Professional Documents
Culture Documents
Department of Radiation Oncology, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan;
b,d,
Chang-Ming Chen a,
National Yang-Ming University, Taipei, Taiwan; c Management Ofce for Health Data, China Medical University Hospital, Taichung, Taiwan; d Division of Hematology-Oncology,
Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan; e Department of Otolaryngology-Head & Neck Surgery, Tri-Service General Hospital,
National Defense Medical Center, Taipei, Taiwan; f Departments of Public Health, China Medical University, Taichung, Taiwan; g Graduate Institute of Clinical Medicine Science and
School of Medicine, College of Medicine, China Medical University, Taichung, Taiwan; h Department of Nuclear Medicine and PET Center, China Medical University Hospital,
Taichung, Taiwan
a r t i c l e
i n f o
Article history:
Received 17 August 2012
Accepted 26 August 2012
Available online 14 September 2012
Keywords:
Otitis media
Nasopharyngeal cancer
Cohort study
a b s t r a c t
Purpose: To determine whether the diagnosis of otitis media (OM) in adults is associated with an
increased risk for the subsequent development of nasopharyngeal cancer (NPC) using a nationwide population-based retrospective study.
Methods and materials: We selected 13,513 adult patients that had been previously diagnosed with OM
between 2000 and 2005 from the Taiwan Longitudinal Health Insurance Database 2000 as the study
cohort, and randomly extracted the data of 135,130 participants matched by sex, age, and baseline year
for the comparison cohort. The follow-up period was terminated upon developing NPC, withdrawal from
the national health insurance system, or the end of 2009. Cumulative incidences and hazard ratios (HRs)
of NPC development were determined.
Results: The subsequent NPC incidence rates in the OM and comparison cohorts were 6.41 and 0.58 per
10000 person-years, respectively (adjusted HR, 11.04; 95% CI, 7.685.87; P < 0.0001). The NPC risk for
males was signicantly higher than that for females (adjusted HR = 3.24; 95% CI, 2.164.85). In both
female and male patients, the diagnosis of OM was associated with a signicantly increased risk for
NPC (adjusted HR, 11.91 vs. 10.78, respectively). Among the OM cohort, 62 participants were subsequently diagnosed with NPC, with 71% of them occurring within 1 year following the diagnosis of OM.
However, even after 5-year follow-up, the OM cohort still displayed a higher risk for NPC (adjusted
HR = 2.50). Stratied by the frequency of OM episodes, more than one episode per year had a signicantly
greater risk of developing NPC, compared with the comparison cohort (HR = 29.22; 95% CI, 20.1942.27).
Conclusion: We found that adult OM is a warning sign for the development of NPC in Taiwan, with
approximately an 11-fold higher risk for adult OM patients. We recommend that OM patients undergo
follow-up examinations for at least 5 years. To extrapolate our ndings, further studies are warranted
in other areas in which NPC is endemic.
2012 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 104 (2012) 338342
of the opening of the Eustachian tube by the tumor or adenoid tissue may lead to OM with effusion. Thus, much emphasis is placed
on the exclusion of NPC in adult patients with OM. However, it remains unclear whether adult OM is an indicator of subsequent
NPC, and whether OM patients should undergo subsequent regular
medical examinations as a high-risk group.
NPC shows a distinct geographical and racial distribution. It is
rare in most parts of world, but common in southern China, Hong
Kong, and Taiwan. According to the 2008 cancer report released by
the Taiwan Department of Health, the incidence of NPC was 9.99
per 100000 for men and 3.47 per 100000 for women. It is the ninth
most common cause of cancer-related death for men and the 14th
for women in Taiwan. The Taiwan National Health Insurance (NHI)
program was initiated in 1996, with 97% of the hospitals and clinics throughout Taiwan under contract with the system by the end
of 1996 [2]. By 1998, the health care of almost 99% of the population of Taiwan was covered by the NHI. The NHI patient records
provide a unique opportunity to examine our hypothesis that a
diagnosis of OM in adults is associated with an increased risk for
subsequent development of NPC using a nationwide populationbased cohort study.
