A Disproportionality Analysis of the Adverse Effect Profiles of Methimazole and Propylthiouracil in Patients with Hyperthyroidism Using the Japanese Adverse Drug Event Report Database

Abstract

Background:

Antithyroid drugs (ATDs) are frequently used to achieve euthyroidism in patients with hyperthyroidism. ATDs cause characteristic common and rare adverse events; however, comprehensive comparisons between methimazole (MMI) and propylthiouracil (PTU) in terms of adverse events are limited.

Methods:

In this study, we thoroughly explored adverse events in association with MMI and PTU use with a disproportionality analysis using the Japanese Adverse Drug Event Report (JADER) database and evaluated the prevalence of MMI and PTU prescriptions using the National Database of Health Insurance Claims and Specific Health Checkups (NDB) Open Data Japan. We analyzed 3271 cases of MMI use and 1029 cases of PTU use with respect to 9789 preferred terms (PTs) for adverse events registered in the JADER database by calculating and comparing reporting odds ratios (RORs).

Results:

We found that 8 PTs, including agranulocytosis (p < 0.0001, 4.01-fold), aplasia cutis congenita (p < 0.0001, 123.22-fold), and exomphalos (p = 0.0002, 22.17-fold), demonstrated significantly higher RORs (more than 4-fold) for MMI use than for PTU use. Nineteen PTs, including anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (p < 0.0001, 29.84), rapidly progressive glomerulonephritis (p < 0.0001, 6.44), and pulmonary alveolar hemorrhage (p < 0.0001, 7.77), had RORs for PTU use more than four times those for MMI use. NDB Open Data Japan showed more frequent PTU prescriptions than MMI prescriptions for women of reproductive age.

Conclusions:

This large-scale study confirmed that a variety of congenital malformations were identified as having significantly high RORs for MMI use, while diseases related to ANCA-associated vasculitis were specific to PTU.

Introduction

Antithyroid drugs (ATDs), also known as thionamides, consist of methimazole (MMI), carbimazole, and propylthiouracil (PTU), and they are used to achieve and maintain euthyroidism in patients with hyperthyroidism by blocking the organification of iodine. ATDs are known to cause a wide range of side effects, including agranulocytosis, liver dysfunction, skin reactions, anti-neutrophil cytoplasmic antibody (ANCA)–associated vasculitis, arthritis-related symptoms, such as myalgia, arthralgia, and polyarthritis, aplastic anemia, thrombocytopenia, gastrointestinal effects, cholestasis, insulin-autoimmune syndrome, pancreatitis, and fetal congenital anomalies.^1

–4 Skin reactions, arthralgia, and gastrointestinal effects are not fatal, but they are commonly experienced by patients taking ATDs. Although agranulocytosis reportedly occurs in only 0.1–0.5% of patients receiving ATD therapy,⁵ it can be fatal and should be prevented.

Because most ATD-induced agranulocytosis is observed within a few months after the first administration of an ATD,⁶ monitoring the neutrophil count once every 2 weeks until 2 months after the first administration is recommended for patients using ATDs in the Japanese guidelines for Graves' disease treatment.

Agranulocytosis is a well-known and severe adverse event caused by ATDs. However, the incidence rates of ATD-induced agranulocytosis vary by study because the occurrence of agranulocytosis is uncommon. The estimated frequencies of ATD-related side effects other than agranulocytosis are 1–5% for arthralgia, 1–5% for gastrointestinal effects, 1–2% for polyarthritis, and 2–6% for fever and skin reactions.^1,7 Less frequent ATD-related side effects include an abnormal sense of taste or smell, ANCA-associated vasculitis, cholestasis, hepatitis, cholestatic jaundice, Stevens–Johnson syndrome, hypoprothrombinemia, insulin-autoimmune syndrome, sialadenitis, thrombocytopenia, aplastic anemia, and pancreatitis.^1,4 These less frequent side effects have been further categorized as “rare side effects” and “very rare side effects” in previous reports; however, the definitions differ by study.^1,4,8,9

The difference in the frequency of side effects between MMI and PTU remains highly controversial. The teratogenic effects of MMI compared with PTU have been widely indicated and the possible involvement of the teratogenicity of PTU in the adverse effects on the head and neck region and urinary tract has been suggested.^{10

–21} However, the details of the effects of MMI and PTU on each type of malformation are not yet fully understood. Furthermore, comprehensive analyses and comparative studies of a wide range of adverse events caused by ATDs have been limited.

The Japanese Adverse Drug Event Report (JADER) database contains all pharmacovigilance data that have been spontaneously reported to the Pharmaceuticals and Medical Devices Agency (PMDA) by pharmaceutical industries and medical care providers since April 2004. More than 94% of the reports in the JADER database are from medical professionals, indicating high reliability in pharmacovigilance studies.²² As of October 2021, the database contained 705,294 case reports of patients with adverse drug reactions. A total of 3271 patients with MMI use and 1029 patients with PTU use were registered, and among them, >100 experienced fetal adverse events. To the best of our knowledge, these data constitute the largest scale database. In this study, we studied the association of MMI or PTU use with every registered adverse event using the JADER database.

Materials and Methods

Study design and participants

This study did not require institutional review board approval because the JADER database is an anonymized database open to the public. This study attempted to examine the correlation between every type of adverse event and the ATDs available in Japan based on information contained in the JADER database. The JADER database contains data on reports in a format based on the E2B (M2) guideline of the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use. For this study, data recorded in the JADER database from April 2004 to September 2021 were downloaded from the PMDA website. We extracted data from the following files: patient demographic information (DEMO) files, drug information (DRUG) files, adverse events and patient outcomes (REAC) files, and disorder history (HIST) files. Each file type was linked to a specific identification number.

Statistical analysis

We assessed the association between MMI and PTU use and every type of adverse event reported in the database. In the JADER database, the disease names are listed according to the preferred terms (PTs) provided by the Medical Dictionary for Regulatory Activities/Japanese version (MedDRA/J). We also analyzed the prescription data in the 6th version of the National Database of Health Insurance Claims and Specific Health Checkups (NDB) Open Data Japan. These data contain prescription information from all Japanese medical practitioners' receipts for health insurance claims in fiscal year 2019. Cases that lacked essential data were removed from the analysis. The associations of each drug with every PT contained in the JADER database were assessed by calculating the reporting odds ratio (ROR), which is an established parameter in pharmacovigilance research. RORs were calculated using a two-by-two contingency table.

The RORs are expressed as point estimates with 95% confidence intervals [CIs]. Haldane–Anscombe 1/2 correction was applied to calculate RORs when the ROR could not be obtained because one or more 0s were present in the 2 × 2 contingency table used to calculate RORs. Correlations were considered significant when the lower bound of the two-sided CI of the ROR for the risk of the PT exceeded 1. Next, we compared the RORs between the MMI and PTU groups for the PTs with 3 or more patients in each group and assessed the difference with a z test.

