INT J TUBERC LUNG DIS 5(1):65–69
© 2001 IUATLD
The management of anti-tuberculosis drug-induced
hepatotoxicity
K. Tahaoğlu, G. Ataç, T. Sevim, T. Törün, Ö. Yazıcıoğlu, G. Horzum, I. Gemci, A. Öngel,
N. Kapaklı, E. Aksoy
.
Süreyyapaşa Center for Chest Disease and Thoracic Surgery, Istanbul, Turkey
S E T T I N G : A tuberculosis ward in a chest disease teach-
ing hospital.
O B J E C T I V E : To compare the efficacy of two different
retreatment protocols on hepatotoxicity recurrence in
tuberculosis treatment.
D E S I G N : In a prospective, randomised study, 45 patients
with new tuberculosis developed hepatotoxicity after
anti-tuberculosis treatment. Patients in Group I (n  20)
were retreated with a drug regimen consisting of iso-
niazid, rifampicin, ethambutol and streptomycin adminis-
tered by gradually increasing the number and dosage of
the drugs. Patients in Group II (n  25) were retreated
with the same regimen (isoniazid, rifampicin, pyrazina-
mide and ethambutol) in the same dosages throughout.
TODAY, the most effective chemotherapy regimen
for new tuberculosis cases is a combination of iso-
niazid (H), rifampicin (R), pyrazinamide (Z) and eth-
ambutol (E) or streptomycin (S). These are well toler-
ated, safe drugs with few side-effects.1 However, H,
R and Z each have hepatotoxic side-effects which
increase when the drugs are combined.1,2 Although
anti-tuberculosis drug-induced hepatotoxicity is well
known, there is no agreement on the clinical approach
for cases in whom hepatotoxicity has developed.1–3
In this study, 45 cases in whom hepatotoxicity had
developed were randomized into two groups, in
which patients were retreated with two different
retreatment protocols and compared with regard to
recurrence of hepatotoxicity.
MATERIALS AND METHODS
Patients
The study was performed in the tuberculosis ward of
a chest diseases and chest surgery teaching hospital
between 1992 and 1998. The study comprised 37
patients with new pulmonary tuberculosis and eight
with tuberculosis pleurisy who had anti-tuberculosis
Correspondence
. to: Dr Kemal Tahaoğ lu, Süreyyapaşa Göğ üs Kalp Damar Hastalıkları Eğ itim Hastanesi, Posta kod: 81554
Maltepe, Istanbul, Turkey. Tel: (90) 216 411 7197. Fax : (90) 216 411 7198. e-mail: kemaltaha@hotmail.com
Article submitted 4 May 2000. Final version accepted 22 September 2000.
.
SUMMARY
R E S U L T S : Hepatotoxicity recurred in respectively zero
and six (24%) patients in Groups I and II (P  0.021).
Of the six patients with recurrence of hepatitis, one
could not be followed up. The other five received the
same retreatment protocol as Group I. By the end of
retreatment, all patients were cured.
C O N C L U S I O N : The recurrence rate of hepatotoxicity in
the retreatment of tuberculosis is higher in the reintro-
duction of a full-dose regimen including pyrazinamide,
which causes more hepatotoxicity than gradual reintro-
duction of a regimen without pyrazinamide.
K E Y W O R D S : anti-tuberculosis therapy; hepatotoxicity
drug-induced hepatotoxicity. No patient with sero-
logical evidence of acute infection with hepatitis virus
was included in the study. All of the patients were hos-
pitalised for at least the first 2 months for retreatment.
Criteria for diagnosis of tuberculosis
For the diagnosis of pulmonary tuberculosis at least
two positive sputum specimens for acid-fast bacilli by
microscopy and/or culture positivity for Mycobacte-
rium tuberculosis were required. Tuberculous pleu-
risy was diagnosed by detection of caseating granu-
lomas in histopathological examination of tissue
specimens taken by parietal pleural needle biopsy.
Drug regimens
Treatment was planned as recommended by our
national tuberculosis control programme; the total
treatment period was 9 months, with an initial phase
of 2 months consisting of HRZ and E (or S), followed
by a continuation phase of 7 months consisting of
HR. Treatment was given daily, and at least the initial
phase was given during hospitalisation. Drug dosages
were as follows: H: 300 mg/day; R: 600 mg/day; Z:
1500 mg/day; E: 1500 mg/day; S: 1000 mg/day.