Patients and methods
Data source
The Taiwan National Health Research Institute established and
managed the National Health Insurance Research Database (NHIRD)
which includes the reimbursement claim data for the Taiwan NHI
program. All personal identication information is encrypted before
being released to the public to protect patient privacy.
Our research used the Longitudinal Health Insurance Database
(LHID), a subset of the NHIRD. LHID is composed of historical claim
data for one million claimants randomly sampled from the total insured population between 1996 and 2000. Anonymous identication numbers are used to link each claimants demographic
information, including sex, birth date, occupation, residential area,
and registry of medical services. The disease diagnosis that was
used in our study was dened by the International Classication
of Diseases, Ninth Revision, Clinical Modication (ICD-9-CM) from
outpatient data, inpatient data, and the registry of catastrophic illness. Our study was approved by the Ethics Review Board of the
China Medical University (CMU-REC-101-012).
Study population
Our study used a population-based retrospective approach. The
OM cohort included participants that were initially diagnosed with
OM (ICD-9-CM 381.0-381.4 and ICD-9-CM 382) between 2000 and
2005, and the baseline was set as the date of the initial OM diagnosis. Ten comparison cohort participants were randomly selected for
each OM cohort participant. Comparison cohort participants were
matched by sex, age, and baseline year. The event of the study
was dened as subsequent NPC based on the diagnosis code,
ICD-9-CM 147, from the registry of catastrophic illness. We excluded patients with a history of cancer before the baseline year,
and those who were aged 20 years or under during the baseline
year. The follow-up period was terminated upon developing NPC,
withdrawal from the insurance system, or the end of 2009.
The demographic data included sex, age, occupation, and residence area. Occupation was classied into three groups: White collar, blue collar, and others. The urbanization of Taiwan cities
grouped into seven levels that were based on the following indices:
(1) Population density (people/km2); (2) the population ratio of
different educational levels; (3) the population ratio of elderly persons; (4) the population ratio of agriculture workers and the number of physicians per 100000 people [3]. The subjects in levels 5, 6
and 7 were small so that these levels were combined into level 4.
Level 1 was considered to represent the highest degree of urbanization and level 4 represented the lowest.
evaluated by the KaplanMeier method, and the differences between the incidence curves were evaluated by the log-rank test.
The Coxs proportional hazards regression model, adjusted for potential confounding factors, was used to estimate the hazard ratio
(HR) and condence interval (CI) for the OM cohort and the comparison cohort. The average OM frequency was calculated as the
total number of OM diagnoses during the follow-up period divided
by the follow-up duration in years. The average OM frequency was
separated into 3 groups by percentile (33rd percentile and 66th
percentile). To measure the association between the average OM
frequency and the risk of NPC, we estimated the risk in every level
of average OM frequency, and the OM frequency was considered a
continuous variable to evaluate the trends using the Coxs proportional hazards regression model.
Data management and analysis were performed using SAS version 9.1 software (SAS Institute, Cary, NC, USA), and the cumulative
incidence curve was plotted using R software (R Foundation for
Statistical Computing, Vienna, Austria). P values for the two-tailed
tests that were less than .05 were considered to represent signicant differences among the data sets.
Results
Our study evaluated 13,513 OM participants and 135,130 comparison cohort participants between 2000 and 2005 (Table 1). The
average age (47.5 y) and sex ratio were identical between the two
cohorts. In both cohorts, approximately 30% of the participants
lived in the highest urbanization level, and approximately 50%
were classied as white-collar.
The NPC incidence rate in the OM cohort was 6.41 per 10000
person-years, and was approximately 11-fold higher than the
NPC incidence rate in the comparison cohort (0.58 per 10000 person-years; Table 2). The NPC cumulative incidence curve showed
that the OM cohort had a signicantly higher risk for NPC than
the comparison cohort (P value for log-rank test <0.0001; Fig. 1).
