- Email: [email protected]
Assessment of statin-induced interstitial pneumonia in patients treated for hyperlipidemia using a health insurance claims database in Japan
Accepted Manuscript Assessment of statin-induced interstitial pneumonia in patients treated for hyperlipidemia using a health insurance claims database in Japan Kenji Momo, Akiko Takagi, Atsuko Miyaji, Masayoshi Koinuma PII:
S1094-5539(17)30205-5
DOI:
10.1016/j.pupt.2018.04.003
Reference:
YPUPT 1721
To appear in:
Pulmonary Pharmacology & Therapeutics
Received Date: 18 August 2017 Revised Date:
25 February 2018
Accepted Date: 4 April 2018
Please cite this article as: Momo K, Takagi A, Miyaji A, Koinuma M, Assessment of statin-induced interstitial pneumonia in patients treated for hyperlipidemia using a health insurance claims database in Japan, Pulmonary Pharmacology & Therapeutics (2018), doi: 10.1016/j.pupt.2018.04.003. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Original Article
1 2
Assessment of statin-induced interstitial pneumonia in patients treated for
4
hyperlipidemia using a health insurance claims database in Japan
RI PT
3
5
Kenji Momo1)*, Akiko Takagi1), Atsuko Miyaji2) and Masayoshi Koinuma1)
6 7
SC
1) Faculty of Pharmaceutical Sciences, Teikyo Heisei University, 4-21-2, Nakano, Nakanoku, Tokyo, 164-8530, Japan. 2) MIYAJI NAIKA clinic of internal medicine, 1-35-9,
9
M AN U
8
Kamitakada, Nakano-ku, Tokyo, 164-0002, Japan.
10 11 12 13
15
*
Corresponding author: [email protected]; Tel: +81-3-5860-4249
TE D
14
This manuscript has 2,489 words (3 tables and 1 figure) without abstract.
17
Running head: Assessment of statin-induced interstitial pneumonia in Japan
18
Keywords: Statin, interstitial pneumonia, claims database, hyperlipidemia
AC C
EP
16
1
ACCEPTED MANUSCRIPT ABSTRACT
2
Purpose: This study aimed to determine the frequency and risk factors for statin-induced
3
interstitial pneumonia (IP).
4
Method: We conducted a retrospective cohort study using a large Japanese health insurance
5
claims database. We determined the statin-induced IP incidence in patients treated with
6
statins for hyperlipidemia (n = 194,814) with 12-month screening and 3-month observation
7
periods. Statin-induced IP was defined as: (1) diagnosis with IP (ICD-10 codes: J70.2-J70.4,
8
J84.1, and J84.9) within 3 months after starting statins; (2) steroid administration starts after
9
starting statins; (3) undergoing laboratory tests for sialylated carbohydrate antigen KL-6 or
10
pulmonary surfactant protein-D; and (4) undergoing high-resolution computed tomography
11
(HRCT). Risk factors for IP were defined as presence of lung-related diseases including lung
12
cancer and IP (ICD-10 codes: A16, J43–46, 60–70, and 80–89) that were known to the risk
13
factors inducing IP during the screening period.
14
Results: Cohort 1 had no IP-inducing risk factors; based on lung-related disease history, we
15
identified 4 cases (male/female: 0/4, 61 ± 2.5 years) and 46,574 controls (male/female:
16
29,677/16,897, 51.3 ± 9.5 years). In cohort 1, all cases were female and average age was
17
older than that of controls (p < 0.01). Cohort 2 had lung-related disease history that were
18
known to the risk factors inducing IP; we identified 25 cases (male/female: 11/14, 52.8 ± 11.3
19
years) and 4,005 controls (male/female: 2,305/1,700, 51.0 ± 10.4 years). IP incidence was
20
higher in cohort 2 than in cohort 1, who had no IP risk factors (0.6% vs. 0.009%, p < 0.01).
21
The adjusted case/control odds ratio in cohort 2 was 3.8-fold (1.7–8.5) in patients who had
22
taken atorvastatin, respectively.
23
Discussion: We clarified the incidence (0.009% and 0.6% in patients without and with lung-
24
related disease history that were known to the risk factors inducing IP, respectively) and risk
25
factors for statin-induced IP (elderly females without lung-related disease history;
AC C
EP
TE D
M AN U
SC
RI PT
1
2
ACCEPTED MANUSCRIPT 1
atorvastatin administration in those with lung-related disease history). Physicians and
2
pharmacists should pay close attention to female patients starting atorvastatin, especially
3
those with past histories of lung-related diseases that were known to the risk factors for IP.
AC C
EP
TE D
M AN U
SC
RI PT
4
3
ACCEPTED MANUSCRIPT 1 2 3
INTRODUCTION
4
by hyperplasia of the alveolar wall caused by interstitium inflammation1). This IP
5
accompanies pulmonary fibrosis as a secondary effect of IP, and this leads stenosis of
6
pulmonary alveolus and reduction of tissue resilience1). The symptoms of drug-induced IP are
7
fever, dry cough, and respiratory distress, which frequently lead to fatal outcome1).
8
Reportedly, drug-induced IP results from the use of various drugs, including anticancer
9
agents, amiodarone, methotrexate, or biologic agents2–7).
SC
RI PT
The pathology of interstitial pneumonia (IP), respiratory impairment, is accompanied
In the Canadian adverse reaction newsletter in March 2010, Health Canada
11
announced, “statins and intestinal lung disease” and reported 8 patients with statin-induced
12
interstitial lung disorders, 2 of whom were diagnosed with IP8). This report corroborated with
13
the trends mentioned in some other reports. In Japan, the Pharmaceuticals and Medical
14
Devices Agency (PMDA) also released the revised information for package inserts of statins
15
on the risk of statin-induced IP9). The suspected mechanism of statin-induced interstitial lung
16
diseases (ILD), including IP, was attributed to the mediation by the inhibition of
17
phospholipase as part of the statin effect in lipid metabolism10), impact of mitochondrial
18
metabolism and interference with energy metabolism on the muscles11), and effect of the
19
immune mediation such that oxygen free radicals cause cell injury and inflammation11).
TE D
EP
AC C
20
M AN U
10
In contrast, some studies have also reported a negative correlation between statin and
21
ILD, including IP12). However, the frequency of ILD, including IP, is reportedly higher in
22
Japanese compared to Caucasian individuals because of ethnic differences arising from their
23
genetic variations, although this mechanism remains unclear13,14). Thus, the incidence and
24
risk factors for statin-induced IP lack adequate evidence. This study aims to assess the
25
frequency and risk factors for stain-induced IP using an extensive health insurance claims
26
database in Japan.
4
ACCEPTED MANUSCRIPT 1 2
METHODS
3
Data source We conducted this research using the large health insurance claims database developed
5
by Japan Medical Data Center (JMDC) Co. Ltd., Tokyo, Japan. JMDC collects medical and
6
pharmacy claims from more than 50 occupation-based public health insurance agencies for
7
corporate employees and their family members. As of August 2016, the database included
8
3,600,000 recipients aged 0–74 years, representing 2.0% of the Japanese population.
9
Statin-prescribing patterns and retrospective cohort study to detect IP
SC
RI PT
4
The statin-prescribing pattern analysis involved 194,814 patients who were prescribed
11
statins (atorvastatin, rosuvastatin, pitavastatin, simvastatin, pravastatin, and fluvastatin) from
12
Jan 2004 to Jan 2016. The retrospective cohort study to detect statin-induced IP was designed
13
with a 12-month screening period and a 3-month observation period. Collected data included
14
age, sex, duration of statin use, laboratory investigation, and presence of other diseases and
15
comorbidities in 62,969 patients treated with statins for hyperlipidemia (Fig. 1).