66 The International Journal of Tuberculosis and Lung Disease
Treatment outcomes
Cure: a sputum smear-positive patient who is smear-
negative at completion of treatment.
Treatment completed: an extra-pulmonary tuber-
culosis patient who has completed treatment.
Diagnostic criteria for drug induced hepatotoxicity
Normalisation of liver functions after withdrawal of
all anti-tuberculosis drugs, and at least one of the fol-
lowing criteria present for the diagnosis of drug-
induced hepatotoxicity: a rise to five times the normal
levels (40U/L) of serum aspartate aminotransferase
(AST) and/or alanin aminotransferase (ALT); a rise in
the level of serum total bilirubin over 1.5 mg/dl; any
increase in AST and/or ALT above pretreatment levels,
together with anorexia, nausea, vomiting and jaundice.
The same criteria were used to determine the
recurrence of hepatotoxicity during retreatment.
Laboratory monitoring
Serum AST, ALT, bilirubin, total protein, albumin
and globulin levels of all patients were measured
before treatment. These tests were performed once a
month in patients with a history of high alcohol
intake and/or a history of viral hepatitis. In other
cases, liver enzymes and bilirubin were checked when
patients had symptoms of hepatitis and jaundice.
After withdrawal of treatment, serum transami-
nases were measured twice weekly until they returned
to normal levels. Regular laboratory monitoring was
performed during the retreatment protocol. Unless
there were symptoms of hepatitis, transaminases and
bilirubin were measured twice in the first 2 weeks,
once in the following 2 weeks, and once every 2
weeks in the second month. After the second month,
laboratory measurements were performed only in the
case of symptoms suggestive of hepatotoxicity.
STUDY DESIGN
Patients who developed anti-tuberculosis drug-induced
hepatotoxicity were randomized into two groups. The
risk factors for hepatotoxicity were compared statisti-
cally to ensure that there was no increased susceptibility
to hepatotoxicity in either group. The risk factors eval-
uated for hepatotoxicity were age, sex, alcohol con-
sumption, hepatitis markers, radiological extension of
the disease in the lungs, pretreatment serum albumin
level, diabetes mellitus, additional hepatotoxic drug
use, low body weight and low body mass index (BMI).
Low body weight was considered as 10% lower than
normal for sex and height; values of 20 and below were
considered as low BMI.4,5 For hypoalbuminemia, 3.5
mg/dl was accepted as the normal level, which was con-
sidered in our laboratory as the value of the lowest nor-
mal serum albumin. Alcohol consumption was consid-
ered a risk factor in those who had drunk 6 units
(48 g ethanol) per day for more than one year.6
Patients were classified according to radiological
extension of the disease. The disease was considered
extensive if cavities totalled 15 cm in diameter and/
or moderately dense infiltrates involved 75% of
lung fields on chest X-ray.7 Pleurisy cases were con-
sidered as diseases which were not extensive.
Before anti-tuberculosis chemotherapy, serum AST,
ALT, total protein, albumin, globulin, urea, creatinin
tests and total blood count were performed in all
patients. Serum AST, ALT, bilirubin, total protein, albu-
min and globulin levels of all patients were measured in
our biochemistry laboratory by Technicon RAXT ana-
lyzer. Hepatitis markers and anti-human immunodefi-
ciency virus (HIV) were analyzed by using ELISA
immunoassay kits. HIV antibody (anti-HIV) and hepa-
titis B surface Antigen (HbsAg) were tested in all cases.
Because of technical difficulties before 1995, other viral
hepatitis markers such as IgM hepatitis B core antibody
(Anti-HbcIgM), IgG hepatitis B core antibody (Anti-
HbcIgG), antibody hepatitis B surface antigen (Anti-
HBs), IgM antibody hepatitis A (Anti-HAVIgM) and
antibody hepatitis C (Anti-HCV) were not tested in 20
patients (11 in Group I and nine in Group II).
