Subsequent risk of nasopharyngeal carcinoma among patients with allergic rhinitis: A nationwide population-based cohort study.

Subsequent risk of nasopharyngeal carcinoma among patients with allergic rhinitis:

A nationwide population-based cohort study

Kuen-Tze Lin, MD,

Wen-Yen Huang, MD,

Che-Chen Lin, MSc,

Yee-Min Jen, MD, PhD,

Chun- Shu Lin, MD,

Cheng-Hsiang Lo, MD,

Chia-Hung Kao, MD

INTRODUCTION

Allergic rhinitis is the most common chronic disorder of the respiratory tract and includes symptoms of sneezing, rhinorrhea, nasal obstruction, nasal itching, postnasal drip, and cough. ^1,2 The economic burden of allergic rhinitis and its complications are considerable. ^3,4 Allergic rhinitis is also associated with a lower quality of life. ^5,6 The prevalence of allergic rhinitis varies worldwide. The prevalence

of allergic rhinitis is extremely high (24.2% to

43.0%) and continues to increase in Taiwan, ^1,7,8 which is an island country with a population of approximately 23 million, the majority of whom are ethnic Chinese. ⁷ Nasopharyngeal cancer (NPC) is a malignant tumor of the nasopharynx, the narrow tubular passage behind the nasal cavity. The incidence of NPC demonstrates a marked geographic variation. It is rare in most parts of the world but common in Southern China, Hong Kong, and Taiwan. ^8,9 Based on the 2009 cancer registry annual report released by the Taiwan Department of Health, the incidence of NPC was 9.99 per 100,000 for men and 2.98 per 100,000 for women. Thus, in Taiwan, NPC is the

11th most common cause of cancer-related death for men and the 16th for women.

Because allergic rhinitis is such a common condition,

any association between allergic rhinitis and an increased

risk of NPC is a public health concern in Taiwan. Previous

studies have shown that the risk of NPC is higher in

patients with chronic local infection/inflammation of the

aerodigestive tract in the head and neck, such as sinusitis,

otitis media, and tonsillitis. ^10–13 However, the association

between allergic rhinitis and subsequent NPC is less known and even less studied.

Taiwan initiated its National Health Insurance (NHI) program in 1996, and 97% of the hospitals and clinics throughout Taiwan were under contract with the system by the end of 1996. By 1998, the NHI provided health care for nearly 99% of the population of Taiwan. ¹⁴ The NHI has made the Taiwan National Health Insurance Research Database (NHIRD) available to researchers in Taiwan, and this database has been extensively used in epidemiologic studies. ^11,15,16

In this research, we conducted a large nationwide retrospective cohort study to examine the risk of NPC after

diagnosis of allergic rhinitis.

MATERIALS AND METHODS Data source

The Taiwan government instituted the Taiwan NHI program, a single-payer and universal insurance plan, in

1996. By 1998, the program covered nearly 99% of the citizens of Taiwan. The NHIRD, which contains the

annual registration files and original claims data for reimbursement, is managed by the National Health Research

Institute (NHRI). To protect patient privacy, all personal identification information is encrypted before the data are released for research.

We used the Longitudinal Health Insurance Database (LHID), which is a sub-dataset of the NHIRD. The almost 23.75 million insured people in the period of 1996 to 2000 were used as the database pool. The NHRI assigned a random number for each person by Knuth and Park and Miller’s method. The NHRI random sampled 1 million insured people into the LHID. According to the NHRI report, there was no statistically significant difference between LHID and NHIRD in age, sex, annual birth rate, and health costs. The NHRI created a scrambled, anonymous identification number to combine each person’s

information, including sex, birth date, and registry of

medical services.

In this study, we collected disease histories from inpatient, out-patient, and catastrophic illness registry files.

The disease diagnoses were based on the International

Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM).

Study population

We conducted a population-based retrospective cohort study to clarify the relationship between allergic rhinitis and developing NPC risk. We established an allergic rhinitis cohort of patients with newly diagnosed allergic rhinitis (ICD-9-CM 477) from 2000 to 2005. The index date was set on a half year of allergic rhinitis diagnosis. The comparison cohort consisted of persons without any allergic rhinitis diagnosis in the LHID randomly 2-fold frequency- matched according to sex, age (per 5 years), and

year of index date. This study excluded patients with cancer occurrence before the index date, and aged <20 years.

The follow-up period was terminated upon developing NPC (ICD-9-CM 147; based on data obtained from the catastrophic illness registry file), withdrawal from the insurance program, or December 31, 2010.

