Predisposing Factors in Hepatitis Induced by IsoniazidRifampin Treatment of Tuberculosis12 CAROLA GRONHAGEN-RISKA, P.-E. HELLSTROM, and B. FROSETH
SUMMARY. Seventy-five patients who developed mild hepatic reactions (serum transaminase concentrations of 45 to 149 units per liter) and 50 patients who showed more serious liver damage (serum transaminase values > 150 units per liter) were compared with 261 consecutive patients who had no liver reactions during treatment with rifampin and isoniazid. Generally, liver toxicity occurred in 18 per cent of patients receiving combined anti-tuberculous drug therapy. Small -increases in transaminase occurred in 14 per cent of the patients; large increases occurred in 4 per cent. Elderly women comprised a risk group. Among patients exhibiting a more serious hepatic lesion (transaminase values > 150 units per liter), alcoholics, mostly men, formed another risk group, together with other patients with a history of previous liver or biliary disease. Of 261 patients who did not develop a liver reaction, 57 per cent were slow INH acetylators. In this study, the groups with small and large increases in transaminase were clearly separated; in the former group there was no preponderance of phenotype, whereas in the latter group, slow acetylators clearly dominated among early (first 4 weeks of treatment) hepatic reactions (P < 0.01). Studies of single-drug regimens of isoniazid have shown that neither slow nor rapid acetylation has any causal influence on isoniazid-induced hepatitis. Because the metabolism of rifampin is independent of the acetylation process, rifampin and isoniazid in combination seem to cause a toxic hepatitis that differs from the hepatitis induced by either drug separately.
Introduction
Rifampin ( RMP) and isoniazid (INH), often combined with a third antituberculous drug, present an effective and much used regimen in current treatment of tuberculosis. The hepatic toxicity of INH was noted quite early (1-3), but because most drugs combined with it were far more toxic, this finding did not receive the attention it deserved. Hepatic injury occurring during therapy with RMP was reported immediately after this drug had been adopted. Previous liver disease was noted to be a predisposing factor (4). (Received in original form July 19, 1977 and in revised form April 28,1978) iFrom Mjolbolsta Hospital, Chest Department, 10350 Mjolbolstasjukhus, Finland. 2 Supported in part by the Finnish Association against Tuberculosis and by the Nummela Foundation.
Later, there was a report of the increased risk of liver reaction when RMP was combined with INH as opposed to ethambutol (EMB), which is not toxic to the liver (5). A study made at our hospital, however, did not show any difference in the frequency of liver reaction among patients receiving RMP-INH or RMP-EMB therapy (6). The potentially serious side effects of INH on the liver became obvious when this drug was used experimentally in large prophylactic studies. In the United States, in a study including approximately 13,000 patients, 13 died of hepatic failure (7); in most cases, INH was believed to be the triggering factor. The International Union against Tuberculosis (IUAT) arranged a similar chemoprophylactic study in Europe. More than 20,000 persons received INH daily. There were no deaths, but hepatitis occurred in 2.8 to 7.7 per 1,000 persons, depending on age. Morbidity increased with age. Patients with pre-
AMERICAN REVIEW OF RESPIRATORY DISEASE, VOLUME 118, 1978
461
462
GRONHAGEN-RISKA, HELLSTROM, AND FROSETH
vious liver disease comprised a risk group. Most reactions occurred at the beginning of treatment (8). Mitchell and associates (9) have reported the occurrence of asymptomatic, spontaneously receding, and quite small increases in transaminase in approximately 20 per cent of all patients receiving INH therapy prophylactically. The INH acetylation rate was not important. At our clinic, we have made a study of hepatic injury occurring during combined RMP-INH therapy. We were particularly interested in comparing small, usually spontaneously receding, increases in transaminase and large increases, which were often combined with serious symptoms and demands for changes in therapy. We wanted to know which factors, if any, differentiate a seemingly harmless hepatic reaction from a potentially dangerous one. Materials and Methods
The patients studied (all of whom had pulmonary tuberculosis), received various doses of RMP intermittently (not once weekly) or daily. In the final analysis, we did not separate these groups, because other studies have not shown that different RMP regimens would alter the occurrence of hepatic reactions (6, 10). All patients received 300 mg of INH daily. In most patients, a third drug was used that was not considered toxic to the liver (EMB, streptomycin, or capreomycin). Age, sex, time of onset of the hepatic reactions, earlier liver disease, or concurring factors potentially damaging to the liver were recorded. The INH acetylator phenotype was determined using documented methods (11-13) before or after treatment when liver function tests were normal. At the beginning of treatment, serum transminase concentrations, alkaline phosphatase, and bilirubin were measured once weekly; later, they