藥學雜誌電子報110期
110
Vol. 28 No. 1
Mar. 31 2012
中華民國一○一年三月三十一日出版

Analysis of the Adverse Drug Reactions Causing Hospitalization in A Regional Teaching Hospital in Taiwan


Hua-Sin Chen, Jian-Ming Lai, Hsi-Hui Chen, Wen-Hsiang Ting, Guan-Yu Lu
Department of Pharmacy, Kuang Tien General Hospital

Abstract

Objective: Analysis of the adverse drug reactions (ADRs) causing hospitalization in a regional teaching hospital in Taiwan.

Methods: Adverse drug reaction events from March 2008 to October 2009 were collected through the spontaneous reports or retrieved from International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) coding system (code: 960–979, 995.2). All of the ADR events were reviewed by the clinical pharmacists.

Results: During this period, totally 51,866 patients admitted to internal medicine ward. Clinical pharmacists reviewed 334 patients (0.6%). Among them, 261 patients were suicide or drug abuse cases so excluded from this study. The other 73 patients (0.14%) were diagnosed as ADRs related admissions. Type A reaction or augmented pharmacologic effect was the major cause, which accounted for 74.0% (54 patients). They were classified as the following: side effects (28 cases, 38.4%), overdoses (16 cases, 21.9%), drug-drug or drug-food interactions (7 cases, 9.6%), and others (3 cases, 4.1%). There were 19 patients associated with type B reaction or bizarre effect. ADRs related admissions were sex independent, 58.9% of the patients were ≧ 66 years old and taking multiple medicines (average 5.6 medicines). In addition, we found that the most trouble medicines were digoxin (9 cases), warfarin (9 cases), and phenytoin (8 cases).

Conclusions: Most of the ADRs related admissions were caused by side effects, overdoses and drug-drug or drug-food interactions. They are theoretically preventable. In this study, we re-emphasized that drugs with narrow therapeutic index should be closely watched for the ADR events.

Key Words: adverse drug reaction, clinical pharmacist, hospital admission

Introduction

According to WHO 2000 causality criteria, ADR event was any noxious, unintended, and undesired effect of a medication, which occurred at doses normally used in human for prophylaxis, diagnosis, or therapy of disease. There were some ADR related admissions studies around the world.1-3 Firstly, a review of 14 Australian reports published from 1988 through 1996, ADR admissions were 2.4-3.6%, and 32-69% of which were regarded as preventable events.1 Secondly, a prospective observational study of 18,820 patients admissions in two hospitals found that 6.5% of admissions were ADR related in Merseyside.2 Among these ADR events, 72% judged as avoidable. Thirdly, a meta-analysis of 39 prospective studies between 1966 and 1996 from US hospitals, the incidence of serious and fatal ADR events related admissions were 4.7% and 0.13%, respectively.3 Additionally, ADR events had been regarded as a major public health problem since they represented a sizable percentage of morbidity, mortality, an economic burden, and wasted medical resources in hospital. During 1995 and 2000, health care cost was attributed to drug-related morbidity and mortality estimated to increase from $76.6 billion to more than $177.4 billion in the United States.4

In particular, there were some serious ADR events related human genomic variation had researched around the world in recent years.5 Considerable pharmacogenomic research had focused on understanding the molecular mechanism behind ADR events and finding biomarker that identified population at risk. Notably, human leukocyte antigen (HLA) class I antigen B-several variations of the HLA-B allele were associated with adverse reactions to certain drugs. For example, two specific versions of this gene were related to increased drug sensitivity among the Han Chinese population. Individual who had HLA-B*1502 gene variant (except Japanese) was highly associated with hypersensitivity skin disorder called Stevens-Johnson syndrome in response to carbamazepine.6 Another version, HLA-B*5801 gene variant, was associated with an increased risk of severe cutaneous adverse reactions in people treated with allopurinol.7

Besides the WHO 2000 causality criteria, ADR events would be classified by causality and severity. According to the causing, ADR events would be classified as types A-F reaction.8 The Naranjo causality scale was a commonly method for adjusting to judge the causality of ADR events.9 The ADR event was assigned to a probability category from the total score as follows: probability was assessed by total score, with a score of 0 or inferior indicating a doubtful association, 1-4 indicating a possible association, a score of 5-8 indicating a probable association, and a score of 9 or greater indicating a definite association between medication administration and the ADR event.9 Furthermore, ADR events were further classified for severity. The severity of ADR event was determined by using the modified Hartwig and Seigel scale.10 Severity was assigned with the following standard definitions routinely used in the facility's ADR program: minor (no antidote, therapy, or prolongation of hospital stay required), moderate (required a change in drug therapy or treatment or prolongation of hospital stay by at least one day), severe (potentially life-threatening and caused permanent injury or required intensive medical care), or lethal (directly or indirectly contributed to the death of the patient).10

