1 pharmaceuticals associated with skin cancer [19].
Past history of risk factors for hematologic cancer were hepatitis C virus infection, HIV/AIDS, bone marrow disorders, polycythemia vera, myelodysplastic syndromes, and other malignant neoplasms related to lymph, hematopoietic stem cells, radiation therapy, and the use of other group 1 pharmaceuticals associated with hematologic cancer [20].
Past history of risk factors for bladder cancer were urinary tract infection diseases, kidney and bladder stones, urothelial carcinoma, and radiation therapy (S3 Table) [21].
A p-value less than 0.05 was considered statistically significant. All statistical analyses were performed with SAS ver. 9.4 and SAS Enterprise Guide (SAS Institute Inc.).
1. Study subjects
For group 1 immunosuppressive drugs, a total of 789,668 and 3,780,245 patients were included after excluding 1,228 and 1,439 patients among patients with clinical indications of azathioprine and cyclosporine, respectively. For group 1 antineoplastic drugs, a total of 112,605/198,299/967,097/31,417 and 149,439 patients were included after excluding 1,322/425/2,119/1,914 and 134 patients among patients with clinical indications of chlorambucil, melphalan, cyclophosphamide, busulfan, and thiotepa, respectively. For group 1 skin ointments, a total of 1,610,848 patients were included after excluding 104 patients among patients with clinical indications of methoxsalen (Fig. 1).
2. General characteristics
For group 1 immunosuppressive drugs, a total of 5,635 and 19,354 ever-users of azathioprine and cyclosporine were identified in patients with clinical indications of azathioprine and cyclosporine, respectively. The median follow-up periods were 10.0 and 14.3 years, and mean age were 39.5 and 36.7 in ever-users of azathioprine and cyclosporine, respectively. For group 1 antineoplastic drugs, a total of 108, 476, and 2,907 ever-users of chlorambucil, melphalan, and cyclophosphamide were identified in patients with clinical indications of chlorambucil, melphalan, and cyclophosphamide. The median follow-up periods were 0.5, 0.4, and 11.0 years, and mean age were 60.5, 70.6, and 50.1 in ever-users of chlorambucil, melphalan, and cyclophosphamide, respectively. For group 1 skin ointments, a total of 2,470 ever-users of methoxsalen were identified in patients with clinical indications of methoxsalen. The median follow-up period was 15.0 years, and mean age was 39.8 in ever-users of methoxsalen (Tables 1 and 2).
Since prescription records were not found among patients with clinical indications of busulfan and thiotepa, general characteristics of only non-users of busulfan and thiotepa were presented in S4 Table.
3. Cancer risk in association with ever-use
The fully adjusted HR of skin cancer in ever-users of azathioprine, cyclosporine, and methoxsalen were 4.63 (95% CI, 2.91 to 7.39), 2.30 (95% CI, 1.79 to 2.95), and 2.32 (95% CI, 1.36 to 3.95), respectively. The fully adjusted HR of hematologic cancer in ever-users of azathioprine, cyclosporine, cyclophosphamide, chlorambucil, and melphalan were 3.15 (95% CI, 2.41 to 4.13), 2.96 (95% CI, 2.59 to 3.40), 3.83 (95% CI, 3.20 to 4.59), 3.51 (95% CI, 2.53 to 4.87), and 16.31 (95% CI, 13.41 to 19.85), respectively. The fully adjusted HR of bladder cancer in ever-users of cyclophosphamide was 2.69 (95% CI, 1.92 to 3.78) (Table 3).
4. Cancer risk in association with lifetime cumulative dose
The fully adjusted HR of skin cancer in association with lifetime cumulative dose in low-dose and high-dose subgroups were 1.91 (95% CI, 0.71 to 5.10) and 9.74 (95% CI, 6.67 to 14.23) for azathioprine, 1.90 (95% CI, 1.28 to 2.84) and 3.50 (95% CI, 2.81 to 4.35) for cyclosporine, and 1.35 (95% CI, 0.43 to 4.18) and 3.08 (95% CI, 1.70 to 5.58) for methoxsalen, respectively. The fully adjusted HR of hematologic cancer in association with lifetime cumulative dose in low-dose and high-dose subgroups were 1.57 (95% CI, 1.11 to 2.22) and 3.39 (95% CI, 2.76 to 4.17) for azathioprine, 1.33 (95% CI, 0.99 to 1.80) and 6.75 (95% CI, 6.12 to 7.46) for cyclosporine, 0.85 (95% CI, 0.56 to 1.29) and 8.90 (95% CI, 7.96 to 9.96) for cyclophosphamide, 3.91 (95% CI, 2.87 to 5.32) and 2.88 (95% CI, 2.17 to 3.81) for chlorambucil, and 21.17 (95% CI, 18.46 to 24.27) and 15.75 (95% CI, 13.80 to 17.98) for melphalan, respectively. The fully adjusted HR of bladder cancer in association with lifetime cumulative dose in low-dose and high-dose subgroups were 2.56 (95% CI, 1.41 to 4.63) and 2.75 (95% CI, 1.83 to 4.15) for cyclophosphamide (Table 4, Fig. 2).
5. Cancer risk in association with lifetime administered duration
The fully adjusted HR of skin cancer in association with lifetime administered duration in low-duration and high-duration subgroups were 1.95 (95% CI, 0.73 to 5.22) and 9.58 (95% CI, 6.56 to 13.99) for azathioprine, 1.95 (95% CI, 1.32 to 2.89) and 3.48 (95% CI, 2.79 to 4.33) for cyclosporine, and 1.33 (95% CI, 0.43 to 4.13) and 3.10 (95% CI, 1.71 to 5.62) for methoxsalen, respectively. The fully adjusted HR of hematologic cancer in association with lifetime administered duration in low-duration and high-duration subgroups were 1.75 (95% CI, 1.25 to 2.44) and 3.23 (95% CI, 2.62 to 3.99) for azathioprine, 1.46 (95% CI, 1.10 to 1.94) and 6.68 (95% CI, 6.05 to 7.38) for cyclosporine, 1.29 (95% CI, 0.91 to 1.81) and 8.53 (95% CI, 7.60 to 9.56) for cyclophosphamide, 4.08 (95% CI, 2.99 to 5.55) and 2.81 (95% CI, 2.12 to 3.72) for chlorambucil, and 20.53 (95% CI, 17.91 to 23.53) and 16.09 (95% CI, 14.09 to 18.37) for melphalan, respectively. The fully adjusted HR of bladder cancer in association with lifetime administered duration in low-duration and high-duration subgroups were 3.18 (95% CI, 1.80 to 5.61) and 2.48 (95% CI, 1.63 to 3.77) for cyclophosphamide (Table 5, Fig. 2).
6. Cancer risk in association with single dose
The fully adjusted HR of skin cancer in association with single dose in low-dose and high-dose subgroups were 5.50 (95% CI, 3.30 to 9.17) and 7.61 (95% CI, 4.69 to 12.34) for azathioprine, 2.63 (95% CI, 1.99 to 3.49) and 3.26 (95% CI, 2.51 to 4.23) for cyclosporine, and 3.44 (95% CI, 1.90 to 6.23) and 1.15 (95% CI, 0.37 to 3.58) for methoxsalen, respectively. The fully adjusted HR of hematologic cancer in association with single dose in low-dose and high-dose subgroups were 2.09 (95% CI, 1.58 to 2.76) and 3.13 (95% CI, 2.49 to 3.95) for azathioprine, 2.51 (95% CI, 2.07 to 3.04) and 6.82 (95% CI, 6.12 to 7.59) for cyclosporine, 1.89 (95% CI, 1.44 to 2.48) and 8.45 (95% CI, 7.51 to 9.51) for cyclophosphamide, 3.06 (95% CI, 2.27 to 4.14) and 3.49 (95% CI, 2.62 to 4.64) for chlorambucil, and 18.28 (95% CI, 15.97 to 20.93) and 17.69 (95% CI, 15.48 to 20.22) for melphalan, respectively. The fully adjusted HR of bladder cancer in association with single dose in low-dose and high-dose subgroups were 3.98 (95% CI, 2.56 to 6.18) and 1.84 (95% CI, 1.09 to 3.10) for cyclophosphamide (Table 6, Fig. 2).
Cancer risks in association with high dose and duration of group 1 pharmaceuticals are presented in S5-S13 Tables.
This study identified the associations between group 1 pharmaceuticals and risk of related cancers. Group 1 immunosuppressive drugs were found to be associated with increased risk of skin and hematologic cancer and group 1 antineoplastic drugs were found to be associated with increased risk of hematologic and bladder cancer. In addition, significant dose-response relationship between lifetime cumulative dose, lifetime administered duration, single dose of group 1 pharmaceuticals, and risk of related cancers were identified.
The risk of cancer among ever-users of group 1 pharmaceuticals and the risk of cancer in association with lifetime cumulative dose, lifetime administered duration, and single dose are mostly consistent with previous observational studies included in monographs by IARC working group [1,2,5,6]. In addition, a systematic review and meta-analysis study, and retrospective studies using nationwide population-based database on association between group 1 pharmaceuticals and related cancers presented similar trends of results [12,20,22,23]. However, the risk of hematologic cancer in association with melphalan and chlorambucil was found to be much higher than the risk from previous studies [24-27].
The overall increasing trend in risk of cancer according to median and tertile of ever-users were identified in lifetime cumulative dose, lifetime administered duration, and single dose of group 1 pharmaceuticals. However, in the case of chlorambucil and melphalan, a substantial decrease in risk of hematologic cancer in association with lifetime cumulative dose and lifetime administered duration was observed in the high dose and duration subgroup compared to the low dose and duration subgroup. In addition, in the case of cyclophosphamide, a decrease in risk of bladder cancer in association with single dose was observed in the high-dose subgroup compared to the low-dose subgroup.
These kinds of trends were also observed when ever-users were classified into tertile. In the case of chlorambucil and melphalan, the risk of hematologic cancer in association with lifetime cumulative dose, lifetime administered duration, and single dose in intermediate dose and duration subgroup was higher than the risk of cancer in high dose and duration subgroup. In addition, in the case of azathioprine, the risk of hematologic cancer in association with lifetime administered duration and single dose were decreasing in intermediate dose and duration subgroup compared to low dose and duration subgroup, but increasing again in high dose and duration subgroup.
In the case of azathioprine, Ingvar et al. [28] reported similar trends to our result, with the risk of cutaneous squamous cell carcinoma in association with accumulated dose showing higher risk in intermediate dose subgroup than high dose subgroup. This may be explained by various factors including genetic heterogeneity and the depletion of susceptible at lower doses [29].