were measured once every second or third month or when toxic symptoms occurred. All but 2 patients included in our study had normal transaminase values on initiation of treatment. We set the limit between high and low serum glutamic oxaloacetate transaminase (SGOT) and/or serum glutamic pyruvate transaminase (SGPT) as 150 units per liter at 37° C; the normal maximum in our laboratory is 40 units per liter. We analyzed 75 patients with a small increase in transaminase (SGOT and/or SGPT, 45 to 149 units per liter); 45 of them were from a group of 319 consecutive patients. The remaining 30 patients were collected unselectively from a limited area of our district. These 75 "mild reactors" were compared with all 50 patients from our hospital who developed large increases in transaminase (SGOT and/or SGPT > 150 units per liter) out of a total of 1,545
patients receiving anti-tuberculous treatment between 1968, when RMP was introduced, until the end of 1976. Eleven patients in whom reasons other than RMP-INH treatment were strongly suspected of causing a liver reaction, i.e., infectious hepatitis or simultaneous high alcohol consumption, were excluded. Nine of these were known alcoholics who showed increases in transaminase only on readmissions and whose values returned to normal during continued treatment in the hospital. The chi-square test was used in all statistical calculations. Results
Some kind of liver damage was noted in 18 per cent of 319 consecutive patients receiving RMPINH therapy. In 45 patients (14 per cent) the increases in transaminase were small, and in the remaining 13 patients (4 per cent), the increases were large (SGOT and/or SGPT > 150 units per liter). Of 319 patients studied, 261 had no hepatic reactions; these patients form our comparison group. Fifty-nine per cent (154 subjects) of this group were men, and 57 per cent (148 subjects) were slow INH acetylators. There was a difference in the distribution of slow and fast INH acetylators between men and women, but the difference was not significant. The comparison group was divided into 3 age groups: 60 years of age. Among 261 patients without signs of hepatic injury, women were distributed equally among all age groups, but there was an increased frequency of men (65) in the group 40 to 59 years of age. Of 75 patients who developed a mild hepatic reaction (SGOT and/or SGPT of 45 to 149 units per liter), only 3 complained of diffuse symptoms such as fatigue or epigastric discomfort. Of 50 patients showing a liver reaction of SGOT and/ or SGPT >150 units per liter, however, one died of acute hepatic necrosis. Microscopy revealed a massive centrilobular necrosis. Another patient developed acute ileus, but recovered on discontinuation of treatment. Eighteen other patients (36 per cent) had various symptoms, such as fatigue, loss of appetite, epigastric pain or discomfort, nausea, or enlarged, tender liver. Both types of reaction occurred mainly during the first 8 weeks of RMP-INH treatment; 47 of 75 mild reactions (63 per cent) and 32 of 50 high transaminase values (64 per cent) occurred within this time. In both groups, nearly one half of the cases occurred within 4 weeks. The rest of the reactions were evenly distributed throughout the whole period of treatment, 9 to 24 months. Thir-
463
ISONIAZID-RIFAMPIN TREATMENT OF TUBERCULOSIS
TABLE 1 SEX D I S T R I B U T I O N AMONG PATIENTS E X H I B I T I N G INCREASES IN TRANSAMINASE DURING ANTITUBERCULOUS THERAPY WITH R I F A M P I N A N D ISONIAZID Men Liver Function A. SGOT and/or SGPT, 4 5 - 1 4 9 units/liter &. S-GOT and/or S-GPT, > 150 units/liter C. Normal
(no.) 34/75 26/50 154/261
Women
(%)
(no.)
(%)
45 52 59
41/75 24/50 107/261
55 48 41
*For comparison of A with C, P < 0.01. For B compared to C, there was no significant difference.
reaction occurred in patients more than 60 years of age; the expected proportion was only 34 per cent. The female patients accounted for this discrepancy. Among more serious liver reactions there was a decrease in morbidity among women less than 40 years of age. The INH acetylator phenotype distribution among patients with a large increase in transaminase was studied (table 2). Of the 50 patients belonging to this group, 45 were phenotyped; 33 (73 per cent) were slow acetylators. T h e difference from the control distribution was significant.
teen of 50 patients had a large increase in transaminase (SGOT and/or SGPT >150 units per liter) after a previous mild hepatic reaction (SGOT and/or SGPT 45 to 149 units per liter) that continued during 23 to 210 days of uninterrupted RMP-INH therapy. More than one half of the mild reactors were women (table 1). The same tendency prevailed among patients who showed a large increase in liver transaminase, but the difference, compared with the control group, was not significant. The risk of a toxic liver reaction increased with age (figure 1). Nearly one half of both types of
RATIO
NUMBER IN GROUP TOTAL NUMBER
EIZ] MEN WOMEN
0.50 0.45 0.40 0.35 0.30 0.25
V~l
0.20 0.15 0.10 0.05
M 4> LTI
V
w y
3
-
-
U/1
i/n
w
> 60
150
150 149
w
n o rma1
w ,w/
45- 149
~
n o rma1
FUNCTION
mv\
40-59
AGE
LIVER
A
Fig. 1. Age distribution of patients exhibiting increases in serum transaminase during therapy with rifampin and isoniazid.