Previous study conducted by Christel et al. for ADRs in older Australians had shown that the ADR-related hospital stays increased from 2.5 per 1000 person-years in 1981 to 12.9 per 1000 person-years in 2002, and the largest increases occurred in those aged ≧ 80 years old.11 But the incidence of ADR related hospitalization was rare reported in Taiwan. Therefore, we aimed to analyze the ADR related admission to prevent further ADR events in our hospital.

Methods

The ADR events were recorded through the spontaneous reports or used ICD-9-CM coding system (code: 960–979, 995.2). According to WHO 2000 causality criteria, we investigated the ADR events in patients led to hospital from 1 March 2008 to 31 October 2009, but expired one on arrival and ADR events during hospitalization were excluded. The assessment of causality was then performed for all the cases using the Naranjo's algorithm to divide into classification of doubtful, possible, probable, and definite. The severity of ADR events was determined by using the modified Hartwig and Seigel scale which was divided into minor, moderate, severe, or lethal. All cases were evaluated by the clinical pharmacists and approved by ADRs team in our hospital.

Results

As shown in Figure 1, data collected from 1 March 2008 to 31 October 2009 including 51,866 patients admitted to internal medicine ward. Clinical pharmacists reviewed 334 patients (0.6%) who suspected to be the ADR related admissions. Among them, 261 patients were suicide or drug abuse case so excluded from this study. The remaining 73 (0.14%) were diagnosed as ADR related admissions, the other 261 were excluded from this study secondary to suicide or drug abuse. Type A (or augmented pharmacologic effect) and type B (or bizarre effect) accounted for 54 (74.0%) and 19 (26.0%), respectively. Type A consists of side effects (28 cases, 38.4%), overdoses (16 cases, 21.9%), drug-drug/drug-food interactions (7 cases, 9.6%), and others (3 cases, 4.1%). Type B consists of allergies (18 cases, 24.6%) and idiosyncrasy (1 case, 1.4%). The population characteristics were described in Table 1. Sex distribution was not difference for the ADR events. Number of medications for patients ADR related admissions were slightly greater than outpatients with average 4.7 versus 3.1. As shown in Table 2, it indicated that correlation between the patients with ADR related admissions and taking multiple medicines. The probability and severity of ADR related admission were shown in Table 3. For the assessment of the study population by number medications, most of patients receive concurrent medications ≤ 6. The populations of severity and probability were not significant difference between scheduled medications ≤ 6 and > 6. The probability analysis results were probable (69.9%), possible (23.3%), definite (4.1%), and doubtful (2.7%). The severity was severe (74.0%), moderate (17.8%), minor (5.5%), and lethal (2.7%) as shown in Table 4. Finally, we also analyzed the symptoms and doubtful medications leading to ADR events as shown in Table 5 and Table 6, respectively. The most frequent disorder was neuro-toxicity (24.7%), followed by skin disorder (21.9%), and cardiovascular disorder (17.8%). The leading pharmacological category of ADR related hospitalization was drugs acting on the nervous system (31.5%), followed by cardiovascular-renal drugs (28.8%), and metabolic and nutrient agents (13.7%). The most doubtful medications were digoxin (9 cases), warfarin (9 cases), and phenytoin (8 cases).

 

Figure 1. Types of adverse drug reactions.

 

Table 1. Characteristics of the study population.

Characteristic

Cases (N = 73)

No of case

%

Male/Female

35/38

47.9/52.1

Age (y/o)

1-94 (average 63.0)

≤ 18

2

2.8

19-44

13

17.8

45-65

15

20.5

≥ 66

43

58.9

Multiple medicine (average)

admission 4.7 vs. outpatient 3.1

< 65 y/o (n =27) 3.1

 

≥ 66 y/o (n =43) 5.6

 

Urgent admission

   

through emergency department

56

76.7

through outpatient department

17

23.3

 

Table 2. Correlation between the ADR events and multiple medicine.

