Scand J Rheumatol 2014;43:301–306

301

Increased risk of cancer for patients with ankylosing spondylitis: a nationwide population-based retrospective cohort study L-M Sun1*, C-H Muo2, J-A Liang3,4*, S-N Chang5, F-C Sung2,6, C-H Kao4,7 1 Department of Radiation Oncology, Zuoying Branch of Kaohsiung Armed Forces General Hospital, Kaohsiung, 2Management Office for Health Data, China Medical University Hospital, Taichung, 3Department of Radiation Oncology, China Medical University Hospital, Taichung, 4Graduate Institute of Clinical Medicine Science and School of Medicine, College of Medicine, China Medical University, Taichung, 5Department of Medical Research, Taichung Veterans General Hospital, 6Graduate Institute of Clinical Medical Science, China Medical University College of Medicine, Taichung, and 7Department of Nuclear Medicine PET Centre, China Medical University Hospital, Taichung, Taiwan

Objectives: Few studies have investigated the relationship between ankylosing spondylitis (AS) and other inflammatory spondyloarthritis and subsequent cancer. The aim of this study was to determine whether AS is associated with cancer risk. Method: We used data from the National Health Insurance (NHI) system of Taiwan to investigate this association. The AS cohort included 4133 patients, and each patient was randomly frequency matched with four persons without AS based on sex, age, and entry year (control cohort). We conducted a Cox proportional hazards regression analysis to estimate the influence of AS on cancer risk. Results: Among patients with AS, the overall risk of developing cancer was 38% higher than that of people without AS, and the difference was significant [adjusted hazard ratio (HR) 1.38, 95% confidence interval (CI) 1.18–1.60]. This phenomenon held true even when we analysed males and females separately. The risk of developing lung or head and neck cancer among patients with AS was significantly higher; and risks for liver, bladder, and uterus cancers were marginally significantly higher. Conclusions: This nationwide population-based cohort study shows that Taiwanese patients with AS have a higher risk of developing cancer, particularly lung or head and neck cancer.

Ankylosing spondylitis (AS) is a chronic systemic inflammatory disorder of the axial skeleton. Although this disorder has an overall prevalence of 0.04–0.9%, men are more affected than women (1–3). The aetiology of AS remains unclear but is known to involve the interaction of genetic and environmental factors. Immune mechanisms probably play a key role in its pathophysiology, and AS is thought to be type of autoimmune diseases (4–7). Researchers have investigated the association between cancer and autoimmune diseases for several years, and various reports have suggested an increased cancer risk in some autoimmune diseases (7–10). The role of chronic inflammation in the development of cancer has also been a well-studied topic in epidemiological and experimental studies (10–13). Many AS patients are relatively young, and this has aroused major socioeconomic interest (14). The clinical Chia-Hung Kao, Graduate Institute of Clinical Medicine Science and School of Medicine, College of Medicine, China Medical University, No. 2, Yuh-Der Road, Taichung 404, Taiwan. E-mail: [email protected] *These authors contributed equally to this work. Accepted 6 November 2013

features of AS, including the anatomical distribution of the arthritis, type of joint ossification, extra-articular manifestations, and sex distribution, differ from those of rheumatoid arthritis (15). This implies that the suggested positive association between rheumatoid arthritis and cancer may not appear in patients with AS. Several researchers have studied the increased cancer risk in patients with AS subjected to radiation treatment (16, 17). However, few studies have focused on the possible relationship between AS itself and malignancy (15, 18, 19). To clarify this issue, we conducted a large population-based retrospective cohort study to investigate the relationship between AS and subsequent cancer risk in Taiwanese patients. The extended sample in this study also includes patients with other inflammatory spondyloarthritis. The original database was derived from the National Health Insurance (NHI) system in Taiwan. Method Data sources The Taiwan NHI programme has provided compulsory universal health insurance in Taiwan since 1995. The Taiwan National Health Research Institute (NHRI)