After adjusting for potential confounders, the HR of subsequent
NPC in the OM cohort was 11.04 (95% CI, 7.6815.87), compared
with the comparison cohort. We also applied sensitivity analysis
to measure the NPC risk in the study population throughout the
follow-up duration. While 71% of the NPC events occurred in the
OM cohort within 1 year of OM diagnosis, approximately 10% of
the NPC events in the comparison cohort occurred within the same
period. These results suggest that the OM cohort displayed a significantly increased risk of NPC, compared with the comparison
cohort.
Table 1
Baseline demographic status and comorbidity compared between Comparison and
otitis media cohorts.
Variable
Statistical analysis
We used the chi-square test for category variables and the t-test
for continuous variables to assess the difference in baseline demographic characteristics between the OM cohort and the comparison
cohort participants. The total NPC incidence and the demographicspecic NPC incidence was calculated per 10000 person-years. The
cumulative NPC incidence curves for the study cohorts were also
340
339
t-Test.
Comparison group
N = 135130 (%)
N = 13513 (%)
47.5 (15.7)
47.5 (15.7)
50790 (37.6)
30450 (22.5)
53890 (39.9)
5079 (37.6)
3045 (22.5)
5389 (39.9)
74120 (54.9)
61010 (45.1)
7412 (54.9)
6101 (45.1)
p-Value
0.99
1.0000
1.0000
0.0012
41119
39396
24003
30610
(30.4)
(29.2)
(17.8)
(22.7)
3950
3872
2522
3169
(29.2)
(28.7)
(18.7)
(23.5)
0.0005
70160 (51.9)
46206 (34.2)
18764 (13.9)
6918 (51.2)
4827 (35.7)
1768 (13.1)
Table 2
Incidence of nasopharyngeal cancer and multivariate Cox proportional hazards regression analysis measured hazard ratio for study cohort.
Variable
Total
Time
>1
>2
>3
>4
>5
lag
year
years
years
years
years
Comparison group
HR (95% CI)a
Event
PYs
Rate
Event
PYs
Rate
55
948726
0.58
62
96778
6.41
11.10 (7.7215.96)
11.04 (7.6815.87)
49
45
40
31
23
947824
945119
940786
935001
859021
0.52
0.48
0.43
0.33
0.27
18
14
13
11
6
96730
96579
96308
95813
88225
1.86
1.45
1.35
1.15
0.68
3.58
3.03
3.15
3.43
2.51
3.56
3.01
3.14
3.41
2.50
(2.096.14)
(1.665.51)
(1.695.90)
(1.736.83)
(1.026.17)
(2.076.11)
(1.655.48)
(1.685.87)
(1.716.79)
(1.026.14)
Crude HR.
Model adjusted for age, sex, urbanization level or occupation.
Discussion
OM primarily occurs in childhood. Medical studies of OM have
overwhelmingly focused on the epidemiology, the etiology, the
immunology, and the management of OM in children. The most
Table 3
Demographic-specic incidence of nasopharyngeal cancer and multivariate Cox proportional hazards regression analysis measured hazard ratio for study cohort.
Variable
Age group
640
4150
>50
Sex
Female
Male
Urbanization level
1
2
3
4
Occupation
White collar
Blue collar
Others
Comparison group
HR (95% CI)a
4.54
8.06
7.29
10.45 (5.2820.68)
9.94 (5.1719.10)
12.60 (7.1222.29)
10.55 (5.3320.88)
9.96 (5.1819.15)
12.56 (7.1022.22)
Ref
1.65 (1.032.66)
1.26 (0.801.98)
53747
43030
3.35
10.23
11.88 (5.9923.57)
10.78 (7.0316.55)
11.91 (6.0023.63)
10.78 (7.0316.55)
Ref.
3.24 (2.164.85)
17
14
13
18
28355
27806
18131
22486
6.00
5.03
7.17
8.00
8.33 (4.4015.79)
10.02 (4.7721.01)
11.07 (4.9624.70)
18.95 (8.5142.17)
8.25 (4.3515.64)
9.92 (4.7320.81)
11.08 (4.9624.73)
19.31 (8.6742.99)
Ref.