Multiple medicine

ADR events

(N = 73)

Accumulation (%)

1

10

13.7

2

5

20.5

3

6

28.7

4

9

41.0

5

10

54.7

6

20

82.1

7

6

90.3

8

2

93.1

9

1

94.4

>9

1

95.8

unknown

3

100.0

average # of medication: 4.7

 

Table 3. Assessment of the study population by number of medications used before hospitalization.

Variable

Cases (N = 73)

≤ 6 medications

N = 60

> 6 medications

N = 10

unknown

N = 3

No of case

%

No of case

%

No of case

%

Probability

           

Doubtful

0

0.0

1

10.0

1

33.3

Possible

15

25.0

2

50.0

0

0.0

Probable

42

70.0

7

70.0

2

66.7

Definite

3

5.0

0

0.0

0

0.0

Severity

           

Minor

4

6.7

0

0.0

0

0.0

Moderate

8

13.3

4

40.0

1

33.3

Severe

47

78.3

6

60.0

1

33.3

Lethal

1

1.7

0

0.0

1

33.3

 

Table 4. The study population of severity and probability.

Probability

Severity

Doubtful

Possible

Probable

Definite

Total (%)

Minor

0

1

3

0

4 (5.5)

Moderate

0

3

9

1

13 (17.8)

Severe

2

13

38

1

54 (74.0)

Lethal

0

0

1

1

2 (2.7)

Total (%)

2 (2.7)

17 (23.3)

51 (69.9)

3 (4.1)

73 (100%)

 

Table 5. Disorder categories lead to ADR related admission.

Category

Cases (N = 73)

No of case

%

Neurotoxicity

18

24.7

Skin disorder

16

21.9

Cardiovascular disorder

13

17.8

Gastrointestinal disorder

10

13.7

Haematological disorder

7

9.6

Hypoglycaemia

3

4.1

Hepatotoxicity

2

2.7

Nephrotoxicity

1

1.4

Pulmonary toxicity

1

1.4

Shock

1

1.4

Fever

1

1.4

 

Table 6. Pharmacological categories lead to ADR events.

Category

Cases (N = 73)

No of case

%

Drugs acting on the nervous system

23

31.5

Cardiovascular-renal drugs

21

28.8

Metabolic and nutrient agents

10

13.7

Hematological agents

9

12.3

Anti-infective agents

5

6.8

Unclassified agents

2

2.7

Anti-neoplastic agents

1

1.4

Gastrointestinal agents

1

1.4

Respiratory tract drugs

1

1.4

 

Discussion

ADR related hospital admissions were all treated by physicians immediately, but they rarely spontaneous reported to clinical pharmacists. ICD-9 coding system can assist clinical pharmacist to catch some of the missing events.

In our study, 73 patients (0.14%) of all admitted patients over a 20-month period were ADR related. The incidence of ADR related admission was lower than other previous ADR reports, probably due to different study criteria or physicians are still reluctant to report ADRs in our hospital.

The elderly patients have higher ADR events rate, which are coincident with previous study.14 This can make sense that the older patients always have more complicated long-term illnesses and taking multiple medicines (average 5.6 medications) to control their diseases. Most of the ADR related admissions are coming in through the emergency department (76.7%), this may indicate that the ADR related admissions were more severe than other ADR events (see Table 1). For the concurrent medication use, although multiple medicines is one of the risk factors of ADRs and which had been confirmed by many previous studies. Our data indicated that patients with ADR related admissions only have slightly increased medication use, and about 13.7% developed ADR with only one medication use (see Table 2). Most of these medications are drugs acting on the nervous system or cardiovascular-renal drugs (see Table 6), and are associated narrow therapeutic index. We recommend when patients first time take these medications need aggressive education and alert physician to monitor these medications closely to prevent ADR events. There were two ADR events (2.7%) caused death during study period. One patient died probably secondary to sulfamethoxazole/trimethoprim induced toxic epidermal necrolysis (TEN) and the other one probably caused by cardiovascular disorder induced by digoxin.

For the national ADR reporting system in Taiwan during 2008,15 the leading pharmacological category is drugs acting on the nervous system (28.3%), followed by anti-infective agents (21.7%), and anti-neoplastic agents (12.4%). And the most doubtful medications are ketorolac (245 cases), diclofenac (189 cases), and docetaxel (136 cases). Some of the distributions of pharmacological category and medication are different from our study, it is uncertain that these differences are population related, need further research to follow up.

