107

Pharmacoepidemiologic study of potential drug

interactions in outpatients of a university hospital in

Thailand

B. Janchawee*

, W. Wongpoowarak

, T. Owatranporn

and

V. Chongsuvivatwong§

*Department of Pharmacology, Faculty of Science, Prince of Songkla University, Hat Yai, Thailand,

Department of Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, Prince of Songkla

University, Hat Yai, Thailand, Pharmaceutical Department, Songklanagarind Hospital, Prince of Songkla

University, Hat Yai, Thailand and §Epidemiology Unit, Faculty of Medicine, Prince of Songkla University,

Hat Yai, Thailand

S U M M A R Y

Background: Drug–drug interaction is a potential cause of adverse drug reactions. The incidence of such drug interactions in university hospitals in Thailand is unknown.

Purpose: To estimate the rate of potential drug– drug interactions in outpatients of a typical Thai university hospital, and to identify risk factors for such interactions in Thai patients.

Methods: One-year outpatients’ prescription data were retrieved from the hospital computer records. Potential drug interactions were identi-fied using the existing drug-interaction database system. Potential interactions within a specific prescription and involving drugs prescribed 1-, 3- and 7-day earlier were searched for. Possible associations between occurrence of an interaction and a patient’s age and gender and the number of items on the prescription were explored.

Results: The overall rate of potential drug inter-actions was 27Æ9% with a maximal value of 57Æ8% at the Department of Psychiatry. The rate of the most potentially significant interactions was 2Æ6%, being the highest in the Department of Medicine (6Æ0%), with isoniazid vs. rifampin as the most common interacting combination. The rate increased with the patient’s age and pre-scription size (P = 0Æ000). The odd’s ratio of hav-ing at least one potential drug interaction was 1Æ8

(64Æ2%) when age increased by 20 years

(P = 0Æ000) and 2Æ8 (165Æ7%) when another drug was added (P = 0Æ000). The rate of potential drug interactions was the same for both genders. The rate of potential drug interactions detected across prescriptions was higher than within prescrip-tions and was dependent on the time interval between prescriptions.

Conclusions: Potential drug interactions were common in our sample of patients. The rate of such interactions increased with the number of drugs prescribed and the patient’s age.

Keywords: drug interactions, hospital, outpa-tients, prescriptions, rate

I N T R O D U C T I O N

Drug interactions in patients receiving multi-drug therapy are of wide concern. Such interactions are an important cause of adverse drug reactions and may lead to an increased risk of hospitalization and higher health care costs (1–6). Studies conducted in various, mainly western, countries report rates of potential drug–drug interactions ranging from approximately 1 to 52% (7–20). The incidence of actual occurrence of drug interactions has been reported to be much smaller ranging from 0 to 1Æ3% (10, 20). Differences in methods used, including criteria for data collection, study periods and target population, contribute to these discrepancies.

In Thailand, a number of short-term studies report on potential interactions among selected groups of drugs or patients (21–24), or on those with the most potential clinical significance (25, 26).

Received 6 November 2003, Accepted 7 July 2004

Correspondence: Benjamas Janchawee PhD, Department of Pharmacology, Faculty of Science, Prince of Songkla University, Hat Yai 90112, Thailand. Tel./fax: +66 74 446678; e-mail: [email protected]

These reports suggest rates of up to 53%. One of these studies, covering all drugs, reported a 16% rate of potential interactions (27). The incidence of clinical drug interactions has not been reported.

One problem in the study of rates of drug interactions in prescriptions in Thailand is the lack of appropriate drug databases and drug interaction software, so that all interactions have to be identi-fied by hand and objectivity and exhaustive cov-erage cannot be readily validated. Additionally, only interactions within single prescriptions have been searched for in published studies. In a pre-vious study (28), we described the development of a computerized drug-interaction database that can be used by both the public and within Thai hos-pitals. This computerized system enables the rapid and accurate detections of potential drug interac-tions from prescription data.

The aim of this study was to estimate the rate of potential drug interactions in outpatients attending a university hospital in Thailand and to identify risk factors for such interactions.

