2021 Volume 44 Issue 10 Pages 1499-1505
A major adverse effect of benzbromarone is hepatotoxicity. Therefore, periodic liver function tests are required at least for the first 6 months of benzbromarone administration. However, it is not clear whether the relevant blood tests are implemented appropriately. Here, we performed a cross-sectional survey of the implementation status of liver function tests in patients who were newly prescribed benzbromarone, using the Japanese large claims database. Male patients who were newly prescribed benzbromarone from January 2010 to December 2016 were included. We targeted patients who continued benzbromarone during the observation period (up to 180 d from the start of administration). The primary endpoint was the proportion of patients in whom periodic liver function tests were implemented. A periodic liver function test was defined as one or more liver function tests performed during both 1–90 and 91–180 d of initial benzbromarone administration. We labeled the tests as a “periodic test” or “non-periodic test” based on whether periodic liver function tests were performed or not, respectively. Furthermore, factors influencing non-periodic test were analyzed. Periodic testing was implemented only in 28.7% of patients. Additionally, factors such as number of hospital beds ≤19 (compared to 100–199 beds) and duration of the first prescription of benzbromarone were associated with non-periodic testing. Our study revealed that periodic liver function tests are not performed sufficiently in Japan. Thus, clinicians prescribing benzbromarone should be educated about the test. Our blood-test-based approach should be applied to other drugs and countries in future research.
For more than 30 years, benzbromarone, a uricosuric drug, has been prescribed for the treatment of gout.1) It inhibits the tubular reabsorption of urate by human proximal tubule urate transporter 1,2,3) leading to increased urate elimination via urine. Although the clinical efficacy of benzbromarone in reducing serum urate has been well established,4–6) it has major adverse effects such as hepatotoxicity. This adverse effect was first reported in the Netherlands in 1994.7) Subsequently, there were reports of three Japanese patients with fulminant hepatitis, including a case of death.8–10) In 2008, it was estimated that fulminant hepatitis occurred in 1 per 17000 patients in Europe.11) Although the underlying mechanism is not well established, some hypotheses have been proposed; for example, hepatotoxicity may be caused by metabolite formation and mitochondrial toxicity.12–14)
Because benzbromarone causes hepatotoxicity, it has not been approved in the United States. Moreover, it has been widely withdrawn from the market in European countries.15,16) Benzbromarone is still marketed in Japan; nonetheless, in February 2000, a “yellow letter” (a healthcare professional letter with important safety information about drugs and devices, requiring immediate communication from a regulatory agency) was issued.17) The yellow letter and package insert of benzbromarone state that periodic liver function tests must be performed at least for the first 6 months of benzbromarone administration.17,18) Furthermore, an alert from the Pharmaceuticals and Medical Devices Agency specifically indicated the frequency of blood tests to be more than one every 3 months within the first 6 months of benzbromarone administration.19) This is because benzbromarone-associated hepatic disorder occurs between the start of treatment and 6 months (the periodic tests are also recommended after the first 6 months).19) Stamp et al. reported that abnormal liver function test results were obtained in 20.5% of the enrolled patients within 6 months from the start of benzbromarone treatment.4)
Nasuhara et al. investigated the implementation status of blood tests, in accordance with the instructions in the package insert, in patients who received thiamazole in a single hospital in Japan.20) They reported that periodic blood tests that met the requirements stipulated in the package insert warning were implemented in 15.7% of the patients (45/287 patients).20) However, to the best of our knowledge, there is no other surveillance study. Thus, further research on other drugs is necessary for safe pharmacotherapy. Moreover, although it is important that clinicians should carefully monitor the liver function after initiation of benzbromarone, it is not clear whether liver function tests are implemented appropriately in Japan. In this study, we performed a cross-sectional survey of the current implementation status of liver function tests in patients who were newly prescribed benzbromarone, using the Japanese large claims database. In addition, the occurrence of liver injury was investigated in the eligible patients.
For this epidemiological study, we employed the JMDC claims database (JMDC Inc., Tokyo, Japan), which comprises details of approximately 7.3 million individuals.21) This database stores health insurance claims and health examination results of patients in Japan. It contains de-identified and individual-level data of employees from large companies and their families. Therefore, the database has limited information of patients aged ≥65 years and has no data of patients aged ≥75 years. Information on several indices can be obtained from the database, including encrypted personal identifiers, year and month of birth, sex, period over which the data were obtained, diagnoses, consultations, drugs, and procedures. Diagnoses are registered based on the International Classification of Diseases, Tenth Revision (ICD-10) codes. Drugs are recorded based on the Anatomical Therapeutic Chemical Classification System (ATC) codes.