If the z test for the RORs between the MMI and PTU groups showed p < 0.05, statistical significance was considered. We previously reported that deleting cases of concomitant use from the database could lessen the impact of the confounding influences of concomitant drugs; thus, in the secondary analysis, we deleted cases with concomitant MMI and PTU use and repeated the process.²³ The data were analyzed using R software (version 4.1.1; The R Foundation for Statistical Computing, Vienna, Austria).

Results

In the JADER database, 705,294 patients were registered. None of patients lacked identification numbers or generic names of the drugs used; thus, we could analyze all of them to calculate RORs. The participant flow is shown as a diagram (Fig. 1). A total of 3271 patients received MMI therapy, and 1029 patients received PTU therapy. In addition, the demographic characteristics of the whole population and each treatment group are given in Table 1. A total of 9789 PTs for adverse events were registered in the database. MMI had significant RORs for 186 PTs, and PTU had significant RORs for 143 PTs. PTs that demonstrated significant RORs and ROR point estimates higher than 5 for ATDs are given in Tables 2 and 3. Among these PTs, 96 were associated with 3 or more cases of either MMI or PTU use.

FIG. 1.

Study population flow diagram. ATDs, antithyroid drugs; MMI, methimazole; ROR, reporting odds ratio; PTU, propylthiouracil.

Table 1.

Population Characteristics

	Overall			MMI			PTU
	All	Male (n = 321,834)	Female (n = 305,483)	All	Male (n = 679)	Female (n = 2310)	All	Male (n = 132)	Female (n = 763)
Sex (male, %)	51.30			22.72			14.75
Age (years)	61.03 ± 21.10	61.75 ± 20.39	60.40 ± 21.52	48.87 ± 20.92	50.89 ± 20.95	49.45 ± 19.90	40.96 ± 19.89	45.76 ± 22.58	41.63 ± 18.25

Patients without age information were excluded.

MMI, methimazole; PTU, propylthiouracil.

Table 2.

Preferred Terms That Had Point Estimates of Reporting Odds Ratios Higher Than 5 for Methimazole Use

PT	Overall		MMI		ROR [CI]
PT	Cases (n)	Noncases (n)	Cases (n)	Noncases (n)	ROR [CI]
Adverse reactions
Agranulocytosis	3965	701,329	991	2280	102.17 [94.05–110.99]
Antineutrophil cytoplasmic antibody-positive	29	705,265	7	3264	68.43 [29.21–160.31]
Antineutrophil cytoplasmic antibody-positive vasculitis	429	704,865	83	3188	52.80 [41.45–67.26]
Polyarthritis	174	705,120	31	3240	46.96 [31.80–69.35]
Insulin autoimmune syndrome	138	705,156	24	3247	45.51 [29.26–70.77]
Goiter	66	705,228	11	3260	43.07 [22.52–82.35]
Microscopic polyangiitis	54	705,240	8	3263	37.41 [17.65–79.33]
Thyrotoxic crisis	72	705,222	9	3262	30.74 [15.28–61.86]
Granulocytopenia	790	704,504	86	3185	26.90 [21.45–33.74]
Polyarteritis nodosa	31	705,263	3	3268	23.02 [6.99–75.74]
Basedow's disease	199	705,095	18	3253	21.46 [13.20–34.87]
Rapidly progressive glomerulonephritis	177	705,117	16	3255	21.43 [12.80–35.86]
Abnormal thyroid function test	49	705,245	4	3267	19.10 [6.87–53.14]
Tonsillitis	131	705,163	10	3261	17.79 [9.33–33.93]
Eosinophilic myocarditis	41	705,253	3	3268	16.96 [5.23–54.96]
Systemic lupus erythematosus	397	704,897	25	3246	14.53 [9.68–21.81]
Lupus nephritis	61	705,233	3	3268	11.11 [3.48–35.48]
Aplastic anemia	507	704,787	24	3247	10.74 [7.12–16.20]
Reversible cerebral vasoconstriction syndrome	94	705,200	4	3267	9.55 [3.51–26.01]
Duodenal ulcer perforation	170	705,124	7	3264	9.23 [4.33–19.69]
Decreased granulocyte count	514	704,780	20	3251	8.74 [5.58–13.68]
Cholestasis	568	704,726	21	3250	8.29 [5.35–12.83]
Nephropathy	213	705,081	7	3264	7.31 [3.44–15.53]
Skin exfoliation	334	704,960	10	3261	6.64 [3.54–12.47]
Hyperthyroidism	1334	703,960	36	3235	6.01 [4.31–8.38]
Congestive cardiomyopathy	112	705,182	3	3268	5.91 [1.88–18.62]
Vascular device infection	117	705,177	3	3268	5.65 [1.80–17.80]
Nephritis	160	705,134	4	3267	5.51 [2.04–14.87]
Esophagitis	282	705,012	7	3264	5.47 [2.58–11.60]
Teratogenic effects
Congenital intestinal malformation	14	705,280	14	3257	6249.79 [372.73–104,794.55]^a
Choanal atresia	4	705,290	4	3267	1933.65 [104.08–35,924.11]^a
Aplasia cutis congenita	42	705,252	41	3230	8911.12 [1225.37–64,803.46]
Exomphalos	36	705,258	34	3237	3686.86 [885.37–15,352.86]
Vitello-intestinal duct remnant	19	705,275	17	3254	1833.80 [423.50–7940.55]
Intestinal fistula	34	705,260	23	3248	451.92 [220.12–927.83]
Congenital hyperthyroidism	21	705,273	12	3259	287.21 [120.93–682.10]
Esophageal atresia	16	705,278	9	3262	276.70 [102.99–743.43]
Skull malformation	11	705,283	5	3266	179.12 [54.64–587.22]
Congenital hypothyroidism	31	705,263	11	3260	118.44 [56.70–247.38]
Ventricular septal defect	99	705,195	11	3260	26.91 [14.37–50.42]
Patent ductus arteriosus	70	705,224	4	3267	13.02 [4.74–35.75]
Threatened abortion	63	705,231	3	3268	10.74 [3.37–34.26]

Haldane–Anscombe 1/2 correction applied.

PT, preferred term; ROR, reporting odds ratio.

Table 3.

