Clinical management of anti-tuberculosis related cutaneous adverse drug reactions based on reintroduction

Anti-tuberculosis (TB) treatment (ATT), including rifampicin (RMP), isoniazid (INH), pyrazinamide (PZA) and ethambutol (EMB) can be complicated by potentially life-threatening adverse drug reactions (ADRs). ¹ The prevalence of cutaneous adverse drug reactions (CADR) during ATT ranges from 0.6% to 7.5%.^1–4 Furthermore, the global migration of populations has resulted in the importation of multidrug-resistant tuberculosis strains, necessitating prompt and appropriate treatment with a risk of adverse effects, particularly CADR.⁵ This study aimed to assess the spectrum of CADR in TB patients undergoing ATT, evaluate the clinical management strategies involving reintroduction, and review the outcomes, providing valuable insights into effective approaches.

We retrospectively evaluated the medical records of all adult patients hospitalized in our department (1 January 2006–31 December 2019) and who presented with CADR during ATT. Patients were identified through the department's database. ATT was administered for both confirmed and suspected cases of TB, and all patients who received ATT were included, even if the diagnosis of TB was later reconsidered. Patients without skin lesions were excluded. TB was microbiologically confirmed or supported by clinical, anatomopathological and radiological evidence, confirmed by treatment response. Two dermatologists (GM, EC) clinically evaluated the CADR. They were classified⁵ as follows: maculopapular exanthema (MPE) with and without severity criteria,⁶ urticaria, toxic epidermal necrolysis (TEN), Stevens-Johnson syndrome (SJS) and drug rash with eosinophilia and systemic symptoms (DRESS), assessed with REGISCAR criteria.⁷ Scoring system for RegiSCAR (Registration of serious cutaneous adverse reaction) DRESS validation is based on the RegiSCAR study group criteria. This system was used to classify suspected cases into definite (score of 6 and above), probable (score of 4 and 5), possible (score of 2 and 3) and no DRESS (score of < 2).

Causality of a CADR with a drug was determined based on the reintroduction of the drug and the French pharmacovigilance method.⁸ All patients received full-dose anti-TB drugs reintroduced successively. Sequential reintroduction was mainly done one by one or two by two with close monitoring among hospitalized patients. Drug causality assessment was based on either the recurrence of cutaneous symptoms after reintroduction of the culprit drug or the absence of cutaneous symptoms after the first two to three drugs were reintroduced. After identification, the culprit drug was switched to another anti-TB drug to resume effective treatment rapidly.

Of the 1041 patients who received ATT, CADR was diagnosed in 13 patients (1.25%). Of these 13 patients, 11 were female, and the mean age was 54.6 years (±18). Seven were Africans and six were Europeans (Table 1). Four had a so old history of tuberculosis that there was no information on past treatment, despite all of them were born in France.

A total of 10 patients were diagnosed with TB, including lung involvement in two (n° 8, 12), lymph node involvement in one (n° 9), uveitis in one (n° 5) and disseminated forms in six (n° 2,3,4,6,7,13). Three patients received presumptive treatment but were not ultimately diagnosed with TB (n° 1,10,11). Microbiological confirmation was achieved in seven patients, and clinical confirmation in three. Two patients had drug-resistant TB (low-level resistance to INH and MDR-TB). In total, 11 patients received standard ATT with INH, RMP, EMB and PZA. In one patient INH was changed for moxifloxacin (MXF) due to low-level resistance (n° 8), and the patient with MDR-TB received a combination of antibiotic drugs (n°7). Among the 13 ATT-related CADR, MPE was observed in 10 patients, urticaria in one and DRESS in two (attributed to EMB and RMP after reintroduction). All patients, except n° 11, were treated with topical steroids. Mean time to reintroduction of ATT was 4 days (±1.9). Reintroduction was incomplete (i.e. reintroduction of all the initial AT drugs was not achieved) in three patients. In two patients (n° 10,11) the diagnosis of TB was reconsidered before the end of reintroduction, whereas one had multidrug-resistant TB (n° 7). Causality assessment identified a drug in 10 patients but was uncertain in two patients (n° 6,13) whereas no drug was identified in one patient (n° 12).

RMP was involved in five cases, PZA in 4 cases and EMB in one case. Switching to MXF was done in nine cases. Out of 10 TB patients, nine were considered cured after 24 months, while one patient died from cardiac failure after 4 months of ATT due to a pre-existing heart condition.

This study confirms the low prevalence of ATT-related CADR, predominantly MPE. Our findings support the reintroduction of full-dose anti-TB drugs with close monitoring, as it helps prevent the emergence of drug resistance. Two cases of DRESS were identified, including one attributed to EMB, a rare but documented CADR.⁹ No other potentially life-threatening CADR were observed. ATT involves simultaneous administration of three to four anti-TB drugs, making it challenging to pinpoint the specific culprit drug in CADR cases. While skin tests may be useful, resuming ATT often cannot wait for their results. Our experience encourages us to say that sequential reintroduction of anti-TB drugs enabled rapid identification of the culprit drug, its rapid replacement with safe continuation of ATT limiting the risk of drug resistance.

RMP followed by PZA were the most frequently incriminated drugs, which is consistent with previous studies^1,3 no incentive to combine them when reintroduced 2 by 2. Based on our findings, we recommend reintroducing ATT drugs sequentially, starting with INH, followed by EMB, PZA and finally RMP if reintroduced individually.

Reintroduction of the culprit drug led to CADR recurrence in most cases, except one, with no additional adverse effects observed, even in cases of DRESS. This is challenging, as reintroduction of culprit drug after DRESS is generally contraindicated, but because of the absence of multiple therapeutic alternatives and the need to quickly resume effective treatment to avoid resistance, reintroduction under supervision may be justified.⁹ In our study, ATT was stopped at the onset of DRESS, and patients were closely monitored in the hospital during the successive reintroductions. The culprit drug was discontinued when DRESS recurred. Our study had limitations, including the inclusion of all patients on ATT, even those without a confirmed TB diagnosis. Two out of three patients with other diseases experienced severe CADR. Therefore, rechallenge should only be considered for patients with highly probable or confirmed TB. The retrospective design and monocentric recruitment limited our study's generalizability. The incidence of CADR may be underestimated, but we deemed unreported cases as likely benign and not requiring treatment changes.

In conclusion, we suggest that PZA and RMP should be reintroduced last in case of reintroduction of anti-TB drugs one-by-one, and PZA and RMP should not be combined in case of reintroduction two-by-two. The two modes of reintroduction could be compared.

Acknowledgments

The authors would like to thank Oula Itani, MD, and Victor Fabre, for their help with English editing. Thanks to Benedicte Lebrun-Vignes, MD and Alexandra Aubry, MD PhD, for their adds regarding pharmacovigilance and microbiological data, respectively.

Author contributions

Substantial contributions to the conception or design of the work: Quentin Beytout, Nagisa Godefroy, Eric Caumes—Acquisition, analysis or interpretation of data for the work: Quentin Beytout, Nagisa Godefroy, Gentiane Monsel, Stéphane Jaureguiberry, Eric Caumes—Drafting the work or revising it critically for important intellectual content: Quentin Beytout, Eric Caumes—Agreement to be accountable for all aspects of the work: Quentin Beytout, Nagisa Godefroy, Gentiane Monsel, Stéphane Jaureguiberry, Eric Caumes.

Funding

None.

Conflict of interest

None.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References