464
GRONHAGEN-RISKA, HELLSTROM, AND FROSETH
TABLE 2 ISONIAZID A C E T Y L A T O R PHENOTYPE D I S T R I B U T I O N AMONG PATIENTS DEVELOPING INCREASES IN T R A N S A M I N A S E D U R I N G T H E R A P Y W I T H ISONIAZID (INH) A N D R I F A M P I N
Slow INH Acetylators Liver Function
Rapid INH Acetylators
JnoT)
(%T
A. SGOT and/or SGPT, 4 5 - 1 4 9 units/liter
38
51
37
B. SGOT and/or SGPT, > 150 units/liter
33
73
148
57
C. Normal
(no.)
(%)
Slow and Rapid I N H Acetylators (no.)
(%)
75
100
12
49 27
45
100
113
43
261
100
For comparison of B with C# P < 0.05. For A compared to C, difference not significant.
Many workers have previously noted the early occurrence of liver reactions during therapy. We found a relationship between this phenomenon and the phenotype, but only in the group showing a large increase in serum transaminase (table 3). Slow acetylators were more at risk, especially during the first 4 weeks of treatment. After 8 weeks of treatment, there was no longer a difference between slow and rapid acetylators. Patients with a small increase in transaminase showed no relationship between time of reaction and phenotype. None of our patients was visibly icteric. Nine patients had slight bilirubinemia, the highest value being 5.1 mg per 100 ml. The normal maximal value for our laboratory is 1.2 mg per 100 ml. One of the patients was a mild reactor, and the remainder were in the group with a large increase in transaminase. Increases in alkaline phosphatase were more common, but showed the same pattern. Previous liver disease or some other factor damaging to the liver, such as alcoholism, obviously predispose a patient to toxic hepatitis
(table 4). The IUAT study in Europe (8) of the hepatotoxicity of INH showed that concurrent and previous biliary disorders were risk factors. Patients with this kind of medical history were therefore studied. All patients with previous hepatic or biliary disorder had normal transaminase values on initiation of treatment. Only alcoholics had slightly increased values. The exact number of alcoholics among the 261 patients who showed no hepatic reaction was not known. Earlier studies of our hospital's tuberculous patients showed alcoholism in approximately 12 per cent (14). It should be pointed out again that patients whose hepatic reactions in all probability were solely or primarily due to alcohol were excluded from this study, (5 alcoholics from the group with a small increase and 4 from the group with a large increase in transaminase). Of 50 patients with SGOT and/or SGPT value >150 units per liter, 25 had previously had some hepatobiliary disturbance, whereas only 3 of 75 mild reactors had a similar background. Furthermore, in the former group, 12 of 13 alcoholics were men; in the liver disease group, there was
TABLE 3 RELATIONSHIP BETWEEN ISONIAZID A C E T Y L A T O R PHENOTYPE A N D T I M E OF OCCURRENCE OF LIVER REACTION AFTER I N I T I A T I O N OF T H E R A P Y W I T H I S O N I A Z I D (INH) A N D R I F A M P I N
Liver Function A. SGOT and/or SGPT, 4 5 - 1 4 9 units/liter
B. SGOT and/or SGPT, > 150 units/liter
C. Normal
Time (weeks) 150 units/liter
Previous Hepatic or Biliary Disease Men 1 5
Women 1
Alcoholism Men
Women
1 7
12
Whole Group
1
75 50
For A compared to B, P < 0.001.
no sex preponderance. The acetylator phenotype distribution was equal in the "previous liver disease" and healthy groups. Discussion
Approximately one fifth of patients who receive RMP-INH therapy develop some kind of hepatic reaction. Elderly women form a special risk group, as is usually the case in toxic or allergic reactions (15). Iwainsky and associates (16) have reported higher concentrations of RMP among women, which could account for RMP-induced side effects. Other workers already mentioned (6, 10) have not reported any dose dependence of hepatic reactions, apart from once weekly regimens, which were not used in this study. Decreasing the doses of RMP would thus be of questionable benefit. One half of the patients in our study who developed large increases in transaminase either were alcoholics or had some kind of liver or biliary disease, previously or at the time of study. All but 2 had normal liver function tests when admitted to the hospital, which stresses the importance of adequate information on the patients. Most of the alcoholics were men, which accounts for the lack of significance of the preponderance of women among the group with large increases in transaminase (table 1). Nearly 40 per cent of the patients with increases in transaminase >150 units per liter had abdominal or general symptoms. Although most of these hepatic reactions receded spontaneously, we consider them noteworthy, because patients feeling more ill from treatment than from tuberculosis lose their motivation to take drugs. Slow INH acetylators were prone to develop large increases in transaminase, especially during the first 4 weeks of treatment, during which time one half of the reactions occurred. This fact probably stresses the importance of unmetabolized INH in early toxic hepatitis. The RMP is deacetylated, a process in no way related