Immunosuppressive drugs, by inducing long-term systemic immunosuppression, may facilitate the infection of oncogenic viruses or opportunistic infections, thereby initiating carcinogenesis and further promoting the carcinogenic process initiated due to the weakened immune system, thus potentially increasing the risk of specific cancers [30]. Among group 1 pharmaceuticals, five antineoplastic drugs are all alkylating agents, which damage both DNA and RNA or impair DNA function. By themselves, they can induce a decrease in leukocytes and other components, lowering immune resistance and increasing the risk of infection. This, in turn, promotes the opportunistic infections related to oncogenic viruses [31]. The compromised immune system and the reduced recovery of normal blood cells due to chemotherapy, in combination with complex regimens and radiotherapy, are likely to significantly increase the risk of cancer. Immunosuppressive drugs and group 1 pharmaceuticals, being long-term medications with cumulative dosage accumulation, could lead to an increased risk of cancer.
Chlorambucil and melphalan are both bifunctional alkylating agents that create intrastrand and interstrand DNA cross-links, acting as potent non-specific anticancer agents. Among them, melphalan is known not to exhibit a linear relationship with cancer cell apoptosis in response to dosage increases in its pharmacokinetic properties. Instead, the response to the drug is initially delayed, followed by a reaction, and then reaches a point where no further response is observed, following a logarithmic pattern. However, it cannot be assumed that no dose-response relationship will be observed for secondary hematologic cancer based on melphalan’s pharmacokinetic properties. Melphalan administration varies between intravenous and oral routes depending on the disease [32].
In a study published by Travis et al. [27], the risk of hematologic cancer in female ovarian cancer patients was reported as relative risk (RR) 4.0 for platinum, RR 8.1 for other platinum-based drugs, while for melphalan, it was significantly higher at RR 20.8. For intravenous melphalan, the RR was 22.9, compared to an RR of 9.0 for oral melphalan, indicating a significant difference depending on the method of administration [27]. The average intravenous dose was 140 mg, while for oral administration, it was 425 mg, confirming that higher average doses and longer durations of administration progressively increase the risk.
The dose-response relationship for chlorambucil appears unclear, with observations indicating a decrease in risk at higher doses. This could likely be attributed to the method of use for chlorambucil. For example, in treating diseases such as chronic lymphocytic leukemia, chlorambucil is initially administered at a starting dose, then escalated to the maximum dose. Once the maximum response is achieved, treatment is discontinued rather than maintained at a lower dose [33]. As chlorambucil is primarily administered orally, the issue of dosage confusion due to the method of administration is expected to be minimal.
When clinical indications improve or exacerbate, the therapeutic dosage is accordingly adjusted. However, due to the properties of NHIS data, it is challenging to detect the improvement or exacerbation of clinical indications [7]. In addition, understanding changes in treatment dosages necessitates a preliminary grasp of the clinical progression. Hence, it is necessary to utilize a patient cohort built from real clinical data to identify the confounding effects on the association between drug and cancer caused by multiple factors, including the improvement or exacerbation of clinical indications and the subsequent adjustments in treatment dosages.
This study has several limitations and strengths. Study was based on the NHIS database to establish a retrospective customized cohort. In the process of acquiring customized data, the variables on pharmaceuticals were classified as sensitive information, making it impossible to apply for the entire group 1 pharmaceuticals data. Application was only possible in parts where the major ingredient code was available. Therefore, it was not feasible to analyze all group 1 pharmaceuticals within the cohort.
Since study population from NHIS database could only be extracted based on the diagnostic definition of diseases (by type of diseases and year of diagnosis) in the first hand, general population without these clinical indications who might have used group I pharmaceutical could not be considered in this study. Hence, to capture the most extent of group 1 pharmaceutical use as possible, we have thoroughly reviewed IARC monographs that best summarizes the characteristics of the representative population group using group 1 pharmaceuticals. In addition, systematic reviews on association between cancer and each group I pharmaceutical was conducted to comprise as much population with various clinical indications as possible in this study.
Comparison of risk was only made between users and non-users in combined clinical indications, and not in individual clinical indications. Since baseline risk between each clinical indications differ, this may introduce selection bias when comparing the risk between non-users and users of each group 1 pharmaceutical. Further analysis and comparison of risk between patients with the same clinical indications is required for reducing the possibility of confounding effects caused by difference in severity of clinical indications and selection bias caused when selecting study participants from different population.
According to previous retrospective cohort studies on group 1 pharmaceutical and associated cancers, the use of immunosuppressive drugs such as azathioprine and cyclosporine required at least 3 years of latency period as follow-up for incidence of skin cancer in lung and heart transplant recipients [14,34], and required at least 3 years of latency period as follow-up for incidence of skin and hematologic cancers in patients with autoimmune disease, such as ulcerative colitis, rheumatic arthritis, and psoriasis [11,35]. The follow-up periods for azathioprine, cyclosporine, cyclophosphamide [36], and methoxsalen [37] were long enough to observe incidence of skin and hematologic cancers (10.0 years, 14.3 years, 11.0 years, and 15.0 years, respectively) in this study.
However, since the use of antineoplastic drugs such as chlorambucil required at least more than a year of latency period as follow-up for incidence of hematologic cancer in patients with polycythemia vera [38], and melphalan required at least more than a year of latency period as follow-up for incidence of hematologic cancer in breast and ovarian cancer patients [26,27], the median follow-up periods for chlorambucil and melphalan were insufficient to adequately observe the incidence of hematologic cancers. (0.5 years and 0.4 years, respectively) in this study.
For each clinical indication, and even within the same clinical indication, the actual single dose amount and unit (mg/kg, mg, mg/m2, μg/kg, mg/mL) vary depending on the patient’s age, weight, clinical condition, and course, as well as the route of administration (e.g., oral or intravenous). While the dataset provides the quantity of a single dose and the number of daily doses, it lacks information on the standard dose and its unit, leading to possible errors in estimating the total cumulative dose. However, since all exposure variables are extracted from the same data source and the measurement unit of exposure variables are uniform, this reduces the possibility of introducing nondifferential bias [39,40].
Based on the NHIS database for a retrospective cohort study, selection bias becomes a concern if records or information regarding outcome variables are more extensively preserved or identified in either ever-users or non-users. In this study, patients with clinical indications were initially extracted, followed by the extraction of ever-users to establish the retrospective cohort. Since all subjects were monitored in the same manner for cancer incidence, the possibility of selection bias could be minimized to the greatest extent possible.
Group 1 pharmaceuticals are primarily used for rare and intractable clinical indications such as cancer and autoimmune diseases, where the patients often cannot discontinue use [41]. Above all, a surveillance system for these subjects is necessary. Discussions are also needed on various monitoring approaches, including a voluntary surveillance system based on post-marketing reports, a patient registry system for those who must use the medication due to their clinical indications, and an automated big data surveillance system utilizing the NHIS database [42,43]. The surveillance system should integrate monitoring using big data (e.g., data mining techniques), surveillance through information systems, epidemiological investigations, and causality assessments after the detection of risk signals, followed by subsequent actions and feedback [44-46].
For patients continuing chemotherapy and those who must persistently use immunosuppressants for solid organ transplantation, who are not using WHO-listed essential pharmaceuticals, it is necessary within clinical departments to discuss and establish protocols that ensure patients are fully informed about the associated cancer risk. Patients should receive a comprehensive explanation of the clinical significance and be fully informed about related information, allowing both the physician and patient to make informed choices regarding the use of group 1 pharmaceuticals regimens and alternative interventions. In addition, discussions should also be held regarding the provision of financial support for regular cancer screenings for the patients, as well as the establishment of a compensation system following health screenings [47].
In conclusion, while group 1 pharmaceuticals have already been classified as carcinogenic by the IARC, this study has also confirmed that ever-users of these pharmaceuticals have a higher risk of cancer compared to the non-users in Korea. Given that these pharmaceuticals are WHO-listed essential pharmaceuticals and may be difficult to replace, it is imperative to continuously analyze risks and discuss the surveillance system. An automated surveillance system should be established from the data source to continuously provide feedback and discussions, ultimately reflecting these insights into actual policy decisions.
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
crt-2024-1201_S1_Table.pdf
crt-2024-1201_S2_Table.pdf
crt-2024-1201_S3_Table.pdf
crt-2024-1201_S4_Table.pdf
crt-2024-1201_S5_Table.pdf
crt-2024-1201_S6_Table.pdf
crt-2024-1201_S7_Table.pdf
crt-2024-1201_S8_Table.pdf
crt-2024-1201_S9_Table.pdf
crt-2024-1201_S10_Table.pdf
crt-2024-1201_S11_Table.pdf
crt-2024-1201_S12_Table.pdf
crt-2024-1201_S13_Table.pdf