© 2014 Informa Healthcare on license from Scandinavian Rheumatology Research Foundation DOI: 10.3109/03009742.2013.863969

www.scandjrheumatol.dk

302

manages the Taiwan NHI research database (NHIRD), and releases data for research purposes. The NHIRD includes complete outpatient visits, hospital admissions, prescriptions, and disease status for nearly 99% of Taiwan’s 23.74 million citizens. The details of the NHIRD have been described in a previous study (20). The data analysed in this study included claims of 1 million beneficiaries randomly selected from all insurants in 2000. The data from these beneficiaries in the NHRID are representative of the entire population because the NHRI reports no statistically significant differences in the distributions of age, sex, or health care expenditures between the subset of the NHIRD and all enrollees. We also used the catastrophic illness certificate database to identify cancer patients. This database consists of catastrophic illness certificates issued for cancer, histological, or cytological evidence of such disease. We confirm that all data were deidentified and analysed anonymously. This study was approved by the Ethics Review Board at China Medical University (CMU-REC-101-012).

Study subjects We identified newly diagnosed AS and other inflammatory spondyloarthritis patients in the claims database between 2000 and 2005 using the International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9CM) code 720. Patients were excluded if they had a diagnosis of a malignant tumour (ICD-9-CM 140-208) before the index date. Based on the same exclusion criteria, the non-AS control persons were 4.1 frequency matched with sex, age, and entry year to the case group. We analysed data from 20 665 people.

Comorbidities and end-point We recorded major comorbidities such as hypertension (ICD-9-CM 401–405), diabetes mellitus (ICD-9-CM 250), hyperlipidaemia (ICD-9-CM 272), ischaemic heart disease (ICD-9-CM 410–414), chronic obstructive pulmonary disease (ICD-9-CM 496), asthma (ICD-9-CM 493), and stroke (ICD-9-CM 430–438) at baseline. Because the NHIRD lacked smoking data, we used smoking-associated diseases (such as chronic obstructive pulmonary disease and asthma) instead of smoking in the analyses; a similar method has been used by others (21). Both the AS patient cohort and the comparison cohort were followed from the date of enrolment to the end of 2010, and upon diagnosis of cancer. The follow-up time (in person-years) was calculated for each person until the cancer was diagnosed or censored. The date of censoring was defined as the date the study patient died during the follow-up period, the date of last withdrawal from the NHI, or the date of follow-up termination.

www.scandjrheumatol.dk

L-M Sun et al

Statistical analysis We compared the differences in categorical variables between two groups, such as age, sex, and comorbidities, using χ 2 tests. Using the same variables, we also calculated the incidence density with person-years in the study cohort and comparison cohort. We calculated the incidence rate (IR) of cancer for each variable, and performed a Cox proportional hazards regression analysis to evaluate the risk of developing cancer in AS patients. We estimated the cumulative IR using the Kaplan-Meier method, and used the log-rank test to investigate the differences in the risk of cancer between two cohorts. We performed all analyses using SAS version 9.2 for Windows (SAS Institute Inc, Cary, NC, USA). A two-tailed p-value < 0.05 was considered statistically significant.

Results Table 1 shows the demographic characteristics and medical comorbidities of the AS and non-AS cohorts. We found no significant difference in age and sex between the two cohorts. The AS patients had a higher proportion of hypertension (27.6% vs. 22.7%, p < 0.0001), diabetes (18.4% vs. 13.0%, p < 0.0001), hyperlipidaemia (9.17% vs. 7.83%, p = 0.005), ischaemic heart disease (15.9% vs. 10.4%, p < 0.0001), chronic obstructive pulmonary disease (29.6% vs. 20.8%, p < 0.0001), and asthma (6.02% vs. 4.21%, p < 0.0001) compared to the non-AS patients. Table 2 shows the incidence rates (IRs) and hazard ratios (HRs) of AS-associated cancer based on sociodemographic status. The follow-up results show a higher cancer incidence in the AS cohort than in the non-AS cohort [7.02 vs. 5.20 per 1000 person-years; adjusted HR Table 1. Demographics between AS and non-AS groups. AS (n = 4133) Gender, n (%) Male Female Age (years), mean (SD) Age (years), n (%) < 35 35–49 50–64  65 Comorbidity, n (%) Hypertension Diabetes Hyperlipidaemia Ischaemic heart disease Chronic obstructive pulmonary disease Asthma Stroke χ 2, *t test.