We considered demographic characteristics and NPCassociated comorbidities as confounding factors in this

study.

Based on inpatient and outpatient files, the comorbidities included hypertension (ICD-9-CM 401-405), diabetes (ICD-9-CM 250), coronary artery disease (CAD; ICD-9- CM 410-414), atopic dermatitis (ICD-9-CM 691), and asthma (ICD-9-CM 493).

Statistical analysis

To describe the structure of the allergic rhinitis and comparison cohorts, we showed the mean and SD for

continuous variables and the count and percentage for category variables. The t test for continuous variables and

the chi-square test for category variables were used to examine the differences between the 2 cohorts. The results demonstrate a cumulative NPC incidence and

demographic-specific and comorbidity-specific NPC incidence

for the allergic rhinitis and comparison cohorts.

Compared to the comparison cohort, the Cox proportional hazards regression model with adjusted potential confounding factors was applied to the estimate hazard ratio

(HR) and the confidence interval (CI) for the allergic rhinitis cohort. The average number of allergic rhinitis visits

was measured as the total frequency of allergic rhinitis visits during the follow-up time divided by the follow-up duration (years). The allergic rhinitis cohort was divided into 3 sub-cohorts based on allergic rhinitis visits: <2 times per year, 2 to 4 times per year, and _4 times per years. We used 2 methods to assess the relationship between increasing allergic rhinitis visits and developing NPC risk. In the first method, we estimated the HR for each frequency level of average allergic rhinitis visits, and treated allergic rhinitis visits as continuous variables to test the trend by using the Cox proportional hazards regression. In the second method, we calculated the number of allergic rhinitis visits in each follow-up year and

treated this value as a time-dependent covariate in the Cox proportional hazards regression model to estimate the association between the frequencies of allergic rhinitis visits and developing NPC risk.

We used SAS 9.3 software (SAS Institute, Cary, NC) to manage and analyze the data. The significance level was set at < .05 for 2-sided testing of the p value.

RESULTS

We established a 67,532 allergic rhinitis patient cohort and a 135,064 individual comparison cohort with the same mean age (42.9 years) and sex ratio (55.4% female;

Table 1). The proportion of comorbidity in the allergic rhinitis cohort was substantially greater than that in the

comparison cohort.

The NPC incidence in the allergic rhinitis cohort was 1.14 per 10,000 person-years. In the comparison cohort, the NPC incidence was 0.53 per 10,000 person-years (Table 2). The NPC incidence in the allergic rhinitis

cohort was nearly 2.16-fold higher than that in the comparison

cohort. After adjusting for the possible confounding factors of the study, the allergic rhinitis cohort had a 2.33-fold higher risk of developing NPC than did the comparison cohort (HR 5 2.33; 95% CI 5 1.59–3.40).

Compared to the comparison cohort, the patients with allergic rhinitis had a similar risk of developing NPC in each age group. In the female population, the patients with allergic rhinitis had a nearly 3-fold higher risk of developing NPC compared with the comparison persons (HR 5 3.02; 95% CI 5 1.47–6.22); in the male population, the patients with allergic rhinitis had only a 2-fold

higher risk of developing NPC than did the comparison cohort (HR 5 2.06; 95% CI 5 1.31–3.25).

Table 2 also shows the comorbidity-specific developing NPC incidence and estimated HR for both study cohorts.

Compared to the comparison cohort, the allergic rhinitis cohort was significantly associated with a higher risk of developing NPC when both cohorts were without each comorbidity. Especially of the study population without any comorbidity, the patients with allergic rhinitis still had a 2.43-fold higher risk of developing NPC risk than did the comparison person (HR 5 2.43; 95% CI 5 1.54–3.82).

Table 3 shows the relationship between allergic rhinitis visit frequencies and developing NPC risk. There was no difference between the NPC risk in comparison persons and patients with <2 times the average allergic rhinitis visits per year. However, patients with 2 to 4 times or _4

times the average allergic rhinitis visits per year were significantly associated with increasingly developing NPC

risk (HR 5 5.25 and 14.80, respectively). These results also show that the increasing average frequencies of allergic rhinitis visits might be associated with increased NPC

risk (p value for trend < .0001). Moreover, after timedependent modifiers, an increase in allergic rhinitis visit

frequency was associated with an 11% increase in the risk of developing NPC (HR 5 1.11; 95% CI 5 1.08–

1.14).

