Documente Academic
Documente Profesional
Documente Cultură
Rannakoe J
Lehloenya1,2 and
Keertan Dheda*2,3,4
1
Division of Dermatology, Department
of Medicine, University of Cape Town,
Western Cape, South Africa
2
Lung Infection and Immunity Unit,
Division of Pulmonology, Department
of Medicine & UCT Lung Institute,
University of Cape Town,
Western Cape, South Africa
3
Institute of Infectious Diseases and
Molecular Medicine, University of
Cape Town, Western Cape,
South Africa
4
Department of Infection,
UCL Medical School, London, UK
*Author for correspondence:
Tel.: +27 21 406 7650
Fax: +27 21 406 7651
keertan.dheda@uct.ac.za
www.expert-reviews.com
First- and second-line anti-tuberculosis drugs are associated with a diverse presentation of
cutaneous adverse drug reactions (CADR), ranging from mild to life threatening. An individual
drug can cause multiple types of CADR, and a specific type of CADR can be due to any
anti-tuberculosis drug, which can make the management of tuberculosis (TB) following CADR
challenging. The higher incidence of TB and CADR in HIV-infected persons makes TB-associated
CADR a burgeoning problem for clinicians, particularly in high HIV-prevalence settings.
This review discusses the pathogenesis, epidemiology, clinical presentation, diagnosis and
management of TB-associated CADR. Clinical controversies including its impact on treatment
outcomes, challenges in restarting optimal anti-tuberculosis therapy and the timing of highly
active antiretroviral therapy initiation in those with HIV coinfection are also discussed. Finally,
gaps in the current knowledge of TB-associated CADR have been identified and a research
agenda has been proposed.
KEYWORDS: adverse reactions management skin therapy tuberculosis
Anti-tuberculosis drugs are associated with significant adverse drug reactions (ADRs), which
can make the treatment of tuberculosis (TB) a
challenge. ADRs can broadly be divided into
two types. Type A reactions are predictable,
dose-dependent and are the most common. Type
B ADRs are unpredictable, dose-independent
and comprise 1520% of all ADRs. These
include immunologically mediated drug hypersensitivity or nonimmune-mediated idiosyncratic reactions [1] . Common ADRs associated
with anti-tuberculosis drugs include hepatitis,
cutaneous adverse drug reactions ( CADR),
nausea/vomiting, influenza-like illness and
arthralgia [2] . In this review, CADR due to
anti-tuberculosis drugs, which are usually type
B reactions, will be focused on.
CADR can either be confined to only the skin
or be part of a multisystem disorder and may
complicate anti-tuberculosis therapy with firstand many second-line agents with a wide variety
of clinical presentations. These include morbiliform eruptions, StevensJohnson syndrome
(SJS), drug hypersensitivity syndrome (DHS),
cutaneous vasculitis, lichenoid drug eruption
10.1586/ERI.12.13
ISSN 1478-7210
475
Review
Morbiliform (measles-like) drug eruption or maculopapular exanthems are the most common presentation of a CADR, accounting
for 95% of all cases [40] . The initial presentation, 714 days after
initial exposure to the offending drug, is erythematous macules
and papules starting centrally and spreading peripherally. In
severe cases, the lesions become confluent leading to erythroderma. In a great majority of cases, morbiliform drug eruptions
are self-limiting and treatment can continue uninterrupted [41] .
However, maculopapular exanthem can be the initial presentation
of more serious reactions such as SJS and DHS [23] . Worsening of
the rash, accompanied by systemic symptoms or mucositis, are
usually early indicators of severe disease and warrant stopping
treatment.