M E T H O D S

Study design

A retrospective large-scale study was undertaken, using computerized data held at the Prince of Songkla University hospital. All drug groups, patient age ranges and morbidity types were included.

Study population

Prescription data of outpatients visiting the hospi-tal over a 12-month period (1 January–31 December 2000) were included. The data consisted of drug names, dispensing dates, departments of the pre-scribing doctors and patients’ birth date, gender and hospital number. To maintain confidentiality, the patient hospital number was encrypted by a hospital programmer prior to handing the data to the investigators.

Study protocol

Potential drug–drug interactions were detected using our previously developed computerized drug-interaction database system (28). It contains

information of drugs available locally and list of over 1700 drug–drug interactions cited in three sources (29–31). The detection algorithm was based on a structured query language, described else-where (32).

Searches were performed both within prescrip-tions and across successive prescripprescrip-tions for each patient during periods of 1-, 3- and 7-day. For within prescription searches, a single prescription is defined as: for one patient, from one doctor, in 1 day. For 1-, 3- and 7-day across prescription detection, a single prescription is defined as: for one patient, within 1, 3 and 7 day(s), respectively.

Potential drug–drug interactions were expressed as the number of interacting drug pairs, number of prescriptions with interactions or rate. The rate was calculated as: (the number of prescriptions with potential drug interactions)/(number of prescrip-tions with two or more drugs · 100). The whole population was used to investigate any relation-ship between rate of potential drug interactions and patients’ age and gender, and number of drug items on the prescription. A sample of 1020, ran-domly selected prescriptions, was used to estimate the rates, expressed as odd’s ratios.

Statistical analysis

Demographic data of patients and prescriptions were presented as mean, standard deviation and percentage. Logistic regression analysis was used to determine the odd’s ratio. Probability (P) values of 0Æ05 or less were considered statistically signifi-cant. All statistical analyses were performed by using the statistical package SPSSSPSS version 10Æ0 for

Windows.

R E S U L T S

During a 1-year period, 124 528 patients received

prescriptions. Their average age was

39Æ0 ± 22Æ4 years (range: 0–100 years). Most patients were adults (20–39Æ9 years; 30Æ8%) and middle-aged (40–59Æ9 years; 27Æ4%). The ratio of male to female was 40Æ5 : 59Æ5%.

Nearly half of all patients (44Æ9%) visited the hospital only once during the whole year and received only one prescription. The others visited more frequently and had from two to 53 prescrip-tions/person/year. Some of these patients had

more than one prescription dispensed by more than one department on the same day. Average visit intervals ranged approximately from 7 to 68 days.

There were in total 360 418 prescriptions, of which 71Æ8% or 258 951 prescriptions were mul-tiple, containing two or more drugs, and these are considered to carry a risk of drug interactions. Most of these (30Æ8% or 79 837 prescriptions) were dispensed from the Department of Medicine. Pre-scription size ranged from 1 to 15 with an average of 2Æ5 ± 1Æ4 items per prescription. The number of large prescriptions (five items or more) increased with increasing age (Fig. 1).

Potential drug interactions detected within a single prescription expressed as either the number of interacting drug pairs or the number of pre-scriptions with interactions at each significance level are shown in Table 1. There were in total 612 interacting drug pairs in 72 296 prescriptions with potential interactions. Forty-eight interacting drug pairs (out of 6859 prescriptions) were labelled as significance-l interactions. The overall rate of potential drug interactions was 27Æ9% (72 296/ 258 951). Those of significance-1 level accounted for 2Æ6%. The most common significance-1 inter-acting drug pair was isoniazid–rifampin (2507 prescriptions), while the second most common was digitalis glycosides–loop diuretics, involving 1330 prescriptions (Table 2). Significance-1 interactions were highest in the Department of Medicine (4787 prescriptions or 6Æ0%). The most common inter-acting combination prescribed in this department was again isoniazid–rifampin (INH 100 mg–rif-ampicin 300 mg). The second most common was digitalis glycosides–loop diuretics (lanoxin 0Æ25 mg–furosemide 40 mg).