Ethical Statement

The Institutional Review Board of Seoul National University Hospital approved this study (IRB No. E-2102-155-1199). The Institutional Review Board of National Evidence-based healthcare Collaborating Agency (NECA) approved this study (IRB No. NECAIRB21-003). The Institutional Review Board of Seoul National University Hospital waived the requirement for informed consent as this study did not involve personally identifiable information of research participants and did not use human-derived biological materials.

Author Contributions

Conceived and designed the analysis: Lim W, Lee NR, Shin HG, Yu SY, Park SK.

Collected the data: Lim W, Lee NR, Shin HG, Yu SY, Park SK.

Contributed data or analysis tools: Lee NR, Shin HG, Yu SY, Park SK.

Performed the analysis: Lim W.

Wrote the paper: Lim W.

Conflicts of Interest

Conflict of interest relevant to this article was not reported.

Funding

This study was supported by National Evidence-based healthcare Collaborating Agency (NA21-003).

Fig. 1.
Flow chart of selecting study subjects in National Health Insurance Service database.
crt-2024-1201f1.jpg
Fig. 2.
Hazard ratio (95% confidence interval [CI]) for users relative to non-users of group 1 pharmaceuticals.
crt-2024-1201f2.jpg
Table 1.
General characteristics of patients with indication of azathioprine, cyclosporine, and cyclophosphamide
Azathioprine
 Cyclosporine
 Cyclophosphamide
Non-users (n=784,033) Ever-users (n=5,635) p-value Non-users (n=3,760,891) Ever-users (n=19,354) p-value Non-users (n=964,190) Ever-users (n=2,907) p-value
FU (yr), median (IQR) 9.0 (8.9) 10.0 (9.1) < 0.01 12.5 (7.8) 14.3 (6.6) < 0.01 8.9 (8.5) 11.0 (9.9) < 0.01
Age (yr), mean±SD 50.6±18.4 39.5±16.8 < 0.01 32.8±22.6 36.7±19.0 < 0.01 50.6±18.5 50.1±15.8 0.15
Sex
 Men 287,355 (36.7) 2,599 (46.1) < 0.01 1,734,123 (46.1) 9,917 (51.2) < 0.01 360,734 (37.4) 689 (23.7) < 0.01
 Women 496,678 (63.3) 3,036 (53.9) 2,026,768 (53.9) 9,437 (48.8) 603,456 (62.6) 2,218 (76.3)
Family history of cancer
 No 645,890 (82.4) 4,639 (82.3) 0.85 3,154,367 (83.9) 16,250 (84.0) 0.74 787,999 (81.7) 2,341 (80.5) 0.10
 Yes 138,143 (17.6) 996 (17.7) 606,524 (16.1) 3,104 (16.0) 176,191 (18.3) 566 (19.5)
Risk for skin cancera)
 No 783,752 (100) 5,634 (100) 0.73 3,760,450 (100) 19,353 (100) 0.73 - -
 Yes 281 (0.0) 1 (0.0) 441 (0.0) 1 (0.0) - -
Risk for hematologic cancerb)
 No 782,801 (99.8) 5,621 (99.8) 0.09 3,759,554 (100) 19,318 (99.8) < 0.001 963,568 (99.9) 2,903 (99.9) 0.12
 Yes 1,232 (0.2) 14 (0.2) 1,337 (0.0) 36 (0.2) 622 (0.1) 4 (0.1)
Risk for bladder cancerc)
 No - - - - 940,207 (97.5) 2,825 (97.2) 0.25
 Yes - - - - 23,983 (2.5) 82 (2.8)
Cigarette smoking
 Never 577,297 (73.6) 4,033 (71.6) < 0.01 2,563,819 (68.2) 12,348 (63.8) < 0.01 709,285 (73.6) 2,329 (80.1) < 0.01
 Past 94,473 (12.1) 790 (14.0) 449,699 (12.0) 2,396 (12.4) 116,723 (12.1) 301 (10.4)
 Current 112,263 (14.3) 812 (14.4) 747,373 (19.9) 4,610 (23.8) 138,182 (14.3) 277 (9.5)

Values are presented as number (%) unless otherwise indicated. FU, follow up; IQR, interquartile range; SD, standard deviation.

a) The risk factors for skin cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS) or skin cancer in past disease codes,

b) The risk factors for hematologic cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: hepatitis C virus infection, HIV/AIDS, bone marrow disorder or hematologic diseases (polycythemia), myelodysplastic disease, other and unspecified neoplasm of lymphoid, hematopoietic and related tissues or radiotherapy in past disease codes,

c) The risk factors for bladder cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: urinary infections, kidney and bladder stones, urothelial carcinomas, or radiotherapy in past disease codes.

Table 2.
General characteristics of patients with indication of methoxsalen, chlorambucil, and melphalan
Methoxsalen
 Chlorambucil
 Melphalan
Non-users (n=1,608,378) Ever-users (n=2,470) p-value Non-users (n=112,497) Ever-users (n=108) p-value Non-users (n=197,823) Ever-users (n=476) p-value
FU (yr), median (IQR) 10.7 (8.8) 15.0 (3.9) < 0.01 8.0 (9.4) 0.5 (0.7) < 0.01 7.0 (8.0) 0.4 (0.6) < 0.01
Age (yr), mean±SD 25.7±24.4 39.8±19.2 < 0.01 47.9±19.3 60.5±20.2 < 0.01 47.8±18.0 70.6±8.5 < 0.01
Sex
 Men 748,418 (46.5) 1,142 (46.2) 0.77 33,789 (30.0) 67 (62.0) < 0.01 62,078 (31.4) 250 (52.5) < 0.01
 Women 859,960 (53.5) 1,328 (53.8) 78,708 (70.0) 41 (38.0) 135,745 (68.6) 226 (47.5)
Family history of cancer
 No 1,338,436 (83.2) 2,099 (85.0) 0.02 90,870 (80.8) 87 (80.6) 0.95 154,952 (78.3) 378 (79.4) 0.57
 Yes 269,942 (16.8) 371 (15.0) 21,627 (19.2) 21 (19.4) 42,871 (21.7) 98 (20.6)
Risk for skin cancera)
 No 1,608,027 (100) 2,470 (100) > 0.99 - - - -
 Yes 351 0 - - - -
Risk for hematologic cancerb)
 No - - 111,086 (98.8) 88 (81.5) < 0.001 197,349 (99.8) 473 (99.4) 0.11
 Yes - - 1,411 (1.3) 20 (18.5) 474 (0.2) 3 (0.6)
Cigarette smoking
 Never 1,063,464 (66.1) 1,648 (66.7) 0.01 87,201 (77.5) 68 (63.0) 0.001 152,600 (77.1) 352 (74.0) 0.06
 Past 222,181 (13.8) 292 (11.8) 11,793 (10.5) 20 (18.5) 19,489 (9.9) 62 (13.0)
 Current 322,733 (20.1) 530 (21.5) 13,503 (12.0) 20 (18.5) 25,734 (13.0) 62 (13.0)

Values are presented as number (%) unless otherwise indicated. FU, follow up; IQR, interquartile range; SD, standard deviation.

a) The risk factors for skin cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS) or skin cancer in past disease codes,

b) The risk factors for hematologic cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: hepatitis C virus infection, HIV/AIDS, bone marrow disorder or hematologic diseases (polycythemia), myelodysplastic disease, other and unspecified neoplasm of lymphoid, hematopoietic and related tissues or radiotherapy in past disease codes.

Table 3.
HR (95% CI) for ever-users relative to non-users of group 1 pharmaceuticals
Outcome Exposre Cohort
 No. of cancers IR IRR HRa) (95% CI) HRb) (95% CI) p-value
No. Person-year
Skin cancer Azathioprine
 Non-user 783,900 6,652,263 1,220 0.02 1.00 1.00 < 0.01
 Ever-user 5,635 52,820 32 0.06 3.3 3.01 (2.12-4.27) 4.63 (2.91-7.39)
Cyclosporine
 Non-user 3,760,754 40,290,898 5,766 0.01 1.00 1.00 < 0.01
 Ever-user 19,354 236,712 106 0.04 3.1 2.73 (2.25-3.30) 2.30 (1.79-2.95)
Methoxsalen
 Non-user 1,608,315 14,985,310 1,459 0.01 1.00 1.00 < 0.01
 Ever-user 2,470 33,212 14 0.04 4.4 4.05 (2.39-6.86) 2.32 (1.36-3.95)
Hematologic cancer Azathioprine
 Non-user 783,892 6,651,255 4,977 0.07 1.00 1.00 < 0.01
 Ever-user 5,633 52,768 124 0.24 3.1 2.85 (2.39-3.41) 3.15 (2.41-4.13)
Cyclosporine
 Non-user 3,760,732 40,289,651 11,423 0.03 1.00 1.00 < 0.01
 Ever-user 19,353 236,684 451 0.19 6.7 5.93 (5.40-6.51) 2.96 (2.59-3.40)
Cyclophosphamide
 Non-user 963,445 7,849,680 19,042 0.24 1.00 1.00 < 0.01
 Ever-user 2,907 28,619 331 1.16 4.8 5.31 (4.77-5.93) 3.83 (3.20-4.59)
Chlorambucil
 Non-user 112,359 889,804 9,538 1.07 1.00 1.00 < 0.01
 Ever-user 108 720 91 12.64 11.8 8.68 (7.06-10.67) 3.51 (2.53-4.87)
Melphalan
 Non-user 197,701 1,415,357 5,280 0.37 1.00 1.00 < 0.01
 Ever-user 474 1,632 461 28.25 75.7 49.53 (44.97-54.55) 16.31 (13.41-19.85)
Bladder cancer Cyclophosphamide
 Non-user 963,498 7,854,646 4,092 0.05 1.00 1.00 < 0.01
 Ever-user 2,907 28,785 34 0.12 2.3 2.50 (1.79-3.51) 2.69 (1.92-3.78)

All incidence rates were presented as percentage. CI, confidence interval; HR, hazard ratio; IR, incidence rate; IRR, incidence rate ratio.

a) Crude HR,

b) Adjusted for age, sex, family history of cancer, smoking and past history of risk factors for each cancer outcome (presented in S3 Table).