16
Definition of patients with IP
TE D
M AN U
10
We focused on 6 statins that were started for the first time in each patient for the
18
treatment of hyperlipidemia. Patients were divided into 2 groups (cohorts 1 and 2). Cohort 1
19
did not have lung-related disease histories that were known to the risk factors inducing IP 13)
20
as the International Statistical Classification of Diseases and Related Health Problems, 10th
21
ed. (ICD-10) codes: A16, J43–46, 60–70, and 80–89. Cohort 2 had lung-related disease
22
history that were known to the risk factors inducing IP.
AC C
EP
17
23
IP identification was defined as follows: (1) diagnosis with IP within 3 months after
24
starting statins (ICD10; J70.2-J70.4, J84.1, and J84.9); (2) starting oral or injected steroid
25
administration within 3 months after starting statins; (3) having undergone laboratory tests for
5
ACCEPTED MANUSCRIPT 1
Krebs von den Lungen-6 or surfactant protein-D; and (4) with undergoing high-resolution
2
computed tomography (HRCT) in the observation periods. The protocol for this observational
3
study was approved by the Ethics Committee of Teikyo Heisei University.
4
Data analysis Data are expressed as medians with ranges, means ± standard deviations or odds ratio
6
(OR) and 95% confidence interval (CI). Incidence of statin-induced IP between the 2 cohorts
7
was compared using the chi-squared test. Risk factors (sex, type of statin used, and presence
8
of disease) for IP occurrence in cases vs. controls were analyzed using Fisher’s exact test or
9
OR (95% CI). Age was compared using Welch’s t-test. In cohort 2, multivariate logistic
10
regression was used to calculate OR and 95% confidence intervals (95% CIs) after
11
controlling simultaneously for potential confounders that with p < 0.2 between cases and
12
controls in the univariate analyses. The data analysis was used JMP 13® (SAS Institute Inc.,-
13
Cary,-NC,-US). Differences were considered significant when p < 0.05.
M AN U
SC
RI PT
5
15 16
RESULTS
TE D
14
A total of 194,814 patients (male/female: 113,432/81,382) were treated with statins
18
for hyperlipidemia according to the JMDC health insurance claims database. The prescribing
19
patterns were divided into inpatients and outpatients (3.1% vs. 96.9%, respectively) (Table 1).
20
The most highly used statins were rosuvastatin (n = 75,798, 45,305,382 person days, 598 ±
21
535 days) and atorvastatin (n = 64,307, 41,025,223 person days, 638 ± 589 days).
22
Comorbidities were diabetes mellitus (29.3%), hypertension (46.5%), and cerebrovascular
23
diseases (9.8%). Other anti-hyperlipidemic drugs administered were fibrates (5.9%) and
24
ezetimibe (6.1%).
25
AC C
EP
17
Statin-induced IP was compared between the case and control groups in cohorts 1
6
ACCEPTED MANUSCRIPT and 2 (Tables 2 and 3), all diagnosed with IP (ICD10: J70.2-J70.4, J84.1, and J84.9), within 3
2
months after starting a statin. In cohort 1, all cases were female (n = 4). Average age was
3
higher among cases than among controls (61.0 ± 2.5 vs. 51.3 ± 9.5 years, p < 0.01). In cohort
4
2, age was not significantly different between cases and controls (52.8 ± 11.3 vs. 51.0 ± 10.4,
5
N.S.). The tendency for the difference in sex for the occurrence of IP after administration of
6
statins was also observed in cohort 2, but the incidence was not reached significant difference
7
in females (56%) than in males (44%) (OR: 1.7 [0.8–3.8]).
RI PT
1
For multi-variate data analyses, variables considered in the models were “sex
9
(female)”, “administration of atorvastatin”, “comorbidity of diabetes mellitus and liver
10
diseases” that were p < 0.2 in the univariate analyses. The “comorbidity of Esophagus,
11
stomach, and duodenum diseases” was excluded because of the confounding factor as
12
steroids treatment. To adjust these factors, the adjusted OR for these factors such as female
13
(adjust OR: 1.8 [0.8-4.0]), patients treated with atorvastatin (adjust OR: 3.8 [1.7-8.5]) and
14
comorbidity of diabetes mellitus (adjust OR: 2.5 [1.1–5.6]) and liver diseases (adjust OR: 1.9
15
[0.8-4.4]) were observed. In these factors, atrvastatin administration and diabetes milletus
16
were higher in cases compared to controls in cohort 2.
TE D
M AN U
SC
8
In cohort 2, with lung-related diseases history that we known to the risk factors
18
inducing IP, a higher incidence of IP after starting statins as compared to cohort 1 was
19
observed in this study (0.009 %: 4/46,574 vs. 0.6%: 25/4,005, p < 0.01).
AC C
20
EP
17
In addition, in our observation period, the patients underwent respiratory function test
21
was 1 and 7 cases in cohort 1 (females, rosuvastatin) and cohort 2 (male/female: 3:4,
22
rosuvastatin/atorvastatin: 3:4). Although the number of patients were small, similar risk
23
tendency for female and atorvastatin were observed as compared to that of without definition
24
for respiratory functions in cohort 2.
25
7
ACCEPTED MANUSCRIPT 1
DISCUSSION In this study, we determined the incidence (0.6% and 0.009% with and without lung-
3
related disease history that were known to the risk factors inducing IP, respectively) and risk
4
factors (presence of lung-related disease history) for statin-induced IP in patients treated with
5
statins. This was achieved by analyzing data from a health insurance claims database in
6
Japan.
RI PT
2
The mean age of our sample was 54.6 years, which was younger than that of the national
8
census for diseases (35% of patients with hyperlipidemia were over 70 years) conducted by
9
the Ministry of Health, Labour and Welfare in Japan (Table 1).15, 16 This differences may
10
have been due to the fact that our data were from medical and pharmacy claims from public
11
health insurance agencies for corporate employees and their family members. Thus, the
12
number of individuals in our sample at or above retirement age (i.e., 60 years) was
13
insufficient. However, the statin-prescribing patterns in our study for males (58%), those with
14
diabetes mellitus (29%), and those with hypertension (46%) showed similar trends to those of
15
individuals under 60 years treated with anti-hyperlipidemic drugs in the database from the
16
specific health checkup conducted by the Ministry of Health, Labour and Welfare in Japan
17
(males: 66%, diabetes mellitus: 18%, hypertension: 46%).
18
data are reliable for individuals under 60 years of age in Japan.
