G Model

ANL-2025; No. of Pages 5

Auris Nasus Larynx xxx (2015) xxxxxx

Contents lists available at ScienceDirect

Auris Nasus Larynx

journal homepage: www.elsevier.com/locate/anl

Factors associated with prolonged duration of post-tympanoplasty

local treatment in adult chronic otitis media patients: A retrospective
observational study using a Japanese inpatient database
Sayaka Suzuki a,*, Hideo Yasunaga b, Hiroki Matsui b, Kiyohide Fushimi c,
Tatsuya Yamasoba a
a
Department of Otolaryngology, Head and Neck Surgery, Faculty of Medicine, The University of Tokyo, Tokyo 1130033, Japan
b
Department of Clinical Epidemiology and Health Economics, School of Public Health, The University of Tokyo, Tokyo 1130033, Japan
c
Department of Health Policy and Informatics, Tokyo Medical and Dental University Graduate School of Medicine, Tokyo 1130034, Japan

A R T I C L E I N F O A B S T R A C T

Article history: Objective: The occurrence of persistent infection following tympanoplasty has been reported in many
Received 17 July 2015 studies, and it is important to know the risks for site infection after tympanoplasty. In this study, we
Accepted 8 September 2015 aimed to explore the factors affecting early wound complications after tympanoplasty for chronic otitis
Available online xxx
media.
Methods: We conducted a retrospective cohort study using the Diagnosis Procedure Combination
Keywords: database. Data on a total of 13,094 adult patients from 420 acute-care hospitals who received
Tympanoplasty
tympanoplasty for chronic otitis media from 2010 to 2013 were extracted. The duration (days) of
Chronic otitis media
Local wound treatment
postsurgical local wound treatment was measured as an outcome, because this duration was assumed to
Early complications be prolonged by the existence of wound infection. The associations between treatment duration and
background characteristics (age, sex, body mass index, smoking status, diabetes mellitus, use of
antithrombotic agents, with or without cholesteatoma, duration of anesthesia, academic hospital or not,
and hospital volume) were assessed by multivariable linear regression analyses, fitted with a generalized
estimating equation to adjust for within-hospital clustering.
Results: The median treatment duration in each hospital was 8 days (interquartile range: 711). Factors
significantly associated with longer treatment duration were: older age (0.2 days for 10-year increase),
use of antithrombotic agents during hospitalization (1.8 days), and prolonged duration of anesthesia (vs.
<120 min of anesthesia, additional 1, 2, 3, and 4 days for 120179, 180239, 240299, and !300 min of
anesthesia, respectively). Body mass index and smoking status were not significantly associated with
treatment duration.
Conclusions: Older age, antithrombotic agents during hospitalization, and longer anesthesia time were
independently associated with early local wound complications after tympanoplasty for chronic otitis
media.
! 2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction ossicular chain for improvement of hearing function, and repair

Tympanoplasty (TP) is a safe and common procedure for
chronic otitis media (COM) with or without cholesteatoma to
remove inflammatory lesions of the middle ear, reconstruct the
* Corresponding author at: Department of Otolaryngology, Head and Neck
Surgery, Faculty of Medicine, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku,
Tokyo 1130033, Japan. Tel.: +81 3 3815 5411; fax: +81 3 3814 9486.
E-mail address: sayasuzuki-tky@umin.ac.jp (S. Suzuki).
perforations of the tympanic membrane [1,2].
A postauricular incision is usually made to harvest an
autologous cartilageperichondrium graft or temporal fascia graft.
Early postsurgical local complications, such as wound infection,
otorrhea from the outer ear canal caused by infection of the middle
ear or mastoid, and graft failure, are troublesome, because they can
cause excess healthcare resource utilization for postsurgical local
treatment, as well as patient discomfort and inconvenience [3].
Numerous studies assessing the occurrence of persistent
infection following TP have produced inconsistent findings, arising

http://dx.doi.org/10.1016/j.anl.2015.09.006
0385-8146/! 2015 Elsevier Ireland Ltd. All rights reserved.