to the INH metabolism. Mitchell and Jollows (17) have suggested a relation between rapid acetylation and INH-induced hepatitis; INH metabolites are postulated to be the triggering factors. Speculating along the same lines, others have suggested that early hepatitis in patients treated with RMP-INH is caused by enzyme induction of RMP, and consequently, increased production of toxic INH metabolites (18). In the IUAT study (8), however, there was no preponderance of either slow or rapid acetylators. In addition, other workers have found no change in the halflife of either RMP or INH when these 2 drugs are taken together (19, 20). A high concentration of INH in slow acetylators could account for a hypersensitivity reaction by the liver. The patients in this study, however, showed no other signs of such a reaction (fever, rash, eosinophilia, decrease in blood pressure, and so on). Dave and co-workers, (21), using the human lymphocyte transformation test to study patients with INH-induced hepatitis, found no signs of hypersensitivity. The patients in this study received both INH and RMP. Thus, hepatitis caused by a combination of these drugs seems to differ from hepatitis caused by either drug separately, because the speed of acetylation of INH is of some importance. Another explanation is that when measuring the rate of acetylation, some unknown property of the liver has been determined that in slow acetylators predisposes the patient to a significant liver reaction. All patients receiving RMP-INH therapy should be checked regularly for liver dysfunction. Special attention should be paid to elderly women, alcoholics, patients with previous liver or biliary disease, as well as slow INH acetylators. References
1. Randolph, H., and Joseph, S.: Toxic hepatitis with jaundice occurring in a patient treated
466
GRONHAGEN-RISKA, HELLSTROM, AND FROSETH
with isoniazid, JAMA, 1953,152,38. 2. Gellis, S. N., and Murphy, R. V.: Hepatitis following isoniazid, Dis Chest, 1955,28,462. 3. Merrit, A. D., and Fetter, B. F.: Toxic hepatic necrosis (hepatitis) due to isoniazid, Ann Intern Med, 1959,50, 804. 4. Lesobre, R., Ruffino, J., Teyssier, L., Archard, F., and Brefort, G.: Les icteres au cours du traitement par le rifampicine, Rev Tuberc, 1969, 33, 393. 5. Lees, A. W., Allan, G. W., Smith, J., Tyrrel, W. F., and Fallon, R. J.: Toxicity from rifampicin plus isoniazid and rifampicin plus ethambutol therapy, Tubercle, 1971,52,182. 6. Mattson, K.: Side effects of rifampicin, Scand J Respir Dis, 1973 (Supplement 82, p. 1). 7. Black, M., Mitchell, J. R., Zimmerman, H. J., Ishak, K. G., anl Epler, G. R.: Isoniazid-associated hepatitis in 114 patients, Gastroenterology, 1975,69,289. 8. Riska, N.: Hepatitis cases in INH treated groups and in a control group, Bull Int Un Tuberc, 1976, 51, 203. 9. Mitchell, J. R., Long, M. W., Thorgeirsson, V. P., and Jollows, V. J.: Acetylation rates and monthly liver function tests during one year of isoniazid preventive therapy, Chest, 1975, 68,181. 10. Aquinas, M., Allan, W. G. L., Horsfall, P. A. L., Jenkins, P. K., Wong, H. Y., Girling, D., Tall, R., and Fox, W.: Adverse reactions to daily and intermittent rifampicin regimens for pulmonary tuberculosis in Hong Kong, Br Med J, 1972, 1, 765. 11. Jenne, J. W.: Studies of human patterns of isoniazid metabolism using an intravenous fall-off technique with a chemical method, Am Rev
Respir Dis, 1960,52,1. 12. Maher, J. R., Whiney, J. S., Chambers, J. S., and Stratonis, D. J.: The quantitative determination of isoniazid and para-aminosalicyclic acid in body fluids, Am Rev Tuberc, 1957, 76, 852. 13. Schroder, H.: Simplified method for determining acetylator phenotype, Br Med J, 1972, 3, 506. 14. Riska, H.: Supervised ambulatory treatment of tuberculosis with brief initial hospitalization, Scand J Respir Dis, 1973 (Supplement 83, p. 1). 15. Hurwitz, N.: Predisposing factors in adverse reactions to drugs, Br Med J, 1969,1,536. 16. Iwainsky, H., Winsel, K., Werner, E., and Eule, H.: On the pharmacokinetics of rifampicin during treatment with intermittent administration, Scand J Respir Dis, 1976,57,5. 17. Mitchell, J. R., and Jollows, D. J.: Metabolic activation of drugs to toxic substances, Gastroenterology, 1975, 68, 392. 18. Pessayre, D., Bentata, M., Degott, C , Nouel, O., Miguet, J-P., Rueff, B., and Benhamou, J-P.^ Isoniazid-rifampicin fulminant hepatitis, Gastroenterology, 1977, 72,284. 19. Boman, G.: Serum concentration and half-life of rifampicin after simultaneous oral administration of aminosalicylic acid or isoniazid, Eur J Clin Pharmacol, 1974, 7,369. 20. Acocella, G., Bonollo, L., Garimoldi, M., Mainardi, M., Tenconi, L., and Nicolis, F.: Kinetics of rifampicin and isoniazid administered alone and in combination to normal subjects and patients with liver disease, Gut, 1972,13,47. 21. Dove, J. T., Chaparas, S. D., and Hedrick, S. R.: Failure to demonstrate transformation of lymphocytes of patients with isoniazid-associated hepatitis, Am Rev Respir Dis, 1972, 106, 485.