Table 4.
HR (95% CI) for lifetime cumulative dose (median) relative to non-users of group 1 pharmaceuticals
Outcome Exposre Cohort
 No. of cancers IR IRR HRa) (95% CI) HRb) (95% CI) p-value
No. Person-year
Skin cancer Azathioprine
 Non-users 783,900 6,652,263 1,220 0.02 1.00 1.00 < 0.01
 < 647 mg 2,817 22,438 4 0.02 1.0 0.95 (0.36-2.53) 1.91 (0.71-5.10)
 > 647 mg 2,818 30,382 28 0.09 5.0 4.45 (3.06-6.47) 9.74 (6.67-14.23)
Cyclosporine
 Non-users 3,760,754 40,290,898 5,766 0.01 1.00 1.00 < 0.01
 < 122 mg 9,677 114,674 24 0.02 1.5 1.31 (0.88-1.96) 1.90 (1.28-2.84)
 > 122 mg 9,677 122,038 82 0.07 4.7 3.95 (3.17-4.91) 3.50 (2.81-4.35)
Methoxsalen
 Non-users 1,608,315 14,985,310 1,459 0.01 1.00 1.00 < 0.01
 < 16 mg 1,226 15,992 3 0.02 1.9 1.85 (0.60-5.75) 1.35 (0.43-4.18)
 > 16 mg 1,244 17,220 11 0.06 6.6 5.98 (3.30-10.82) 3.08 (1.70-5.58)
Hematologic cancer Azathioprine
 Non-users 783,892 6,651,255 4,977 0.07 1.00 1.00 < 0.01
 < 644 mg 2,816 22,416 32 0.14 1.9 1.85 (1.30-2.61) 1.57 (1.11-2.22)
 > 644 mg 2,817 30,352 92 0.30 4.1 3.85 (3.14-4.74) 3.39 (2.76-4.17)
Cyclosporine
 Non-users 3,760,732 40,289,651 11,423 0.03 1.00 1.00 < 0.01
 < 114 mg 9,670 114,593 43 0.04 1.3 1.22 (0.90-1.65) 1.33 (0.99-1.80)
 > 114 mg 9,683 122,091 408 0.33 11.8 10.31 (9.34-11.38) 6.75 (6.12-7.46)
Cyclophosphamide
 Non-users 963,445 7,849,680 19,042 0.24 1.00 1.00 < 0.01
 < 210 mg 1,446 15,220 22 0.14 0.6 0.59 (0.39-0.89) 0.85 (0.56-1.29)
 > 210 mg 1,461 13,399 309 2.31 9.5 9.25 (8.27-10.35) 8.90 (7.96-9.96)
Chlorambucil
 Non-users 112,359 889,804 9,538 1.07 1.00 1.00 < 0.01
 < 341 mg 54 392 41 10.46 9.8 9.24 (6.80-12.56) 3.91 (2.87-5.32)
 > 341 mg 54 328 50 15.24 14.2 13.66 (10.34-18.04) 2.88 (2.17-3.81)
Melphalan
 Non-users 197,701 1,415,357 5,280 0.37 1.00 1.00 < 0.01
 < 243 mg 237 649 224 34.51 92.5 64.85 (56.68-74.18) 21.17 (18.46-24.27)
 > 243 mg 237 983 237 24.11 64.6 52.19 (45.80-59.47) 15.75 (13.80-17.98)
Bladder cancer Cyclophosphamide
 Non-users 963,498 7,854,646 4,092 0.05 1.00 1.00 < 0.01
 < 210 mg 1,448 15,232 11 0.07 1.4 1.31 (0.73-2.38) 2.56 (1.41-4.63)
 > 210 mg 1,459 13,553 23 0.17 3.3 3.09 (2.05-4.66) 2.75 (1.83-4.15)

All incidence rates were presented as percentage. CI, confidence interval; HR, hazard ratio; IR, incidence rate; IRR, incidence rate ratio.

a) Crude HR,

b) Adjusted for age, sex, family history of cancer, smoking and past history of risk factors for each cancer outcome (presented in S3 Table).

Table 5.
HR (95% CI) for lifetime administered duration (median) relative to non-users of group 1 pharmaceuticals
Outcome Exposre Cohort
 No. of cancers IR IRR HRa) (95% CI) HRb) (95% CI) p-value
No. Person-year
Skin cancer Azathioprine
 Non-users 783,900 6,652,263 1,220 0.02 1.00 1.00 < 0.01
 < 322 days 2,817 22,203 4 0.02 1.0 0.96 (0.36-2.57) 1.95 (0.73-5.22)
 > 322 days 2,818 30,617 28 0.09 5.0 4.41 (3.03-6.41) 9.58 (6.56-13.99)
Cyclosporine
 Non-users 3,760,754 40,290,898 5,766 0.01 1.00 1.00 < 0.01
 < 49 days 9,673 115,064 25 0.02 1.5 1.36 (0.92-2.01) 1.95 (1.32-2.89)
 > 49 days 9,681 121,648 81 0.07 4.7 3.93 (3.16-4.90) 3.48 (2.79-4.33)
Methoxsalen
 Non-users 1,608,315 14,985,310 1,459 0.01 1.00 1.00 < 0.01
 < 4 days 1,224 15,931 3 0.02 1.9 1.86 (0.60-5.79) 1.33 (0.43-4.13)
 > 4 days 1,246 17,281 11 0.06 6.6 5.95 (3.29-10.77) 3.10 (1.71-5.62)
Hematologic cancer Azathioprine
 Non-users 783,892 6,651,255 4,977 0.07 1.00 1.00 < 0.01
 < 325 days 2,816 22,181 35 0.16 2.1 2.04 (1.46-2.85) 1.75 (1.25-2.44)
 > 325 days 2,817 30,587 89 0.29 3.9 3.70 (3.00-4.56) 3.23 (2.62-3.99)
Cyclosporine
 Non-users 3,760,732 40,289,651 11,423 0.03 1.00 1.00 < 0.01
 < 49 days 9,675 115,094 47 0.04 1.4 1.32 (0.99-1.76) 1.46 (1.10-1.94)
 > 49 days 9,678 121,590 404 0.33 11.7 10.30 (9.33-11.37) 6.68 (6.05-7.38)
Cyclophosphamide
 Non-users 963,445 7,849,680 19,042 0.24 1.00 1.00 < 0.01
 < 70 days 1,453 15,174 33 0.22 0.9 0.88 (0.63-1.24) 1.29 (0.91-1.81)
 > 70 days 1,454 13,445 298 2.22 9.1 8.90 (7.94-9.98) 8.53 (7.60-9.56)
Chlorambucil
 Non-users 112,359 889,804 9,538 1.07 1.00 1.00 < 0.01
 < 69 days 54 374 41 10.96 10.2 9.52 (7.01-12.94) 4.08 (2.99-5.55)
 > 69 days 54 346 50 14.45 13.5 13.19 (9.99-17.42) 2.81 (2.12-3.72)
Melphalan
 Non-users 197,701 1,415,357 5,280 0.37 1.00 1.00 < 0.01
 < 32 days 237 711 224 31.50 84.4 60.98 (53.31-69.76) 20.53 (17.91-23.53)
 > 32 days 237 921 237 25.73 69.0 54.82 (48.10-62.47) 16.09 (14.09-18.37)
Bladder cancer Cyclophosphamide
 Non-users 963,498 7,854,646 4,092 0.05 1.00 1.00 < 0.01
 < 70 days 1,449 15,100 12 0.08 1.5 1.45 (0.82-2.55) 3.18 (1.80-5.61)
 > 70 days 1,458 13,685 22 0.16 3.1 2.93 (1.93-4.46) 2.48 (1.63-3.77)

All incidence rates were presented as percentage. CI, confidence interval; HR, hazard ratio; IR, incidence rate; IRR, incidence rate ratio.

a) Crude HR,

b) Adjusted for age, sex, family history of cancer, smoking and past history of risk factors for each cancer outcome (presented in S3 Table).