Please cite this article in press as: Suzuki S, et al. Factors associated with prolonged duration of post-tympanoplasty local treatment in
adult chronic otitis media patients: A retrospective observational study using a Japanese inpatient database. Auris Nasus Larynx (2015),
http://dx.doi.org/10.1016/j.anl.2015.09.006
G Model
ANL-2025; No. of Pages 5
2 S. Suzuki et al. / Auris Nasus Larynx xxx (2015) xxxxxx
from differences in the definition of infection, timing of evaluation,
and study period, as well as relatively small sample sizes [35]. To
our knowledge, there have been no studies on the duration of
postsurgical local treatment for site infection. Furthermore, the
factors associated with prolonged duration of such treatment
remain unknown.
Using a nationwide inpatient database in Japan, the present
study aimed to explore the factors affecting early complications
after TP for COM. We adopted the duration of postsurgical local
treatment as an outcome, as an alternative to the severity of
postsurgical infection to be treated.
2. Methods
2.1. Ethical considerations
We have obtained study approval from the Institutional Review
Board of The University of Tokyo. Because of the anonymous nature
of the data, the requirement for informed consent was waived.
2.2. Data source
We extracted patient data from the Diagnosis Procedure
Combination (DPC) database, a national inpatient database in
Japan. The DPC database includes administrative claims data and
discharge abstract data collected for all inpatients (approximately
6.8 million patients in 2012) discharged from more than
1000 participating acute-care hospitals across Japan. The details
are described elsewhere [6].
For each patient, the database includes the following informa-
tion: age; sex; patient characteristics (weight, height, Brinkman
Index); main diagnoses, comorbidities at admission, and compli-
cations after admission encoded by International Statistical
Classification of Diseases (ICD)-10 codes and text data in Japanese;
surgical interventions and medical procedures encoded with
original Japanese codes; daily records of drug administration;
duration of anesthesia; type of hospital (academic or non-
academic); length of stay; and discharge status.
All 82 academic hospitals in Japan are obliged to participate in
the DPC database, while the participation of community hospitals
is voluntary. The codes related to surgeries, procedures, medica-
tions, and anesthesia are almost complete because they are
compulsory for healthcare cost reimbursement. To maximize the
accuracy of these data, the physicians in charge are required to
record the information on diagnoses and therapies with reference
to the patients medical charts [7,8].
2.3. Patient selection and characteristics
Data for patients with diagnoses of otitis media (ICD-10 codes:
H65x, H66x, H67x) on admission who underwent TP between July
2010 and March 2013 (33 months in total) were extracted.
With the aim of selecting a homogeneous target population with
COM, we excluded patients with the following comorbidities: (i)
otitis media possibly accompanying acute inflammation in middle
or inner ear (ICD-10 codes: H66.0, H830); (ii) benign or malignant
tumor (D140, D180, D33.3, D43.3, Q85.0, Cxx); (iii) fracture of skull
or facial bones (S02x); and (iv) congenital malformation of ear
(Q16x, Q17x). We also excluded: (v) patients who received surgical
wound treatment for other surgical sites to selectively identify local
wound treatment for TP; (vi) patients who received revision surgery,
including second-look procedure for cholesteatoma, because of
potentially impaired wound healing after second or subsequent
surgery; and (vii) patients under 19 years of age.
The following background characteristics of the patients were
assessed: age; sex; body mass index (BMI; kg/m2); smoking status
Please cite this article in press as: Suzuki S, et al. Factors associated with prolonged duration of post-tympanoplasty local treatment in