When drug-induced hepatotoxicity was detected,
all anti-tuberculosis drugs were withdrawn. After
hepatotoxicity-related symptoms had disappeared
and laboratory findings had returned to normal lev-
els, anti-tuberculosis treatment was reintroduced in
all study patients. The groups started treatment again
on two different retreatment protocols. In Group I (n 
20), drug regimens not containing Z were started
gradually as follows: day 1, S 1000 mg/day and E
1500 mg/day; day 3, S 1000 mg/day, E 1500 mg/day
and H 100 mg/day; day 6, S 1000 mg/day, E 1500 mg/
day and H 200 mg/day; day 9, S 1000 mg/day, E 1500
mg/day and H 300 mg/day; day 12: S 1000 mg/day, E
1500 mg/day, H 300 mg/day and R 150 mg/day; day
15, S 1000 mg/day, E 1500 mg/day, H 300 mg/day
and R 300 mg/day; day 18, S 1000 mg/day, E 1500
mg/day, H 300 mg/day and R 450 mg/day.
Group II (n  25) was retreated with the same
drug regimen as previously, i.e., H 300 mg/day, R
600 mg/day, Z 1500 mg/day and E 1500 mg/day,
with no change.
Data analysis
For comparison of Groups I (n  20) and II (n  25),
2 or Fisher’s exact tests were used for qualitative risk
factors and Student’s t-test was used for quantitative
risk factors. Fisher’s exact test was used to compare
the two retreatment protocols with regard to recur-
rence of hepatotoxicity. Statistical analyses were per-
formed using SPSS version 6.0 (SPSS Inc, Chicago IL)
on a personal computer.
RESULTS
There were 30 (77%) male and 15 (33%) female
patients, with ages ranging from 15 to 76 years (mean
 The management of anti-tuberculosis drug-induced hepatotoxicity 67

Table 1 Mean values of highest calculated AST, ALT and bilirubin levels in 45 patients,
and statistical comparison of Groups I and II

All cases Group I Group II

(n  45) (n  25) (n  20) P*
AST [U/L] 0.16
Mean  SD 281.77  210.84 330.95  177.36 242.44  230.16
(min–max) (60–1025) (86–540) (49–1025)
ALT [U/L] 0.60
Mean  SD 259.46  189.49 276.20  173.07 246.08  204.19
(min–max) (42–897) (42–570) (48–897)
Total bilirubin [mg/dl] 0.26
Mean  SD 1.55  1.58 1.84  1.48 1.31  1.65
(min–max) (0.2–6.96) (0.8–6.96) (0.2–6.77)
* Comparison of Group I and II.
AST  aspartate aminotransferase; ALT  alanin aminotransferase.

39.96  16.45 years). A total of 37 (83%) had active whom hepatitis recurrence occurred, one patient
pulmonary tuberculosis and eight (17%) had tuber- could not be followed up. The other five cases were
culous pleurisy. given the retreatment regimen as for Group I, and
The highest values in hepatic function tests hepatotoxicity did not recur. Retreatment in all 44
recorded when hepatotoxicity developed and the sta- cases was completed in 9 months, and all were cured.
tistical analysis of the values are shown in Table 1.
There were no statistically significant differences
DISCUSSION
between Groups I and II with respect to liver damage.
The time from start of tuberculosis chemotherapy Hepatotoxicity is one of the main side-effects of the
until the development of hepatotoxicity (latency drugs used in tuberculosis treatment. Once this side-
period) was a mean of 17.35  18.67 (6–102) days; effect has occurred, all drugs are withdrawn and
the time from withdrawal of the drugs to resolution retreatment is started only when all biochemical
of hepatotoxicity was a mean of 18.71  11.36 (4– markers have returned to normal levels. In order to
58) days. avoid recurrence of hepatotoxicity with this retreat-
Table 2 displays the comparison of risk factors in ment procedure, it seems logical that there should be
both groups. Radiologically extensive disease in some drug changes in the second treatment regimen.
