Hypersensitivity reactions with first-line antituberculosis drugs and outcomes of rapid desensitizations

Background Data about drug hypersensitivity reactions with first-line antituberculosis drugs and their management is limited. Rapid drug desensitization seems to be an appropriate management. Objective Evaluate the efficacy of the rapid desensitization protocols in patients who had a reaction phenotypically compatible with immediate-type drug hypersensitivity with first-line antituberculosis drugs and identify possible risk factors of breakthrough reactions during the protocols. Methods This is a retrospective study of active tuberculosis patients who had a reaction phenotypically compatible with immediate-type drug hypersensitivity with first-line antituberculosis drugs and underwent desensitization with the drugs used during the reaction. Characteristics of drug hypersensitivity and breakthrough reactions, and outcomes of rapid desensitizations were recorded. Results One hundred and seventy-nine patients were included in the study. Most of the initial reactions (n = 132, 73.7%) occurred within the first week of treatment and were mild (n = 146, 81.6%). A total of 690 desensitizations were performed. Desensitizations were successfully completed without any breakthrough reaction in 103 (57.5%) patients and in 29 of 36 (80.6%) patients after a breakthrough reaction. The overall success of desensitizations were found to be 95% (132 of 139 patients). Most of the breakthrough reactions (84%) were mild. Sixteen patients had breakthrough reactions with multiple drugs. Although pyrazinamide was the most common culprit of breakthrough reactions and had the lowest desensitization success, it had the highest rate of a single breakthrough reaction (p < 0.001). Timing of the initial reaction and concomitant breakthrough reaction with ethambutol were found to have increased the risk for breakthrough reaction caused by rifampicin (p = 0.017 and p = 0.010 respectively). Conclusion The rapid desensitization protocols used in this study provide a successful and effective management of the patients with a reaction phenotypically compatible with immediate-type drug hypersensitivity with first-line antituberculosis drugs.

Tuberculosis is a serious public health threat unless it is treated properly.Until the SARS-CoV-2 (COVID-19) pandemic, it was the leading cause of death from a single infectious agent, ranking above HIV/AIDS. 1The main principle of antituberculosis treatment is to use the safest and most effective drug combination regularly for an adequate time.Treatment of drug-susceptible tuberculosis disease requires combination therapy with first-line antituberculosis drugs, including isoniazid (H), rifampicin (R), pyrazinamide (Z), and ethambutol (E).All 4 drugs are used in the initiation phase of treatment for 2 months, and then H and R are used in the maintenance phase of treatment for 4 months. 20][11] However, in the management of antituberculosis DHR, even if the culprit first-line antituberculosis drug is detected, it is not easy to remove it from therapy because there is a limited number of alternative drugs, and second-line antituberculosis drugs are less effective, more expensive, and have more side effects than first-line antituberculosis drugs, and also switching to second-line therapy prolongs the duration of therapy. 1,124][15][16][17][18] Detection of the culprit drugs with drug provocation testing will take a long time in such cases.Rapid and appropriate management of patients with antituberculosis drug hypersensitivity is important because prolonged hospitalization and failure of using regular or standard drug combinations are identified as risk factors for drug resistance. 19,20In the management of antituberculosis DHR there are some suggestions about desensitization with all drugs used during hypersensitivity reaction rather than detecting the culprit drug. 21,22 previously defined 6-8 step readministration protocols in patients who had immediate-type DHR to first-line antituberculosis drugs within 1 h after drug intake.Eleven of the 13 patients tolerated the drugs without any breakthrough reaction (BTR).In the remaining 2 patients who had BTRs during protocols, drug tolerance was achieved by modifying the protocol. 23e aim of this study is to evaluate the efficacy and safety of the previously defined protocols in a larger patient population who were phenotypically compatible with immediate type DHR with first-line antituberculosis drugs regardless of having a reaction within 1 h and also identify possible risk factors of BTRs during the protocols.

Study population
This is a retrospective observational study of active tuberculosis patients who were phenotypically compatible with immediate-type DHR with first-line antituberculosis drugs and underwent desensitization with the drugs used during the DHR from January 2010 to December 2022 in our Allergy and Clinical Immunology Clinics.The study was approved by the local ethics committee (approval number 2012-KAEK-15/2614).

Inclusion
criterias of patients were 1 phenotypically compatible with immediate-type DHR with first-line antituberculosis drugs and 2 chronologically developed a reaction within 12 h after the last dose and 3 desensitized to first-line antituberculosis drug used during DHR.Exclusion criterias of patients were 1 phenotypically compatible with delayed-type DHR and 2 insufficient medical records.
In this study, phenotypic immediate-type DHR is defined as a reaction with signs and symptoms such as cutaneous symptoms (pruritus, flushing, urticaria, angioedema), rhinitis, conjunctivitis, respiratory symptoms (dyspnea, bronchospasm), gastrointestinal symptoms (abdominal pain, nausea, vomiting, diarrhea), decrease in blood pressure or anaphylaxis irrespective of chronology.Patients who had a reaction with these signs and symptoms for up to 12 h were included in the study.
Phenotypic delayed-type DHR is defined as a reaction with signs and symptoms such as maculopapular eruption, fixed drug eruption, drug reaction with eosinophilia and systemic symptoms, acute generalized exanthematous pustulosis, symmetrical drug-related intertriginous and flexural exanthemas, Stevens-Johnson syndrome, toxic epidermal necrolysis, serum sickness-like reactions, organ specific delayed drug reactions or vasculitis which were exclusion criteria of the study.