17
This finding suggests that our
AC C
EP
TE D
M AN U
SC
7
19
Risk factors for statin-induced IP were being an elderly female in the group without
20
lung-related disease history that were known as the IP inducing factors, and being with
21
diabetes mellitus or having taken atorvastatin in the group with lung-related disease history
22
(Tables 2 and 3). In cohort 2, female sex did not reached statistical significance, though the
23
ratio for the female sex obtained higher in case (control vs. case: 42.4% and 56%). Some
24
reports have suggested a relationship between statin blood concentration and IP onset5, 3, 6). In
25
Japan, the proportion of individuals with a body mass index over 25% was higher in males
8
ACCEPTED MANUSCRIPT 1
than in females (14.4% vs. 5.6%, respectively) in the data from the specific health checkup
2
conducted by the Ministry of Health, Labour and Welfare17. This suggests that small stature
3
in females, and especially elderly females, may partially explain the risk for statin-induced IP
4
onset, due to the blood concentration according to the dose for mg/kg of statins. The biological mechanism for atorvastatin-induced IP is unknown. However, the
6
frequency of statin-induced IP in patients treated with atorvastatin was higher compared to
7
that of other statins, according to the spontaneous adverse effects reporting system in Japan
8
(atorvastatin, rosuvastatin, pravastatin, pitavastatin, fluvastatin, and simvastatin: 68, 39, 29,
9
22, 15, and 12 cases, respectively) (data not shown). The results of both the spontaneous
10
adverse report systems in Japan and our results (large claims data) were consistent. Results
11
suggest that further molecular or biological research will be required to clarify the
12
mechanisms for the higher risk for IP onset by atorvastatin.
M AN U
SC
RI PT
5
The incidence of statin-induced IP in this study cohort 1 (without lung-related disease
14
history that were known to the risk factors inducing IP) was low (0.009%) (Table 2). In
15
cohort 2 (with lung-related disease history that were known to the risk factors inducing IP),
16
the frequency for IP was higher than cohort 1 (0.009% vs. 0.6%, p<0.01), but lower than the
17
well known ILD including IP inducing drug, gefitinib (0.6% vs. 4.0-5.8%). Because in the
18
case with ILD including IP caused by gefitinib were reported fetal outcomes approximately
19
30%, monitoring for IP need for the patients treated with statins (Table 3)4, 5). Although
20
JMDC claims data could not detect the fatal event “death,” initial symptoms such as cough
21
and out of breath should be monitored within 3 months of starting statins in patients with
22
lung-related-disease that were known to the risk factors inducing IP past history.
AC C
EP
TE D
13
23
This study has some limitations that depending claim database. First, the definition of IP
24
considered in this study might be insufficient for the following reasons: (1) our claim data
25
could not obtain the clinical data; and (2) some of disease types were mixed in our definition
9
ACCEPTED MANUSCRIPT of “Lung-related disease history that were known to the risk factors inducing IP,” such as the
2
history of pulmonary fibrosis or, for example, combined pulmonary fibrosis and emphysema.
3
Second, we did not observe statistically significant differences in the risk factors for female
4
sex and atorvastatin when respiratory function test were added to the definition of IP.
5
Although cohort 1 (n = 1) and cohort 2 (n = 7) had a small sample size for the analysis, we
6
observed the similar tendency for induced IP in female sex (4 out of 7 cases) and atorvastatin
7
administration (4 out of 7 had atorvastatin) when considered lung function.
RI PT
1
In conclusion, there was a higher incidence of statin-induced IP in patients with lung-
9
related disease history that were known to the risk factors inducing IP. Particularly among
10
patients with risk factors such as female sex, elderly and atorvastatin administration,
11
physicians and caregivers should monitor patients in the initial 3 months after starting statins,
12
and consider individual patient risk.
13
M AN U
SC
8
ACKNOWLEDGMENT
15
We gratefully acknowledge the contribution of the JMDC Co. Ltd., especially Makiko
16
Kaneko and Nauta Yamanaka, JMDC Co. Ltd. who advised us in this study.
19 20
EP
18
Competing interests: None declared.
AC C
17
TE D
14
10
ACCEPTED MANUSCRIPT 1
Figure Legend
2
Figure 1. Case identification flow in cohorts 1 and 2 in patients treated with statins according
4
to the JMDC health insurance claims database.
5
ICD-10, International Statistical Classification of Diseases and Related Health Problems, 10th
6
ed.; JMDC, Japan Medical Data Center; KL-6, sialylated carbohydrate antigen KL-6; SP-D,
7
pulmonary surfactant protein-D, IP; Interstitial Pneunomia.
AC C
EP
TE D
M AN U
SC
8
RI PT
3
11
ACCEPTED MANUSCRIPT 1
REFERENCES
2
1.
Itoh Y, Sendo T, Oishi R. Folia Pharmacol. Jpn, 127: 425-432 (2006).
3
2.
Kudoh S, Kato H, Nishiwaki Y, Fukuoka M, Nakata K, Ichinose Y, Tsuboi M, Yokota S, Nakagawa K, Suga M; Japan Thoracic Radiology Group, Jiang H, Itoh Y, Armour A,
5
Watkins C, Higenbottam T, Nyberg F. Interstitial lung disease in Japanese patients with
6
lung cancer: a cohort and nested case-control study. Am J Respir Crit Care Med.
7
177:1348-57 (2008).
and Thomas Berghaus. Drug Induced Interstitial Lung Disease. The Open Respiratory
9
Medicine Journal. 6: 63-74 (2012).
10 11
Martin Schwaiblmair, Werner Behr, Thomas Haeckel, Bruno Märkl, Wolfgang Foerg,
SC
3.
4.
M AN U
8
RI PT
4
Nakagawa K, Kudoh S, Ohe Y, Johkoh T, Ando M, Yamazaki N, Seki A, Takemoto S, Fukuoka M. Postmarketing surveillance study of erlotinib in Japanese patients with
13
non-small-cell lung cancer (NSCLC): an interim analysis of 3488 patients
14
(POLARSTAR). J Thorac Oncol. 7: 1296-303 (2012).
15
5.
TE D
12
Kudoh S, Kato H, Nishiwaki Y, Fukuoka M, Nakata K, Ichinose Y, Tsuboi M, Yokota S, Nakagawa K, Suga M. Japan Thoracic Radiology Group, Jiang H, Itoh Y, Armour A,
17
Watkins C, Higenbottam T, Nyberg F., Interstitial lung disease in Japanese patients with
18
lung cancer: a cohort and nested case-control study. Am J Respir Crit Care Med.177:
19
1348-1357 (2008). 6.
AC C
20
EP
16
Yamada Y, Shiga T, Matsuda N, Hagiwara N, Kasanuki H. Incidence and predictors of
21
pulmonary toxicity in Japanese patients receiving low-dose amiodarone. Circ J. 71:
22
1610-1616 (2007).
23
7.
rheumatoid arthritis. Nat Rev Rheumatol. 6: 644-652 (2010).
24 25
Takeuchi T, Kameda H. The Japanese experience with biologic therapies for
8.
Statins and interstitial lung disease. Health Canada, Canadian Adverse Reaction
12
ACCEPTED MANUSCRIPT Newsletter 20 (4), Oct 2010.
1 2
9.
Revisions of PRECAUTIONS in drug package inserts, Phrmaceuticals and Medical
3
Devices Agency (PMDA), http://www.pmda.go.jp/safety/info-services/
4
drugs/calling-attention/revision-of-precautions/0206.html#4 (Dec/20/2017 access).
6
10. Lantuejoul S, Brambilla E, Brambilla C, Devouassoux G. Statin-induced fibrotic
RI PT
5
nonspecific interstitial pneumonia. Eur Respir J. 19: 577-580 (2002).
11. Fernández AB, Karas RH, Alsheikh-Ali AA. Thompson PD. Statins and interstitial lung
8
disease: a systematic review of the literature and of food and drug administration
9
adverse event reports. Chest. 134: 824-830 (2008).
11
12. Saad N, Suissa S, Camus P, Ernst P. Statins and the risk of interstitial lung disease: a
M AN U
10
SC
7
cohort study. Thorax. 68: 361-364 (2013).
