Methods
Ethical consideration
All studies were conducted and designed in accordance with the
Declaration of Helsinki using KNHIS-NSC data from Korea’s health
insurance service. This study also
passed the Institutional Review Board of [removed for blind peer
review] (IRB number 2019-04-010). Written informed consent was not
required because of the nature of the study, as the KNHIS-NSC data are
anonymized for research purposes.
Database
KNHIS is Korea’s health insurance service, established by the Korean
government in 1963, and since 1989, almost everyone in Korea has been
enrolled. This database includes information such as the identification
number of the individual, sex, age, residential area, and income
quintiles, as well as diagnostic code, treatment history, prescription
details including medication, and cost. KNHIS collects data by
anonymizing it with the person-id to replace the 13-digit identification
number of each individual. In KNHIS, the Korean Standard Classification
of Disease (KCD) is used as the diagnostic code, similar to the WHO
International Classification of Disease (ICD-10). The data used in this
study consisted of approximately 1,025,340 individuals randomly selected
from the total health insurance database in the period 2002 to 2013. The
selected sample cohort of approximately 1 million individuals was
divided into age (18 categories), sex (2 categories), income level (41
categories), and residential area (17 categories), and is representative
of the approximately 50 million people in Korea.
Study cohort
We took the washout period for antihistamine in 2002 to be 1 year. In
other words, patients who did not take antihistamine throughout 2002
were assigned to the patient group and the control group.
The antihistamine (AH) group consisted of patients who were diagnosed
with allergic rhinitis (AR) between January 1, 2003 and December 31,
2003, taking at least three days of antihistamine medication. The
inclusion criteria for the AH group were as follows: (1) patients
diagnosed with KCD codes J301, J302, J303, J304, and (2) allergen skin
test or
multiple
allergen simultaneous test (MAST), and (3) patients taking at least one
days of antihistamine medication.
Exclusion criteria were as follows: (1) individuals who had been
diagnosed with acute sinusitis or chronic sinusitis in 2002; or (2) who
had undergone endoscopic sinus surgery, septoplasty, or turbinoplasty in
2002; or (3) who took antihistamines in 2002 or (4) who had died in
2002.
In order to wash out the effect of sinusitis on antihistamines, patients
taking antihistamines in 2002 were excluded from the AH group, and
patients with sinusitis in 2003 were included in the AH group.
In the control group (non-antihistamine; non-AH group) of patients who
did not take antihistamines, their age, sex, residential area, and
household income were not significantly different from the AH group. The
1:4 propensity score matching was performed to obtain the non-AH group
in 2003.10 As a result, the control group (non-AH) had
1988 individuals and the AH group had 497.
Outcome variables and operational definition of chronic rhinosinusitis
(CRS) and antihistamine
Our study was largely divided into an antihistamine (AH) group and a
non-antihistamine (non-AH) control group in AR patients, and the
occurrence of sinusitis was checked. Antihistamine included the
following drugs: 1st generation: chlorpheniramine,
hydroxyzine hydrochloride, mequitazine, piprinhydrinate;
2nd generation: azelastine hydrochloride, bepotastine
besilate, emedastine difumarate, ketotifen fumarate, olopatadine
hydrochloride, desloratadine, fexofenadine hydrochloride,
levocetirizine, ebastine. The detailed sex, age, residential area, and
socioeconomic status of the study population were obtained from the
database. Patients were divided into four groups: those younger than 20
years, those aged between 20 and 40, those aged 40 to 60, and those aged
60 or older. The residential areas were divided into urban (large cities
– Seoul, Pusan, Daegu, Incheon, Daejeon, Gwangju, Ulsan) and rural (the
rest of the region). Income level was divided into low (lower than 30th
percentile), middle (30th to 70th percentile), and high (over 70th
percentile). The definition of end point in this study was the
occurrence of sinusitis. Thus, the elapsed time between the date of
first visit to the hospital and the date of sinusitis during the
follow-up period was calculated as naturally. Patients were censored
after December 31, 2013 if no sinusitis occurred. The risk of developing
sinusitis between the AH and non-AH groups was compared according to
age, gender, residential area, and income level. As a subgroup analysis
in the AH group, we defined as “1 week” if the consecutive duration of
antihistamine usage is less than 1 week, “2 week” if the duration is
less than 2 week, “1 month” if the duration is less than 4 week, and
“over 1 month” if the duration is equal or more than 4 week .
Statistical analysis
Data analysis was conducted between January 2019 and March 2019. The
Kaplan–Meier survival curve was used to examine the difference in
survival function between the study groups.11 We also
used log-rank tests to compare the difference between the two groups. To
determine whether antihistamine use increased sinusitis occurrence, we
used Cox proportional hazard regression analysis which adjusted for
other predictor variables to obtain the hazard ratio (HR) and 95%
CI.12 All statistical analyses were performed using R
version 3.53 and Stata version 14.