The rates of potential drug interactions in the different departments within the same prescrip-tions are shown in Table 3. The overall rate of potential interactions was highest in the Depart-ment of Psychiatry, viz. 57Æ8%, whilst the number of prescriptions with interactions was highest in the Department of Medicine (33 823 prescriptions) with the rate of 42Æ4%. Compared with the ‘within’ prescription data, the ‘across’ prescriptions rate increased with time interval between prescriptions (Table 3).

The number of prescriptions with interactions and the rate of potential drug interactions in rela-tion to age range of patients are shown in Fig. 2. The number of prescriptions containing two or more drugs and those with interactions were highest among patients aged 40–59Æ9 years (viz. 74 900 and 23 862 prescriptions, respectively). The rate increased with age and was highest (42Æ5%) in the 80 years or older age group. Logistic regression showed that the odd’s ratio was 1Æ029 [95% CI: 1Æ021–1Æ037, P = 0Æ000, P = 1/(1 + e 2Æ67)0Æ03 Age)] when the patient’s age increased by 1 year. The rate of potential interactions increased with size of prescription, reaching almost 100% with prescrip-tions containing eight or more drug items (Fig. 3). With inclusion of all prescriptions, the relationship between the rate of potential interactions and

0·1 1 10 100

0–19 20–39 40–59 60–79 80 or more

Patient age groups (years)

Number of pr escriptions (%) 1 2–4 5–7 8 or more

Fig. 1. Percentage of prescriptions with various sizes in

different age groups.

Table 1. Number and rate of potential interactions and

their significance Significance Number of interacting drug pairs Number of prescriptions

with interactions Rate (%)

1 48 6859 2Æ6 2 167 23 793 9Æ2 3 88 12 538 4Æ8 4 183 23 848 9Æ2 5 126 44 476 17Æ2 Total 612 111 514a _27Æ9

Total (any significance) 72 296

Significance rating is defined by the drug interactions informa-tion source used for the database system (1 = major severity, established documentation; 2 = moderate severity, established documentation; 3 = minor severity, established documentation; 4 = major to moderate severity, possible documentation; 5 = minor or any severity, possible or unlikely documentation).

logarithm of prescription size is shown in Fig. 4. Logistic regression analysis of rate and size of prescriptions showed that the odd’s ratio was

2Æ831 [95% CI: 2Æ427–3Æ301, P = 0Æ000, P = 1/(1 + e4Æ36)1Æ04 Prescription size)] when the prescription size increased by one.

Table 2. The five most common potential drug interactions with significance-1

Interacting drug pairs Prescribing drug pairs Number of prescriptions

Isoniazid–Rifampin INH 100 mg–Rifampicin 300 mg 2507

Digitalis glycosides–Loop diuretics Lanoxin 0Æ25 mg–Furosemide 40 mg 1330

Methotrexate–NSAIDs MTX 2Æ5 mg–Ibuprofen 400 mg 728

Digitalis glycosides–Thiazide diuretics Lanoxin 0Æ25 mg–Amiloride/HCTZ 627

Methotrexate–Sulfonamides MTX 2Æ5 mg–Salazopyrin EN 500 mg 405

Prescribing drugs pairs shown here are only the most common combinations.

Table 3. Rate % (number of

pre-scriptions with interactions) of potential interactions found either within or across prescriptions Department

Within prescriptions

Across prescriptions

1-Day 3-Day 7-Day

Anaesthesiology 22Æ6 (924) 24Æ2 (1094) 24Æ9 (1178) 25Æ7 (1309) Community Medicine 33Æ4 (3762) 34Æ3 (4069) 34Æ5 (4182) 35Æ3 (4410) Dentistry 5Æ0 (39) 6Æ8 (86) 6Æ6 (91) 7Æ4 (114) Medicine 42Æ4 (33 823) 43Æ7 (37 267) 44Æ1 (38 398) 44Æ9 (40 328) Obstetrics and Gynaecology 15Æ0 (2770) 15Æ6 (3275) 16Æ0 (3479) 16Æ7 (3827) Ophthalmology 21Æ0 (2074) 20Æ2 (2750) 21Æ2 (3038) 22Æ0 (3386) Orthopaedic Surgery 31Æ8 (6012) 33Æ9 (7043) 34Æ3 (7334) 35Æ2 (7831) Otolaryngology 12Æ1 (4072) 13Æ5 (4974) 14Æ0 (5268) 15Æ0 (5845) Pediatrics 13Æ1 (4556) 13Æ7 (4938) 14Æ2 (5218) 15Æ0 (5721) Psychiatry 57Æ8 (6049) 59Æ9 (6748) 60Æ0 (6892) 60Æ3 (7139) Radiology 20Æ2 (2393) 20Æ9 (2715) 21Æ8 (2986) 23Æ4 (3482) Surgery 23Æ3 (5822) 24Æ2 (6701) 24Æ9 (7127) 26Æ0 (7870) Totals 27Æ9 (72 296) 28Æ8 (81 660) 29Æ3 (85 191) 30Æ1 (91 262) 0 20000 40000 60000 80000 100000 120000 0–19·9 20–39·9 40–59·9 60–79·9 80 or more