SUMMARY. Seventy-five patients who developed mild hepatic reactions (serum transaminase concentrations of 45 to 149 units per liter) and 50 patients who showed more serious liver damage (serum transaminase values > 150 units per liter) were compared with 261 consecutive patients who had no liver reactions during treatment with rifampin and isoniazid. Generally, liver toxicity occurred in 18 per cent of patients receiving combined anti-tuberculous drug therapy. Small -increases in transaminase occurred in 14 per cent of the patients; large increases occurred in 4 per cent. Elderly women comprised a risk group. Among patients exhibiting a more serious hepatic lesion (transaminase values > 150 units per liter), alcoholics, mostly men, formed another risk group, together with other patients with a history of previous liver or biliary disease. Of 261 patients who did not develop a liver reaction, 57 per cent were slow INH acetylators. In this study, the groups with small and large increases in transaminase were clearly separated; in the former group there was no preponderance of phenotype, whereas in the latter group, slow acetylators clearly dominated among early (first 4 weeks of treatment) hepatic reactions (P < 0.01). Studies of single-drug regimens of isoniazid have shown that neither slow nor rapid acetylation has any causal influence on isoniazid-induced hepatitis. Because the metabolism of rifampin is independent of the acetylation process, rifampin and isoniazid in combination seem to cause a toxic hepatitis that differs from the hepatitis induced by either drug separately.
Introduction
Rifampin ( RMP) and isoniazid (INH), often combined with a third antituberculous drug, present an effective and much used regimen in current treatment of tuberculosis. The hepatic toxicity of INH was noted quite early (1-3), but because most drugs combined with it were far more toxic, this finding did not receive the attention it deserved. Hepatic injury occurring during therapy with RMP was reported immediately after this drug had been adopted. Previous liver disease was noted to be a predisposing factor (4). (Received in original form July 19, 1977 and in revised form April 28,1978) iFrom Mjolbolsta Hospital, Chest Department, 10350 Mjolbolstasjukhus, Finland. 2 Supported in part by the Finnish Association against Tuberculosis and by the Nummela Foundation.
Later, there was a report of the increased risk of liver reaction when RMP was combined with INH as opposed to ethambutol (EMB), which is not toxic to the liver (5). A study made at our hospital, however, did not show any difference in the frequency of liver reaction among patients receiving RMP-INH or RMP-EMB therapy (6). The potentially serious side effects of INH on the liver became obvious when this drug was used experimentally in large prophylactic studies. In the United States, in a study including approximately 13,000 patients, 13 died of hepatic failure (7); in most cases, INH was believed to be the triggering factor. The International Union against Tuberculosis (IUAT) arranged a similar chemoprophylactic study in Europe. More than 20,000 persons received INH daily. There were no deaths, but hepatitis occurred in 2.8 to 7.7 per 1,000 persons, depending on age. Morbidity increased with age. Patients with pre-
AMERICAN REVIEW OF RESPIRATORY DISEASE, VOLUME 118, 1978
461
462
GRONHAGEN-RISKA, HELLSTROM, AND FROSETH
vious liver disease comprised a risk group. Most reactions occurred at the beginning of treatment (8). Mitchell and associates (9) have reported the occurrence of asymptomatic, spontaneously receding, and quite small increases in transaminase in approximately 20 per cent of all patients receiving INH therapy prophylactically. The INH acetylation rate was not important. At our clinic, we have made a study of hepatic injury occurring during combined RMP-INH therapy. We were particularly interested in comparing small, usually spontaneously receding, increases in transaminase and large increases, which were often combined with serious symptoms and demands for changes in therapy. We wanted to know which factors, if any, differentiate a seemingly harmless hepatic reaction from a potentially dangerous one. Materials and Methods
The patients studied (all of whom had pulmonary tuberculosis), received various doses of RMP intermittently (not once weekly) or daily. In the final analysis, we did not separate these groups, because other studies have not shown that different RMP regimens would alter the occurrence of hepatic reactions (6, 10). All patients received 300 mg of INH daily. In most patients, a third drug was used that was not considered toxic to the liver (EMB, streptomycin, or capreomycin). Age, sex, time of onset of the hepatic reactions, earlier liver disease, or concurring factors potentially damaging to the liver were recorded. The INH acetylator phenotype was determined using documented methods (11-13) before or after treatment when liver function tests were normal. At the beginning of treatment, serum transminase concentrations, alkaline phosphatase, and bilirubin were measured once weekly; later, they were measured once every second or third month or when toxic