Study PopulationPatients who were newly prescribed benzbromarone from January 2010 to December 2016 were included. In addition, we enrolled only male patients because most patients with gout are male.22,23) To detect new prescriptions of benzbromarone, we screened data from 24 months before the prescription of benzbromarone for each patient. If a patient was prescribed benzbromarone multiple times, only the first administration was evaluated in this study. If there was a lapse of more than 7 d in the administration, it was considered the end of administration.24) We observed patients for up to 180 d (i.e., 6 months) after the start of administration (observation period). The exclusion criteria were as follows: (1) patients who were registered in the JMDC claims database within 24 months before the start of benzbromarone (to detect new benzbromarone prescriptions); (2) patients who were prescribed benzbromarone occasionally (because it was difficult to calculate the prescription period); and (3) in-patients. The occasional use of benzbromarone was detected by the “occasional use flag” in the JMDC claims database.
Data CollectionBenzbromarone was identified using ATC code M04AB03. A liver function test was considered to be implemented when the following three tests were performed on the same day: (1) aspartate aminotransferase; (2) alanine aminotransferase; and (3) gamma-glutamyl transpeptidase. Almost all patients with fulminant hepatitis presented elevated levels of the above laboratory indices.7–10,17) To evaluate benzbromarone-induced hepatotoxicity, the following diagnoses were identified based on the ICD-10 codes: (1) fulminant hepatitis (ICD-10 codes: K711 and B199), (2) liver impairment (ICD-10 codes: K729, K746, and K769) and (3) acute drug induced hepatitis (ICD-10 code: K712). In addition, we classified these as cases of definite or suspected diagnosis based on the “suspected disease flag” in the JMDC claims database. Moreover, the baseline data of patients, including age, sex, clinical departments that prescribed benzbromarone, number of hospital beds, proportion of patients in whom liver function tests were implemented, daily dose of benzbromarone, duration of the first prescription, and total duration of prescription during the observation period, were collected. Baseline was defined as the day of the first benzbromarone prescription. Clinical departments were identified based on text codes. The five clinical departments with the highest frequency of prescriptions were extracted. Duration was defined as the total number of prescription days and followed up to 180 d after the start of administration (i.e., the observation period). The number of hospital beds was categorized as ≤19, 20–99, 100–199, 200–299, 300–499, and ≥500 beds.
OutcomesTo evaluate the implementation status of periodic liver function test, we targeted patients who continued benzbromarone administration during the observation period among the enrolled patients. The primary endpoint was the proportion of patients in whom periodic liver function tests were implemented. Periodic liver function test was defined as one or more liver function tests performed during both 1–90 (i.e., within 3 months) and 91–180 d (i.e., 4–6 months) after the start of administration, based on the alerts in Japan.17–19) In this study, we refer to the categories as “periodic test” or “non-periodic test” based on whether periodic liver function tests were performed or not, respectively. Additionally, the detailed status of implementation of liver function test was investigated, that is, the proportion of tests conducted at least once in the following periods after administration: (1) 1–90, (2) 91–180, and (3) 1–180 d. Moreover, factors influencing non-periodic testing were analyzed.
To investigate the status of liver injury (safety analysis), we targeted the following subjects among the enrolled patients: (1) patients who were not diagnosed with fulminant hepatitis, liver impairment, or acute drug induced hepatitis within 2 years before starting benzbromarone and (2) patients who underwent liver function tests at least once during administration. Among these patients, the proportion of those with fulminant hepatitis, liver impairment, and acute drug induced hepatitis was evaluated during the observation period. In addition, the onset time of these adverse drug reactions (ADRs) was classified from 0 to 6 months after the start of administration because the disease name (i.e., ICD-10 codes) can be specified only on a monthly basis in the claims database.
Data AnalysisCategorical variables were compared using Fisher’s exact test. Continuous variables were compared using Mann–Whitney U test or Student t-test. Mann–Whitney U test was performed if variables were non-normally distributed, as determined using Shapiro–Wilk test.
To analyze the factors that affect the non-periodic test, multiple logistic regression analysis was performed. As potential risk factors, we analyzed the baseline data shown in the section of data collection (except for the total duration of prescription during the observation period and proportion of implementation of liver function tests). Results with a p-value ≤0.05 were considered significant. All statistical analyses were performed using JMP 14® software (SAS Institute, Inc., Cary, NC, U.S.A.).
EthicsAs data were completely de-identified, and the institutional review board of the Faculty of Pharmaceutical Sciences of Hokkaido University waived the requirement for informed consent.