13. ”Yakuzaisei Haisyougai no Shindan/ Chiryou no Tebiki” (Guidance for diagnose and
13
treatment for Interstitial Lung Disease), The Japanese Respiratory Society. issue 1
14
(2013).
TE D
12
14. Azuma A. “Nihonjin ni Ooi Yakuzaiseihaisyougai” (Interstitial lung disease is
16
frequently occurred in Japanese), Jpn J Clin Pharmacol Ther. 39: 29S-30S (2008).
17
Proceedings.
19 20
15. Summary of Patient Survey,2014, Ministry of Health, Labour and Welfare, Japan
AC C
18
EP
15
http://www.mhlw.go.jp/toukei/saikin/hw/kanja/14/dl/kanja.pdf 16. The National Health and Nutrition Survey in Japan, 2014, Ministry of Health, Labour
21
and Welfare, Japan
22
http://www.mhlw.go.jp/bunya/kenkou/eiyou/dl/h26-houkoku.pdf
23
17. The database for the specific health checkup, 2016 Ministry of Health, Labour and
24
Welfare, Japan. http://www.mhlw.go.jp/bunya/shakaihosho/iryouseido01/info02a-
25
2.html
13
ACCEPTED MANUSCRIPT Table 1. Statin-prescribing pattern analysis in the JMDC health insurance claims database Variable
Value
Age [mean (SD)]
54.6 (9.9)
Sex (male) [n (%)]
113,432 (58.2)
Sex (female) [n (%)]
81,382 (41.8) inpatients
Rosuvastatin [n]
2,483
administration period [person days]
29,117
administration period [mean (SD)]
outpatients
RI PT
Statins
75,733
45,276,265
598 (535) 2,128
64,228
SC
Atorvastatin [n] administration period [person days]
27,773
administration period [mean (SD)]
40,997,450
638 (589)
80
administration period [person days] administration period [mean (SD)] Pravastatin [n] administration period [person days] administration period [mean (SD)] Simvastatin [n]
M AN U
Pitavastatin [n]
13,108
22,750,683
576 (542)
635
37,055
8,141
21,719,702
586 (588)
TE D
administration period [person days] administration period [mean (SD)] Fluvastatin [n]
39,496
administration period [person days]
EP
administration period [mean (SD)]
53
8,844
772
5,171,520
585 (589) 119
6,232
1,667
3,637,263
583 (603)
Comorbidities (ICD-10)
57,142 (29.3)
Hypertension [n (%)]
90,567 (46.5)
cerebrovascular disease [n (%)]
19,158 (9.8)
ischemic heart disease [n (%)]
25,605 (13.1)
AC C
Diabetes mellitus [n (%)]
Anti-hyperlipidemia drugs (ATC) Fibrate [n (%)]
11,422 (5.86)
Ezetimibe [n (%)]
11,898 (6.11)
Nicomol/Nicotrol [n (%)]
78 (0.04)
Cholestyramine [n (%)]
1 2 3
1,062 (0.55)
Probucol [n (%)] 514 (0.26) ICD-10; International Statistical Classification of Diseases and Related Health Problems, 10th ed.; JMDC, Japan Medical Data Center; SD, standard deviation; ATC, Anatomical Therapeutic Chemical Classification
14
ACCEPTED MANUSCRIPT Table 2. Statin-induced interstitial pneumonia: case–control comparison of patients treated with statins in cohort 1 Cohort 1
Control
Case
N
46,574
4
51.3 (9.5)
61 (2.5)
29,677 (63.7)
Sex (male) [n (%)] Sex (female) [n (%)]
Inpatients [n] Outpatients [n]
16,897 (36.3)
4 (100)
754
0
45,820
SC
Statins [n]
2
Atorvastatin
11,762
1
Pitavastatin
8,369
1
M AN U
16,603
7,370
0
1,355
0
1,115
0
16,480
1
14,652
0
Esophagus, stomach, and duodenum diseases
9,816
3
Liver diseases
8,436
1
Simvastatin Fluvastatin Comorbidities (ICD-10) [n] Hypertensive diseases
6 7 8
EP
ICD-10; International Statistical Classification of Diseases and Related Health Problems, 10th ed.; SD, standard deviation
AC C
5
TE D
Diabetes mellitus
2 3 4
4
Rosuvastatin
Pravastatin
1
0 (0)
RI PT
Age [mean (SD)]
9 10 11 12 13 15
ACCEPTED MANUSCRIPT 1
Table 3. Statin-induced interstitial pneumonia: case–control comparison of patients treated with statins in cohort 2 Cohort 2
OR Crude
Inpatients [n]
4,005 51.0 (10.4) 2305 (57.6) 1700 (42.4) 96
25 52.8 (11.3) 11 (44.0) 14 (56.0) 4
Outpatients [n]
3,909
21
Rosuvastatin
1,426
9
1.0
[0.4–2.3]
Atorvastatin
1,005
14
3.8
[1.7–8.4]
Pitavastatin
797
1
0.2
[0.02–1.2]
Pravastatin
584
1
0.2
[0.03–1.8]
Simvastatin
112
0
-
Fluvastatin
81
0
-
7
0.7
Sex (male) [n (%)] Sex (female) [n (%)]
1.7
6 7 8
[95％ ％CI]
1.8
[0.8-4.0]
3.8
[1.7-8.5]
[0.3–1.7]
EP
2.5 [1.1-5.6] Diabetes mellitus 1,407 15 2.8 [1.2–6.2] Esophagus, stomach, and 1,258 24 52.4 [7.1–387.8] duodenum diseases 1.9 [0.8-4.4] Liver diseases 805 9 2.2 [1.0–5.1] CI, confidence interval; ICD-10; International Statistical Classification of Diseases and Related Health Problems, 10th ed.; RR, relative risk; SD, standard deviation
AC C
5
1,438
TE D
Hypertensive diseases
M AN U
Statins [n]
Comorbidities (ICD-10) [n]
[0.8－3.8]
SC
Age [mean (SD)]
OR Adjust
RI PT
Case
N
2 3 4
[95％ ％CI]
Control
9 10 11 12
16
ACCEPTED MANUSCRIPT 1
Figure 1 JMDC health insurance database (2005 Jan – 2016 Jan)
2
Administered statins for hyperlipidemia (n = 194,814)
3 4 5 6 7
RI PT
Lack of health insurance data for the screening period (12 months) and observation period (3 months) (n = 131,845)
With below diseases in the screening periods (n=6,875)
8
・Other respiratory diseases principally affecting the
9
・Tuberculosis (ICD10:A15-19)
SC
interstitium (ICD-10: J80-84)
Cohort 2
・Influenza and pneumonia (ICD-10:J12-J18)
10
M AN U
・Chronic lower respiratory diseases (ICD-10:J43-46) ・Lung diseases due to external agents (ICD-10:J60-J70) ・Other related diseases (ICD10:J85-99)
TE D
With IP diagnosis (ICD-10: J70.2J70.4, J84.1, and J84.9) and administered oral or injection steroids and KL-6 or SP-D and HRCT in the observation period (n = 25)
Case
EP
Cohort 1
AC C
With IP diagnosis (ICD10: J J70.2-J70.4, J84.1, and J84.9) and administered oral or injection steroids and KL-6 or SP-D and HRCT in the observation period (n = 4)
Without IP diagnosis (ICD10: J70.2-J70.4, J84.1, and J84.9) in the observation period (n = 4,030)
Control
Without IP diagnosis (ICD10: J70.2-J70.4, J84.1, and J84.9) in the observation period (n = 46,574)
Control
Case
17
S1094-5539(17)30205-5
DOI:
10.1016/j.pupt.2018.04.003
Reference:
YPUPT 1721
To appear in:
Pulmonary Pharmacology & Therapeutics
Received Date: 18 August 2017 Revised Date:
25 February 2018
Accepted Date: 4 April 2018
Please cite this article as: Momo K, Takagi A, Miyaji A, Koinuma M, Assessment of statin-induced interstitial pneumonia in patients treated for hyperlipidemia using a health insurance claims database in Japan, Pulmonary Pharmacology & Therapeutics (2018), doi: 10.1016/j.pupt.2018.04.003. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Original Article
1 2
Assessment of statin-induced interstitial pneumonia in patients treated for
4
hyperlipidemia using a health insurance claims database in Japan
RI PT
3
5
Kenji Momo1)*, Akiko Takagi1), Atsuko Miyaji2) and Masayoshi Koinuma1)
6 7
SC
1) Faculty of Pharmaceutical Sciences, Teikyo Heisei University, 4-21-2, Nakano, Nakanoku, Tokyo, 164-8530, Japan. 2) MIYAJI NAIKA clinic of internal medicine, 1-35-9,
9
M AN U
8
Kamitakada, Nakano-ku, Tokyo, 164-0002, Japan.