Group I and female sex in Group II were found to be This study revealed that the rate of recurrence of
statistically significant risk factors for the progression hepatotoxicity in the retreatment of tuberculosis is
of hepatotoxicity (P  0.001 and P  0.036, respec- lower when a regimen that does not include pyrazin-
tively). All patients were HIV negative. Serological amide is gradually reintroduced than when the full-
markers for hepatitis viruses in the patients are shown dose regimen including pyrazinamide is given again
in Table 3. (P  0.021). This evidence was further enhanced by
While there was no recurrence of hepatitis after the fact that patients who had had hepatotoxicity
retreatment in Group I, it did develop in six cases recurrence with the retreatment regimen containing
(24%) following retreatment in Group II. The differ- pyrazinamide showed no recurrence (even in a third
ence with respect to recurrence of hepatotoxicity was period of anti-tuberculosis treatment) when they were
statistically significant (P  0.021). Of the six cases in gradually reintroduced to a regimen without pyrazin-
amide. According to these results, the recurrence of
hepatotoxicity seemed to be related to the use of a
Table 2 Risk factors for hepatotoxicity in Groups I and II full-dose regimen with pyrazinamide.
There were some risk factors for hepatotoxicity
Group I Group II
(n  20) (n  25)
Risk factors n (%) n (%) P Table 3 Serologic markers for hepatitis viruses in the patients
Age 50 7 (35) 6 (24) 0.418
Female sex 3 (15) 11 (44) 0.036 Group I (n  20) Group II (n  25)
Hepatitis
Alcohol use 3 (15) 1 (4) 0.223 markers Positive Negative Positive Negative P
Extensive disease 9 (45) 0 0.001
Hypoalbuminemia 13 (65) 9 (36) 0.053 HbsAg* 0 20 0 25 —
Diabetes mellitus 2 (1) 1 (4) 0.415 AntiHBsAg† 1 13 1 10 0.696
Low body weight 11 (55) 13 (52) 0.841 AntiHBcIgG† 1 13 4 7 0.250
Low BMI 7 (35) 8 (32) 0.832 AntiHBcIgM† 0 14 0 11 —
Additional hepatotoxic drugs 2 (1) 3 (12) 0.608 AntiHAV† 0 14 0 11 —
Paracetamol 1 2 AntiHCV† 0 14 0 11 —
Chloropropamide 1 1
* Tested in 45 cases.
BMI  body mass index. † Tested in 14 cases in Group I and 11 cases in Group II.
68 The International Journal of Tuberculosis and Lung Disease
development among the 45 patients. Although these risk
factors showed homogeneous distribution between the
groups (nine patients in Group I, and 11 in Group II)
(Table 2), their contribution to the difference in hepato-
toxicity recurrence rates between the groups cannot be
ruled out. Similarly, different doses of rifampicin in the
two groups (although both are within the well known
therapeutic range of 450 mg/day in Group I and 600
mg/day in Group II) may be kept in mind when inter-
preting the different hepatotoxicity rates of the groups.
Gradually introducing the drugs by giving them in
increasing number and dosage, has been considered
by some authors to be one of the reasons for a suc-
cessful retreatment protocol.2,8,9 A similar observa-
tion (in patients who also did not receive pyrazina-
mide) was made in this study. This effect is believed to
be the avoidance of hypersensitivity reaction6,10 due
to the administration of all drugs at once, as seen in
some patients in Group II.
Although there is no reported controlled study, there
are different opinions on how to perform retreatment
after hepatotoxicity has resolved.8,11–13 The aim of the
management of hepatotoxicity in these patients must be
to maintain cure of the disease, without liver damage.
Clinical and laboratory monitoring throughout retreat-
ment is of great importance in order to detect hepato-
toxicity. In our study, we placed great emphasis on clin-
ical monitoring, and symptoms of hepatotoxicity were
meticulously observed during daily visits. Laboratory
monitoring was also performed during retreatment.
Careful selection of the drugs to be used in retreat-
ment is of paramount importance. Using drugs which
are hepatotoxic and whose effect increases when they
are used in combination is complicated. However,
isoniazid and rifampicin are essential in tuberculosis
treatment, and in regimens which do not contain
rifampicin the treatment success rate is compromised.