Of 12211 patients administered benzbromarone from January 2010 to December 2016, 4051 male patients were enrolled in this study. Among them, 3253 patients discontinued benzbromarone during the observation period. Thus, 798 patients were included in the evaluation of implementation status of liver function tests (Fig. 1).
As shown in Fig. 2, 28.7% (n = 229) of the patients underwent periodic liver function tests. In addition, the proportions of patients in whom tests were implemented at least once during 1–90, 91–180, and 1–180 d after starting benzbromarone were 45.2% (n = 361), 40.6% (n = 324), and 57.1% (n = 456), respectively.
Periodic liver function test was defined as one or more liver function tests performed on both 1–90 and 91–180 d. Days 1–90, 91–180, and 1–180 reflect the proportion of patients in whom liver function test was implemented at least once in each period after the start of benzbromarone.
Table 1 presents a comparison of the characteristics of patients who underwent periodic and non-periodic testing. Significant differences were observed with respect to age, clinical department, number of hospital beds, number of implementations of liver function tests, and daily dose.
| Parameter | Periodic test group (n = 229) | Non-periodic test group (n = 569) | p-Value | |
|---|---|---|---|---|
| Age (years), median (IQR) [range] | 49 (42.5–57) [12–74] | 52 (46–58) [16–74] | 0.002‡* | |
| Clinical department of prescription of benzbromarone, n (%) | 0.004†* | |||
| Internal medicine | 153 (66.8) | 426 (74.9) | ||
| Orthopedic surgery | 34 (14.8) | 52 (9.14) | ||
| General surgery | 2 (0.87) | 21 (3.69) | ||
| Gastroenterology | 3 (1.31) | 11 (1.93) | ||
| Urology | 7 (3.06) | 7 (1.23) | ||
| Other departments | 30 (13.1) | 52 (9.14) | ||
| Number of hospital beds, n (%) | 0.010†* | |||
| ≤19 beds | 146 (63.8) | 434 (76.3) | ||
| 20–99 beds | 16 (6.99) | 23 (4.04) | ||
| 100–199 beds | 18 (7.86) | 34 (5.98) | ||
| 200–299 beds | 12 (5.24) | 18 (3.16) | ||
| 300–499 beds | 16 (6.99) | 33 (5.80) | ||
| ≥500 beds | 21 (9.17) | 27 (4.75) | ||
| Number of implementations of liver function tests, 0, 1, 2, ≥3 times, n (%) | ||||
| 1–90 d | 0 | 0 (0) | 437 (76.8) | <0.001* |
| 1 | 144 (62.9) | 121 (21.3) | ||
| 2 | 60 (26.2) | 9 (1.58) | ||
| ≥3 | 25 (10.9) | 2 (0.35) | ||
| 91–180 d | 0 | 0 (0) | 474 (83.3) | <0.001†* |
| 1 | 150 (65.5) | 84 (14.8) | ||
| 2 | 55 (24.0) | 10 (1.76) | ||
| ≥3 | 24 (10.5) | 1 (0.18) | ||
| 1–180 d | 0 | 0 (0) | 342 (60.1) | <0.001†* |
| 1 | 0 (0) | 205 (36.0) | ||
| 2 | 112 (48.9) | 19 (3.34) | ||
| ≥3 | 117 (51.1) | 3 (0.53) | ||
| Daily dose of benzbromarone, n (%) | <0.001†* | |||
| 12.5 mg | 14 (6.11) | 4 (0.70) | ||
| 25 mg | 86 (37.6) | 186 (32.7) | ||
| 50 mg | 122 (53.3) | 349 (61.3) | ||
| 75 mg | 0 (0) | 1 (0.18) | ||
| 100 mg | 7 (3.06) | 26 (4.57) | ||
| 150 mg | 0 (0) | 3 (0.53) | ||
| Duration of the first prescription of benzbromarone (d), median (IQR) [range] | 30 (21–36) [5–90] | 30 (28–40) [3–126] | 0.089‡ | |
IQR: interquartile range, † Fisher’s exact test, ‡ Mann–Whitney U test. * p-values ≤0.05 were considered statistically significant. All continuous variables were non-normally distributed, as determined by Shapiro–Wilk test.