Table 6.
HR (95% CI) for single dose (median) relative to non-users of group 1 pharmaceuticals
Outcome Exposre Cohort
 No. of cancers IR IRR HRa) (95% CI) HRb) (95% CI) p-value
No. Person-year
Skin cancer Azathioprine
 Non-users 783,900 6,652,263 1,220 0.02 1.00 1.00 < 0.01
 < 1.7 mg 2,817 27,338 15 0.05 3.0 2.71 (1.63-4.51) 5.50 (3.30-9.17)
 > 1.7 mg 2,818 25,482 17 0.07 3.6 3.41 (2.12-5.51) 7.61 (4.69-12.34)
Cyclosporine
 Non-users 3,760,754 40,290,898 5,766 0.01 1.00 1.00 < 0.01
 < 1.2 mg 9,670 117,302 49 0.04 2.9 2.54 (1.92-3.37) 2.63 (1.99-3.49)
 > 1.2 mg 9,684 119,410 57 0.05 3.3 2.89 (2.22-3.75) 3.26 (2.51-4.23)
Methoxsalen
 Non-users 1,608,315 14,985,310 1,459 0.01 1.00 1.00 < 0.01
 < 2.9 mg 1,235 17,042 11 0.06 6.7 6.09 (3.37-11.03) 3.44 (1.90-6.23)
 > 2.9 mg 1,235 16,170 3 0.02 1.9 1.82 (0.59-5.63) 1.15 (0.37-3.58)
Hematologic cancer Azathioprine
 Non-users 783,892 6,651,255 4,977 0.07 1.00 1.00 < 0.01
 < 1.7 mg 2,816 27,328 50 0.18 2.4 2.33 (1.76-3.08) 2.09 (1.58-2.76)
 > 1.7 mg 2,817 25,440 74 0.29 3.9 3.75 (2.98-4.72) 3.13 (2.49-3.95)
Cyclosporine
 Non-users 3,760,732 40,289,651 11,423 0.03 1.00 1.00 < 0.01
 < 1.2 mg 9,670 117,328 105 0.09 3.2 2.83 (2.34-3.43) 2.51 (2.07-3.04)
 > 1.2 mg 9,683 119,356 346 0.29 10.2 9.14 (8.21-10.17) 6.82 (6.12-7.59)
Cyclophosphamide
 Non-users 963,445 7,849,680 19,042 0.24 1.00 1.00 < 0.01
 < 1.0 mg 1,453 14,941 52 0.35 1.4 1.42 (1.08-1.86) 1.89 (1.44-2.48)
 > 1.0 mg 1,454 13,678 279 2.04 8.4 8.17 (7.25-9.19) 8.45 (7.51-9.51)
Chlorambucil
 Non-users 112,359 889,804 9,538 1.07 1.00 1.00 < 0.01
 < 2.8 mg 54 401 43 10.72 10.0 9.76 (7.23-13.17) 3.06 (2.27-4.14)
 > 2.8 mg 54 319 48 15.05 14.0 13.01 (9.8-17.28) 3.49 (2.62-4.64)
Melphalan
 Non-users 197,701 1,415,357 5,280 0.37 1.00 1.00 < 0.01
 < 7.5 mg 236 803 228 28.39 76.1 57.94 (50.72-66.2) 18.28 (15.97-20.93)
 > 7.5 mg 238 829 233 28.11 75.3 57.35 (50.27-65.43) 17.69 (15.48-20.22)
Bladder cancer Cyclophosphamide
 Non-users 963,498 7,854,646 4,092 0.05 1.00 1.00 < 0.01
 < 1.0 mg 1,453 14,903 20 0.13 2.6 2.46 (1.58-3.81) 3.98 (2.56-6.18)
 > 1.0 mg 1,454 13,882 14 0.10 1.9 1.83 (1.08-3.09) 1.84 (1.09-3.10)

All incidence rates were presented as percentage. CI, confidence interval; HR, hazard ratio; IR, incidence rate; IRR, incidence rate ratio.

a) Crude HR,

b) Adjusted for age, sex, family history of cancer, smoking and past history of risk factors for each cancer outcome (presented in S3 Table).