Drug hypersensitivity syndrome
Review
Grade 1: mild
Grade 2: moderate
Grade 3: severe
Grade 4: life
threatening
Grade 5:
death
Allergic reaction
Transient flushing
or rash;
drug fever <38C;
intervention not
indicated
Intervention indicated;
responds promptly to
symptomatic treatment;
prophylactic medication
indicated for 24 h
Prolonged; recurrence of
symptoms following initial
improvement; hospitalization
indicated for clinical sequelae
Life-threatening
consequences;
urgent intervention
indicated
Death
WCC low
<2.01.0 109 /l
<1.0 109 /l
Neutrophil count
decreased
<1.00.5 109 /l
<0.5 109 /l
Platelets
<7550 109 /l
<5025 109 /l
<25 109 /l
Fever
38.039.0C
>39.040.0C
>40C for 24 h
Death
Pruritus
Intense or widespread;
Mild or localized;
topical intervention intermittent; skin changes
from scratching; oral
indicated
intervention indicated;
limiting instrumental ADL
Intense or widespread;
constant; limiting self-care ADL
or sleep; oral corticosteroid or
immunosuppressive therapy
indicated
Rash
maculo-papular
Macules/papules
covering <10%
BSA with or
without symptoms
Macules/papules covering
1030% BSA with or
without symptoms;
limiting instrumental ADL
Macules/papules covering
>30% BSA with or without
symptoms; limiting self-care
ADL
Face edema
Localized facial
edema
Pain of skin
Mild pain
Severe pain;
limiting self-care ADL
ALT increased
>5.020.0 ULN;
>5 ULN for >2 weeks
>20 ULN
Death
AST increased
>5.020.0 ULN;
>5 ULN for >2 weeks
>20 ULN
Death
Creatinine
>1.53.0 ULN;
>1.53.0 baseline
>3.0 baseline;
3.06.0 ULN
>6.0 ULN
Death
ADL: Activities of daily living; ALT: Alanine transaminase; AST: Aspartate aminotransferase; BSA: Body surface area; LLN: Lower limit of normal; ULN: Upper limit of
normal; WCC: White-cell count.
Reproduced with permission from [38].
Lichenoid drug eruptions initially present as itchy small pink macules that gradually progress to become firm violaceous, flat-topped,
polygonal and scaly papules. In some cases, the lesions persist as
macules, only increasing in size with continuing exposure to the
offending drug. The mucous membranes, particularly buccal and
genital mucosae, are favored sites with the characteristic white lace
477
Review
Figure 1. Anti-tuberculosis therapy-associated cutaneous adverse drug reactions. (A) Toxic epidermal necrolysis showing blisters
and epidermal stripping; (B) hyperpigmented round multiple lesions of fixed drug eruption; (C) purplish scaly plaques of lichenoid drug
eruption; (D) palpable purpura and blisters of drug-induced vasculitis; and (E) erythematous morbiliform lesions of drug hypersensitivity
syndrome.
frequently on the lower limbs. Depending on severity of the reaction, the purpura can progress to become hemorrhagic blisters
and ulcerate. As with all immune-complex mediated vasculitides,
internal organ involvement, particularly involvement of the kidney, has to be excluded. Antituberculous therapy-associated
vasculitis is rare, but has been reported with rifampicin and
pyrazinamide [6062] .
There are other forms of CADR that are very rarely associated with anti-tuberculosis therapy, including urticaria and
photodermatitis [8,63,64] .
Cutaneous vasculitis
The diagnosis of drug-induced cutaneous vasculitis can be suspected clinically and its hallmark is palpable purpura, most
478
The pathogenesis of hypersensitivity reactions to all drugs, including anti-tuberculosis drugs, is not fully understood. However, these
reactions are known to involve many distinct immune mechanisms
and there are numerous hypotheses that have been proposed. The
Expert Rev. Anti Infect. Ther. 10(4), (2012)
Review
Table 2. Summary and clinical features of different types of tuberculosis-associated cutaneous adverse
drug reactions.
Type of CADR
Clinical features
Morbiliform drug
eruption
Starts 714 days after the drug exposure as erythematous macules and papules. Rarely life threatening but
may signal a more severe reaction
Drug hypersensitivity
syndrome
Onset is approximately 3 weeks after drug exposure presenting as fever, edema, and a progressive
erythematous macules and papules often associated with eosinophilia and transaminitis
StevensJohnson
syndrome and toxic
epidermal necrolysis
At onset often confused with upper respiratory tract infection, which progresses to erythematous macules
and papules, targetoid lesions, epidermal detachment and mucositis. There is usually painful erythema and
blisters of the palms and soles
Initially appears as solitary or numerous round, well-circumscribed, erythematous itchy macules that resolve
with persistent hyperpigmentation recurring in exactly the same sites on re-exposure to the offending drug
Presents as itchy pink macules and purple flat-topped, polygonal and scaly papules. In mucous membranes,
there is the characteristic white lacy pattern
Cutaneous vasculitis
Usually found on the lower limbs where it presents as palpable purpura, hemorrhagic blisters and ulcers. Can
be associated with systemic disease, particularly involving the kidney
A drug provocation test (DPT), defined as a controlled administration of a drug in order to diagnose drug hypersensitivity
www.expert-reviews.com
479
Review
both approaches increase the risk of inducing additional and possibly a fatal CADR. Thus, the risk of reintroduction reactions
must be balanced against the morbidity and mortality associated
with less effective regimens. There are limited data to guide the
reintroduction (rechallenge) of drugs in TB-associated CADR.