The British Thoracic Society emphasises the impor-
tance of rifampicin resistance and suggests that such
cases should be regarded as multidrug-resistant
tuberculosis until all susceptibilities are established.13
On the other hand, in a regimen that does not contain
rifampicin, the duration of treatment is prolonged to
12–18 months instead of 6–9 months. Withdrawal of
a drug such as isoniazid or rifampicin because hepa-
totoxicity developed during the initial phase1 may
cause incomplete initial phase treatment in countries
such as ours, where primary drug resistance is high,
leading to acquired drug resistance.14,15
There is some debate on whether or not to use pyr-
azinamide in retreatment after hepatotoxicity. The
French health authorities and Merck-Sharp-Dohme
have recommended that pyrazinamide should not be
used in retreatment after hepatotoxicity.3,16 Fatal
hepatic necrosis caused by pyrazinamide has also
been reported.9,16–18
In conclusion, there was no recurrence of hepatotox-
icity in patients with gradual reintroduction of a
retreatment regimen without pyrazinamide, while 24%
of patients with immediate, full-dose reintroduction of
retreatment regimens with pyrazinamide did experience
recurrence. Most tuberculosis cases in whom hepato-
toxicity has developed can be successfully treated with
regimens including isoniazid and rifampicin; however it
should be kept in mind that these cases require careful
clinical and laboratory monitoring.
Acknowledgements
The authors. wish to thank to Semih Halezeroglu, MD, and Profes-
sor Oktay I mecik for their valuable scientific suggestions on the
manuscript, and Günay Can, MD, for evaluating statistical data.
References
1 Schaberg T. The dark side of antituberculosis therapy: adverse
events involving liver function. Eur Respir Rev 1995; 4: 1247–
1249.
2 Ormerod L P. Hepatotoxicity of antituberculosis drugs. Tho-
rax 1996; 51: 111–113.
3 Thompson N P, Caplin M E, Hamilton M I, et al. Antitubercu-
losis medication and the liver: dangers and recommendations
in management. Eur Respir J 1995; 8: 1384–1388.
4 Russell M R. Nutritional Assessment. In: Wyngaarden J B,
Smith L H, Bennett J C, eds. Cecil Textbook of Medicine. 19th
ed. Philadelphia: WB Saunders, 1989: 1151–1154.
5 Roche A F, Siervogel R M, Chumlea W C. Grading body fat-
ness from limited anthropometric data. Am J Clin Nutr 1981;
34: 2828–2831.
6 Pande J N, Singh S P N, Khilnani G C, Khilnani S, Tandon R K.
Risk factors for hepatotoxicity from antituberculosis drugs: a
case-control study. Thorax 1996; 51: 132–136.
7 Goble M, Iseman M D, Madsen L A, Waite D, Ackerson Z,
Horsburgh R. Treatment of 171 patients with pulmonary tu-
berculosis resistant to isoniazid and rifampin. N Engl J Med
1993; 328: 527–532.
8 O’Brien R J. Hepatotoxic reaction to antituberculous drugs:
adjustment to therapeutic regimen. JAMA 1991; 265: 3323.
9 Antituberculous therapy and acute liver failure. [Editorials]
Lancet 1995; 345: 1170–1172.
10 Steele M A, Burk R F, DesPrez R M. Toxic hepatitis with iso-
niazid and rifampin. Chest 1991; 99: 465–471.
11 World Health Organization. Treatment of tuberculosis: Guide-
lines for national programmes. 2nd ed. Geneva: WHO, 1997;
25–39.
12 American Thoracic Society. Treatment of tuberculosis and tu-
berculosis infection in adults and children. Am J Respir Crit
Care Med 1994; 149: 1359–1374.
13 Joint Tuberculosis Committee of the British Thoracic Society.
Chemotherapy and management of tuberculosis in the United
Kingdom: recommendations 1998. Thorax 1998; 53: 536–548.
14 Tahao ğlu K, Kızkın Ö, Karagöz T, Tor M, Partal M, Şado ğlu T.
High initial and acquired drug resistance in pulmonary tuber-
culosis in Turkey. Tubercle Lung Dis 1994; 75: 324–328.
15 Tahaoğlu K, Kızkın Ö, Hatipoğlu T, Güngör G, Partal M.
Alarming drug resistance rates for tuberculosis in Sürreyapaşa
Hospital, Turkey. Int J Antimicrobial Agents 1996; 7: 261–263.