adult chronic otitis media patients: A retrospective observational study using a Japanese inpatient database. Auris Nasus Larynx (2015),
http://dx.doi.org/10.1016/j.anl.2015.09.006
(current or ex-smoker, non-smoker); diabetes mellitus (DM; ICD-
10 codes: E10E14) on admission; middle ear cholesteatoma (H71,
cholesteatoma written in Japanese); use of antithrombotic
agents during hospitalization; duration of anesthesia; and type
of hospital (academic or non-academic). Antithrombotic agents
contained antiplatelet agents (aspirin, cilostazol, ticlopidine,
clopidogrel, sarpogrelate, beraprost, icosapentate) and anticoagu-
lant agents (warfarin, dabigatran, edoxaban, rivaroxaban, apix-
aban). Hospital volume (HV) was calculated based on the annual
number of TPs performed during the study period in each hospital.
We trisected hospitals into three HV groups ("27, 2851, !52).
According to the World Health Organization definitions, we
classified BMI as underweight (<18.50 kg/m2), normal weight
(18.5024.99 kg/m2), or overweight and obese (!25.00 kg/m2)
[9]. Normal weight was separated into lownormal (18.50
22.99 kg/m2) and highnormal (23.0024.99 kg/m2).
2.4. Outcome measurement
The primary outcome was duration (days) of postoperative
treatment for surgical wound and middle ear (treatment
duration). Postoperative treatment included sterilization or gauze
dressing for surgical wound, or irrigation of the external ear.
2.5. Statistical analysis
The continuous variables were compared using a t-test. To
investigate the correlations between the categorical variables, we
calculated Spearmans rho correlation coefficients. We also
checked the variance inflation factors for each independent
variable, to explore the multicollinearity between the independent
variables. A variance inflation factor of >10 was defined to indicate
multicollinearity. Multivariable linear regression analyses were
performed to investigate the relationships between patient
characteristics and duration of postoperative treatment for
surgical wound and middle ear. Because the study was based on
a multicenter retrospective design, data were structured hierar-
chically, including patient level and hospital level. Therefore, the
patient characteristics or physician practice patterns within the
same hospital could have been clustered. To adjust for such within-
hospital clustering, we fitted a generalized estimating equation to
the multivariable linear regression analyses [10].
The threshold for significance was set at p < 0.05. All statistical
analyses were performed using Statistical Package for Social
Sciences software version 20.0 (IBM SPSS Corp., Armonk, NY, USA).
3. Results
We identified 16,486 patients who underwent TP for otitis
media during the study period. Among them, a total of
3392 patients were excluded for the following reasons: acute
suppurative otitis media (n = 2); labyrinthitis (n = 3); any type of
malignancy (n = 186); benign middle ear tumor (n = 9); glomus
tympanicum tumor (n = 6); benign neoplasm of cranial nerves
(n = 2); neoplasm with uncertain or unknown behavior of cranial
nerves (n = 19); fracture of skull or facial bones (n = 5); congenital
malformation of ear (n = 124); surgical wound treatment for other
surgical site (n = 453); revision surgery (n = 86); and age under
19 years (n = 3019). No patients with diagnoses of neurofibroma-
tosis or von Recklinghausen disease were identified. Finally, we
selected 13,094 eligible patients from 420 hospitals.
Table 1 shows the patient characteristics. The mean (#SD) age
was 54.8 # 16.2 years. The sex ratio was close to one. More than one-
quarter of the patients were overweight or obese. Among the total
patients, 60.8% had cholesteatoma, 6.8% had DM, and 1.8% received
antithrombotic agents during hospitalization. Most of the TPs were
G Model
ANL-2025; No. of Pages 5