Traditional contraindications to rechallenge include pregnancy,
those with significant comorbidity, and those at increased risk due
to severe or life-threatening reactions such as the bullous reactions, erythroderma, drug hypersensitivity syndrome, anaphylaxis, systemic vasculitis and drug-induced autoimmune disease.
There are obvious limitations to these recommendations, based
on expert opinion, and further studies are required to establish
their validity (BOX 1) . Exclusion of these patients to rechallenge may
place them at unacceptably high risk of tuberculosis-associated
morbidity and mortality. It is also crucial, if possible, to perform
drug sensitivity testing before rechallenge is attempted. This
will avoid exposing the patient to a potentially life-threatening
reaction to a drug that will not benefit them.
How rechallenge should occur is another contentious issue.
Whether rechallenge should occur at full dose or with incrementally increasing doses remain unclear. Sharma et al., in a welldesigned study, showed that full dose rechallenge versus dose
escalation is not associated with more reintroduction reactions
in anti-tuberculosis drug-induced hepatitis [97] .
Furthermore, the sequence in which the drugs are reintroduced
is still contentious. Some authors suggest that the drugs least likely
to cause a reaction should be reintroduced first, whereas others suggest that the most effective drugs, namely rifampicin and isoniazid,
should be reintroduced first to minimize the risks of suboptimal
therapy [52,74,98] . In cases reported in the literature, the sequence of
reintroduction was arbitrarily selected [5] . The mortality and morbidity associated with recurrence of CADR (on rechallenge) compared to that associated with suboptimal anti-tuberculosis therapy
is unknown. To further complicate matters, there is disagreement
on which of the currently used drugs are most frequently implicated
as causing CADR. Some authors suggest that pyrazinamide is the
most frequent offender, whereas others suggest streptomycin [6,22] .
Yee et al. found rifampicin to be the most frequent offender in HIVinfected compared to non-HIV; infected patients (hazard ratio:
2.42.9, 95% CI: 0.319.5, for isoniazid; hazard ratio: 8.0, 95%
CI: 1.443, for rifampicin; and hazard ratio: 2.1, 95% CI: 0.317.1,
for pyrazinamide) [6] . Our own study in Cape Town, in a predominantly HIV-infected population, found rifampicin to cause reintroduction reactions in 57% of 23 patients who had DPT to first-line
TB drugs, supporting the findings of Yee et al. [6,38] .
When and how to rechallenge in the setting of TB-associated
CADR remains to be conclusively determined. In a retrospective study performed in Cape Town, 22 out of 23 reintroduction
reactions occurred within 72 h of reintroducing the offending
drug. Itch, hepatitis and fever were the most frequent reactions in
23 patients occurring in 48, 39 and 35% of the patients, respectively [38] . These findings were made in hospitalized patients with
strict monitoring and the drugs were withdrawn at the earliest sign
of a reintroduction reaction and before worsening of laboratory
and clinical parameters. The severity of the original hepatitis does
www.expert-reviews.com
Review
not seem to affect the likelihood of developing a subsequent reintroduction reaction. Larger prospective studies are now needed
to guide clinicians performing DPT in TB-associated CADR.
Until all these issues are resolved, clinical decision making will
be constrained by lack of evidence-based data.
Clinicians should be aware that reintroduction reactions to
multiple drugs during DPT have been reported for many classes
of drugs, including anti-tuberculosis drugs [43,69,99101] . In a retrospective study in Zimbabwe, 13% of patients reacted when rechallenged with multiple anti-tuberculosis drugs [22] . On the other
hand, in our own study, four out of 23 patients who developed
reintroduction reactions on DPT reacted to two anti-tuberculosis drugs each. The reason for multiple drug-related reactions is
unclear. However, possibilities include: viral infections such as
HIV, EpsteinBarr virus, cytomegalovirus and human herpes
virus 6, where there is a significant immune stimulation facilitating an enhanced response to multiple drugs in susceptible
individuals; crossreactivity of antigenic epitopes with structurally similar drugs (unlikely except with fluoroquinolones and
aminoglycosides) [43] and a true multiple drug hypersensitivity
syndrome. Multiple drug hypersensitivity syndrome is defined
as a well documented, repeated and clearly immune-mediated
reaction to structurally unrelated drugs, possibly due to a deficient
tolerance mechanism against small chemical compounds.
HAART initiation in HIV-TB coinfected patients & CADR
In a recently published retrospective study of patients, with antituberculosis therapy-associated CADR, seen between 2001 and
2009 in Cape Town, only one out of 60 HIV-infected patients
were on HAART when they developed the initial drug reaction
[38] . However in 2011, HAART coverage has increased considerably. The issue of which drugs to implicate if CADR develops in
coinfected patients and the timing and sequence of HAART reintroduction is complex. When CADR develops in HIV-TB coinfected patients on HAART, all drug therapy is usually stopped.