symptoms occurred. All but 2 patients included in our study had normal transaminase values on initiation of treatment. We set the limit between high and low serum glutamic oxaloacetate transaminase (SGOT) and/or serum glutamic pyruvate transaminase (SGPT) as 150 units per liter at 37° C; the normal maximum in our laboratory is 40 units per liter. We analyzed 75 patients with a small increase in transaminase (SGOT and/or SGPT, 45 to 149 units per liter); 45 of them were from a group of 319 consecutive patients. The remaining 30 patients were collected unselectively from a limited area of our district. These 75 "mild reactors" were compared with all 50 patients from our hospital who developed large increases in transaminase (SGOT and/or SGPT > 150 units per liter) out of a total of 1,545
patients receiving anti-tuberculous treatment between 1968, when RMP was introduced, until the end of 1976. Eleven patients in whom reasons other than RMP-INH treatment were strongly suspected of causing a liver reaction, i.e., infectious hepatitis or simultaneous high alcohol consumption, were excluded. Nine of these were known alcoholics who showed increases in transaminase only on readmissions and whose values returned to normal during continued treatment in the hospital. The chi-square test was used in all statistical calculations. Results
Some kind of liver damage was noted in 18 per cent of 319 consecutive patients receiving RMPINH therapy. In 45 patients (14 per cent) the increases in transaminase were small, and in the remaining 13 patients (4 per cent), the increases were large (SGOT and/or SGPT > 150 units per liter). Of 319 patients studied, 261 had no hepatic reactions; these patients form our comparison group. Fifty-nine per cent (154 subjects) of this group were men, and 57 per cent (148 subjects) were slow INH acetylators. There was a difference in the distribution of slow and fast INH acetylators between men and women, but the difference was not significant. The comparison group was divided into 3 age groups: 60 years of age. Among 261 patients without signs of hepatic injury, women were distributed equally among all age groups, but there was an increased frequency of men (65) in the group 40 to 59 years of age. Of 75 patients who developed a mild hepatic reaction (SGOT and/or SGPT of 45 to 149 units per liter), only 3 complained of diffuse symptoms such as fatigue or epigastric discomfort. Of 50 patients showing a liver reaction of SGOT and/ or SGPT >150 units per liter, however, one died of acute hepatic necrosis. Microscopy revealed a massive centrilobular necrosis. Another patient developed acute ileus, but recovered on discontinuation of treatment. Eighteen other patients (36 per cent) had various symptoms, such as fatigue, loss of appetite, epigastric pain or discomfort, nausea, or enlarged, tender liver. Both types of reaction occurred mainly during the first 8 weeks of RMP-INH treatment; 47 of 75 mild reactions (63 per cent) and 32 of 50 high transaminase values (64 per cent) occurred within this time. In both groups, nearly one half of the cases occurred within 4 weeks. The rest of the reactions were evenly distributed throughout the whole period of treatment, 9 to 24 months. Thir-
463
ISONIAZID-RIFAMPIN TREATMENT OF TUBERCULOSIS
TABLE 1 SEX D I S T R I B U T I O N AMONG PATIENTS E X H I B I T I N G INCREASES IN TRANSAMINASE DURING ANTITUBERCULOUS THERAPY WITH R I F A M P I N A N D ISONIAZID Men Liver Function A. SGOT and/or SGPT, 4 5 - 1 4 9 units/liter &. S-GOT and/or S-GPT, > 150 units/liter C. Normal
(no.) 34/75 26/50 154/261
Women
(%)
(no.)
(%)
45 52 59
41/75 24/50 107/261
55 48 41
*For comparison of A with C, P < 0.01. For B compared to C, there was no significant difference.
reaction occurred in patients more than 60 years of age; the expected proportion was only 34 per cent. The female patients accounted for this discrepancy. Among more serious liver reactions there was a decrease in morbidity among women less than 40 years of age. The INH acetylator phenotype distribution among patients with a large increase in transaminase was studied (table 2). Of the 50 patients belonging to this group, 45 were phenotyped; 33 (73 per cent) were slow acetylators. T h e difference from the control distribution was significant.
teen of 50 patients had a large increase in transaminase (SGOT and/or SGPT >150 units per liter) after a previous mild hepatic reaction (SGOT and/or SGPT 45 to 149 units per liter) that continued during 23 to 210 days of uninterrupted RMP-INH therapy. More than one half of the mild reactors were women (table 1). The same tendency prevailed among patients who showed a large increase in liver transaminase, but the difference, compared with the control group, was not significant. The risk of a toxic liver reaction increased with age (figure 1). Nearly one half of both types of
RATIO
NUMBER IN GROUP TOTAL NUMBER
EIZ] MEN WOMEN
0.50 0.45 0.40 0.35 0.30 0.25
V~l
0.20 0.15 0.10 0.05
M 4> LTI
V
w y
3
-
-
U/1
i/n
w
> 60
150
150 149
w
n o rma1
w ,w/
45- 149
~
n o rma1
FUNCTION
mv\
40-59
AGE
LIVER
A
Fig. 1. Age distribution of patients exhibiting increases in serum transaminase during therapy with rifampin and isoniazid.