The results of multivariate logistic regression analysis are shown in Table 2. Considering the distribution of the number of patients in Table 1, the daily dose of benzbromarone was used in this analysis as a binary variable, i.e., more or less than 50 mg.
| Characteristics | Odds ratio | 95% CI | p-Value |
|---|---|---|---|
| Age† (years) | 1.02 | 1.002–1.03 | 0.025* |
| Clinical department of prescription of benzbromarone‡ | |||
| Internal medicine (reference) | 1.00 | ||
| Orthopedic surgery | 0.51 | 0.31–0.84 | 0.008* |
| General surgery | 3.06 | 0.70–13.4 | 0.139 |
| Gastroenterology | 1.42 | 0.38–5.32 | 0.607 |
| Urology | 0.30 | 0.10–0.89 | 0.031* |
| Other departments | 0.79 | 0.47–1.34 | 0.388 |
| Number of hospital beds‡ | |||
| ≤19 beds | 2.28 | 1.21–4.32 | 0.011* |
| 20–99 beds | 1.05 | 0.42–2.61 | 0.917 |
| 100–199 beds (reference) | 1.00 | ||
| 200–299 beds | 1.01 | 0.38–2.63 | 0.991 |
| 300–499 beds | 1.29 | 0.55–3.03 | 0.552 |
| ≥500 beds | 0.76 | 0.33–1.78 | 0.532 |
| Daily dose of benzbromarone ≥50 mg | 1.60 | 1.15–2.23 | 0.005* |
| Duration of the first prescription of benzbromarone† (d) | 1.01 | 1.004–1.02 | 0.005* |
95% CI: 95% confidence interval. † Odds ratio indicates odds per single unit increase. ‡ Odds ratio is for comparison with reference value. * Results with p-values ≤0.05 were considered statistically significant.
As risk factors for the non-periodic test, aging, daily dose of benzbromarone ≥50 mg, and long-term duration of the first prescription were extracted. In addition, the factors of clinical departments of orthopedic surgery and urology showed lower odds ratio compared to the factor of internal medicine department, and a higher odds ratio was obtained for general surgery (although the results were not significant). Moreover, having ≤19 hospital beds showed a higher odds ratio compared to having 100–199 beds.
Safety AnalysisAmong the enrolled patients (n = 4051), 833 patients were selected for safety analysis (Fig. 1). Fulminant hepatitis and acute drug induced hepatitis were not detected in any patient. However, 6.36% (n = 53) of the patients were assigned the disease name of liver impairment during the observation period. Among these, 15 were cases of a definite diagnosis of liver impairment and 38 were cases of a suspected diagnosis. This ADR occurred most frequently in the 1 month after initiation of benzbromarone administration, including suspected diagnosis (Fig. 3). The suspected diagnosis did not change to a definite diagnosis of liver impairment in any patient during the observation period. When the patients were classified as those with liver impairment (n = 53) and without liver impairment (n = 780), the proportions of patients who discontinued benzbromarone therapy during the observation period were 73.6% (39/53 patients) and 59.7% (466/780 patients), respectively (p = 0.058, compared using Fisher’s exact test). When only cases of definitive diagnosis of liver impairment were classified, these proportions were 80.0% (12/15 patients) and 60.3% (493/818 patients), respectively (p = 0.181, compared using Fisher’s exact test).
Liver impairment was detected by the International Classification of Diseases, 10th Revision (ICD-10) code of K729, K746, and K769. A definite or suspected diagnosis was determined based on the “suspected disease flag” in the JMDC claims database.
In this study, we aimed to perform a cross-sectional survey of the implementation status of liver function tests in patients who were newly prescribed benzbromarone, using the Japanese large claims database. We found that periodic liver function tests are not performed sufficiently in Japan.
Owing to the reported serious cases of benzbromarone-induced hepatotoxicity, a periodic liver function test is recommended within 6 months after the start of administration.8–10,17–19) On the contrary, some recent studies have reported that there were no serious adverse effects related to benzbromarone.4,25) In the safety analysis of our study, there was no case of fulminant hepatitis. As the frequency of fulminant hepatitis is very low (estimated 1 per 17000 patients), it is possible that the occurrence of fulminant hepatitis could not be detected in the previous studies (including our study), with only a few hundred cases.4,25) Even if fulminant hepatitis does not occur, liver impairment frequently occurs after the start of benzbromarone treatment. In our study, the proportion of patients with the newly assigned disease name of “liver impairment” was 6.36%, including those with a suspected diagnosis. In Japan, clinicians often make a suspected diagnosis to levy medical fees for performing blood tests and prescriptions. In our study, 15 patients (1.80%) received a definite diagnosis of liver impairment. As this result was based on the proportion of patients assigned the disease name of liver impairment, it was difficult to verify whether it was an adverse effect of benzbromarone. However, our result (Fig. 3) is consistent with that of previous studies and alerts, which reported that benzbromarone-induced hepatotoxicity is likely to occur early after benzbromarone administration.4,17–19) Thus, performing a periodic liver function test is essential for safe benzbromarone therapy. Moreover, the proportion of patients who discontinued benzbromarone therapy tended to be higher among those who received a definitive diagnosis of liver impairment. Although the actual reason of discontinuation is unknown, the occurrence of liver impairment may be associated with discontinuation of benzbromarone.