Age ranges (years)

Number of pr escriptions 0 5 10 15 20 25 30 35 40 45 Rate (%)

Total prescriptions Prescriptions with > 1 drug

Prescriptions with interactions Rate

Fig. 2. Number of prescriptions with interactions and

rate of potential drug interactions among different age ranges. 1 10 100 1000 10000 100000 1000000 2–4 5–7 8 or more Prescription size Number of pr escriptions 0 10 20 30 40 50 60 70 80 90 100 Rate (%)

Prescriptions with > 1 drug Prescriptions with interactions Rate

Fig. 3. Number of prescriptions with interactions and

rate of potential drug interactions among different sizes of prescriptions.

The rates of potential drug interactions in pre-scriptions for male and female patients were equal (viz. 27Æ9%).

D I S C U S S I O N

This study revealed that the rate of potential drug interactions in outpatients was 27Æ9% overall and 2Æ6% for the most potentially significant combina-tions. These values cannot be directly compared with those reported previously because of the dif-ferences in the study design. However, a retro-spective study of drug–drug interactions in outpatients of a provincial hospital in Thailand reported the rate of potential interactions during 2 weeks among all drug groups involving 3119 prescriptions as being 15Æ9% (27). Two studies involving a 3-month period of data collection and investigating 40 pairs of potentially important or major significant drug interactions reported the rate of less than 0Æ5% (25, 26).

Our finding clearly shows that the rate of potential drug interactions is directly proportional to the patient’s age. A higher rate was seen among older patients. This corresponds to other studies reporting that potential drug interactions were common in elderly people who were on multi-drug regimen (9, 10, 12, 14, 19). Our study shows that the odd’s ratio of having potential drug interaction was 1Æ029 at every 1-year period or 1Æ786 at every 20-year increment of age, and that the probability of having at least one drug inter-action increased by 1Æ5 or 64Æ2%. None of previous studies has directly demonstrated such a

relationship nor resulted in a numerical prediction figure.

The rate of potential drug interactions was also directly related to prescription size. This result is similar to that established by Weerutamasen (27), who studied potential drug interactions in pre-scriptions of patients in all age ranges during 2-week period and revealed that the rate increased by the number of drug items. Results from other studies indirectly support this finding. Reviewing medical record, Mitchell et al. (8) showed that the percentage of patients with potential drug interac-tions increased with number of drug prescribed per patient. A result from Nolan and O’Malley (33) indicated that the percentage of elderly patients being prescribed potentially interacting combina-tions increased greatly with the number of medi-cations given. Our study showed that the odd’s ratio of having potential drug interactions was 2Æ831, and that the probability increased by 165Æ7%, as the number of prescribed drug increased by one. Another study (12) reported a smaller probability that, while the number of concurrent medications per patients increased by one, the probability of having at least one drug–drug interaction increased by 22Æ4%.

In this study, regression analysis showed that there is a slight correlation between patient age and prescription size. In addition, the proportion of large size prescriptions tended to increase with increasing age. On the other hand, most of the large size prescriptions were given to older patients. These data correspond to the works reported by Muranaka et al. (13) that the number of drugs in any one prescription increased with the age of the patient, and by Stewart and Cooper (34) that elderly patients used more medications than younger patients and that the trend of increasing drug use continued to 80 years of age. Thus the higher rate of potential drug interactions in old age seen in this study is probably because of the higher number of prescribed medications. Our data sug-gests that prescription size is a clear predictor of potential drug interactions.