464
GRONHAGEN-RISKA, HELLSTROM, AND FROSETH
TABLE 2 ISONIAZID A C E T Y L A T O R PHENOTYPE D I S T R I B U T I O N AMONG PATIENTS DEVELOPING INCREASES IN T R A N S A M I N A S E D U R I N G T H E R A P Y W I T H ISONIAZID (INH) A N D R I F A M P I N
Slow INH Acetylators Liver Function
Rapid INH Acetylators
JnoT)
(%T
A. SGOT and/or SGPT, 4 5 - 1 4 9 units/liter
38
51
37
B. SGOT and/or SGPT, > 150 units/liter
33
73
148
57
C. Normal
(no.)
(%)
Slow and Rapid I N H Acetylators (no.)
(%)
75
100
12
49 27
45
100
113
43
261
100
For comparison of B with C# P < 0.05. For A compared to C, difference not significant.
Many workers have previously noted the early occurrence of liver reactions during therapy. We found a relationship between this phenomenon and the phenotype, but only in the group showing a large increase in serum transaminase (table 3). Slow acetylators were more at risk, especially during the first 4 weeks of treatment. After 8 weeks of treatment, there was no longer a difference between slow and rapid acetylators. Patients with a small increase in transaminase showed no relationship between time of reaction and phenotype. None of our patients was visibly icteric. Nine patients had slight bilirubinemia, the highest value being 5.1 mg per 100 ml. The normal maximal value for our laboratory is 1.2 mg per 100 ml. One of the patients was a mild reactor, and the remainder were in the group with a large increase in transaminase. Increases in alkaline phosphatase were more common, but showed the same pattern. Previous liver disease or some other factor damaging to the liver, such as alcoholism, obviously predispose a patient to toxic hepatitis
(table 4). The IUAT study in Europe (8) of the hepatotoxicity of INH showed that concurrent and previous biliary disorders were risk factors. Patients with this kind of medical history were therefore studied. All patients with previous hepatic or biliary disorder had normal transaminase values on initiation of treatment. Only alcoholics had slightly increased values. The exact number of alcoholics among the 261 patients who showed no hepatic reaction was not known. Earlier studies of our hospital's tuberculous patients showed alcoholism in approximately 12 per cent (14). It should be pointed out again that patients whose hepatic reactions in all probability were solely or primarily due to alcohol were excluded from this study, (5 alcoholics from the group with a small increase and 4 from the group with a large increase in transaminase). Of 50 patients with SGOT and/or SGPT value >150 units per liter, 25 had previously had some hepatobiliary disturbance, whereas only 3 of 75 mild reactors had a similar background. Furthermore, in the former group, 12 of 13 alcoholics were men; in the liver disease group, there was
TABLE 3 RELATIONSHIP BETWEEN ISONIAZID A C E T Y L A T O R PHENOTYPE A N D T I M E OF OCCURRENCE OF LIVER REACTION AFTER I N I T I A T I O N OF T H E R A P Y W I T H I S O N I A Z I D (INH) A N D R I F A M P I N
Liver Function A. SGOT and/or SGPT, 4 5 - 1 4 9 units/liter
B. SGOT and/or SGPT, > 150 units/liter
C. Normal
Time (weeks) 150 units/liter
Previous Hepatic or Biliary Disease Men 1 5
Women 1
Alcoholism Men
Women
1 7
12
Whole Group
1
75 50
For A compared to B, P < 0.001.
no sex preponderance. The acetylator phenotype distribution was equal in the "previous liver disease" and healthy groups. Discussion
Approximately one fifth of patients who receive RMP-INH therapy develop some kind of hepatic reaction. Elderly women form a special risk group, as is usually the case in toxic or allergic reactions (15). Iwainsky and associates (16) have reported higher concentrations of RMP among women, which could account for RMP-induced side effects. Other workers already mentioned (6, 10) have not reported any dose dependence of hepatic reactions, apart from once weekly regimens, which were not used in this study. Decreasing the doses of RMP would thus be of questionable benefit. One half of the patients in our study who developed large increases in transaminase either were alcoholics or had some kind of liver or biliary disease, previously or at the time of study. All but 2 had normal liver function tests when admitted to the hospital, which stresses the importance of adequate information on the patients. Most of the alcoholics were men, which accounts for the lack of significance of the preponderance of women among the group with large increases in transaminase (table 1). Nearly 40 per cent of the patients with increases in transaminase >150 units per liter had abdominal or general symptoms. Although most of these hepatic reactions receded spontaneously, we consider them noteworthy, because patients feeling more ill from treatment than from tuberculosis lose their motivation to take drugs. Slow INH acetylators were prone to develop large increases in transaminase, especially during the first 4 weeks of treatment, during which time one half of the reactions occurred. This fact probably stresses the importance of unmetabolized INH in early toxic hepatitis. The RMP is deacetylated, a process in no way related