HAART is recommenced after the reintroduction of anti-TB
therapy. Determining the offending drug in this setting is further
complicated by the overlapping adverse effect profiles of HAART,
anti-tuberculosis therapy and drugs used for the prophylaxis or
treatment of other opportunistic infections that are often initiated during this period [102] . It is recommended that antiretroviral
therapy be initiated soon (within 28 weeks) after starting antituberculosis therapy because early initiation of HAART lowers the incidence of TB-associated deaths [103,104,201] . It is also
known that anti-tuberculosis therapy-associated CADR occurs
most frequently in the first 100 days after initiation of therapy [33] .
This period also coincides with restoration of host immunity and
development of TB-associated immune reconstitution syndrome
(IRIS) [105] . Thus, diagnostic uncertainty may arise (infection vs
IRIS vs ADR), particularly at low CD4 counts. In addition, it is
hypothesized that DHS itself may be a form of IRIS [106] .
Conclusion
The incidence of TB-associated CADR may vary because of disease burden, prescribing practices, ethnic differences, comorbidities
481
Review
Trials that are currently underway will lead to more robust data
on anti-tuberculosis-associated CADR and guide development
guidelines on management of these patients. HIV plays an important role in both TB and CADR, and studies have to be done in
both high and low HIV burden settings so that management can
be adapted appropriately. The most promising and exciting developments are likely to come from the use of in vitro tests to establish causality in CADR, particularly those that measure cytokine
secretion, upregulation of activation markers or gene expression. This will help avoid the time consuming and potentially
risky process of drug provocation testing. The increasing use of
Key issues
Cutaneous adverse drug reactions (CADR) may complicate anti-tuberculosis therapy with all first- and several second-line
antituberculous agents with a wide variety of clinical presentations.
HIV infection is associated with an increased risk of tuberculosis-associated CADR.
The incidence of tuberculosis-associated CADR is unknown because of inconsistent reporting in the literature. Harmonization of case
definitions, severity grading, study design and an awareness of population differences are required.
Different types of CADR carry different risks of mortality and morbidity, and it is prudent for clinicians to recognize them and manage
them accordingly.
Interruption of anti-tuberculosis therapy is frequent in CADR and this may be associated with increased mortality, morbidity and
possibly the amplification of drug resistance. The significance of these associations is unclear and more studies are needed.
Initiation of highly active antiretroviral therapy, anti-tuberculosis therapy and prophylaxis for opportunistic infections concurrently
complicates the identification of the offending drugs following TB-associated CADR.
The limited efficacy and high toxicities of second-line anti-tuberculosis drugs often make it necessary to reintroduce first-line agents following
tuberculosis-associated CADR. This carries a risk of recurrence of CADR, which must be balanced against the risk of suboptimal antituberculosis therapy. The optimum balance between these two risks is patient specific and further studies are needed to clarify this issue.
The drug provocation test is considered the gold standard to identify the offending drug. It may be positive to multiple anti-tuberculosis drugs.
There are many unanswered questions including the sequence of reintroduction of anti-tuberculosis drugs, dosing schedules and the
merits of sequential full dose rechallenge versus sequential escalating dose rechallenge. Several appropriately designed studies will be
required to address these questions.
482
References
12
13
10
11
www.expert-reviews.com
14
15
16
17
Park JJ, Choi YD, Lee JB, Kim SJ, Lee SC,
Won YH et al. Psoriasiform drug eruption
induced by anti-tuberculosis medication:
potential role of plasma-cytoid dendritic cells.
Acta Derm. Venereol. 90(3), 305306 (2010).
Chintu C, Luo C, Bhat G, Raviglione M,
DuPont H, Zumla A. Cutaneous
hypersensitivity reactions due to
thiacetazone in the treatment of
tuberculosis in Zambian children infected
with HIV-I. Arch. Dis. Child. 68(5),
665668 (1993).
18
19
20
21
22
23
Review
25
26
27
28
29
30
31
32
33
34
483
Review
35
36
47
58
59
60
61
62
37
38
49
63
50
64
51
65
52
66
67
68
69
70
71
39
40
41
42
43
48
44
45
484
53
54
55
56
57
73
74
75
82
93
83
94
84
95
96
97
98
99
100
101
102
85
86
76
87
88
89
90
80
91
81
77
78
79
www.expert-reviews.com
Review
92
485
Review
103
104
486
105
106
Website
201
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.