Our study shows that the highest overall rate of potential drug interactions was that of the Department of Psychiatry (57Æ8%), followed by that of the Department of Medicine (42Æ4%). Generally, the highest rate is expected to be of the Department of Medicine because of wider range of drugs used

0 20 40 60 80 100 120 0 0·2 0·4 0·6 0·8 1 1·2 1·4

Log (prescription size)

Rate (%)

Fig. 4. Relationship between rate of potential

and, as shown in this study, the highest number of prescriptions at risk of drug interactions. When the number of interacting drug pairs are taken into account, it was found that overall, and those with the label significance-1 were the highest in the Department of Medicine, viz. 80 670 and 5984 pairs, respectively. In the Department of Psychi-atry, those numbers were 9431 and 157 pairs, respectively. The number of interacting drug pairs per prescription with interaction was 2Æ4 for the Department of Medicine and 1Æ6 for the Depart-ment of Psychiatry. As calculation of the rate in this study was based on the number of prescriptions with interactions, the rate and the number of interacting drug pairs should be simultaneously considered.

No difference in the rate of potential drug interactions between male and female patients was observed in this study. This suggests that the pat-tern of drug use in this population was similar in both genders.

It was observed that some patients received more than one prescription during each hospital visit or received a different one earlier during a certain period of time. In this study, we detected any potential interactions both within and across prescriptions, whilst in previous studies only interactions within a single prescription were investigated. In clinical practice, interactions across prescriptions can occur and the rate may well be higher compared with that of within prescriptions. According to prescription data retrieved from the computer system of the hospital, only drug name and strength of drug preparation were shown (e.g. Inderal 40 mg tablets), but not the duration of drug treatment. Even though the amount of drug pre-scribed and how the drug was taken, were recor-ded by the hospital, estimation of the duration of drug treatment from the data needs further pro-gramming. However, assuming that a patient had been taking any drug continuously prior to receiving another drug in the next period, for example 1-, 3- or 7-day, we have shown that the rate of drug interactions was higher and dependent on the period of time.

This study revealed a relatively high rate of the potentially most significant drug interactions. These included the commonly prescribed combi-nation regimen, isoniazid and rifampin, for tuber-culosis treatment. The risk of hepatotoxicity is

increased in patients receiving both drugs (35, 36). Other interacting drug pairs may also cause serious adverse events. Loop- or thiazide-diuretic-induced electrolyte disturbances, may predispose to digita-lis-induced arrhythmias (37, 38). NSAIDs (39–41) and sulfonamides (42–44) may increase methot-rexate toxicity, especially when given concomit-antly with a high dose of methotrexate was used. The clinical significance of such interactions depends on many factors such as drug dosage, period of concurrent drug use and extent of patient monitoring by physicians.

C O N C L U S I O N S

Potential drug interactions are frequent among outpatients prescribed multiple medications. The rate is directly related to number of drugs pre-scribed and patient age. Older patients are prone to drug interactions partly because they receive more medications. With close management, drug inter-actions often need not have clinically important adverse consequences.

A C K N O W L E D G E M E N T S

The study was supported financially by the Health System Research Institute (HSRI), Thailand and the World Health Organization. We thank Asst. Prof. Dr Sorayut Vasiknanonte, Associate Dean for Medical Informatics, Faculty of Medicine, Prince of Songkla University and the programmers of the Computer Unit, Songklanagarind Hospital, for providing the available prescription data. We also thank Mrs Naowanit Trisdikhun, Head of Phar-macy Department, Songklanagarind Hospital and the Director of Songklanagarind Hospital. The support of the HSRI Network, Prince of Songkla University as to the coordination and cooperation throughout the research has been highly appreci-ated.

R E F E R E N C E S

1. Boston Collaborative Drug Surveillance Program (1972) Adverse drug interactions. The Journal of the American Medical Association, 220, 1238–1239. 2. May FE, Stewart RB, Cluff LE (1977) Drug

interac-tions and multiple drug administration. Clinical Pharmacology and Therapeutics, 22, 322–328.