to the INH metabolism. Mitchell and Jollows (17) have suggested a relation between rapid acetylation and INH-induced hepatitis; INH metabolites are postulated to be the triggering factors. Speculating along the same lines, others have suggested that early hepatitis in patients treated with RMP-INH is caused by enzyme induction of RMP, and consequently, increased production of toxic INH metabolites (18). In the IUAT study (8), however, there was no preponderance of either slow or rapid acetylators. In addition, other workers have found no change in the halflife of either RMP or INH when these 2 drugs are taken together (19, 20). A high concentration of INH in slow acetylators could account for a hypersensitivity reaction by the liver. The patients in this study, however, showed no other signs of such a reaction (fever, rash, eosinophilia, decrease in blood pressure, and so on). Dave and co-workers, (21), using the human lymphocyte transformation test to study patients with INH-induced hepatitis, found no signs of hypersensitivity. The patients in this study received both INH and RMP. Thus, hepatitis caused by a combination of these drugs seems to differ from hepatitis caused by either drug separately, because the speed of acetylation of INH is of some importance. Another explanation is that when measuring the rate of acetylation, some unknown property of the liver has been determined that in slow acetylators predisposes the patient to a significant liver reaction. All patients receiving RMP-INH therapy should be checked regularly for liver dysfunction. Special attention should be paid to elderly women, alcoholics, patients with previous liver or biliary disease, as well as slow INH acetylators. References
1. Randolph, H., and Joseph, S.: Toxic hepatitis with jaundice occurring in a patient treated
466
GRONHAGEN-RISKA, HELLSTROM, AND FROSETH
with isoniazid, JAMA, 1953,152,38. 2. Gellis, S. N., and Murphy, R. V.: Hepatitis following isoniazid, Dis Chest, 1955,28,462. 3. Merrit, A. D., and Fetter, B. F.: Toxic hepatic necrosis (hepatitis) due to isoniazid, Ann Intern Med, 1959,50, 804. 4. Lesobre, R., Ruffino, J., Teyssier, L., Archard, F., and Brefort, G.: Les icteres au cours du traitement par le rifampicine, Rev Tuberc, 1969, 33, 393. 5. Lees, A. W., Allan, G. W., Smith, J., Tyrrel, W. F., and Fallon, R. J.: Toxicity from rifampicin plus isoniazid and rifampicin plus ethambutol therapy, Tubercle, 1971,52,182. 6. Mattson, K.: Side effects of rifampicin, Scand J Respir Dis, 1973 (Supplement 82, p. 1). 7. Black, M., Mitchell, J. R., Zimmerman, H. J., Ishak, K. G., anl Epler, G. R.: Isoniazid-associated hepatitis in 114 patients, Gastroenterology, 1975,69,289. 8. Riska, N.: Hepatitis cases in INH treated groups and in a control group, Bull Int Un Tuberc, 1976, 51, 203. 9. Mitchell, J. R., Long, M. W., Thorgeirsson, V. P., and Jollows, V. J.: Acetylation rates and monthly liver function tests during one year of isoniazid preventive therapy, Chest, 1975, 68,181. 10. Aquinas, M., Allan, W. G. L., Horsfall, P. A. L., Jenkins, P. K., Wong, H. Y., Girling, D., Tall, R., and Fox, W.: Adverse reactions to daily and intermittent rifampicin regimens for pulmonary tuberculosis in Hong Kong, Br Med J, 1972, 1, 765. 11. Jenne, J. W.: Studies of human patterns of isoniazid metabolism using an intravenous fall-off technique with a chemical method, Am Rev
Respir Dis, 1960,52,1. 12. Maher, J. R., Whiney, J. S., Chambers, J. S., and Stratonis, D. J.: The quantitative determination of isoniazid and para-aminosalicyclic acid in body fluids, Am Rev Tuberc, 1957, 76, 852. 13. Schroder, H.: Simplified method for determining acetylator phenotype, Br Med J, 1972, 3, 506. 14. Riska, H.: Supervised ambulatory treatment of tuberculosis with brief initial hospitalization, Scand J Respir Dis, 1973 (Supplement 83, p. 1). 15. Hurwitz, N.: Predisposing factors in adverse reactions to drugs, Br Med J, 1969,1,536. 16. Iwainsky, H., Winsel, K., Werner, E., and Eule, H.: On the pharmacokinetics of rifampicin during treatment with intermittent administration, Scand J Respir Dis, 1976,57,5. 17. Mitchell, J. R., and Jollows, D. J.: Metabolic activation of drugs to toxic substances, Gastroenterology, 1975, 68, 392. 18. Pessayre, D., Bentata, M., Degott, C , Nouel, O., Miguet, J-P., Rueff, B., and Benhamou, J-P.^ Isoniazid-rifampicin fulminant hepatitis, Gastroenterology, 1977, 72,284. 19. Boman, G.: Serum concentration and half-life of rifampicin after simultaneous oral administration of aminosalicylic acid or isoniazid, Eur J Clin Pharmacol, 1974, 7,369. 20. Acocella, G., Bonollo, L., Garimoldi, M., Mainardi, M., Tenconi, L., and Nicolis, F.: Kinetics of rifampicin and isoniazid administered alone and in combination to normal subjects and patients with liver disease, Gut, 1972,13,47. 21. Dove, J. T., Chaparas, S. D., and Hedrick, S. R.: Failure to demonstrate transformation of lymphocytes of patients with isoniazid-associated hepatitis, Am Rev Respir Dis, 1972, 106, 485.