  • 1. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Pharmaceuticals. IARC Monogr Eval Carcinog Risks Hum. 2012;100:1–401.
  • 2. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Some drugs and herbal products. IARC Monogr Eval Carcinog Risks Hum. 2016;108:7–419. PubMedPMC
  • 3. IARC Monographs Vol 126 Group. Carcinogenicity of opium consumption. Carcinogenicity of opium consumption. Lancet Oncol. 2020;21:1407–8. ArticlePubMed
  • 4. Fu EL. Target trial emulation to improve causal inference from observational data: what, why, and how? J Am Soc Nephrol. 2023;34:1305–14. ArticlePubMedPMC
  • 5. Some pharmaceutical drugs. IARC Monogr Eval Carcinog Risk Chem Hum. 1980;24:1–337. PubMed
  • 6. Some antineoplastic and immunosuppressive agents. IARC Monogr Eval Carcinog Risk Chem Hum. 1981;26:1–411. PubMed
  • 7. Cheol Seong S, Kim YY, Khang YH, Heon Park J, Kang HJ, Lee H, et al. Data resource profile: the National Health Information Database of the National Health Insurance Service in South Korea. Int J Epidemiol. 2017;46:799–800. PubMedPMC
  • 8. Jo S, Jun DB, Park S. Impact of differential copayment on patient healthcare choice: evidence from South Korean National Cohort Study. BMJ Open. 2021;11:e044549ArticlePubMedPMC
  • 9. Yoo KB, Ahn HU, Park EC, Kim TH, Kim SJ, Kwon JA, et al. Impact of co-payment for outpatient utilization among Medical Aid beneficiaries in Korea: a 5-year time series study. Health Policy. 2016;120:960–6. ArticlePubMed
  • 10. Kim J, Hyun HJ, Choi EA, Yoo JW, Lee S, Jeong N, et al. Diabetes, metformin, and lung cancer: retrospective study of the Korean NHIS-HEALS database. Clin Lung Cancer. 2020;21:e551–9. ArticlePubMed
  • 11. Lange E, Blizzard L, Venn A, Francis H, Jones G. Disease-modifying anti-rheumatic drugs and non-melanoma skin cancer in inflammatory arthritis patients: a retrospective cohort study. Rheumatology (Oxford). 2016;55:1594–600. ArticlePubMed
  • 12. Tseng HW, Lu LY, Lam HC, Tsai KW, Huang WC, Shiue YL. The influence of disease-modifying anti-rheumatic drugs and corticosteroids on the association between rheumatoid arthritis and skin cancer: a nationwide retrospective case-control study in Taiwan. Clin Exp Rheumatol. 2018;36:471–8. PubMed
  • 13. Coghill AE, Johnson LG, Berg D, Resler AJ, Leca N, Madeleine MM. Immunosuppressive medications and squamous cell skin carcinoma: nested case-control study within the Skin Cancer after Organ Transplant (SCOT) Cohort. Am J Transplant. 2016;16:565–73. ArticlePubMedPMC
  • 14. Hamandi B, Fegbeutel C, Silveira FP, Verschuuren EA, Younus M, Mo J, et al. Voriconazole and squamous cell carcinoma after lung transplantation: a multicenter study. Am J Transplant. 2018;18:113–24. ArticlePubMedPDF
  • 15. Clarke CA, Robbins HA, Tatalovich Z, Lynch CF, Pawlish KS, Finch JL, et al. Risk of merkel cell carcinoma after solid organ transplantation. J Natl Cancer Inst. 2015;107:dju382.ArticlePubMedPMC
  • 16. Hou N, Wang Z, Ling Y, Hou G, Zhang B, Zhang X, et al. Radiotherapy and increased risk of second primary cancers in breast cancer survivors: an epidemiological and large cohort study. Breast. 2024;78:103824.ArticlePubMedPMC
  • 17. Kang EJ, Moon SJ, Lee K, Park IH, Kim JS, Choi YJ. Associations between missing teeth and the risk of cancer in Korea: a nationwide cohort study. BMC Oral Health. 2023;23:418.ArticlePubMedPMCPDF
  • 18. Macacu A, Autier P, Boniol M, Boyle P. Active and passive smoking and risk of breast cancer: a meta-analysis. Breast Cancer Res Treat. 2015;154:213–24. ArticlePubMedPDF
  • 19. Bhat M, Mara K, Dierkhising R, Watt KD. Immunosuppression, race, and donor-related risk factors affect de novo cancer incidence across solid organ transplant recipients. Mayo Clin Proc. 2018;93:1236–46. ArticlePubMed
  • 20. O’Regan JA, Prendeville S, McCaughan JA, Traynor C, O’Brien FJ, Ward FL, et al. Posttransplant lymphoproliferative disorders in Irish renal transplant recipients: insights from a national observational study. Transplantation. 2017;101:657–63. ArticlePubMed
  • 21. Xu Y, Wang H, Zhou S, Yu M, Wang X, Fu K, et al. Risk of second malignant neoplasms after cyclophosphamide-based chemotherapy with or without radiotherapy for non-Hodgkin lymphoma. Leuk Lymphoma. 2013;54:1396–404. ArticlePubMed
  • 22. Na R, Laaksonen MA, Grulich AE, Meagher NS, McCaughan GW, Keogh AM, et al. Iatrogenic immunosuppression and risk of non-Hodgkin lymphoma in solid organ transplantation: a population-based cohort study in Australia. Br J Haematol. 2016;174:550–62. ArticlePubMed
  • 23. Lim W, Moon S, Lee NR, Shin HG, Yu SY, Lee JE, et al. Group I pharmaceuticals of IARC and associated cancer risks: systematic review and meta-analysis. Sci Rep. 2024;14:413.ArticlePubMedPMCPDF
  • 24. Kaldor JM, Day NE, Pettersson F, Clarke EA, Pedersen D, Mehnert W, et al. Leukemia following chemotherapy for ovarian cancer. N Engl J Med. 1990;322:1–6. ArticlePubMed
  • 25. Travis LB, Curtis RE, Stovall M, Holowaty EJ, van Leeuwen FE, Glimelius B, et al. Risk of leukemia following treatment for non-Hodgkin’s lymphoma. J Natl Cancer Inst. 1994;86:1450–7. ArticlePubMed
  • 26. Curtis RE, Boice JD Jr, Stovall M, Bernstein L, Greenberg RS, Flannery JT, et al. Risk of leukemia after chemotherapy and radiation treatment for breast cancer. N Engl J Med. 1992;326:1745–51. ArticlePubMed
  • 27. Travis LB, Holowaty EJ, Bergfeldt K, Lynch CF, Kohler BA, Wiklund T, et al. Risk of leukemia after platinum-based chemotherapy for ovarian cancer. N Engl J Med. 1999;340:351–7. ArticlePubMed
  • 28. Ingvar A, Smedby KE, Lindelof B, Fernberg P, Bellocco R, Tufveson G, et al. Immunosuppressive treatment after solid organ transplantation and risk of post-transplant cutaneous squamous cell carcinoma. Nephrol Dial Transplant. 2010;25:2764–71. ArticlePubMed
  • 29. Vineis P, Alavanja M, Garte S. Dose-response relationship in tobacco-related cancers of bladder and lung: a biochemical interpretation. Int J Cancer. 2004;108:2–7. ArticlePubMed
  • 30. Tie Y, Tang F, Wei YQ, Wei XW. Immunosuppressive cells in cancer: mechanisms and potential therapeutic targets. J Hematol Oncol. 2022;15:61.ArticlePubMedPMCPDF
  • 31. Saini N, Sterling JF, Sakofsky CJ, Giacobone CK, Klimczak LJ, Burkholder AB, et al. Mutation signatures specific to DNA alkylating agents in yeast and cancers. Nucleic Acids Res. 2020;48:3692–707. ArticlePubMedPMCPDF
  • 32. Valdez BC, Hassan M, Andersson BS. Development of an assay for cellular efflux of pharmaceutically active agents and its relevance to understanding drug interactions. Exp Hematol. 2017;52:65–71. ArticlePubMedPMC
  • 33. Yasuda H, Yasuda M, Komatsu N. Chemotherapy for non-Hodgkin lymphoma in the hemodialysis patient: A comprehensive review. Cancer Sci. 2021;112:2607–24. ArticlePubMedPMCPDF
  • 34. Molina BD, Leiro MG, Pulpon LA, Mirabet S, Yanez JF, Bonet LA, et al. Incidence and risk factors for nonmelanoma skin cancer after heart transplantation. Transplant Proc. 2010;42:3001–5. ArticlePubMed
  • 35. Khan N, Abbas AM, Lichtenstein GR, Loftus EV Jr, Bazzano LA. Risk of lymphoma in patients with ulcerative colitis treated with thiopurines: a nationwide retrospective cohort study. Gastroenterology. 2013;145:1007–15. ArticlePubMed
  • 36. Bernatsky S, Clarke AE, Suissa S. Hematologic malignant neoplasms after drug exposure in rheumatoid arthritis. Arch Intern Med. 2008;168:378–81. ArticlePubMed
  • 37. Hannuksela-Svahn A, Pukkala E, Laara E, Poikolainen K, Karvonen J. Psoriasis, its treatment, and cancer in a cohort of Finnish patients. J Invest Dermatol. 2000;114:587–90. ArticlePubMed
  • 38. Finazzi G, Caruso V, Marchioli R, Capnist G, Chisesi T, Finelli C, et al. Acute leukemia in polycythemia vera: an analysis of 1638 patients enrolled in a prospective observational study. Blood. 2005;105:2664–70. ArticlePubMed
  • 39. Haine D, Dohoo I, Dufour S. Selection and misclassification biases in longitudinal studies. Front Vet Sci. 2018;5:99.ArticlePubMedPMC
  • 40. Yland JJ, Wesselink AK, Lash TL, Fox MP. Misconceptions about the direction of bias from nondifferential misclassification. Am J Epidemiol. 2022;191:1485–95. ArticlePubMedPMCPDF
  • 41. Lim W, Sung S, Hong Y, Moon S, Lee S, Kim K, et al. Fraction of cancer attributable to carcinogenic drugs in Korea from 2015 to 2030. Cancer Res Treat. 2025;57:635–48. ArticlePubMedPDF
  • 42. Kim JH, Cho M, Lee JE, Kim J. Cancer incidence attributable to dietary factors in Korea. J Korean Med Assoc. 2025;68:108–20. ArticlePDF
  • 43. Park S, Choi YJ, Park SK, Seo HG. Population attributable fraction as a key measure of primary cancer prevention strategy. J Korean Med Assoc. 2025;68:82–6. ArticlePDF
  • 44. Choi YJ. Cancer attributable to excess body weight in Korea: a focus on primary prevention. J Korean Med Assoc. 2025;68:100–7. ArticlePDF
  • 45. Kim EM, Min J, Kim I. Cancer attributable to occupational factors: a focus on primary prevention. J Korean Med Assoc. 2025;68:121–9. ArticlePDF
  • 46. Ko KP. Cancer attributable to tobacco smoking: a focus on primary prevention. J Korean Med Assoc. 2025;68:91–9. ArticlePDF
  • 47. Kim M, Lim W, Kim K, Bae JS, Lee BJ, Koo BS, et al. A systematic review of economic evaluation of thyroid cancer. Int J Thyroidol. 2022;15:74–104. Article

Figure & Data

REFERENCES

    Citations

    Citations to this article as recorded by  

      • PubReader PubReader
      • ePub LinkePub Link
      • Cite
        CITE
        export Copy Download
        Close
        Download Citation
        Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.
        Format:
        • RIS — For EndNote, ProCite, RefWorks, and most other reference management software
        • BibTeX — For JabRef, BibDesk, and other BibTeX-specific software
        Include:
        • Citation for the content below
        IARC Group 1 Pharmaceuticals and Associated Cancer Risks: A Nationwide Population-Based Cohort Study in Korea
        Close
      • XML DownloadXML Download
      • 0
      • 1
      IARC Group 1 Pharmaceuticals and Associated Cancer Risks: A Nationwide Population-Based Cohort Study in Korea
      Image Image
      Fig. 1. Flow chart of selecting study subjects in National Health Insurance Service database.
      Fig. 2. Hazard ratio (95% confidence interval [CI]) for users relative to non-users of group 1 pharmaceuticals.

      Fig. 1.

      Fig. 2.