3. Prince BS, Goetz CM, Rihn TL, Olsky M (1992) Drug-related emergency department visits and hospital admissions. American Journal of Hospital Pharmacy, 49, 1696–1700.

4. Hamilton RA, Briceland LL, Andritz MH (1998) Frequency of hospitalization after exposure to known drug–drug interactions in Medicaid popula-tion. Pharmacotherapy, 18, 1112–1120.

5. Shad MU, Marsh C, Preskorn SH (2001) The econo-mic consequences of a drug–drug interaction. Journal of Clinical Psychopharmacology, 21, 119–120.

6. McDonnell PJ, Jacobs MR (2002) Hospital admissions resulting from preventable adverse drug reactions. The Annals of Pharmacotherapy, 36, 1331–1336. 7. Anonymous (1972) Risk of drug interaction may

exist in 1 of 13 prescriptions. The Journal of the American Medical Association, 220, 1287–1288. 8. Mitchell GW, Stanaszek WF, Nichols NB (1979)

Documenting drug–drug interactions in ambulatory patients. American Journal of Hospital Pharmacy, 36, 653–657.

9. Gosney M, Tallis R (1984) Prescription of contrain-dicated and interacting drugs in elderly patients admitted to hospital. Lancet, 2, 564–567.

10. Kurfees JF, Dotson RL (1987) Drug interactions in elderly. The Journal of Family Practice, 25, 477–488. 11. Costa AJ (1991) Potential drug interactions in an

ambulatory geriatric population. Family Practice, 8, 234–236.

12. McKenzie LC, Kimberlin CL, Pendergast JF, Berardo DH (1994) Potential drug interactions in a high risk ambulatory elderly population. Journal of Geriatric Drug Therapy, 28, 49–63.

13. Muranaka M, Nagai N, Watanabe M, Hata H, Takenaka K (1994) Investigation of prescriptions given to outpatients in seven Kosei-nenkin hospitals. Japanese Journal of Hospital Pharmacy, 20, 442–453. 14. Bergendal L, Friberg A, Schaffrath A (1995) Potential

drug–drug interactions in 5125 mostly elderly out-patients in Gothenburg, Sweden. Pharmacy World and Science, 17, 152–157.

15. Dwiprahasto I, Kristin E (1995) Drug interaction due to inappropriate prescribing for children under 10 years: an evaluation of prescribing pattern among private practices in Indonesia. Pharmacoepidemiology and Drug Safety, 4, S28.

16. Chan TYK, Lam BSY, Ng NKL, Critchley JAJH (1996) Drug interactions in general medical patients. Phar-macoepidemiology and Drug Safety, 5, S108.

17. Gro¨nroos PE, Irjala KM, Huupponen RK, Scheinin H, Forsstro¨m J, Forsstro¨m JJ (1997) A medication data-base: a tool for detecting drug interactions in hospital. Pharmacoepidemiology and Drug Safety, 53, 13–17.

18. Heininger-Rothbucher D, Bischinger S, Ulmer H, Pechlaner C, Speer G, Wiedermann CJ (2001) Incidence and risk of potential adverse drug interactions in the emergency room. Resuscitation,

49, 283–288.

19. Bjorkman IK, Fastbom J, Schmidt IK, Bernsten CB (2002) Drug–drug interactions in elderly. The Annals of Pharmacotherapy, 36, 1675–1681.

20. Ho YF, Huang SH, Lin HN (2002) Detecting drug– drug interactions in medication profiles of psychi-atric inpatients: a two stage approach. Journal of The Formosan Medical Association, 101, 294–297.

21. Kittiwongsuntorn A (1994) Antacid adverse drug interaction of out-patient prescription at Surin Hos-pital. Medical Journal of Srisaket Surin Buriram Hospi-tals, 9, 268–278.

22. Lertwicha S, Chaimongkol S, Wanapongpisan W (1996) The study of possible drug interaction of

fluoroquinolone in outpatient: Sawanpracharak

Hospital. Region 8 Medical Journal, 4, 157–168. 23. Pajongrak U (1996) Antacid interaction in out-patient

at Banpong Hospital. Region 7 Medical Journal, 15, 479–488.