Preferred Terms That Had Point Estimates of Reporting Odds Ratios Higher Than 5 for Propylthiouracil Use

PT	Overall		Propylthiouracil		ROR [CI]
PT	Cases (n)	Noncases (n)	Cases (n)	Noncases (n)	ROR [CI]
Adverse reactions
Antineutrophil cytoplasmic antibody-positive vasculitis	429	704,865	270	759	1575.30 [1278.88–1940.42]
Antineutrophil cytoplasmic antibody positive	29	705,265	15	1014	744.14 [358.26–1545.62]
Polychondritis	16	705,278	5	1024	312.61 [108.42–901.34]
Goiter	66	705,228	17	1012	241.42 [138.58–420.60]
Rapidly progressive glomerulonephritis	177	705,117	29	1000	137.97 [92.22–206.41]
Microscopic polyangiitis	54	705,240	8	1021	119.95 [56.47–254.79]
Granulomatosis with polyangiitis	21	705,273	3	1026	114.40 [33.65–388.97]
Pachymeningitis	23	705,271	3	1026	102.96 [30.55–347.02]
Abnormal thyroid function test	49	705,245	4	1025	61.07 [21.92–170.12]
Pyoderma gangrenosum	90	705,204	5	1024	40.45 [16.38–99.90]
Systemic lupus erythematosus	397	704,897	19	1010	35.03 [22.01–55.76]
Agranulocytosis	3965	701,329	126	903	25.46 [21.07–30.76]
Granulocytopenia	790	704,504	25	1004	22.90 [15.30–34.27]
Nephritis	160	705,134	5	1024	22.18 [9.08–54.16]
Polyarthritis	174	705,120	5	1024	20.34 [8.34–49.61]
IgA nephropathy	139	705,155	3	1026	15.14 [4.81–47.60]
Membranous glomerulonephritis	196	705,098	4	1025	14.31 [5.31–38.59]
Decreased granulocyte count	514	704,780	10	1019	13.70 [7.31–25.70]
Eosinophilic granulomatosis with polyangiitis	236	705,058	4	1025	11.84 [4.40–31.88]
Pulmonary alveolar hemorrhage	1140	704,154	19	1010	11.80 [7.47–18.65]
Lupus-like syndrome	180	705,114	3	1026	11.63 [3.71–36.47]
Arthralgia	1556	703,738	23	1006	10.48 [6.91–15.89]
Hypothyroidism	1840	703,454	24	1005	9.24 [6.15–13.88]
Cholestasis	568	704,726	5	1024	6.10 [2.53–14.75]
Vasculitis	355	704,939	3	1026	5.85 [1.87–18.25]
Teratogenic effects
Choanal atresia	4	705,290	3	1026	2059.25 [214.02–19,813.90]
Congenital hyperthyroidism	21	705,273	14	1015	1387.70 [558.92–3445.39]
Congenital hypothyroidism	31	705,263	16	1013	741.56 [365.64–1503.97]
Duodenal atresia	6	705,288	3	1026	686.42 [138.38–3404.96]
Goiter congenital	16	705,278	7	1022	535.96 [199.22–1441.90]
Polyhydramnios	25	705,269	5	1024	171.93 [64.41–458.99]
Exomphalos	36	705,258	7	1022	166.33 [72.70–380.56]
Cleft palate	19	705,275	3	1026	128.70 [37.44–442.37]
Ventricular septal defect	99	705,195	14	1015	114.27 [64.71–201.79]
Atrial septal defect	36	705,258	5	1024	110.92 [43.05–285.84]
Intestinal fistula	34	705,260	4	1025	91.61 [32.21–260.50]
Aplasia cutis congenita	42	705,252	4	1025	72.32 [25.76–203.00]
Neurosensory deafness	144	705,150	3	1026	14.60 [4.65–45.90]
Neonatal drug withdrawal syndrome	286	705,008	3	1026	7.27 [2.33–22.73]
Abortion	331	704,963	3	1026	6.28 [2.01–19.59]

IgA, immunoglobulin A.

We conducted a z test for the difference between log (ROR) for MMI and PTU use based on the 96 PTs. We found that among the 96 PTs, 8 had significantly high RORs for MMI (fold change >4, p < 0.05). The PT for adverse reactions with significant p-values (fold-change of RORs vs. PTU) was agranulocytosis (p < 0.0001 [4.01]), and the PTs for teratogenic effects with significant p-values were aplasia cutis congenita (p < 0.0001 [123.22]); exomphalos (p = 0.0002 [22.17]); vitello-intestinal duct remnant (p = 0.003 [22.73]); intestinal fistula (p = 0.014 [4.93]); and congenital intestinal malformation (p = 0.014 [54.68]). We also found that 26 PTs associated with PTU use had significantly high RORs (fold change >4, p < 0.05).

The PTs for adverse reactions with significant p-values (fold-change of RORs vs. MMI) were ANCA-positive vasculitis (p < 0.0001 [29.84]); rapidly progressive glomerulonephritis (RPGN) (p < 0.0001 [6.44]); pulmonary alveolar hemorrhage (p < 0.0001 [7.77]); ANCA positive (p < 0.0001 [10.87]); and goiter (p < 0.0001 [5.61]); and the PT for teratogenic effects with a significant p-value was congenital hypothyroidism (p = 0.0004 [6.26]) (Fig. 2). Supplementary Table S1 provides changes in the number of cases for PTs related to congenital malformations or pregnancy every 4 years and cases of congenital malformations in association with MMI use have decreased since 2016. Data on ATD dosage were lacking for 28.5% of the patients with MMI use and 38.6% of the patients with PTU use. Thus, we analyzed the remaining dose-available data and found that the mean doses of MMI and PTU were 13.25 and 142.65 mg/day, respectively, in the JADER database (Supplementary Fig. S1).

FIG. 2.

Volcano plot of the z test for reporting odds ratios between the methimazole and propylthiouracil groups in the Japanese Adverse Drug Event Report database. (A) Adverse reactions. (B) Teratogenic effects. Black squares, PTs that were >4-fold higher in the PTU group compared with the MMI group in the ROR z test (p < 0.05); black triangles, PTs that were >4-fold higher in the MMI group compared with the PTU group in the ROR z test (p < 0.05); white circles, PTs that were <4-fold higher or not significant (p > 0.05) between the MMI and PTU groups in the ROR z test.

Next, we omitted cases of concomitant MMI and PTU use and reanalyzed them to exclude the overlap between the two drugs. MMI had significant RORs for 176 PTs, and PTU had significant RORs for 114 PTs in this analysis. PTs that indicated ROR point estimates higher than 5 for ATDs are given in Tables 4 and 5. Among these PTs, 78 were associated with 3 or more cases of either MMI or PTU use. The z test revealed that among the 78 PTs, 1 had a significantly high ROR for MMI use and had at least a 4-fold change in the ROR for MMI use compared with that for PTU use (rhabdomyolysis, p = 0.017 [fold-change of RORs vs. PTU 5.66]).

Table 4.