16 Durant F, Bernuau J, Pessayre D, et al. Deleterious influence of
pyrazinamide on the outcome of patients with fulminant or
subfulminant liver failure during antituberculous treatment in-
cluding isoniazid. Hepatology 1995; 21: 929–932.
17 Mitchell I, Wendon J, Fitt S, Williams R. Anti-tuberculous
therapy and acute liver failure. Lancet 1995; 345: 555–556.
18 Danan G, Pessayre D, Larrey D, Benhamau J P. Pyrazinamide
fulminant hepatitis: an old hepatotoxin strikes again. Lancet
1981; 7: 1056–1057.
 The management of anti-tuberculosis drug-induced hepatotoxicity
C A D R E : Salle de tuberculose dans un hôpital tertiaire de
pneumophtisiologie.
O B J E C T I F : Comparer l’efficacité de deux protocoles dif-
férents de reprise de traitement sur la rechute de l’hépa-
totoxicité au cours d’un traitement de tuberculose.
S C H É M A : Etude prospective randomisée. On a observé
une hépatotoxicité après traitement antituberculeux
chez 45 patients avec tuberculose de première atteinte.
Les patients du groupe I (n  20) ont eu comme reprise
de traitement un régime médicamenteux consistant en
isoniazide, rifampicine, éthambutol et streptomycine
administrés sous forme d’augmentation graduelle du
nombre et du dosage des médicaments. Chez les patients
du groupe II (n  25), le régime de reprise de traitement
consistant en isoniazide, rifampicine, pyrazinamide,
MARCO DE REFERENCIA : Servicio de tuberculosis en un
Hospital Universitario de Enfermedades del Tórax.
O B J E T I V O : Comparar la eficacia de dos protocolos dife-
rentes de retratamiento sobre la recurrencia de la hepa-
totoxicidad en el tratamiento de la tuberculosis.
M É T O D O : Estudio prospectivo aleatorio. Existían 405
pacientes con una tuberculosis reciente que presentaron
hepatotoxicidad durante el tratamiento antitubercu-
loso. Los pacientes del grupo I (n  20) fueron retra-
tados con un esquema que contenía isoniacida, rifam-
picina, etambutol y estreptomicina, administrado
mediante el aumento progresivo del número y de las
dosis de las drogas. Los pacientes del grupo II (n  25)
fueron retratados con un esquema que contenía isoni-
acida, rifampicina, pirazinamida y etambutol, adminis-
69
RÉSUMÉ
éthambutol fut administré avec l’ensemble des médica-
ments et le même dosage pendant toute la période de
reprise du traitement.
R É S U L T A T S : On a observé une rechute d’hépatotoxicité
respectivement dans zéro et six (24%) des cas du groupe
I et du groupe II (P  0.021). Parmi les six patients qui
ont eu une rechute d’hépatite, un n’a pas pu être suivi ;
chez les cinq autres, l’on a adopté le protocole de retraite-
ment du groupe I et l’évolution fut favorable. A la fin de
la reprise du traitement, tous les patients étaient guéris.
C O N C L U S I O N : Le taux de rechute de l’hépatotoxicité
après reprise de traitement de la tuberculose est plus élevé
lorsqu’on réintroduit un régime médicamenteux à dose
complète incluant la pyrazinamide que lorsqu’on réintro-
duit de façon graduelle un régime sans pyrazinamide.
RESUMEN
trado con iguales drogas e iguales dosis durante todo el
período de retratamiento.
R E S U L T A D O S : Existió una recurrencia de la hepato-
toxicidad de cero y seis (24%) en los grupos I y II,
respectivamente (P  0,021). De los seis pacientes que
tuvieron una hepatitis recurrente, uno no pudo ser
seguido. Los otros cinco pacientes pasaron al protocolo
del grupo I y todos evolucionaron bien. Al final del
tratamiento, todos curaron.
C O N C L U S I Ó N : Las tasas de recurrencia de la hepato-
toxicidad en el retratamiento de la tuberculosis es más
elevada cuando se introduce un esquema con la dosis
total que incluye pirazinamida que cuando se adminis-
tra un esquema con aumento gradual de las drogas y
sin pirazinamida.