In this study, periodic liver function tests were implemented only in 28.7% of patients (Fig. 2); this value was not surprisingly low. Previously, Nasuhara et al. reported that periodic blood tests that met the requirements on the package insert warning (periodic blood tests four times within 60 d) of thiamazole were implemented in only 15.7% of the cases, which is lower than that of our result.20) This may be because the blood tests for benzbromarone need not be performed as frequently as those for thiamazole. Our results are consistent with those of a previous study reporting that the longer the period from the start of administration, the lower the proportion of implementation of blood test.20) In other words, tests were implemented at least once in 1–90 and 91–180 d in 45.2% and 40.6% of patients, respectively (Fig. 2). In addition, as we only surveyed the status of blood test implementation, the actual purpose of these tests could not be determined. Therefore, some clinicians may not be aware that they must monitor the side effects of benzbromarone even if periodic tests are performed. Thus, the monitoring of liver function after the start of benzbromarone is not sufficient in Japan. Moreover, although only male patients were targeted in this study, we confirmed that the results would be similar if female patients were enrolled (data not shown). Our findings are important because there is a lack of similar studies, except for the study by Nasuhara et al. In addition, to the best of our knowledge, our study is the first to provide generalizable data regarding the status of implementation of blood tests for monitoring ADRs using a large claims database. Thus, our approach of using claims database can be applied to other ADRs and developed further.
As shown in Table 2, patient age was related to non-periodic testing. Due to the occurrence of multiple diseases, polypharmacy (concurrent use of ≥5 drugs) is increasingly common in middle-aged (age 45–64 years) and older individuals (age ≥65 years).26) Therefore, clinicians may lack awareness regarding individual drugs in patients practicing polypharmacy. This may be related to the extraction of age as a risk factor for non-periodic testing. However, as the JMDC claims database does not contain data of patients aged ≥75,21) our results revealed only under 74 years as a risk factor for the non-periodic testing. Thus, further research is required in this regard. Additionally, ≤19 hospital beds (i.e., in clinics) showed a higher odds ratio compared to 100–199 beds; no notable trends were obtained for the other numbers of hospital beds. Small hospitals may not have their own clinical laboratory equipment and require outsourcing. In our previous study, we found that the proportion of patients in whom “therapeutic drug monitoring” for vancomycin (i.e., testing serum vancomycin concentration) was implemented was low in small hospitals.27) Thus, extracting small hospital beds as a risk factor is reasonable. Regarding daily dose of benzbromarone ≥50 mg, we consider that small daily doses showed a causal relationship with the implementation of periodic liver function test, rather than a high dose, which is a risk factor for non-periodic testing. In addition, a long-term duration of the first prescription of benzbromarone showed a higher odds ratio. This may be due to a decrease in the opportunity for performing blood tests under long-term prescription and lack of awareness among clinicians who prescribe benzbromarone, even as a new prescription, for a long duration. In contrast, the differences in the odds ratios between some clinical departments are difficult to interpret. Our results showing a lower odds ratio for “orthopedic surgery and urology” may be related to whether these departments typically treat patients with gout; however, these results require further verification.
Our study had some limitations. First, the duration of benzbromarone therapy was defined as the total number of prescription days. Thus, the actual use of benzbromarone could not be evaluated. Second, the accuracy of diagnosis of liver impairment (i.e., their ICD-10 codes) could not be determined. Third, we did not evaluate the implementation status of liver function tests beyond 6 months after benzbromarone initiation; nevertheless, we did assess implementation status within the 6-month period after treatment initiation, which is when benzbromarone-induced hepatotoxicity is likely to occur. Finally, the latest dataset could not be evaluated because our database contains data only up to June 2017.
In summary, we found that periodic liver function test was implemented in only 28.7% of patients who were newly prescribed benzbromarone in Japan. Several factors, such as the number of hospital beds ≤19 (compared to 100–199 beds) and the duration of the first prescription of benzbromarone, were associated with non-periodic testing. Thus, it is necessary to educate clinicians who prescribe benzbromarone. Moreover, pharmacists should pay attention to prescriptions in patients with the above-mentioned risk factors. Our findings are important for ensuring safe benzbromarone therapy. Additionally, it is suggested that our blood-test-based approach should be applied to other drugs and countries in future research.
The authors declare no conflict of interest.