Preferred Terms That Had Point Estimates of Reporting Odds Ratios Higher Than 5 for Methimazole Use

PT	Overall		Methimazole		ROR [CI]
PT	Cases (n)	Noncases (n)	Cases (n)	Noncases (n)	ROR [CI]
Adverse reactions
Antineutrophil cytoplasmic antibody-positive	29	705,265	5	3026	128.74 [43.12–384.38]
Antineutrophil cytoplasmic antibody-positive vasculitis	431	704,863	52	2979	114.38 [81.96–159.62]
Agranulocytosis	3977	701,317	952	2079	110.77 [101.74–120.59]
Insulin autoimmune syndrome	138	705,156	22	3009	44.97 [28.45–71.07]
Polyarthritis	176	705,118	26	3005	42.42 [27.89–64.52]
Microscopic polyangiitis	54	705,240	6	3025	34.77 [14.73–82.07]
Goiter	66	705,228	6	3025	32.34 [13.76–76.05]
Thyrotoxic crisis	72	705,222	8	3023	29.45 [14.10–61.52]
Granulocytopenia	798	704,496	82	2949	28.52 [22.62–35.96]
Polyarteritis nodosa	31	705,263	3	3028	24.81 [7.54–81.66]
Abnormal thyroid function test	49	705,245	4	3027	22.60 [8.09–63.13]
Basedow's disease	199	705,095	17	3014	21.97 [13.34–36.17]
Tonsillitis	135	705,159	10	3021	19.18 [10.05–36.59]
Eosinophilic myocarditis	41	705,253	3	3028	18.28 [5.64–59.26]
Rapidly progressive glomerulonephritis	178	705,116	10	3021	16.82 [8.84–31.98]
Aplastic anemia	509	704,785	24	3007	11.60 [7.69–17.51]
Reversible cerebral vasoconstriction syndrome	94	705,200	4	3027	10.29 [3.78–28.04]
Systemic lupus erythematosus	405	704,889	16	3015	10.27 [6.22–16.97]
Duodenal ulcer perforation	170	705,124	7	3024	9.96 [4.67–21.23]
Nephropathy	214	705,080	7	3024	7.88 [3.71–16.75]
Cholestasis	570	704,724	18	3013	7.68 [4.80–12.30]
Decreased granulocyte count	534	704,760	16	3015	7.62 [4.63–12.55]
Skin exfoliation	341	704,953	9	3022	6.44 [3.32–12.51]
Congestive cardiomyopathy	112	705,182	3	3028	6.37 [2.02–20.08]
Nephritis	160	705,134	4	3027	6.14 [2.27–16.57]
Vascular device infection	129	705,165	3	3028	6.09 [1.94–19.19]
Hyperthyroidism	1344	703,950	33	2998	5.95 [4.21–8.43]
Esophagitis	284	705,010	7	3024	5.90 [2.78–12.50]
Steroid diabetes	314	704,980	7	3024	5.29 [2.50–11.19]
Ovarian cancer	181	705,113	4	3027	5.23 [1.94–14.11]
Pharyngitis	244	705,050	5	3026	5.04 [2.07–12.22]
Teratogenic effects
Exomphalos	36	705,258	29	3002	13,779.46 [841.73–225,576.49]
Congenital intestinal malformation	14	705,280	12	3019	5805.88 [343.66–98,086.51]
Aplasia cutis congenita	42	705,252	37	2994	8665.87 [1188.58–63,182.46]
Vitello-intestinal duct remnant	19	705,275	15	3016	1743.78 [398.59–7628.74]
Congenital hyperthyroidism	21	705,273	6	3025	1390.88 [167.39–11,556.83]
Intestinal fistula	34	705,260	19	3012	402.13 [191.20–845.74]
Congenital hypothyroidism	32	705,262	9	3022	348.06 [123.81–978.49]
Esophageal atresia	16	705,278	7	3024	231.89 [81.29–661.51]
Skull malformation	11	705,283	5	3026	193.11 [58.90–633.11]
Ventricular septal defect	99	705,195	9	3022	27.48 [13.76–54.88]
Patent ductus arteriosus	70	705,224	4	3027	14.25 [5.19–39.15]
Threatened abortion	63	705,231	3	3028	11.98 [3.75–38.25]

Cases of concurrent propylthiouracil use were omitted.

Table 5.

Preferred Terms That Had Point Estimates of Reporting Odds Ratios Higher Than 5 for Propylthiouracil Use

PT	Overall		Propylthiouracil		ROR [CI]
PT	Cases (n)	Noncases (n)	Cases (n)	Noncases (n)	ROR [CI]
Adverse reactions
Antineutrophil cytoplasmic antibody-positive vasculitis	431	704,863	239	550	2847.40 [2233.51–3630.01]
Antineutrophil cytoplasmic antibody-positive	29	705,265	13	776	1305.26 [556.31–3062.48]
Polychondritis	17	705,277	5	784	406.55 [140.93–1172.83]
Goiter	66	705,228	12	777	251.84 [132.30–479.41]
Rapidly progressive glomerulonephritis	178	705,116	23	766	152.54 [97.55–238.54]
Granulomatosis with polyangiitis	21	705,273	3	786	148.69 [43.71–505.79]
Pachymeningitis	23	705,271	3	786	140.86 [41.60–476.98]
Microscopic polyangiitis	54	705,240	6	783	134.33 [56.79–317.74]
Abnormal thyroid function test	49	705,245	4	785	87.15 [31.14–243.87]
Pyoderma gangrenosum	92	705,202	5	784	53.23 [21.54–131.58]
Agranulocytosis	3977	701,317	87	702	29.98 [23.92–37.57]
Nephritis	160	705,134	5	784	29.61 [12.12–72.37]
Granulocytopenia	798	704,496	21	768	28.05 [18.06–43.55]
Systemic lupus erythematosus	405	704,889	10	779	24.85 [13.21–46.77]
Membranous glomerulonephritis	196	705,098	4	785	18.70 [6.93–50.47]
Pulmonary alveolar hemorrhage	1156	704,138	19	770	15.52 [9.81–24.56]
Eosinophilic granulomatosis with polyangiitis	237	705,057	3	786	11.58 [3.70–36.26]
Decreased granulocyte count	534	704,760	6	783	11.00 [4.90–24.69]
Hypothyroidism	1854	703,440	19	770	9.66 [6.11–15.26]
Arthralgia	1667	703,627	16	773	9.57 [5.82–15.73]
Thrombocytopenic purpura	433	704,861	3	786	6.35 [2.04–19.82]
Acute hepatic failure	482	704,812	3	786	5.62 [1.80–17.52]
Teratogenic effects
Congenital hyperthyroidism	21	705,273	8	781	7182.92 [897.29–57,499.98]
Congenital hypothyroidism	32	705,262	14	775	2111.22 [809.20–5508.22]
Goiter congenital	16	705,278	7	782	896.71 [313.79–2562.51]
Ventricular septal defect	99	705,195	12	777	142.48 [77.19–262.99]
Polyhydramnios	25	705,269	3	786	133.82 [39.68–451.25]
Atrial septal defect	39	705,255	3	786	86.33 [26.34–282.98]
Neurosensory deafness	147	705,147	3	786	18.98 [6.03–59.68]
Neonatal drug withdrawal syndrome	286	705,008	3	786	9.45 [3.02–29.55]
Low neonatal birth weight	454	704,840	3	786	5.97 [1.91–18.62]

Cases of concurrent methimazole use were omitted.