S. Suzuki et al. / Auris Nasus Larynx xxx (2015) xxxxxx 3

Table 1 Table 2
Characteristics of the 13,094 Patients. Results of the multivariable linear regression analyses for treatment duration
(days).
n (%)
Coefficient 95% confidence p-Value
Age (years)
interval
2029 1185 (9.0)
3039 1841 (14.1) Age (10-year increase) 0.2 0.1 to 0.2 <0.001
4049 1468 (11.2) Sex (male) 0.1 $0.06 to 0.24 0.26
5059 2150 (16.4) BMI category
6069 3950 (30.2) "18.49 0.2 $0.01 to 0.50 0.06
7079 2215 (16.9) 18.5022.99 Reference
!80 285 (2.1) 23.0024.99 0.1 $0.07 to 0.28 0.24
Sex (male) 6494 (49.6) !25.00 0.2 $0.00 to 0.35 0.05
BMI category Unspecified $0.6 $1.3 to 0.1 0.10
"18.49 827 (6.3) Smoking category
18.5022.99 5655 (43.2) Non-smoker Reference
23.0024.99 2550 (19.5) Current/ex-smoker $0.03 $0.3 to 0.2 0.80
!25.00 3403 (26.0) Unspecified 0.3 $0.3 to 0.9 0.28
Unspecified 659 (5.0) Diabetes mellitus $0.1 $0.5 to 0.3 0.68
Smoking category Cholesteatoma 0.2 $0.04 to 0.51 0.09
Non-smoker 7736 (59.1) Antithrombotic agents 1.8 0.9 to 2.8 <0.001
Current/ex-smoker 3503 (26.8) Duration of anesthesia (min)
Unspecified 1855 (14.2) "119 Reference
Diabetes mellitus 895 (6.8) 120179 1.5 0.8 to 2.2 <0.001
Cholesteatoma 7926 (60.5) 180239 2.1 1.1 to 3.1 <0.001
Antithrombotic agents 233 (1.8) 240299 2.8 1.6 to 3.9 <0.001
Duration of anesthesia (min) !300 3.6 2.3 to 4.9 <0.001
"119 779 (5.9) Unspecified 2.6 1.1 to 4.2 0.001
120179 2626 (20.1) Academic hospitals $0.6 $1.6 to 0.4 0.25
180239 3679 828.1) Hospital volume
240299 2870 (21.9) "27 Reference
!300 2433 (18.6) 2851 0.3 $0.7 to 1.4 0.50
Unspecified 707 (5.4) !52 0.5 $0.7 to 1.6 0.43
Academic hospitals 7071 (54.0)
BMI indicates body mass index.
Hospital volume
"27 4473 (34.2)
2851 4287 (32.7)
!52 4334 (33.1)
were significantly associated with prolonged duration of treatment
BMI indicates body mass index.
for post-TP wound infection: higher age; use of antithrombotic
agents during hospitalization; and longer anesthesia time. To our
knowledge, this is the first study to investigate the factors
performed under general anesthesia and required more than 2 h. The influencing early postsurgical local complications for COM. In this
median HV was 39 TPs per year (interquartile range (IQR): 1767). study, the patients stayed in hospital and were observed
TP with mastoidectomy (n = 6857) had significantly longer postoperatively for a median period of 10 days. This follow-up
duration of anesthesia than TP without mastoidectomy (n = 6237) period was sufficiently long to capture the length of treatment
(259 # 86 min vs. 205 # 151 min, p = 0.001). The treatment duration duration (median: 8 days).
among the hospitals was distributed with a median of 8 days (IQR: 7 This follow-up duration for a single hospitalization appears
11). The median overall length of stay was 12 days (IQR: 915), and relatively long compared with those in the United States.
the median post-TP length of stay was 10 days (IQR: 814). According to the Organization for Economic Cooperation and
The results of the multivariable linear regression analyses are Development in 2014, the national average length of stay in acute-
shown in Table 2. Longer treatment duration was significantly care hospitals in Japan is 17.5 days, compared with only 4.8 days in
associated with higher age and use of antithrombotic agents the United States [11].
during hospitalization. Although the differences were not signifi- This substantial discrepancy arises because of the differences in
cant, underweight patients and overweight or obese patients were the nature of hospital care between Japan and the United States.
more likely to receive longer treatment than lownormal weight Specifically, most hospitals in Japan provide subsequent nursing
patients. Duration of anesthesia had a significant linear association care as well as early postoperative care for each patient during a
with longer treatment duration. Compared with patients under single hospitalization, while most hospitals in the United States
anesthesia for "2 h, patients under anesthesia for 120179, 180 tend to admit patients to the hospital on the day of surgery and
239, 240299, and !300 min required an additional 1, 2, 3, and discharge them to their home or rehabilitation facilities within
4 days for postsurgical local treatment, respectively. Patients with several postoperative days.
cholesteatoma were more likely to have longer treatment duration, There have been no previous studies regarding the effects of
although the difference was not significant. Smoking habit, DM, BMI on early complications after TP. A recent report on
HV, and academic hospital status were not significantly associated gastrointestinal cancer surgery showed an association between
with treatment duration. The Spearmans rho correlation coeffi- higher BMI and higher incidence of wound infection [12]. In the
cients were all less than 0.4, indicating no significant associations present study, both overweight (BMI > 25.0) and underweight
between the dependent variables. The variance inflation factors (BMI < 18.5) patients received relatively longer duration of local
were all less than 1.5. treatment, but the differences did not reach statistical significance.
Unlike abdominal surgeries, procedural difficulties related to
4. Discussion excess fat and less blood flow in adipose tissue would not
significantly affect wound infections for TP.
The present large-scale nationwide study involving 13,094 TPs A recent meta-analysis across various surgical specialties
for COM between 2010 and 2013 showed that the following factors showed higher risks of healing complications after surgery for