Non-AS (n = 16 532) p-value 1.00

2215 (53.6) 1918 (46.4) 46.5 (17.4)

8860 (53.6) 7672 (46.4) 46.3 (17.5)

1201 (29.1) 1231 (29.8) 1010 (24.4) 691 (16.7)

4804 (29.1) 4924 (29.8) 4040 (24.4) 2764 (16.7)

1141 (27.6) 762 (18.4) 379 (9.17) 657 (15.9) 1225 (29.6)

3744 (22.7) 2148 (13.0) 1294 (7.83) 1718 (10.4) 3445 (20.8)

< 0.0001 < 0.0001 0.005 < 0.0001 < 0.0001

249 (6.02) 87 (2.11)

696 (4.21) 354 (2.14)

< 0.0001 0.89

0.51* 1.00

AS and cancer risk

303

Table 2. Incidence rates (IRs) and hazard ratios (HRs) for cancer. AS

Non-AS

Cancer no. Person-years All Gender Male Female Age (years) < 35 35–49 50–64  65

IR

HR (95% CI), compared to non-AS group

Cancer no. Person-years

226

32 180

7.02

661

127 126

116 110

16 955 15 225

6.84 7.22

340 321

66 939 60 187

7 47 93 79

9677 10 034 7802 4667

0.72 4.68 11.92 16.93

24 133 246 258

37 731 39 453 31 277 18 665

IR

Crude

Mode1 1

Mode1 2

5.20 1.35 (1.16–1.57)** 1.35 (1.16–1.57)** 1.38 (1.18–1.60)** 5.08 1.35 (1.09–1.66)* 1.36 (1.10–1.67)* 5.33 1.36 (1.09–1.68)* 1.34 (1.08–1.66)* 0.64 3.37 7.87 13.82

1.14 (0.49–2.64) 1.39 (1.00–1.94) 1.52 (1.20–1.93)** 1.22 (0.95–1.57)

1.10 (0.47–2.57) 1.36 (0.98–1.91) 1.48 (1.17–1.89)* 1.23 (0.96–1.58)

1.39 (1.13–1.72)* 1.37 (1.10–1.70)* 1.15 (0.49–2.67) 1.38 (0.99–1.93) 1.48 (1.16–1.88)* 1.27 (0.99–1.64)

IR, Incidence rate, per 1000 person-years. Model 1: Adjusted for age, gender, hypertension, hyperlipidaemia, and diabetes. Model 2: Adjusted for age, gender, hypertension, hyperlipidaemia, diabetes, ischaemic heart disease, asthma, and chronic obstructive pulmonary disease. * p < 0.01, ** p < 0.001.

1.35, 95% confidence interval (CI) 1.16–1.57 after controlling for age, gender, hypertension, hyperlipidaemia, and diabetes; HR 1.38, 95% CI 1.18–1.60 after controlling for age, gender, hypertension, hyperlipidaemia, diabetes, ischaemic heart disease, asthma, and chronic obstructive pulmonary disease]. Kaplan-Meier analysis showed that a significant difference in the cumulative cancer IR between the two cohorts (log-rank test p < 0.0001) (Figure 1). Table 2 also shows the sex- and age-stratification analysis of the risk for AS-associated cancer. Compared to patients without AS, the risk of cancer in male and female AS patients was 1.39 (95% CI 1.13–1.72) and 1.37 (95%

0.10 Non-AS

Cumulative incidence

0.08

AS

Log-rank test, p

Increased risk of cancer for patients with ankylosing spondylitis: a nationwide population-based retrospective cohort study.

Few studies have investigated the relationship between ankylosing spondylitis (AS) and other inflammatory spondyloarthritis and subsequent cancer. The...
132KB Sizes 1 Downloads 3 Views