Furthermore, 12 PTs for adverse reactions and 8 PTs for PTU teratogenic effects had significantly high RORs compared with MMI and had at least a 4-fold change in the ROR (ANCA-positive vasculitis, p < 0.0001 [fold-change of RORs vs. MMI 24.89]; RPGN, p < 0.0001 [9.07]; pulmonary alveolar hemorrhage, p < 0.0001 [9.32]; goiter, p = 0.00017 [7.79]; ANCA positive, p = 0.001 [10.14]; and ventricular septal defect, p = 0.0005 [5.19]) (Fig. 3). The abovementioned results were summarized with PTs categorized into standardized MedDRA queries (Supplementary Table S2). When we analyzed the data separately by sex, there were no overall significant sex differences (Supplementary Tables S3–S6). A recent report suggested an association between MMI use and acute pancreatitis.²⁴

FIG. 3.

Volcano plot of the z test for reporting odds ratios between the methimazole and propylthiouracil groups in the Japanese Adverse Drug Event Report database after the exclusion of cases with concurrent use of the other antithyroid drug. (A) Adverse reactions. (B) teratogenic effects. Black squares, PTs that were >4-fold higher in the PTU group compared with the MMI group in the ROR z test (p < 0.05); black triangles, PTs that were >4-fold higher in the MMI group compared with the PTU group in the ROR z test (p < 0.05); white circles, PTs that were <4-fold higher or not significant (p > 0.05) between the MMI and PTU groups in the ROR z test.

Analysis of 14 pancreatitis-related PTs in the JADER database revealed that none of the PTs showed significant RORs, but 11 cases of pancreatitis were found only in patients with MMI use and not in those with PTU use. We next compared our results with those of two previous large-scale observational studies in the context of teratogenic effects (Supplementary Table S7). A Scandinavian study reported 16–18 cases of malformations of the digestive system and a Korean study reported 27 cases of MMI embryopathy.^12,14 On the contrary, in this study, 47 and 107 cases were reported, respectively.

An analysis of the 6th version of the NDB Open Data Japan revealed that 20.2% of all ATD tablets prescribed in Japan in 2019 were PTU tablets and 79.8% were MMI tablets, in 2014, 22.2% were PTU tablets and 77.8% were MMI tablets. Because MMI and PTU doses are often 2 to 3 tablets, the number of tablets could be proportional to the number of patients. Moreover, 59.4% of PTU prescriptions and 37.5% of MMI prescriptions were issued to women aged 20–49 years in 2019, and a similar trend was observed in 2014 (Fig. 4A, B). Thus, PTU was used more frequently than MMI in women of reproductive age. This trend in ATD prescription for women of reproductive age may contribute to the observed higher RORs for congenital malformations in association with PTU use compared with MMI use.

FIG. 4.

Age-specific ratio of methimazole and propylthiouracil prescriptions in Japan. The data were obtained from the (A) 1st version in 2014 and (B) 6th version in 2019 of the National Database of Health Insurance Claims and Specific Health Checkups Open Data Japan.

Discussion

The JADER database contained a large number of rare adverse events associated with ATDs, including 1078 cases of agranulocytosis, 322 cases of ANCA-associated vasculitis, 58 cases of rhabdomyolysis, and >170 cases of congenital malformations. To the best of our knowledge, this study analyzed adverse events of ATDs by using the highest number of cases thus far. In this study, we explored adverse events associated with MMI and PTU use with a disproportionality analysis using the JADER database. We also compared the differences in RORs between MMI and PTU use.

We also excluded cases of patients using both MMI and PTU because these cases can mitigate biases that often appear in this type of analysis.²³ The higher ROR values for MMI and PTU use were largely unaltered before and after the exclusion. Although the significantly higher RORs for PTU use compared with MMI use identified by the z test were similar even after the exclusion of cases of concurrent PTU and MMI use, the associations with most PTs, which were higher for MMI use and included congenital malformations, became insignificant. Because patients who are pregnant or may become pregnant are often switched from MMI to PTU to avoid gestational side effects, both PTU and MMI could be registered as used drugs. Thus, cases of congenital malformations were greatly affected by this exclusion, probably owing to the switching of ATDs.

A systematic review and large-scale observational studies have consistently shown a significant risk for perinatal side effects in pregnant women with MMI use, whereas mixed findings have been reported for this risk with PTU use.^12,14,25 Another study concluded that situs inversus ± dextrocardia and cardiac outflow tract defects were associated with exposure to PTU, whereas choanal atresia, omphalocele, and total situs inversus were associated with exposure to MMI.²⁰ In addition to choanal atresia and exomphalos (omphalocele), the use of MMI showed significantly higher RORs for vitello-intestinal duct remnant, intestinal fistula, and esophageal atresia in this study. MMI may affect the development of the gastrointestinal tract, which results in a wide range of malformations. Of note, the RORs for ventricular septal defects and atrial septal defects had significantly higher values for PTU use, consistent with the abovementioned finding.²⁰

The cardiac outflow tract, ventricular septum, and atrial septum are derived from the endocardial cushion, and PTU might modulate the formation of the endocardial cushion during their developmental periods. The interim report of the Pregnancy Outcomes of Exposure to Methimazole Study (POEM Study) in Japan demonstrated that 5 of 85 neonates in the MMI group had congenital malformations, but none of the 121 neonates in the PTU group had malformations.²⁶ In the POEM study, “methimazole-related congenital malformations” included choanal atresia, esophageal atresia, aplasia cutis congenita, umbilical duct defects, and exomphalos. In this study, the PTs with significantly higher congenital malformation-related RORs for MMI use included congenital gastrointestinal malformation, choanal atresia, and aplasia cutis congenita. Therefore, this study and POEM study demonstrated common congenital malformations.