Please cite this article in press as: Suzuki S, et al. Factors associated with prolonged duration of post-tympanoplasty local treatment in
adult chronic otitis media patients: A retrospective observational study using a Japanese inpatient database. Auris Nasus Larynx (2015),
http://dx.doi.org/10.1016/j.anl.2015.09.006
G Model
ANL-2025; No. of Pages 5
4 S. Suzuki et al. / Auris Nasus Larynx xxx (2015) xxxxxx
both smokers compared with non-smokers and former-smokers
(smoking cessation >4 weeks before surgery) compared with
never-smokers [13]. A previous retrospective study on TP patients
found that smoking was associated with a three-times higher risk
of long-term graft failure [14]. However, our results for treatment
duration did not differ significantly between patients with a
smoking habit (current or former) and non-smokers. We speculate
that the effects of a smoking habit differ between the early and
long-term postoperative periods.
Regarding the effects of DM on postsurgical complications,
previous studies on head and neck cancer surgery have shown
conflicting results [15,16]. In the present study, no significant
associations were found for DM and short-term outcomes after TP.
However, further studies are needed to evaluate both early and late
local complications after TP.
4.1. Antithrombotic agents
Patients taking antithrombotic agents will inevitably increase
in many developed countries that are experiencing aging societies.
Complications caused by continuation of antithrombotic agents
during the perioperative period can vary according to the extent
(minor or major surgery) and site of the invasive procedure. A
recent meta-analysis on cutaneous surgeries reported higher
incidences of postoperative complications, including persistent
oozing, serous oozing, and hematoma, in patients with antith-
rombotic agents than in patients without these agents [17]. A
retrospective analysis on 1144 core needle biopsies in breast
clinics revealed a significant increase for the occurrence of
bruising, but an insignificant effect for hematoma, in patients
with antithrombotic agents compared with patients without these
agents [18]. A large-scale prospective study regarding skin cancer
surgery (5950 lesions) found that higher age (!67 years), warfarin,
and surgery on or around the ear were independent risk factors
for bleeding complications (any type of postoperative bleeding
and hematoma) [19]. The present study revealed that use of
antithrombotic agents during hospitalization resulted in nearly
2 additional days of treatment duration, presumably because of
hematoma or oozing in the early post-TP period. The continuation
or discontinuation of antithrombotic agents should be decided for
each patient, taking the risks and benefits of the medication into
consideration.
4.2. Duration of anesthesia
The DPC database contained neither information on duration of
surgery nor type of TP (type I, II, III, or IV) performed, which are
potentially correlated with the severity of COM and the difficulty of
the procedure. Therefore, we evaluated the duration of anesthesia
as an alternative to examining the type of TP. It is worth
mentioning that the duration of anesthesia had a significant
linear correlation with prolonged treatment duration, even after
adjustment for other background factors. This finding was
concordant with that in a previous study [15].
Procedures for intractable COM would require a longer time to
remove impaired tissue and reconstruct an ossicular chain,
compared with procedures for mild COM. In addition, longer use
of surgical instruments (e.g., hook, retractor) to obtain a sufficient
operating field would cause excess pressure to the surrounding
tissue, leading to edema and venous congestion that can
potentially impair wound healing.
In the present study, treatment duration did not differ
significantly among the three HV groups. A previous controlled
study showed no significant association between surgeon status
and aural wound closure after TP or mastoidectomy [4]. Mean-
while, a recent large-scale database study from Taiwan reported
Please cite this article in press as: Suzuki S, et al. Factors associated with prolonged duration of post-tympanoplasty local treatment in
adult chronic otitis media patients: A retrospective observational study using a Japanese inpatient database. Auris Nasus Larynx (2015),
http://dx.doi.org/10.1016/j.anl.2015.09.006
that higher annual numbers of TPs per hospital or per surgeon were
significantly associated with reduced length of stay and hospitali-
zation costs [20]. Because our data lacked information on the
number of specialists for ear surgery in each hospital, no definite
conclusion could be reached for the association between HV and
treatment duration.
4.3. Limitations
This study had several limitations that need to be acknowl-
edged. First, this was a retrospective cohort study and the
treatment assignment was not random. Unrecorded confounding
factors including severity of preoperative inflammation, type of
incision (postauricular, endaural, transcanal), and use of graft
(temporal, cartilage, none) may have affected the outcome. Second,
comorbidities are generally less accurately recorded in adminis-
trative claims databases than in planned prospective studies.
Third, there may be some patients who discharged earlier or later
than the appropriate hospital discharge because of their social
situation. In such patients, treatment duration would be under-
estimated or overestimated in the current study. Furthermore,
symptoms (otalgia, ear fullness) or signs (otorrhea, wound
swelling, hematoma) might have been under-recorded in the
database. Thus, we focused on treatment duration, as an objective
and accurately recorded outcome, in the present study.
5. Conclusions
Higher age, antithrombotic agents during hospitalization, and
longer anesthesia time were independent risk factors for
prolonged treatment duration, which reflects early postoperative
local wound complications after TP for COM.
Conflicts of interest
None reported.
Acknowledgements
Funding sources: This work was supported by grants for
Research on Policy Planning and Evaluation from the Ministry of
Health, Labour and Welfare, Japan (grant numbers: H27-Policy-
Designated-009 and H27-Policy-Strategy-011).
References
[1] Tos M. Assessment of the results of tympanoplasty. J Laryngol Otol
1972;86:487500.
[2] Tos M. Modification of combined-approach tympanoplasty in attic cholestea-
toma. Arch Otolaryngol 1982;108:7728.
[3] Verschuur HP, de Wever WWH, van Benthem PPG. Antibiotic prophylaxis in
clean and clean-contaminated ear surgery. Cochrane Database Syst Rev 2004.
CD003996.pub2.
[4] Willatt DJ, Durham L, Ramadan MF, Bark-Jones N. A prospective randomized
trial of suture material in aural wound closure. J Laryngol Otol
1988;102:78890.
[5] Khan I, Jan AM, Shahzad F. Middle-ear reconstruction: a review of 150 cases. J
Laryngol Otol 2002;116:4359.
[6] Matsuda S, Fujimori K, Kuwabara K, Ishikawa KB, Fushimi K. Diagnosis
procedure combination as an infrastructure for the clinical study. Asian Pac
J Dis Manag 2011;5:817.
[7] Iwagami M, Yasunaga H, Doi K, Horiguchi H, Fushimi K, Matsubara T, et al.
Postoperative polymyxin B hemoperfusion and mortality in patients with
abdominal septic shock: a propensity-matched analysis. Crit Care Med
2014;42:118793.
[8] Sumitani M, Uchida K, Yasunaga H, Horiguchi H, Kusakabe Y, Matsuda S, et al.
Prevalence of malignant hyperthermia and relationship with anesthetics in
Japan: data from the diagnosis procedure combination database. Anesthesiol-
ogy 2011;114:8490.
[9] WHO Expert Consultation. Appropriate body-mass index for Asian populations
and its implications for policy and intervention strategies. Lancet
2004;363:15763.
G Model
ANL-2025; No. of Pages 5