      IARC Group 1 Pharmaceuticals and Associated Cancer Risks: A Nationwide Population-Based Cohort Study in Korea
      Azathioprine
      		 Cyclosporine
      		 Cyclophosphamide
      Non-users (n=784,033) Ever-users (n=5,635) p-value Non-users (n=3,760,891) Ever-users (n=19,354) p-value Non-users (n=964,190) Ever-users (n=2,907) p-value
      FU (yr), median (IQR) 9.0 (8.9) 10.0 (9.1) < 0.01 12.5 (7.8) 14.3 (6.6) < 0.01 8.9 (8.5) 11.0 (9.9) < 0.01
      Age (yr), mean±SD 50.6±18.4 39.5±16.8 < 0.01 32.8±22.6 36.7±19.0 < 0.01 50.6±18.5 50.1±15.8 0.15
      Sex
       Men 287,355 (36.7) 2,599 (46.1) < 0.01 1,734,123 (46.1) 9,917 (51.2) < 0.01 360,734 (37.4) 689 (23.7) < 0.01
       Women 496,678 (63.3) 3,036 (53.9) 2,026,768 (53.9) 9,437 (48.8) 603,456 (62.6) 2,218 (76.3)
      Family history of cancer
       No 645,890 (82.4) 4,639 (82.3) 0.85 3,154,367 (83.9) 16,250 (84.0) 0.74 787,999 (81.7) 2,341 (80.5) 0.10
       Yes 138,143 (17.6) 996 (17.7) 606,524 (16.1) 3,104 (16.0) 176,191 (18.3) 566 (19.5)
      Risk for skin cancera)
       No 783,752 (100) 5,634 (100) 0.73 3,760,450 (100) 19,353 (100) 0.73 - -
       Yes 281 (0.0) 1 (0.0) 441 (0.0) 1 (0.0) - -
      Risk for hematologic cancerb)
       No 782,801 (99.8) 5,621 (99.8) 0.09 3,759,554 (100) 19,318 (99.8) < 0.001 963,568 (99.9) 2,903 (99.9) 0.12
       Yes 1,232 (0.2) 14 (0.2) 1,337 (0.0) 36 (0.2) 622 (0.1) 4 (0.1)
      Risk for bladder cancerc)
       No - - - - 940,207 (97.5) 2,825 (97.2) 0.25
       Yes - - - - 23,983 (2.5) 82 (2.8)
      Cigarette smoking
       Never 577,297 (73.6) 4,033 (71.6) < 0.01 2,563,819 (68.2) 12,348 (63.8) < 0.01 709,285 (73.6) 2,329 (80.1) < 0.01
       Past 94,473 (12.1) 790 (14.0) 449,699 (12.0) 2,396 (12.4) 116,723 (12.1) 301 (10.4)
       Current 112,263 (14.3) 812 (14.4) 747,373 (19.9) 4,610 (23.8) 138,182 (14.3) 277 (9.5)
      Methoxsalen
      		 Chlorambucil
      		 Melphalan
      Non-users (n=1,608,378) Ever-users (n=2,470) p-value Non-users (n=112,497) Ever-users (n=108) p-value Non-users (n=197,823) Ever-users (n=476) p-value
      FU (yr), median (IQR) 10.7 (8.8) 15.0 (3.9) < 0.01 8.0 (9.4) 0.5 (0.7) < 0.01 7.0 (8.0) 0.4 (0.6) < 0.01
      Age (yr), mean±SD 25.7±24.4 39.8±19.2 < 0.01 47.9±19.3 60.5±20.2 < 0.01 47.8±18.0 70.6±8.5 < 0.01
      Sex
       Men 748,418 (46.5) 1,142 (46.2) 0.77 33,789 (30.0) 67 (62.0) < 0.01 62,078 (31.4) 250 (52.5) < 0.01
       Women 859,960 (53.5) 1,328 (53.8) 78,708 (70.0) 41 (38.0) 135,745 (68.6) 226 (47.5)
      Family history of cancer
       No 1,338,436 (83.2) 2,099 (85.0) 0.02 90,870 (80.8) 87 (80.6) 0.95 154,952 (78.3) 378 (79.4) 0.57
       Yes 269,942 (16.8) 371 (15.0) 21,627 (19.2) 21 (19.4) 42,871 (21.7) 98 (20.6)
      Risk for skin cancera)
       No 1,608,027 (100) 2,470 (100) > 0.99 - - - -
       Yes 351 0 - - - -
      Risk for hematologic cancerb)
       No - - 111,086 (98.8) 88 (81.5) < 0.001 197,349 (99.8) 473 (99.4) 0.11
       Yes - - 1,411 (1.3) 20 (18.5) 474 (0.2) 3 (0.6)
      Cigarette smoking
       Never 1,063,464 (66.1) 1,648 (66.7) 0.01 87,201 (77.5) 68 (63.0) 0.001 152,600 (77.1) 352 (74.0) 0.06
       Past 222,181 (13.8) 292 (11.8) 11,793 (10.5) 20 (18.5) 19,489 (9.9) 62 (13.0)
       Current 322,733 (20.1) 530 (21.5) 13,503 (12.0) 20 (18.5) 25,734 (13.0) 62 (13.0)
      Outcome Exposre Cohort
      		 No. of cancers IR IRR HRa) (95% CI) HRb) (95% CI) p-value
      No. Person-year
      Skin cancer Azathioprine
       Non-user 783,900 6,652,263 1,220 0.02 1.00 1.00 < 0.01
       Ever-user 5,635 52,820 32 0.06 3.3 3.01 (2.12-4.27) 4.63 (2.91-7.39)
      Cyclosporine
       Non-user 3,760,754 40,290,898 5,766 0.01 1.00 1.00 < 0.01
       Ever-user 19,354 236,712 106 0.04 3.1 2.73 (2.25-3.30) 2.30 (1.79-2.95)
      Methoxsalen
       Non-user 1,608,315 14,985,310 1,459 0.01 1.00 1.00 < 0.01
       Ever-user 2,470 33,212 14 0.04 4.4 4.05 (2.39-6.86) 2.32 (1.36-3.95)
      Hematologic cancer Azathioprine
       Non-user 783,892 6,651,255 4,977 0.07 1.00 1.00 < 0.01
       Ever-user 5,633 52,768 124 0.24 3.1 2.85 (2.39-3.41) 3.15 (2.41-4.13)
      Cyclosporine
       Non-user 3,760,732 40,289,651 11,423 0.03 1.00 1.00 < 0.01
       Ever-user 19,353 236,684 451 0.19 6.7 5.93 (5.40-6.51) 2.96 (2.59-3.40)
      Cyclophosphamide
       Non-user 963,445 7,849,680 19,042 0.24 1.00 1.00 < 0.01
       Ever-user 2,907 28,619 331 1.16 4.8 5.31 (4.77-5.93) 3.83 (3.20-4.59)
      Chlorambucil
       Non-user 112,359 889,804 9,538 1.07 1.00 1.00 < 0.01
       Ever-user 108 720 91 12.64 11.8 8.68 (7.06-10.67) 3.51 (2.53-4.87)
      Melphalan
       Non-user 197,701 1,415,357 5,280 0.37 1.00 1.00 < 0.01
       Ever-user 474 1,632 461 28.25 75.7 49.53 (44.97-54.55) 16.31 (13.41-19.85)
      Bladder cancer Cyclophosphamide
       Non-user 963,498 7,854,646 4,092 0.05 1.00 1.00 < 0.01
       Ever-user 2,907 28,785 34 0.12 2.3 2.50 (1.79-3.51) 2.69 (1.92-3.78)
      Outcome Exposre Cohort
      		 No. of cancers IR IRR HRa) (95% CI) HRb) (95% CI) p-value
      No. Person-year
      Skin cancer Azathioprine
       Non-users 783,900 6,652,263 1,220 0.02 1.00 1.00 < 0.01
       < 647 mg 2,817 22,438 4 0.02 1.0 0.95 (0.36-2.53) 1.91 (0.71-5.10)
       > 647 mg 2,818 30,382 28 0.09 5.0 4.45 (3.06-6.47) 9.74 (6.67-14.23)
      Cyclosporine
       Non-users 3,760,754 40,290,898 5,766 0.01 1.00 1.00 < 0.01
       < 122 mg 9,677 114,674 24 0.02 1.5 1.31 (0.88-1.96) 1.90 (1.28-2.84)
       > 122 mg 9,677 122,038 82 0.07 4.7 3.95 (3.17-4.91) 3.50 (2.81-4.35)
      Methoxsalen
       Non-users 1,608,315 14,985,310 1,459 0.01 1.00 1.00 < 0.01
       < 16 mg 1,226 15,992 3 0.02 1.9 1.85 (0.60-5.75) 1.35 (0.43-4.18)
       > 16 mg 1,244 17,220 11 0.06 6.6 5.98 (3.30-10.82) 3.08 (1.70-5.58)
      Hematologic cancer Azathioprine
       Non-users 783,892 6,651,255 4,977 0.07 1.00 1.00 < 0.01
       < 644 mg 2,816 22,416 32 0.14 1.9 1.85 (1.30-2.61) 1.57 (1.11-2.22)
       > 644 mg 2,817 30,352 92 0.30 4.1 3.85 (3.14-4.74) 3.39 (2.76-4.17)
      Cyclosporine
       Non-users 3,760,732 40,289,651 11,423 0.03 1.00 1.00 < 0.01
       < 114 mg 9,670 114,593 43 0.04 1.3 1.22 (0.90-1.65) 1.33 (0.99-1.80)
       > 114 mg 9,683 122,091 408 0.33 11.8 10.31 (9.34-11.38) 6.75 (6.12-7.46)
      Cyclophosphamide
       Non-users 963,445 7,849,680 19,042 0.24 1.00 1.00 < 0.01
       < 210 mg 1,446 15,220 22 0.14 0.6 0.59 (0.39-0.89) 0.85 (0.56-1.29)
       > 210 mg 1,461 13,399 309 2.31 9.5 9.25 (8.27-10.35) 8.90 (7.96-9.96)
      Chlorambucil
       Non-users 112,359 889,804 9,538 1.07 1.00 1.00 < 0.01
       < 341 mg 54 392 41 10.46 9.8 9.24 (6.80-12.56) 3.91 (2.87-5.32)
       > 341 mg 54 328 50 15.24 14.2 13.66 (10.34-18.04) 2.88 (2.17-3.81)
      Melphalan
       Non-users 197,701 1,415,357 5,280 0.37 1.00 1.00 < 0.01
       < 243 mg 237 649 224 34.51 92.5 64.85 (56.68-74.18) 21.17 (18.46-24.27)
       > 243 mg 237 983 237 24.11 64.6 52.19 (45.80-59.47) 15.75 (13.80-17.98)
      Bladder cancer Cyclophosphamide
       Non-users 963,498 7,854,646 4,092 0.05 1.00 1.00 < 0.01
       < 210 mg 1,448 15,232 11 0.07 1.4 1.31 (0.73-2.38) 2.56 (1.41-4.63)
       > 210 mg 1,459 13,553 23 0.17 3.3 3.09 (2.05-4.66) 2.75 (1.83-4.15)
      Outcome Exposre Cohort
      		 No. of cancers IR IRR HRa) (95% CI) HRb) (95% CI) p-value
      No. Person-year
      Skin cancer Azathioprine
       Non-users 783,900 6,652,263 1,220 0.02 1.00 1.00 < 0.01
       < 322 days 2,817 22,203 4 0.02 1.0 0.96 (0.36-2.57) 1.95 (0.73-5.22)
       > 322 days 2,818 30,617 28 0.09 5.0 4.41 (3.03-6.41) 9.58 (6.56-13.99)
      Cyclosporine
       Non-users 3,760,754 40,290,898 5,766 0.01 1.00 1.00 < 0.01
       < 49 days 9,673 115,064 25 0.02 1.5 1.36 (0.92-2.01) 1.95 (1.32-2.89)
       > 49 days 9,681 121,648 81 0.07 4.7 3.93 (3.16-4.90) 3.48 (2.79-4.33)
      Methoxsalen
       Non-users 1,608,315 14,985,310 1,459 0.01 1.00 1.00 < 0.01
       < 4 days 1,224 15,931 3 0.02 1.9 1.86 (0.60-5.79) 1.33 (0.43-4.13)
       > 4 days 1,246 17,281 11 0.06 6.6 5.95 (3.29-10.77) 3.10 (1.71-5.62)
      Hematologic cancer Azathioprine
       Non-users 783,892 6,651,255 4,977 0.07 1.00 1.00 < 0.01
       < 325 days 2,816 22,181 35 0.16 2.1 2.04 (1.46-2.85) 1.75 (1.25-2.44)
       > 325 days 2,817 30,587 89 0.29 3.9 3.70 (3.00-4.56) 3.23 (2.62-3.99)
      Cyclosporine
       Non-users 3,760,732 40,289,651 11,423 0.03 1.00 1.00 < 0.01
       < 49 days 9,675 115,094 47 0.04 1.4 1.32 (0.99-1.76) 1.46 (1.10-1.94)
       > 49 days 9,678 121,590 404 0.33 11.7 10.30 (9.33-11.37) 6.68 (6.05-7.38)
      Cyclophosphamide
       Non-users 963,445 7,849,680 19,042 0.24 1.00 1.00 < 0.01
       < 70 days 1,453 15,174 33 0.22 0.9 0.88 (0.63-1.24) 1.29 (0.91-1.81)
       > 70 days 1,454 13,445 298 2.22 9.1 8.90 (7.94-9.98) 8.53 (7.60-9.56)
      Chlorambucil
       Non-users 112,359 889,804 9,538 1.07 1.00 1.00 < 0.01
       < 69 days 54 374 41 10.96 10.2 9.52 (7.01-12.94) 4.08 (2.99-5.55)
       > 69 days 54 346 50 14.45 13.5 13.19 (9.99-17.42) 2.81 (2.12-3.72)
      Melphalan
       Non-users 197,701 1,415,357 5,280 0.37 1.00 1.00 < 0.01
       < 32 days 237 711 224 31.50 84.4 60.98 (53.31-69.76) 20.53 (17.91-23.53)
       > 32 days 237 921 237 25.73 69.0 54.82 (48.10-62.47) 16.09 (14.09-18.37)
      Bladder cancer Cyclophosphamide
       Non-users 963,498 7,854,646 4,092 0.05 1.00 1.00 < 0.01
       < 70 days 1,449 15,100 12 0.08 1.5 1.45 (0.82-2.55) 3.18 (1.80-5.61)
       > 70 days 1,458 13,685 22 0.16 3.1 2.93 (1.93-4.46) 2.48 (1.63-3.77)
      Outcome Exposre Cohort
      		 No. of cancers IR IRR HRa) (95% CI) HRb) (95% CI) p-value
      No. Person-year
      Skin cancer Azathioprine
       Non-users 783,900 6,652,263 1,220 0.02 1.00 1.00 < 0.01
       < 1.7 mg 2,817 27,338 15 0.05 3.0 2.71 (1.63-4.51) 5.50 (3.30-9.17)
       > 1.7 mg 2,818 25,482 17 0.07 3.6 3.41 (2.12-5.51) 7.61 (4.69-12.34)
      Cyclosporine
       Non-users 3,760,754 40,290,898 5,766 0.01 1.00 1.00 < 0.01
       < 1.2 mg 9,670 117,302 49 0.04 2.9 2.54 (1.92-3.37) 2.63 (1.99-3.49)
       > 1.2 mg 9,684 119,410 57 0.05 3.3 2.89 (2.22-3.75) 3.26 (2.51-4.23)
      Methoxsalen
       Non-users 1,608,315 14,985,310 1,459 0.01 1.00 1.00 < 0.01
       < 2.9 mg 1,235 17,042 11 0.06 6.7 6.09 (3.37-11.03) 3.44 (1.90-6.23)
       > 2.9 mg 1,235 16,170 3 0.02 1.9 1.82 (0.59-5.63) 1.15 (0.37-3.58)
      Hematologic cancer Azathioprine
       Non-users 783,892 6,651,255 4,977 0.07 1.00 1.00 < 0.01
       < 1.7 mg 2,816 27,328 50 0.18 2.4 2.33 (1.76-3.08) 2.09 (1.58-2.76)
       > 1.7 mg 2,817 25,440 74 0.29 3.9 3.75 (2.98-4.72) 3.13 (2.49-3.95)
      Cyclosporine
       Non-users 3,760,732 40,289,651 11,423 0.03 1.00 1.00 < 0.01
       < 1.2 mg 9,670 117,328 105 0.09 3.2 2.83 (2.34-3.43) 2.51 (2.07-3.04)
       > 1.2 mg 9,683 119,356 346 0.29 10.2 9.14 (8.21-10.17) 6.82 (6.12-7.59)
      Cyclophosphamide
       Non-users 963,445 7,849,680 19,042 0.24 1.00 1.00 < 0.01
       < 1.0 mg 1,453 14,941 52 0.35 1.4 1.42 (1.08-1.86) 1.89 (1.44-2.48)
       > 1.0 mg 1,454 13,678 279 2.04 8.4 8.17 (7.25-9.19) 8.45 (7.51-9.51)
      Chlorambucil
       Non-users 112,359 889,804 9,538 1.07 1.00 1.00 < 0.01
       < 2.8 mg 54 401 43 10.72 10.0 9.76 (7.23-13.17) 3.06 (2.27-4.14)
       > 2.8 mg 54 319 48 15.05 14.0 13.01 (9.8-17.28) 3.49 (2.62-4.64)
      Melphalan
       Non-users 197,701 1,415,357 5,280 0.37 1.00 1.00 < 0.01
       < 7.5 mg 236 803 228 28.39 76.1 57.94 (50.72-66.2) 18.28 (15.97-20.93)
       > 7.5 mg 238 829 233 28.11 75.3 57.35 (50.27-65.43) 17.69 (15.48-20.22)
      Bladder cancer Cyclophosphamide
       Non-users 963,498 7,854,646 4,092 0.05 1.00 1.00 < 0.01
       < 1.0 mg 1,453 14,903 20 0.13 2.6 2.46 (1.58-3.81) 3.98 (2.56-6.18)
       > 1.0 mg 1,454 13,882 14 0.10 1.9 1.83 (1.08-3.09) 1.84 (1.09-3.10)
      Table 1. General characteristics of patients with indication of azathioprine, cyclosporine, and cyclophosphamide