The decline in reports of teratogenicity in association with MMI use since 2016 in the JADER database may have been noted by the relevant interim report of the POEM study in 2011. However, as described previously, the ROR for congenital malformations associated with MMI use did not reach statistical significance compared with that for PTU use when patients with concurrent MMI and PTU use were excluded. Further investigations are needed to analyze switching from MMI to PTU during gestational periods to clarify the effects of MMI on congenital malformations in the excluded patients.

Agranulocytosis is a life-threatening complication of ATDs for which the granulocyte count must be assessed routinely. However, whether the risk of agranulocytosis differs between PTU and MMI use remains controversial. The reported incidence of ATD-mediated agranulocytosis is 0.2–0.5%.²⁷ A Danish study showed a higher incidence of agranulocytosis for patients with PTU use (0.27% for PTU use vs. 0.11% for MMI use; p = 0.02), while Japanese and Korean studies indicated approximately twice as many agranulocytosis cases in patients with MMI use compared with those with PTU use.^5,28 Conversely, a meta-analysis demonstrated that the agranulocytosis incidence did not significantly differ between the MMI and PTU groups.²⁹

In this study, the ROR of agranulocytosis for MMI use was approximately four times higher than that for PTU use, and this difference remained significant (at 3.69-fold) even after the elimination of patients with concurrent MMI and PTU use. Because MMI-induced agranulocytosis, unlike PTU-induced agranulocytosis, is dose dependent, the relative risk ratio between MMI and PTU use should be interpreted with caution. The risk of agranulocytosis for MMI and PTU use may depend on a patient's ethnic background. Large-scale studies in other regions are warranted to further determine the safety of ATDs in each country.

Insulin autoimmune syndrome (IAS) produces autoantibodies to human insulin, which results in spontaneous hypoglycemia. MMI is thought to be a major cause of IAS, followed by tiopronin and alpha lipoic acid.³⁰ However, PTU-induced IAS has not yet been reported. In this study, the z test showed a significantly higher ROR for IAS associated with MMI use. Although the difference became insignificant after the exclusion of patients with concurrent PTU and MMI use, 22 cases of IAS were associated with MMI use and no cases were associated with PTU use. This finding supports the idea that IAS is a specific complication of MMI use but not of PTU use. As described in a recent systematic review,³¹ IAS is predominantly observed in Asian countries probably owing to specific HLAs (e.g., HLA-DRB1*0406). Comparisons with the incidence of IAS in other countries may also be of interest.

ANCA-associated vasculitis, also known as ANCA-positive vasculitis, is a disorder characterized by severe, systemic, small-vessel vasculitis. Japanese ANCA-associated vasculitis cases are frequently reported, but the racial/ethnic differences in the prevalence of ANCA-associated vasculitis are unclear.³² The proportion of ANCA-positive cases was estimated at 4–64% in patients with PTU use and 0–16% in patients with MMI use.³³ A postmarketing adverse event report demonstrated that the ratio of the estimated incidence of ANCA-associated vasculitis was 39.2-fold higher for PTU use than for MMI use.³⁴ We found that the ROR of ANCA-associated vasculitis was 29.84-fold higher for PTU use than for MMI use in this study. ANCA-associated vasculitis is related to clinical manifestations, including alveolar hemorrhage, RPGN, membranous nephritis, and pyoderma gangrenosum.^35

–41

A previous study analyzing 16 patients with PTU-induced ANCA-associated vasculitis indicated that more than half of the patients presented with multiple organ complications, particularly kidney-related symptoms.⁴² In this study, we identified 270 cases of ANCA-associated vasculitis with PTU use, and the difference in the RORs between PTU and MMI use for the symptoms related to ANCA-associated vasculitis was less pronounced after the exclusion of patients with concurrent PTU and MMI use, for example, alveolar hemorrhage (ROR: 1.66 for MMI use vs. 15.52 for PTU use; 9.32-fold change; p < 0.0001) and RPGN (ROR: 16.82 for MMI use vs. 152.54 for PTU use; 9.07-fold change; p < 0.0001). Because multiple PTs for one disease could be registered in the JADER database, careful interpretation is required regarding the relevant PT group. ANCA-associated vasculitis without severe symptoms might be predominantly registered as “ANCA-positive vasculitis” in the JADER database.

The incidence of rhabdomyolysis induced by ATDs was reported to be relatively low in previous studies. Four reports demonstrated eight cases of rhabdomyolysis associated with MMI use, but not with PTU use, in patients with Graves' disease.^43

–46 In this study, we identified 43 and 2 cases of rhabdomyolysis in response to MMI and PTU use, respectively, after excluding patient with concurrent MMI and PTU use (ROR: 1.49 for MMI use vs. 0.26 for PTU use; 5.66-fold change; p = 0.016). Periodic confirmation of serum CPK levels may be important for patients using MMI.

An analysis of NDB Open Data Japan, a comprehensive prescription database, showed that prescriptions for 5-mg MMI tablets were 3.86 times more common than prescriptions for 50-mg PTU tablets in terms of the total number of tablets. A Japanese randomized controlled study demonstrated that 30 mg/day MMI, 15 mg/day MMI, and 300 mg/day PTU were equally effective for the initial treatment of mild to moderate hyperthyroidism caused by Graves' disease.⁴⁷ The analysis suggested that PTU was more likely to be prescribed to women of reproductive age. This trend in PTU prescription could cause a bias to increase the ROR of PTU use for PTs related to congenital malformations or pregnancy. It is difficult to compare the total number of adverse events between the MMI and PTU groups because >9000 PTs cannot be analyzed separately for teratogenic effects and nonteratogenic effects.

This study is subject to several limitations. The reports in the JADER database often lacked data on preexisting comorbidities and historical dose data. For example, 83.2% of patients with MMI use and 85.8% of patients with PTU use only had data for a single drug. Therefore, we could not estimate the effects on the results. Evaluation of teratogenic effects in pregnant women was difficult because all PTs were reported as fetal rather than maternal in the database. We applied the z test to compare RORs between PTU and MMI use for each PT, but the comparison method between RORs has not been fully established. Although there are some PT drug combinations with possible inverse causality, they should also be analyzed to avoid biases because of arbitrary omission.

Although all cases should be registered in the database to calculate a representative relative risk, RORs cannot be described as a representative normal risk ratio or an odds ratio when a biased population is registered. Because the JADER database is a spontaneous reporting system, several unavoidable limitations are associated with this type of database analysis. These limitations include missing data, underreporting, overreporting, data registry errors, and a lack of control data. In this study, we did not gather data on coexisting illnesses, drug doses, or durations of exposure. Further analytical observational studies and monitoring must be performed.

Conclusion

We comprehensively evaluated and compared the adverse events associated with the use of MMI and PTU using the JADER database. Eight PTs, including agranulocytosis, had significantly higher RORs for MMI use, and 19 PTs, including ANCA-associated vasculitis, had significantly higher RORs for PTU use. Because the JADER database is a spontaneous reporting system, several limitations must be considered.