Conclusions

From this study we observed that the ADR events caused 0.14% of admissions in our hospital. Most of the ADR events were classified as side effects (38.4%), overdoses (21.9%) and drug-drug or drug-food interactions (9.6%). They are theoretically preventable. Notably, we found a positive correlation between the ADR related admission and the older patients took multiple medicines. Most of the trouble medications are associated with narrow therapeutic index. With the accumulated knowledge of human genomic variation has being used for the development of personalized medicine, which may aid to decrease the number of ADR events and increase the efficacy of medication treatment in the future. In our initial study, we can summarize that when patients use these high risk medications, pharmacists need to closely watched and provide sufficient patient education to prevent ADR events.

Acknowledgement

This work was supported by grants Kuang-97-05 from the Kuang Tien General Hospital Taichung, Taiwan.

References:

1. Roughead EE, Gilbert AL, Primrose JG et al: Drug-related hospital admissions: a review of Australian studies published 1988-1996. Med J Aust 1998; 168: 405-8.

2. Pirmohamed M, James S, Meakin S et al: Adverse drug reactions as cause of admission to hospital: prospective analysis of 18,820 patients. BMJ 2004; 329: 15-9.

3. Lazarou J, Pomeranz BH, Corey PN. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA 1998; 279: 1200-5.

4. Johnson JA, Bootman JL. Drug-related morbidity and mortality. A cost-of-illness model. Arch Intern Med 1995; 155: 1949-56.

5. Ingelman-Sundberg M. Pharmacogenomic Biomarkers for Prediction of Severe Adverse Drug Reactions. N Engl J Med 2009; 358: 637-9.

6. Chung WH, Hung SI, Hong HS et al: Medical genetics: a marker for Stevens-Johnson syndrome. Nature 2004; 428: 486.

7. Hung SI, Chung WH, Liou LB et al: HLA-B*5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. Proc Natl Acad Sci 2005;102:6237.

8. Nebeker JR, Barach P, Samore MH. Clarifying adverse drug events: a clinician's guide to terminology, documentation, and reporting. Ann Intern Med; 140: 795-801.

9. Naranjo CA, Busto U, Sellers EM et al: A method for estimating the probability of adverse drug reactions. Clin Pharmacol Ther 1981; 30: 239-45.

10. Hartwig SC, Siegel J, Schneider PJ. Preventability and severity assessment in reporting adverse drug reactions. Am J Hosp Pharm 1992; 49: 2229-32.

11. Christel LB, C D' Arcy JH, Anthony GS. Adverse drug reactions in older Australians, 1981-2002. Med J Aust 2005; 182: 267-270.

12. National Reporting System of Adverse Drug Reactions in Taiwan. http://adr.doh.gov.tw/default.asp.

摘要

台灣某區域教學醫院導致病人住院之藥物不良反應案例分析

目的:分析台灣中部某區域教學醫院藥物不良反應入院事件。

方法:經自願性回報及病歷回顧 (ICD-9-CM: 960-979, 995.2) 二種方式收集藥物不良反應事件資料。收案條件自2008年3月1日至2009年10月31日因藥物不良反應入院病人。到院後即死亡或不良反應事件係發生於住院期間者,不選入本研究。

結果:因藥物不良反應入院之病人為334人,佔全部入院比例的0.6%。經臨床藥師評定導致入院的藥物不良反應事件73件 (0.14%),將其型態分為A型 (54件) 與B型 (19件),而A型事件有副作用 (28件,38.4%)、藥物過量 (16件,21.9%)、藥物-藥物或藥物-食物間交互作用 (7件,9.6%) 和其他類型 (3件,4.1%)。研究結果發現藥物不良反應事件與性別無關,而與老人和多重用藥有關聯。而在海線地區最具問題的藥物為 digoxin (9件)、warfarin (9件) 與 phenytoin (8件)。

結論:大部分的藥物不良反應事件造成的原因都是屬於副作用、藥物過量、藥物-藥物/藥物-食物間交互作用,而且是可預防性的。本研究亦證實藥物不良反應事件與治療指數狹窄的藥物有關。

作者

光田綜合醫院藥劑部藥師 陳華鑫、賴建名、陳西蕙、丁文香、盧冠宇