      Values are presented as number (%) unless otherwise indicated. FU, follow up; IQR, interquartile range; SD, standard deviation.

      The risk factors for skin cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS) or skin cancer in past disease codes,

      The risk factors for hematologic cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: hepatitis C virus infection, HIV/AIDS, bone marrow disorder or hematologic diseases (polycythemia), myelodysplastic disease, other and unspecified neoplasm of lymphoid, hematopoietic and related tissues or radiotherapy in past disease codes,

      The risk factors for bladder cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: urinary infections, kidney and bladder stones, urothelial carcinomas, or radiotherapy in past disease codes.

      Table 2. General characteristics of patients with indication of methoxsalen, chlorambucil, and melphalan

      Values are presented as number (%) unless otherwise indicated. FU, follow up; IQR, interquartile range; SD, standard deviation.

      The risk factors for skin cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS) or skin cancer in past disease codes,

      The risk factors for hematologic cancer was considered as having at least one of the following risk factors in past diseases before enrollment of cohort: hepatitis C virus infection, HIV/AIDS, bone marrow disorder or hematologic diseases (polycythemia), myelodysplastic disease, other and unspecified neoplasm of lymphoid, hematopoietic and related tissues or radiotherapy in past disease codes.

      Table 3. HR (95% CI) for ever-users relative to non-users of group 1 pharmaceuticals

      All incidence rates were presented as percentage. CI, confidence interval; HR, hazard ratio; IR, incidence rate; IRR, incidence rate ratio.

      Crude HR,

      Adjusted for age, sex, family history of cancer, smoking and past history of risk factors for each cancer outcome (presented in S3 Table).

      Table 4. HR (95% CI) for lifetime cumulative dose (median) relative to non-users of group 1 pharmaceuticals

      All incidence rates were presented as percentage. CI, confidence interval; HR, hazard ratio; IR, incidence rate; IRR, incidence rate ratio.

      Crude HR,

      Adjusted for age, sex, family history of cancer, smoking and past history of risk factors for each cancer outcome (presented in S3 Table).

      Table 5. HR (95% CI) for lifetime administered duration (median) relative to non-users of group 1 pharmaceuticals

      All incidence rates were presented as percentage. CI, confidence interval; HR, hazard ratio; IR, incidence rate; IRR, incidence rate ratio.

      Crude HR,

      Adjusted for age, sex, family history of cancer, smoking and past history of risk factors for each cancer outcome (presented in S3 Table).

      Table 6. HR (95% CI) for single dose (median) relative to non-users of group 1 pharmaceuticals

      All incidence rates were presented as percentage. CI, confidence interval; HR, hazard ratio; IR, incidence rate; IRR, incidence rate ratio.

      Crude HR,

      Adjusted for age, sex, family history of cancer, smoking and past history of risk factors for each cancer outcome (presented in S3 Table).

      Table 1.

      Table 2.

      Table 3.

      Table 4.

      Table 5.

      Table 6.

      Cancer Res Treat : Cancer Research and Treatment
      © Korean Cancer Association.
      Close layer
      TOP