Footnotes

Acknowledgments

The authors acknowledge all contributors to the JADER database. The authors also thank Misa Katayama (Yokohama City University) for her excellent secretarial assistance.

Authors' Contributions

J.S. conceived the concept of this study. M.A. and J.S. wrote the article. M.A. was responsible for data collection and performed the statistical analysis. H.K., T.T., and Y.T. made substantial contributions to the interpretation of data and revised the article for important intellectual content. All authors approved the submission of the article. J.S. is the guarantor of this work and, as such, has full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Author Disclosure Statement

None of the authors have any conflicts of interest to disclose.

Funding Information

J.S. acknowledges support from a Grant-in-Aid for Scientific Research (C) 20K08866 from MEXT of Japan and the Japan IDDM network, Japan Diabetes Foundation, Uehara Memorial Foundation, Mochida Memorial Foundation for Medical and Pharmaceutical Research, Kamome Memorial Foundation of Yokohama City University, Naito Foundation, Astellas Foundation for Research on Metabolic Disorders, Manpei Suzuki Diabetes Foundation, and A*STAR-AMED JOINT CALL for the Strategic International Collaborative Research Program (SICORP).

Supplementary Material

Supplementary Figure S1

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Table S4

Supplementary Table S5

Supplementary Table S6

Supplementary Table S7

References

Cooper

. Antithyroid drugs. N Engl J Med, 2005; 352(9):905–917; doi: 10.1056/NEJMra042972

Sundaresh

, Brito

, Wang

, et al. Comparative effectiveness of therapies for Graves' hyperthyroidism: A systematic review and network meta-analysis. J Clin Endocrinol Metab, 2013; 98(9):3671–3677; doi: 10.1210/jc.2013-1954

Azizi

, Malboosbaf

. Safety of long-term antithyroid drug treatment? A systematic review. J Endocrinol Invest, 2019; 42(11):1273–1283; doi: 10.1007/s40618-019-01054-1

Lane

, Cheetham

. Graves' disease: Developments in first-line antithyroid drugs in the young. Expert Rev Endocrinol Metab, 2020; 15(2):59–69; doi: 10.1080/17446651.2020.1735359

Watanabe

, Narimatsu

, Noh

, et al. Antithyroid drug-induced hematopoietic damage: A retrospective cohort study of agranulocytosis and pancytopenia involving 50,385 patients with Graves' disease. J Clin Endocrinol Metab, 2012; 97(1):E49–E53; doi: 10.1210/jc.2011-2221

Yang

, Zhu

, Zhong

, et al. Characteristics of antithyroid drug-induced agranulocytosis in patients with hyperthyroidism: A retrospective analysis of 114 cases in a single institution in china involving 9690 patients referred for radioiodine treatment over 15 years. Thyroid, 2016; 26(5):627–633; doi: 10.1089/thy.2015.0439

Mazza

, Carlini

, Flecchia

, et al. Long-term follow-up of patients with hyperthyroidism due to Graves' disease treated with methimazole. Comparison of usual treatment schedule with drug discontinuation vs continuous treatment with low methimazole doses: a retrospective study. J Endocrinol Invest, 2008; 31(10):866–872; doi: 10.1007/bf03346433

Kahaly

, Bartalena

, Hegedüs

, et al. 2018 European Thyroid Association Guideline for the Management of Graves' Hyperthyroidism. Eur Thyroid J, 2018; 7(4):167–186; doi: 10.1159/000490384

Wijaya

, Ong-Ramos

. Methimazole-induced aplastic anemia with concomitant hepatitis in a young filipina with Graves' disease. J ASEAN Fed Endocr Soc, 2019; 34(1):99–102; doi: 10.15605/jafes.034.01.16

10.

Andersen

, Andersen

. Antithyroid drugs and birth defects. Thyroid Res, 2020; 13:11; doi: 10.1186/s13044-020-00085-8

11.

Yoshihara

, Noh

, Yamaguchi

, et al. Treatment of graves' disease with antithyroid drugs in the first trimester of pregnancy and the prevalence of congenital malformation. J Clin Endocrinol Metab, 2012; 97(7):2396–2403; doi: 10.1210/jc.2011-2860

12.

Seo

, Kim

, Chung

. Antithyroid drugs and congenital malformations: A Nationwide Korean Cohort Study. Ann Intern Med, 2018; 168(6):405–413; doi: 10.7326/m17-1398

13.

Andersen

, Olsen

, Wu

, et al. Birth defects after early pregnancy use of antithyroid drugs: A Danish nationwide study. J Clin Endocrinol Metab, 2013; 98(11):4373–4381; doi: 10.1210/jc.2013-2831

14.

Andersen

, Knøsgaard

, Olsen

, et al. Maternal thyroid function, use of antithyroid drugs in early pregnancy, and birth defects. J Clin Endocrinol Metab, 2019; 104(12):6040–6048; doi: 10.1210/jc.2019-01343

15.

Chen

, Xirasagar

, Lin

, et al. Risk of adverse perinatal outcomes with antithyroid treatment during pregnancy: A nationwide population-based study. BJOG, 2011; 118(11):1365–1373; doi: 10.1111/j.1471-0528.2011.03019.x

16.

Andersen

, Lönn

, Vestergaard

, et al. Birth defects after use of antithyroid drugs in early pregnancy: A Swedish nationwide study. Eur J Endocrinol, 2017; 177(4):369–378; doi: 10.1530/eje-17-0314

17.

Gianetti

, Russo

, Orlandi

, et al. Pregnancy outcome in women treated with methimazole or propylthiouracil during pregnancy. J Endocrinol Invest, 2015; 38(9):977–985; doi: 10.1007/s40618-015-0281-z

18.

Korelitz

, McNally

, Masters

, et al. Prevalence of thyrotoxicosis, antithyroid medication use, and complications among pregnant women in the United States. Thyroid, 2013; 23(6):758–765; doi: 10.1089/thy.2012.0488

19.

, Rivkees

, Chandra

, et al. Gestational thyrotoxicosis, antithyroid drug use and neonatal outcomes within an integrated healthcare delivery system. Thyroid, 2015; 25(6):698–705; doi: 10.1089/thy.2014.0434

20.

Clementi

, Di Gianantonio

, Cassina

, et al. Treatment of hyperthyroidism in pregnancy and birth defects. J Clin Endocrinol Metab, 2010; 95(11):E337–E341; doi: 10.1210/jc.2010-0652

21.

Andersen

, Olsen

, Wu

, et al. Severity of birth defects after propylthiouracil exposure in early pregnancy. Thyroid, 2014; 24(10):1533–1540; doi: 10.1089/thy.2014.0150

22.