We also used sensitivity analysis to validate the association

between allergic rhinitis occurrence and developing NPC risk in the study population with different follow-up durations (Table 4). These results suggest that the allergic rhinitis cohort was associated with a significantly higher risk of developing NPC compared to the comparison cohort, despite the study population having at least a 3-year follow-up duration. When the follow-up time was _1 year, the allergic rhinitis cohort had a dramatically increased risk of NPC compared with the comparison cohort (HR 5 15.77; 95% CI 5 6.38–38.97). The risk in the allergic rhinitis cohort was not significant when follow-up time was >4 years (HR 5 1.58; 95% CI 5 0.89–2.82). DISCUSSION

To the best of our knowledge, no other large-scale cohort studies have focused on the correlation between allergic rhinitis and NPC. In this study, the allergic rhinitis cohort exhibited a link to the subsequent NPC. We found a 2.33-fold increased risk of NPC among the allergic rhinitis participants after controlling for other critical covariates.

Some may argue that patients with allergic rhinitis and NPC represent overlapping symptoms and that these

increased risks are likely to be attributed to diagnostic confusion or misclassification. Therefore, we applied a sensitivity

analysis. The frequency of allergic rhinitis visits, which refers to allergic rhinitis severity, is correlated strongly with the risk of subsequent NPC. These findings confirm the result that patients with allergic rhinitis increase the risk of subsequent development of NPC in Taiwan.

Previous studies have investigated the correlation

between allergy and subsequent cancer risk. Their results

are discrepant and insufficiently strong to draw firm conclusions.

17–22 Koh et al ²⁰ examined the data from a

population-based cohort of 63,257 Singaporean Chinese, in which 954 cohort participants who had rhinitis or sinusitis developed lung cancer. This indicated a 59% increase in risk compared with those without rhinitis or sinusitis.

Hwang et al ²³ conducted a retrospective cohort study to

investigate whether the diagnoses of allergic rhinitis,

asthma, and atopic dermatitis are associated with an increased overall cancer risk. They found that the overall cancer risks in patients with allergic rhinitis do not significantly increase at a standardized incidence ratio of 1.02

(95% CI 5 0.98–1.05). In a further cancer site-specific

analysis, NPC showed an elevated risk (standardized incidence ratio 5 1.31; 95% CI 5 1.04–1.63). However, this

study did not include comorbidity data or a sensitivity test to adjust for possible confounding factors.

Although the association between allergic rhinitis and NPC is apparent, the mechanism underlying this association is a focus of ongoing research. A possible explanation is that chronic repeated airway stimulation and

inflammation, reduced mucociliary clearance, and epithelial cell changes after the deposition of allergens in the

nasopharynx may promote a malignant change after a certain induction time. ^24,25

The large sample size of this study enhances the statistical power of our results. The participants represented a

wide range of demographic characteristics. Therefore, we could have conducted stratified analyses based on age, sex, and comorbidities. The strengths of this study include its use of population-based data, which are highly representative of the general population. Nearly all of the participants

had a complete follow-up because of the

widespread coverage of the nationwide database, which records outpatient diagnoses and hospitalizations.

There are inherent limitations to the data presented in

this study, and the results should be interpreted with caution.

First, the NHIRD does not contain detailed information regarding status of Epstein–Barr virus infection,

smoking habits, alcohol consumption, dietary factors, or family history of NPC, all of which may be risk factors of NPC. We only include comorbidities that are available in the Taiwan LHID. Therefore, we selected the 3 most

common systemic diseases and the 2 common immune- associated diseases as the comorbidities. Second, the evidence

derived from a retrospective cohort study generally

has a lower statistical quality than that from randomized

trials. This is because of potential biases related to adjustments for confounding variables. Despite our meticulous

study design and control measures for confounding factors, a bias resulting from unknown confounders may

have affected our results. Third, all data in the NHIRD are anonymous. Thus, relevant clinical variables, including imaging results, pathology findings, and serum laboratory data, were unavailable for our study patient cases.

However, the data regarding allergic rhinitis or NPC diagnoses are reliable. Fourth, the follow-up was too short for

a causal association. Fifth, we cannot exclude the possibility of a screening effect because the risk is much

higher when the time lag is shorter than 1 year.

In conclusion, patients with allergic rhinitis might be

associated with subsequent NPC in Taiwan. Those who

had repeated visits for allergic rhinitis had even higher

risk for NPC. Physician should be aware of the link when

assessing patients with allergic rhinitis.