S. Suzuki et al. / Auris Nasus Larynx xxx (2015) xxxxxx 5

[10] Hubbard AE, Ahern J, Fleischer NL, Van der Laan M, Lippman SA, Jewell N, et al.
To GEE or not to GEE: comparing population average and mixed models for
estimating the associations between neighborhood risk factors and health.
Epidemiology 2010;21:46774.
[11] OECD Health Statistics 2014 Frequently Requested Data (Updated November
2014); Available at: http://www.oecd.org/els/health-systems/oecd-
health-statistics-2014-frequently-requested-data.htm [accessed 17.7.15]
[12] Yasunaga H, Horiguchi H, Matsuda S, Fushimi K, Hashimoto H, Ayanian JZ.
Body mass index and outcomes following gastrointestinal cancer surgery in
Japan. Br J Surg 2013;100:133543.
[13] Srensen LT. Wound healing and infection in surgery. The clinical impact of
smoking and smoking cessation: a systematic review and meta-analysis. Arch
Surg 2012;147:37383.
[14] Becvarovski Z, Kartush JM. Smoking and tympanoplasty: implications for
prognosis and the Middle Ear Risk Index (MERI). Laryngoscope 2001;111:
180611.
[15] Schwartz SR, Yueh B, Maynard C, Daley J, Henderson W, Khuri SF. Predictors of
wound complications after laryngectomy: a study of over 2000 patients.
Otolaryngol Head Neck Surg 2004;131:618.
[16] Lotfi CJ, Cavalcanti R de C, Costa e Silva AM, Latorre Mdo R, Ribeiro Kde C,
Carvalho AL, et al. Risk factors for surgical-site infections in head and neck
cancer surgery. Otolaryngol Head Neck Surg 2008;138:7480.
[17] Lewis KG, Dufresne RG. A meta-analysis of complications attributed to antic-
oagulation among patients following cutaneous surgery. Dermatol Surg
2008;34:1604.
[18] Somerville P, Seifert PJ, Destounis SV, Murphy PF, Young W. Anticoagulation
and bleeding risk after core needle biopsy. Am J Roentgenol 2008;191:11947.
[19] Dixon AJ, Dixon MP, Dixon JB. Bleeding complications in skin cancer surgery are
associated with warfarin but not aspirin therapy. Br J Surg 2007;94:135660.
[20] Hou YY, Huang HL, Shi HY. Prevalence and hospital resource utilization in
tympanoplasty and revision tympanoplasty: a population-based comparative
study. Otol Neurotol 2012;33:4005.

Please cite this article in press as: Suzuki S, et al. Factors associated with prolonged duration of post-tympanoplasty local treatment in
adult chronic otitis media patients: A retrospective observational study using a Japanese inpatient database. Auris Nasus Larynx (2015),
http://dx.doi.org/10.1016/j.anl.2015.09.006