Matsuda

, Aoki

, Kawamata

, et al. Bias in spontaneous reporting of adverse drug reactions in Japan. PLoS One, 2015; 10(5):e0126413; doi: 10.1371/journal.pone.0126413

23.

Arai

, Shirakawa

, Konishi

, et al. Bullous pemphigoid and dipeptidyl peptidase 4 inhibitors: A disproportionality analysis based on the Japanese Adverse Drug Event Report Database. Diabetes Care, 2018; 41(9):e130–e132; doi: 10.2337/dc18-0210

24.

Brix

, Lund

, Henriksen

, et al. Methimazole and risk of acute pancreatitis. Lancet Diabetes Endocrinol, 2020; 8(3):187–189; doi: 10.1016/s2213-8587(20)30025-5

25.

Laurberg

, Andersen

. Therapy of endocrine disease: Antithyroid drug use in early pregnancy and birth defects: Time windows of relative safety and high risk?. Eur J Endocrinol, 2014; 171(1):R13–R20; doi: 10.1530/eje-14-0135

26.

Arata

. [Pregnancy outcomes of exposure to methimazole (POEM) study]. Nihon Rinsho, 2012; 70(11):1976–1982.

27.

Kobayashi

, Noh

, Mukasa

, et al. Characteristics of agranulocytosis as an adverse effect of antithyroid drugs in the second or later course of treatment. Thyroid, 2014; 24(5):796–801; doi: 10.1089/thy.2013.0476

28.

Kim

, Yoon

, Jeon

, et al. Characteristics of Korean patients with antithyroid drug-induced agranulocytosis: A multicenter study in Korea. Endocrinol Metab (Seoul), 2015; 30(4):475–480; doi: 10.3803/EnM.2015.30.4.475

29.

, Wu

, Li

, et al. Side effects of PTU and MMI in the treatment of hyperthyroidism: A systematic review and meta-analysis. Endocr Pract, 2020; 26(2):207–217; doi: 10.4158/ep-2019-0221

30.

Uchigata

, Hirata

, Iwamoto

. Drug-induced insulin autoimmune syndrome. Diabetes Res Clin Pract, 2009; 83(1):e19–e20; doi: 10.1016/j.diabres.2008.10.015

31.

Lin

, Chen

, Ning

. Insulin autoimmune syndrome: A systematic review. Int J Endocrinol, 2023; 2023:1225676; doi: 10.1155/2023/1225676

32.

Morita

, Ueda

, Eguchi

. Anti-thyroid drug-induced ANCA-associated vasculitis: A case report and review of the literature. Endocr J, 2000; 47(4):467–470.

33.

Balavoine

, Glinoer

, Dubucquoi

, et al. Antineutrophil cytoplasmic antibody-positive small-vessel vasculitis associated with antithyroid drug therapy: How significant is the clinical problem?. Thyroid, 2015; 25(12):1273–1281; doi: 10.1089/thy.2014.0603

34.

Noh

, Yasuda

, Sato

, et al. Clinical characteristics of myeloperoxidase antineutrophil cytoplasmic antibody-associated vasculitis caused by antithyroid drugs. J Clin Endocrinol Metab, 2009; 94(8):2806–2811; doi: 10.1210/jc.2008-2700

35.

Daimon

, Araki

, Matsuda

, et al. Antineutrophil cytoplasmic antibody-positive rapidly progressive glomerulonephritis caused by propylthiouracil: A case report. Clin Exp Nephrol, 1997; 1(3):225–228; doi: 10.1007/BF02480699

36.

Darben

, Savige

, Prentice

, et al. Pyoderma gangrenosum with secondary pyarthrosis following propylthiouracil. Australas J Dermatol, 1999; 40(3):144–146; doi: 10.1046/j.1440-0960.1999.00346.x

37.

Kudoh

, Kuroda

, Shimamoto

, et al. Propylthiouracil-induced rapidly progressive glomerulonephritis associated with antineutrophil cytoplasmic autoantibodies. Clin Nephrol, 1997; 48(1):41–43.

38.

Ohtsuka

, Yamashita

, Doi

, et al. Propylthiouracil-induced alveolar haemorrhage associated with antineutrophil cytoplasmic antibody. Eur Respir J, 1997; 10(6):1405–1407.

39.

Romas

, Henderson

, Kirkham

. Propylthiouracil therapy: An unusual cause of antineutrophil cytoplasmic antibody associated alveolar hemorrhage. J Rheumatol, 1995; 22(4):803.

40.

Yoshida

, Hanai

, Nakagomi

, et al. Membranous nephropathy with proteinase 3-ANCA-associated vasculitis successfully treated with rituximab. Intern Med, 2021; 60(1):145–150; doi: 10.2169/internalmedicine.4752-20

41.

, Patel

, Ortega-Loayza

. Drug-induced pyoderma gangrenosum: A model to understand the pathogenesis of pyoderma gangrenosum. Br J Dermatol, 2017; 177(1):72–83; doi: 10.1111/bjd.15193

42.

Yang

, Yao

, Dong

, et al. Clinical characteristics and outcomes of propylthiouracil-induced antineutrophil cytoplasmic antibody-associated vasculitis in patients with Graves' disease: A median 38-month retrospective cohort study from a single institution in China. Thyroid, 2017; 27(12):1469–1474; doi: 10.1089/thy.2017.0468

43.

Ito

, Katahira

, Hanakita

, et al. A case of elevation of serum creatine kinase with antithyroid medications for Graves' disease. J Endocrinol Metabol, 2013; 2(6):244–247; doi: 10.4021/jem130w

44.

Mizuno

, Sugiyama

, Nishi

, et al. Elevation of serum creatine kinase in response to medical treatment of Graves' disease in children. Acta Paediatr, 2006; 95(2):243–245; doi: 10.1111/j.1651-2227.2006.tb02214.x

45.

Suzuki

, Ichikawa

, Nagai

, et al. Elevation of serum creatine kinase during treatment with antithyroid drugs in patients with hyperthyroidism due to Graves disease. A novel side effect of antithyroid drugs. Arch Intern Med, 1997; 157(6):693–696.

46.

Kim

, Kim

, Huh

, et al. Elevation of serum creatine kinase during methimazole treatment of Graves disease in a 13-year-old girl and a literature review of similar cases. Ann Pediatr Endocrinol Metab, 2015; 20(2):106–109; doi: 10.6065/apem.2015.20.2.106

47.

Nakamura

, Noh

, Itoh

, et al. Comparison of methimazole and propylthiouracil in patients with hyperthyroidism caused by Graves' disease. J Clin Endocrinol Metab, 2007; 92(6):2157–2162; doi: 10.1210/jc.2006-2135