10 11 12 13
15
*
Corresponding author: [email protected]; Tel: +81-3-5860-4249
TE D
14
This manuscript has 2,489 words (3 tables and 1 figure) without abstract.
17
Running head: Assessment of statin-induced interstitial pneumonia in Japan
18
Keywords: Statin, interstitial pneumonia, claims database, hyperlipidemia
AC C
EP
16
1
ACCEPTED MANUSCRIPT ABSTRACT
2
Purpose: This study aimed to determine the frequency and risk factors for statin-induced
3
interstitial pneumonia (IP).
4
Method: We conducted a retrospective cohort study using a large Japanese health insurance
5
claims database. We determined the statin-induced IP incidence in patients treated with
6
statins for hyperlipidemia (n = 194,814) with 12-month screening and 3-month observation
7
periods. Statin-induced IP was defined as: (1) diagnosis with IP (ICD-10 codes: J70.2-J70.4,
8
J84.1, and J84.9) within 3 months after starting statins; (2) steroid administration starts after
9
starting statins; (3) undergoing laboratory tests for sialylated carbohydrate antigen KL-6 or
10
pulmonary surfactant protein-D; and (4) undergoing high-resolution computed tomography
11
(HRCT). Risk factors for IP were defined as presence of lung-related diseases including lung
12
cancer and IP (ICD-10 codes: A16, J43–46, 60–70, and 80–89) that were known to the risk
13
factors inducing IP during the screening period.
14
Results: Cohort 1 had no IP-inducing risk factors; based on lung-related disease history, we
15
identified 4 cases (male/female: 0/4, 61 ± 2.5 years) and 46,574 controls (male/female:
16
29,677/16,897, 51.3 ± 9.5 years). In cohort 1, all cases were female and average age was
17
older than that of controls (p < 0.01). Cohort 2 had lung-related disease history that were
18
known to the risk factors inducing IP; we identified 25 cases (male/female: 11/14, 52.8 ± 11.3
19
years) and 4,005 controls (male/female: 2,305/1,700, 51.0 ± 10.4 years). IP incidence was
20
higher in cohort 2 than in cohort 1, who had no IP risk factors (0.6% vs. 0.009%, p < 0.01).
21
The adjusted case/control odds ratio in cohort 2 was 3.8-fold (1.7–8.5) in patients who had
22
taken atorvastatin, respectively.
23
Discussion: We clarified the incidence (0.009% and 0.6% in patients without and with lung-
24
related disease history that were known to the risk factors inducing IP, respectively) and risk
25
factors for statin-induced IP (elderly females without lung-related disease history;
AC C
EP
TE D
M AN U
SC
RI PT
1
2
ACCEPTED MANUSCRIPT 1
atorvastatin administration in those with lung-related disease history). Physicians and
2
pharmacists should pay close attention to female patients starting atorvastatin, especially
3
those with past histories of lung-related diseases that were known to the risk factors for IP.
AC C
EP
TE D
M AN U
SC
RI PT
4
3
ACCEPTED MANUSCRIPT 1 2 3
INTRODUCTION
4
by hyperplasia of the alveolar wall caused by interstitium inflammation1). This IP
5
accompanies pulmonary fibrosis as a secondary effect of IP, and this leads stenosis of
6
pulmonary alveolus and reduction of tissue resilience1). The symptoms of drug-induced IP are
7
fever, dry cough, and respiratory distress, which frequently lead to fatal outcome1).
8
Reportedly, drug-induced IP results from the use of various drugs, including anticancer
9
agents, amiodarone, methotrexate, or biologic agents2–7).
SC
RI PT
The pathology of interstitial pneumonia (IP), respiratory impairment, is accompanied
In the Canadian adverse reaction newsletter in March 2010, Health Canada
11
announced, “statins and intestinal lung disease” and reported 8 patients with statin-induced
12
interstitial lung disorders, 2 of whom were diagnosed with IP8). This report corroborated with
13
the trends mentioned in some other reports. In Japan, the Pharmaceuticals and Medical
14
Devices Agency (PMDA) also released the revised information for package inserts of statins
15
on the risk of statin-induced IP9). The suspected mechanism of statin-induced interstitial lung
16
diseases (ILD), including IP, was attributed to the mediation by the inhibition of
17
phospholipase as part of the statin effect in lipid metabolism10), impact of mitochondrial
18
metabolism and interference with energy metabolism on the muscles11), and effect of the
19
immune mediation such that oxygen free radicals cause cell injury and inflammation11).
TE D
EP
AC C
20
M AN U
10
In contrast, some studies have also reported a negative correlation between statin and
21
ILD, including IP12). However, the frequency of ILD, including IP, is reportedly higher in
22
Japanese compared to Caucasian individuals because of ethnic differences arising from their
23
genetic variations, although this mechanism remains unclear13,14). Thus, the incidence and
24
risk factors for statin-induced IP lack adequate evidence. This study aims to assess the
25
frequency and risk factors for stain-induced IP using an extensive health insurance claims
26
database in Japan.
4
ACCEPTED MANUSCRIPT 1 2
METHODS
3
Data source We conducted this research using the large health insurance claims database developed
5
by Japan Medical Data Center (JMDC) Co. Ltd., Tokyo, Japan. JMDC collects medical and
6
pharmacy claims from more than 50 occupation-based public health insurance agencies for
7
corporate employees and their family members. As of August 2016, the database included
8
3,600,000 recipients aged 0–74 years, representing 2.0% of the Japanese population.
9
Statin-prescribing patterns and retrospective cohort study to detect IP
SC
RI PT
4
The statin-prescribing pattern analysis involved 194,814 patients who were prescribed
11
statins (atorvastatin, rosuvastatin, pitavastatin, simvastatin, pravastatin, and fluvastatin) from
12
Jan 2004 to Jan 2016. The retrospective cohort study to detect statin-induced IP was designed
13
with a 12-month screening period and a 3-month observation period. Collected data included
14
age, sex, duration of statin use, laboratory investigation, and presence of other diseases and
15
comorbidities in 62,969 patients treated with statins for hyperlipidemia (Fig. 1).