0.69 (0.421.13)
0.99 (0.591.65)
0.75 (0.451.26)
26
31
5
49880
34481
12417
5.21
8.99
4.03
9.99 (5.8017.21)
13.92 (8.0024.22)
6.49 (2.1219.83)
9.85 (5.7216.97)
14.12 (8.1124.56)
6.53 (2.1319.96)
Ref.
1.63 (1.082.46)
0.99 (0.531.82)
Event
PYs
Rate
Event
PYs
16
18
21
365663
221627
361436
0.44
0.81
0.58
17
18
27
37420
22322
37036
15
40
529038
419688
0.28
0.95
18
44
21
14
11
9
291254
277793
167999
211665
0.72
0.50
0.65
0.43
26
21
8
497557
323083
128085
0.52
0.65
0.62
Rate
a
b
*
Table 4
Incidence of nasopharyngeal cancer and multivariate Cox proportional hazards
regression analysis measured hazard ratio for study cohort by average frequencies of
otitis media diagnosis.
341
Event
55
PYs
Rate
948726
37735
0.58
0.27
0.61
24648
0.81
>1
59
34393
17.15
Model 1
(95% CI)
Ref
0.47 (0.06
3.38)
1.40 (0.34
5.73)
29.26
(20.26
42.25)
Model 2
(95% CI)
Ref
0.47 (0.06
3.40)
1.39 (0.34
5.69)
29.22
(20.19
42.27)
342
[6] Robinson PM. Secretory otitis media in the adult. Clin Otolaryngol Allied Sci
1987;12:297302.
[7] Sham JS, Wei WI, Lau SK, et al. Serous otitis media. An opportunity for early
recognition of nasopharyngeal carcinoma. Arch Otolaryngol Head Neck Surg
1992;118:7947.
[8] Mantovani A. Cancer: inaming metastasis. Nature 2009;457:367.
[9] Coussens LM, Werb Z. Inammation and cancer. Nature 2002;420:8607.
[10] Grivennikov SI, Greten FR, Karin M. Immunity, inammation, and cancer. Cell
2010;140:88399.
[11] Chang
ET, Adami
HO.
The
enigmatic
epidemiology
of
nasopharyngeal
carcinoma. Cancer Epidemiol Biomarkers Prev 2006;15:176577.
[12] Henderson
BE, Louie E, SooHoo
Jing J. Risk factors
associated
with
nasopharyngeal carcinoma. N Engl J Med 1976;295:11016.
[13] Yu MC, Garabrant DH, Huang TB, et al. Occupational and other non-dietary risk
factors for nasopharyngeal carcinoma in Guangzhou, China. Int J Cancer
1990;45:10339.
[14] Yuan JM, Wang XL, Xiang YB, et al. Non-dietary risk factors for nasopharyngeal
carcinoma in Shanghai, China. Int J Cancer 2000;85:3649.
[15] Zhu
K,
Levine
RS,
Brann
EA,
et
al.
Case-control
study
evaluating
homogeneity
and
heterogeneity
of
risk
factors
between
sinonasal
and
nasopharyngeal cancers. Int J Cancer 2002;99:11923.
[16] Bartsch H, Ohshima H, Pignatelli B, et al. Endogenously formed N-nitroso
the
compounds
and
nitrosating
agents
in
human
cancer
etiology.
Pharmacogenetics 1992;2:2727.
[17] Mirvish SS. Role of N-nitroso compounds (NOC) and N-nitrosation in etiology
of gastric, esophageal, nasopharyngeal and bladder cancer and contribution to
cancer of known exposures to NOC. Cancer Lett 1995;93:1748.
[18] Chien YC, Chen JY, Liu MY, et al. Serologic markers of Epstein-Barr virus
infection and nasopharyngeal carcinoma in Taiwanese men. N Engl J Med
2001;345:187782.