24. Wisuthi S, Chansirikarnjana S (1997) Cardiovascular drug interactions in elderly patients. Mahidol Uni-versity Annual Research Abstracts, 24, 112.

25. Supakan M, Jittasarttra P, Sridakul S, Suwanmalee W, Janjaratjit S (1995) A study of potentially important drug, drug interaction on outpatients in regional and general hospitals in the V zones. Med-ical Journal of Srisaket Surin Buriram Hospitals, 10, 145– 171.

26. Pol-udom W, Komkai S (1996) Potentially important drug interaction in outpatients of Maharat Nakhon Ratchasima Hospital. Maharat Nakhon Ratchasima Hospital Medical Bulletin, 20, 165–172.

27. Weerutamasaen P (1994) Incidence rate of couple drugs that might cause drug interactions in out-patients of Chaoprayayommaraj Hospital, Suphan-buri by number of drug items. Bulletin of the Department of Medical Services, 19, 460–464.

28. Janchawee B, Wongpoowarak P, Amnuaypanit C, Wamae M, Ovartlarnporn T, Chongsuvivatwong V (2003) Development and trial of the drug interaction database system. Songklanakarin Journal of Science and Technology, 25, 525–534.

29. Hansten PD, Horn JR (1998) Hansten and Horn’s Managing Clinically Important Drug Interactions. Vancouver, BC, Canada: Applied Therapeutics. 30. Zucchero FJ, Hogan MJ, Schultz CD, Curran JP,

Bremerkamp JP (1999) Evaluations of Drug Interac-tions. St Louis, MO, USA: First DataBank.

31. Tatro DS (2000) Drug Interaction Facts 2000. St Louis, MO, USA: Facts and Comparisons.

32. Wongpoowarak W, Wongpoowarak P (2002) Unified algorithm for real-time detection of drug interaction and drug allergy. Computer Methods and Programs in Biomedicine, 68, 63–72.

33. Nolan L, O’Malley K (1989) The need for a more rational approach to drug prescribing for elderly people in nursing homes. Age and Ageing, 18, 52–56.

34. Stewart RB, Cooper JW (1994) Polypharmacy in the aged practical solutions. Drugs and Aging, 4, 449– 461.

35. Pessayre D, Bentata M, Degott C, Nouel O, Miguet JP, Rueff B, Benhamou JP (1977) Isoniazid-rifampin fulminant hepatitis. A possible consequence of the enhancement of isoniazid hepatotoxicity by enzyme induction. Gastroenterology, 72, 284–289.

36. O’Brien RJ, Long MW, Cross FS, Lyle MA, Snider DE, Jr (1983) Hepatotoxicity from isoniazid and rifampin among children treated for tuberculosis. Pediatrics, 72, 491–499.

37. Leary WP, Reyes AJ (1984) Diuretic-induced mag-nesium losses. Drugs, 28(Suppl. 1), 182–187.

38. Hollifield JW (1986) Thiazide treatment of hyper-tension. Effects of thaizide diuretics on serum potassium, magnesium, and ventricular ectopy. American Journal of Medicine, 80, 8–12.

39. Maiche AG (1986) Acute renal failure due to con-comitant action of methotrexate and indomethacin. Lancet, 1, 1390.

40. Singh RR, Malaviya AN, Pandey JN, Guleria JS (1986) Fatal interaction between methotrexate and naproxen. Lancet, 1, 1390.

41. Stockley IH (1987) Methotrexate–NSAID interac-tions. Drug Intelligence and Clinical Pharmacy, 21, 546. 42. Thomas MH, Gutterman LA (1986) Methotrexate

toxicity in a patient receiving trimethoprim–

sulfamethoxazole. Journal of Rheumatology, 13, 440– 441.

43. Frain JB (1987) Methotrexate toxicity in a patient receiving trimethoprim–sulfamethoxazole. Journal of Rheumatology, 14, 176–177.

44. Govert JA, Patton S, Fine RL (1992) Pancytopenia from using trimethoprim and methotrexate. Annals of Internal Medicine, 117, 877–878.