16
Definition of patients with IP
TE D
M AN U
10
We focused on 6 statins that were started for the first time in each patient for the
18
treatment of hyperlipidemia. Patients were divided into 2 groups (cohorts 1 and 2). Cohort 1
19
did not have lung-related disease histories that were known to the risk factors inducing IP 13)
20
as the International Statistical Classification of Diseases and Related Health Problems, 10th
21
ed. (ICD-10) codes: A16, J43–46, 60–70, and 80–89. Cohort 2 had lung-related disease
22
history that were known to the risk factors inducing IP.
AC C
EP
17
23
IP identification was defined as follows: (1) diagnosis with IP within 3 months after
24
starting statins (ICD10; J70.2-J70.4, J84.1, and J84.9); (2) starting oral or injected steroid
25
administration within 3 months after starting statins; (3) having undergone laboratory tests for
5
ACCEPTED MANUSCRIPT 1
Krebs von den Lungen-6 or surfactant protein-D; and (4) with undergoing high-resolution
2
computed tomography (HRCT) in the observation periods. The protocol for this observational
3
study was approved by the Ethics Committee of Teikyo Heisei University.
4
Data analysis Data are expressed as medians with ranges, means ± standard deviations or odds ratio
6
(OR) and 95% confidence interval (CI). Incidence of statin-induced IP between the 2 cohorts
7
was compared using the chi-squared test. Risk factors (sex, type of statin used, and presence
8
of disease) for IP occurrence in cases vs. controls were analyzed using Fisher’s exact test or
9
OR (95% CI). Age was compared using Welch’s t-test. In cohort 2, multivariate logistic
10
regression was used to calculate OR and 95% confidence intervals (95% CIs) after
11
controlling simultaneously for potential confounders that with p < 0.2 between cases and
12
controls in the univariate analyses. The data analysis was used JMP 13® (SAS Institute Inc.,-
13
Cary,-NC,-US). Differences were considered significant when p < 0.05.
M AN U
SC
RI PT
5
15 16
RESULTS
TE D
14
A total of 194,814 patients (male/female: 113,432/81,382) were treated with statins
18
for hyperlipidemia according to the JMDC health insurance claims database. The prescribing
19
patterns were divided into inpatients and outpatients (3.1% vs. 96.9%, respectively) (Table 1).
20
The most highly used statins were rosuvastatin (n = 75,798, 45,305,382 person days, 598 ±
21
535 days) and atorvastatin (n = 64,307, 41,025,223 person days, 638 ± 589 days).
22
Comorbidities were diabetes mellitus (29.3%), hypertension (46.5%), and cerebrovascular
23
diseases (9.8%). Other anti-hyperlipidemic drugs administered were fibrates (5.9%) and
24
ezetimibe (6.1%).
25
AC C
EP
17
Statin-induced IP was compared between the case and control groups in cohorts 1
6
ACCEPTED MANUSCRIPT and 2 (Tables 2 and 3), all diagnosed with IP (ICD10: J70.2-J70.4, J84.1, and J84.9), within 3
2
months after starting a statin. In cohort 1, all cases were female (n = 4). Average age was
3
higher among cases than among controls (61.0 ± 2.5 vs. 51.3 ± 9.5 years, p < 0.01). In cohort
4
2, age was not significantly different between cases and controls (52.8 ± 11.3 vs. 51.0 ± 10.4,
5
N.S.). The tendency for the difference in sex for the occurrence of IP after administration of
6
statins was also observed in cohort 2, but the incidence was not reached significant difference
7
in females (56%) than in males (44%) (OR: 1.7 [0.8–3.8]).
RI PT
1
For multi-variate data analyses, variables considered in the models were “sex
9
(female)”, “administration of atorvastatin”, “comorbidity of diabetes mellitus and liver
10
diseases” that were p < 0.2 in the univariate analyses. The “comorbidity of Esophagus,
11
stomach, and duodenum diseases” was excluded because of the confounding factor as
12
steroids treatment. To adjust these factors, the adjusted OR for these factors such as female
13
(adjust OR: 1.8 [0.8-4.0]), patients treated with atorvastatin (adjust OR: 3.8 [1.7-8.5]) and
14
comorbidity of diabetes mellitus (adjust OR: 2.5 [1.1–5.6]) and liver diseases (adjust OR: 1.9
15
[0.8-4.4]) were observed. In these factors, atrvastatin administration and diabetes milletus
16
were higher in cases compared to controls in cohort 2.
TE D
M AN U
SC
8
In cohort 2, with lung-related diseases history that we known to the risk factors
18
inducing IP, a higher incidence of IP after starting statins as compared to cohort 1 was
19
observed in this study (0.009 %: 4/46,574 vs. 0.6%: 25/4,005, p < 0.01).
AC C
20
EP
17
In addition, in our observation period, the patients underwent respiratory function test
21
was 1 and 7 cases in cohort 1 (females, rosuvastatin) and cohort 2 (male/female: 3:4,
22
rosuvastatin/atorvastatin: 3:4). Although the number of patients were small, similar risk
23
tendency for female and atorvastatin were observed as compared to that of without definition
24
for respiratory functions in cohort 2.
25
7
ACCEPTED MANUSCRIPT 1
DISCUSSION In this study, we determined the incidence (0.6% and 0.009% with and without lung-
3
related disease history that were known to the risk factors inducing IP, respectively) and risk
4
factors (presence of lung-related disease history) for statin-induced IP in patients treated with
5
statins. This was achieved by analyzing data from a health insurance claims database in
6
Japan.
RI PT
2
The mean age of our sample was 54.6 years, which was younger than that of the national
8
census for diseases (35% of patients with hyperlipidemia were over 70 years) conducted by
9
the Ministry of Health, Labour and Welfare in Japan (Table 1).15, 16 This differences may
10
have been due to the fact that our data were from medical and pharmacy claims from public
11
health insurance agencies for corporate employees and their family members. Thus, the
12
number of individuals in our sample at or above retirement age (i.e., 60 years) was
13
insufficient. However, the statin-prescribing patterns in our study for males (58%), those with
14
diabetes mellitus (29%), and those with hypertension (46%) showed similar trends to those of
15
individuals under 60 years treated with anti-hyperlipidemic drugs in the database from the
16
specific health checkup conducted by the Ministry of Health, Labour and Welfare in Japan
17
(males: 66%, diabetes mellitus: 18%, hypertension: 46%).
18
data are reliable for individuals under 60 years of age in Japan.
17
This finding suggests that our
AC C
EP
TE D
M AN U
SC
7
19
Risk factors for statin-induced IP were being an elderly female in the group without
20
lung-related disease history that were known as the IP inducing factors, and being with
21
diabetes mellitus or having taken atorvastatin in the group with lung-related disease history
22
(Tables 2 and 3). In cohort 2, female sex did not reached statistical significance, though the
23
ratio for the female sex obtained higher in case (control vs. case: 42.4% and 56%). Some
24
reports have suggested a relationship between statin blood concentration and IP onset5, 3, 6). In
25
Japan, the proportion of individuals with a body mass index over 25% was higher in males
8
ACCEPTED MANUSCRIPT 1
than in females (14.4% vs. 5.6%, respectively) in the data from the specific health checkup
2
conducted by the Ministry of Health, Labour and Welfare17. This suggests that small stature
3
in females, and especially elderly females, may partially explain the risk for statin-induced IP
4
onset, due to the blood concentration according to the dose for mg/kg of statins. The biological mechanism for atorvastatin-induced IP is unknown. However, the
6
frequency of statin-induced IP in patients treated with atorvastatin was higher compared to
7
that of other statins, according to the spontaneous adverse effects reporting system in Japan
8
(atorvastatin, rosuvastatin, pravastatin, pitavastatin, fluvastatin, and simvastatin: 68, 39, 29,
9
22, 15, and 12 cases, respectively) (data not shown). The results of both the spontaneous
10
adverse report systems in Japan and our results (large claims data) were consistent. Results
11
suggest that further molecular or biological research will be required to clarify the
12
mechanisms for the higher risk for IP onset by atorvastatin.
M AN U
SC
RI PT
5
The incidence of statin-induced IP in this study cohort 1 (without lung-related disease
14
history that were known to the risk factors inducing IP) was low (0.009%) (Table 2). In
15
cohort 2 (with lung-related disease history that were known to the risk factors inducing IP),
16
the frequency for IP was higher than cohort 1 (0.009% vs. 0.6%, p<0.01), but lower than the
17
well known ILD including IP inducing drug, gefitinib (0.6% vs. 4.0-5.8%). Because in the
18
case with ILD including IP caused by gefitinib were reported fetal outcomes approximately
19
30%, monitoring for IP need for the patients treated with statins (Table 3)4, 5). Although
20
JMDC claims data could not detect the fatal event “death,” initial symptoms such as cough
21
and out of breath should be monitored within 3 months of starting statins in patients with
22
lung-related-disease that were known to the risk factors inducing IP past history.
AC C
EP
TE D
13
23
This study has some limitations that depending claim database. First, the definition of IP
24
considered in this study might be insufficient for the following reasons: (1) our claim data
25
could not obtain the clinical data; and (2) some of disease types were mixed in our definition
9
ACCEPTED MANUSCRIPT of “Lung-related disease history that were known to the risk factors inducing IP,” such as the
2
history of pulmonary fibrosis or, for example, combined pulmonary fibrosis and emphysema.
3
Second, we did not observe statistically significant differences in the risk factors for female
4
sex and atorvastatin when respiratory function test were added to the definition of IP.
5
Although cohort 1 (n = 1) and cohort 2 (n = 7) had a small sample size for the analysis, we
6
observed the similar tendency for induced IP in female sex (4 out of 7 cases) and atorvastatin
7
administration (4 out of 7 had atorvastatin) when considered lung function.
RI PT
1
In conclusion, there was a higher incidence of statin-induced IP in patients with lung-
9
related disease history that were known to the risk factors inducing IP. Particularly among
10
patients with risk factors such as female sex, elderly and atorvastatin administration,
11
physicians and caregivers should monitor patients in the initial 3 months after starting statins,
12
and consider individual patient risk.
13
M AN U
SC
8
ACKNOWLEDGMENT
15
We gratefully acknowledge the contribution of the JMDC Co. Ltd., especially Makiko
16
Kaneko and Nauta Yamanaka, JMDC Co. Ltd. who advised us in this study.
19 20
EP
18
Competing interests: None declared.
AC C
17
TE D
14
10
ACCEPTED MANUSCRIPT 1
Figure Legend
2
Figure 1. Case identification flow in cohorts 1 and 2 in patients treated with statins according
4
to the JMDC health insurance claims database.
5
ICD-10, International Statistical Classification of Diseases and Related Health Problems, 10th
6
ed.; JMDC, Japan Medical Data Center; KL-6, sialylated carbohydrate antigen KL-6; SP-D,
7
pulmonary surfactant protein-D, IP; Interstitial Pneunomia.
AC C
EP
TE D
M AN U
SC
8
RI PT
3
11
ACCEPTED MANUSCRIPT 1
REFERENCES
2
1.
Itoh Y, Sendo T, Oishi R. Folia Pharmacol. Jpn, 127: 425-432 (2006).
3
2.
Kudoh S, Kato H, Nishiwaki Y, Fukuoka M, Nakata K, Ichinose Y, Tsuboi M, Yokota S, Nakagawa K, Suga M; Japan Thoracic Radiology Group, Jiang H, Itoh Y, Armour A,
5
Watkins C, Higenbottam T, Nyberg F. Interstitial lung disease in Japanese patients with
6
lung cancer: a cohort and nested case-control study. Am J Respir Crit Care Med.
7
177:1348-57 (2008).
and Thomas Berghaus. Drug Induced Interstitial Lung Disease. The Open Respiratory
9
Medicine Journal. 6: 63-74 (2012).
10 11
Martin Schwaiblmair, Werner Behr, Thomas Haeckel, Bruno Märkl, Wolfgang Foerg,
SC
3.
4.
M AN U
8
RI PT
4
Nakagawa K, Kudoh S, Ohe Y, Johkoh T, Ando M, Yamazaki N, Seki A, Takemoto S, Fukuoka M. Postmarketing surveillance study of erlotinib in Japanese patients with
13
non-small-cell lung cancer (NSCLC): an interim analysis of 3488 patients
14
(POLARSTAR). J Thorac Oncol. 7: 1296-303 (2012).
15
5.
TE D
12
Kudoh S, Kato H, Nishiwaki Y, Fukuoka M, Nakata K, Ichinose Y, Tsuboi M, Yokota S, Nakagawa K, Suga M. Japan Thoracic Radiology Group, Jiang H, Itoh Y, Armour A,
17
Watkins C, Higenbottam T, Nyberg F., Interstitial lung disease in Japanese patients with
18
lung cancer: a cohort and nested case-control study. Am J Respir Crit Care Med.177:
19
1348-1357 (2008). 6.
AC C
20
EP
16
Yamada Y, Shiga T, Matsuda N, Hagiwara N, Kasanuki H. Incidence and predictors of
21
pulmonary toxicity in Japanese patients receiving low-dose amiodarone. Circ J. 71:
22
1610-1616 (2007).
23
7.
rheumatoid arthritis. Nat Rev Rheumatol. 6: 644-652 (2010).
24 25
Takeuchi T, Kameda H. The Japanese experience with biologic therapies for
8.
Statins and interstitial lung disease. Health Canada, Canadian Adverse Reaction
12
ACCEPTED MANUSCRIPT Newsletter 20 (4), Oct 2010.
1 2
9.
Revisions of PRECAUTIONS in drug package inserts, Phrmaceuticals and Medical
3
Devices Agency (PMDA), http://www.pmda.go.jp/safety/info-services/
4
drugs/calling-attention/revision-of-precautions/0206.html#4 (Dec/20/2017 access).
6
10. Lantuejoul S, Brambilla E, Brambilla C, Devouassoux G. Statin-induced fibrotic
RI PT
5
nonspecific interstitial pneumonia. Eur Respir J. 19: 577-580 (2002).
11. Fernández AB, Karas RH, Alsheikh-Ali AA. Thompson PD. Statins and interstitial lung
8
disease: a systematic review of the literature and of food and drug administration
9
adverse event reports. Chest. 134: 824-830 (2008).
11
12. Saad N, Suissa S, Camus P, Ernst P. Statins and the risk of interstitial lung disease: a
M AN U
10
SC
7
cohort study. Thorax. 68: 361-364 (2013).
13. ”Yakuzaisei Haisyougai no Shindan/ Chiryou no Tebiki” (Guidance for diagnose and
13
treatment for Interstitial Lung Disease), The Japanese Respiratory Society. issue 1
14
(2013).
TE D
12
14. Azuma A. “Nihonjin ni Ooi Yakuzaiseihaisyougai” (Interstitial lung disease is
16
frequently occurred in Japanese), Jpn J Clin Pharmacol Ther. 39: 29S-30S (2008).
17
Proceedings.
19 20
15. Summary of Patient Survey,2014, Ministry of Health, Labour and Welfare, Japan
AC C
18
EP
15
http://www.mhlw.go.jp/toukei/saikin/hw/kanja/14/dl/kanja.pdf 16. The National Health and Nutrition Survey in Japan, 2014, Ministry of Health, Labour
21
and Welfare, Japan
22
http://www.mhlw.go.jp/bunya/kenkou/eiyou/dl/h26-houkoku.pdf
23
17. The database for the specific health checkup, 2016 Ministry of Health, Labour and
24
Welfare, Japan. http://www.mhlw.go.jp/bunya/shakaihosho/iryouseido01/info02a-
25
2.html
13
ACCEPTED MANUSCRIPT Table 1. Statin-prescribing pattern analysis in the JMDC health insurance claims database Variable
Value
Age [mean (SD)]
54.6 (9.9)
Sex (male) [n (%)]
113,432 (58.2)
Sex (female) [n (%)]
81,382 (41.8) inpatients
Rosuvastatin [n]
2,483
administration period [person days]
29,117
administration period [mean (SD)]
outpatients
RI PT
Statins
75,733
45,276,265
598 (535) 2,128
64,228
SC
Atorvastatin [n] administration period [person days]
27,773
administration period [mean (SD)]
40,997,450
638 (589)
80
administration period [person days] administration period [mean (SD)] Pravastatin [n] administration period [person days] administration period [mean (SD)] Simvastatin [n]
M AN U
Pitavastatin [n]
13,108
22,750,683
576 (542)
635
37,055
8,141
21,719,702
586 (588)
TE D
administration period [person days] administration period [mean (SD)] Fluvastatin [n]
39,496
administration period [person days]
EP
administration period [mean (SD)]
53
8,844
772
5,171,520
585 (589) 119
6,232
1,667
3,637,263
583 (603)
Comorbidities (ICD-10)
57,142 (29.3)
Hypertension [n (%)]
90,567 (46.5)
cerebrovascular disease [n (%)]
19,158 (9.8)
ischemic heart disease [n (%)]
25,605 (13.1)
AC C
Diabetes mellitus [n (%)]
Anti-hyperlipidemia drugs (ATC) Fibrate [n (%)]
11,422 (5.86)
Ezetimibe [n (%)]
11,898 (6.11)
Nicomol/Nicotrol [n (%)]
78 (0.04)
Cholestyramine [n (%)]
1 2 3
1,062 (0.55)
Probucol [n (%)] 514 (0.26) ICD-10; International Statistical Classification of Diseases and Related Health Problems, 10th ed.; JMDC, Japan Medical Data Center; SD, standard deviation; ATC, Anatomical Therapeutic Chemical Classification
14
ACCEPTED MANUSCRIPT Table 2. Statin-induced interstitial pneumonia: case–control comparison of patients treated with statins in cohort 1 Cohort 1
Control
Case
N
46,574
4
51.3 (9.5)
61 (2.5)
29,677 (63.7)
Sex (male) [n (%)] Sex (female) [n (%)]
Inpatients [n] Outpatients [n]
16,897 (36.3)
4 (100)
754
0
45,820
SC
Statins [n]
2
Atorvastatin
11,762
1
Pitavastatin
8,369
1
M AN U
16,603
7,370
0
1,355
0
1,115
0
16,480
1
14,652
0
Esophagus, stomach, and duodenum diseases
9,816
3
Liver diseases
8,436
1
Simvastatin Fluvastatin Comorbidities (ICD-10) [n] Hypertensive diseases
6 7 8
EP
ICD-10; International Statistical Classification of Diseases and Related Health Problems, 10th ed.; SD, standard deviation
AC C
5
TE D
Diabetes mellitus
2 3 4
4
Rosuvastatin
Pravastatin
1
0 (0)
RI PT
Age [mean (SD)]
9 10 11 12 13 15
ACCEPTED MANUSCRIPT 1
Table 3. Statin-induced interstitial pneumonia: case–control comparison of patients treated with statins in cohort 2 Cohort 2
OR Crude
Inpatients [n]
4,005 51.0 (10.4) 2305 (57.6) 1700 (42.4) 96
25 52.8 (11.3) 11 (44.0) 14 (56.0) 4
Outpatients [n]
3,909
21
Rosuvastatin
1,426
9
1.0
[0.4–2.3]
Atorvastatin
1,005
14
3.8
[1.7–8.4]
Pitavastatin
797
1
0.2
[0.02–1.2]
Pravastatin
584
1
0.2
[0.03–1.8]
Simvastatin
112
0
-
Fluvastatin
81
0
-
7
0.7
Sex (male) [n (%)] Sex (female) [n (%)]
1.7
6 7 8
[95％ ％CI]
1.8
[0.8-4.0]
3.8
[1.7-8.5]
[0.3–1.7]
EP
2.5 [1.1-5.6] Diabetes mellitus 1,407 15 2.8 [1.2–6.2] Esophagus, stomach, and 1,258 24 52.4 [7.1–387.8] duodenum diseases 1.9 [0.8-4.4] Liver diseases 805 9 2.2 [1.0–5.1] CI, confidence interval; ICD-10; International Statistical Classification of Diseases and Related Health Problems, 10th ed.; RR, relative risk; SD, standard deviation
AC C
5
1,438
TE D
Hypertensive diseases
M AN U
Statins [n]
Comorbidities (ICD-10) [n]
[0.8－3.8]
SC
Age [mean (SD)]
OR Adjust
RI PT
Case
N
2 3 4
[95％ ％CI]
Control
9 10 11 12
16
ACCEPTED MANUSCRIPT 1
Figure 1 JMDC health insurance database (2005 Jan – 2016 Jan)
2
Administered statins for hyperlipidemia (n = 194,814)
3 4 5 6 7
RI PT
Lack of health insurance data for the screening period (12 months) and observation period (3 months) (n = 131,845)
With below diseases in the screening periods (n=6,875)
8
・Other respiratory diseases principally affecting the
9
・Tuberculosis (ICD10:A15-19)
SC
interstitium (ICD-10: J80-84)
Cohort 2
・Influenza and pneumonia (ICD-10:J12-J18)
10
M AN U
・Chronic lower respiratory diseases (ICD-10:J43-46) ・Lung diseases due to external agents (ICD-10:J60-J70) ・Other related diseases (ICD10:J85-99)
TE D
With IP diagnosis (ICD-10: J70.2J70.4, J84.1, and J84.9) and administered oral or injection steroids and KL-6 or SP-D and HRCT in the observation period (n = 25)
Case
EP
Cohort 1
AC C
With IP diagnosis (ICD10: J J70.2-J70.4, J84.1, and J84.9) and administered oral or injection steroids and KL-6 or SP-D and HRCT in the observation period (n = 4)
Without IP diagnosis (ICD10: J70.2-J70.4, J84.1, and J84.9) in the observation period (n = 4,030)
Control
Without IP diagnosis (ICD10: J70.2-J70.4, J84.1, and J84.9) in the observation period (n = 46,574)
Control
Case
17