Patients, disease, DHR, and BTR characteristics
Baseline data including patients' demographics (age, gender, other drug allergies), disease characteristics (site of infection, diagnostic method, case definition), DHR characteristics (culprit drugs, interval between antituberculosis treatment initiation and initial DHR, interval between last administered drug dose and DHR, the symptoms and severity of DHR, any treatment interruption before DHR), BTR characteristics (culprit drug, interval between the last applied step and BTR, the symptoms and severity of BTR), and also continuation to the protocols after BTR and protocol success were recorded as documented in hospital medical records.
The site of tuberculosis infection was classified as pulmonary only, extrapulmonary only, and combined pulmonary and extrapulmonary.The diagnostic method was classified as bacteriologically confirmed if a biological specimen is positive by smear microscopy, culture, or real-time polymerase chain reaction; histologically diagnosed if the tissue biopsy sample is suggestive of caseification; clinically diagnosed if the clinical and radiological findings are suggestive of tuberculosis infection.Case definition was classified as new cases who have never been treated for tuberculosis or have taken antituberculosis drugs for less than 1 month; previously treated cases who have received 1 month or more of antituberculosis drugs.

Desensitization protocols
Desensitization was performed with the protocols we described in our previous study. 23rotocols were started after all symptoms of the DHR resolved and an appropriate washout period of drugs used in the treatment of DHR.No premedication was used before the procedure.Written informed consent was obtained from patients before each desensitization procedure.
The drugs used during DHR were administered using 6-8 step desensitization protocols on consecutive days, as described in Table 1.The regimens were initiated with 7.5, 10, 10, and 5 mg of isoniazid, rifampicin, pyrazinamide, and ethambutol respectively.Drug doses were given in each regimen at a maximum 2.5-fold increase every 30 min.Only the interval between the next to last dose and the last dose was 60 min.All regimens were implemented to ensure tolerance to the therapeutic dosage of each drug before advancing to the next drug on the consecutive day.Each day, before starting a new drug application, the therapeutic dose for the drug or drugs that were administered on previous days was given to the patient.Then, the regimen for the application of the next drug was launched.The procedure was performed under close observation with one-on-one nurse-to-patient care.In case of a BTR, drug administration was suspended, and the reaction was treated according to international guidelines. 9,25After complete resolution of the symptoms, the protocol was continued starting from the last tolerated dose.The effectiveness and safety data of desensitization protocol were recorded.

Statistical analysis
All statistical analyses were performed using the SPSS (statistical package of social sciences) for Windows 16.0 software package.In the evaluation of the data, mean and standard deviation was used for normally distributed data, and median and interquartile range was used for data that did not show a normal distribution.Values and percentages for ratios were determined by descriptive statistical method.In univariate analyses, Chisquare, Fischer, Student's t-test, and Mann-Whitney U tests were used, as appropriate.Multivariate analysis was conducted through a binary logistic regression model and the variables were selected by backward selection with the elimination of variables at a p-value over 0.20.Results was evaluated at a 95% confidence interval.All p values < 0.05 were considered to be statistically significant.
Considering the 26 patients who had a DHR chronologically between 6 and 12 h; 16 patients had urticaria, and 10 patients had pruritus with erythema.Urticaria was accompanied by angioedema in 4 patients, dyspnea in 1 patient, vomiting in 1 patient, and anaphylaxis with hypotension in 1 patient.DHR was on the first day of therapy in 6 (23.1%) and within the first week in 12 (46.2%) of these 26 patients.
Detailed demographics, disease, treatment, and DHR characteristics of patients with and without a BTR are shown in Table 2.No statisticaly significant difference was found in BTR-positive and BTRnegative patients in terms of these variables.
Four patients had known culprit drug before, and desensitization was performed with these drugs, 3 of 4 (75%) patients completed the protocol without any BTR.In 175 (97.8%) patients for whom the culprit drug was unknown, desensitization was performed with all first-line antituberculosis drugs used during the DHR, and 100 of 175 (57.1%) patients completed the protocol without any BTR.There was no significant difference in BTR development in patients with (n ¼ 4) and without (n ¼ 175) known culprit drug before (p ¼ 0.638).
Sixteen patients had BTRs with multiple drugs; 1 patient with all 4 drugs and 15 patients with 2 drugs out of which 5 with RZ, 4 with EZ, and each 2 with HZ, RE, and HR.There was no significant difference in terms of demographics, disease, and DHR characteristics between patients who had BTRs with multiple drugs or not (p > 0.05).The  frequency of a BTR with a single drug was significantly higher in Z than in other drugs (p < 0.001) (Table 4).
The statistically significant variables in univariate analysis of demographics, disease, and DHR characteristics between patients with and without BTRs according to each drug are shown in Table 5. Patients who had BTRs with R were further analyzed to detemine the independent risk factors of BTRs, and a multivariate regression analysis was performed in the model established with treatment interruption before initial DHR, the interval between treatment initiation and DHR (days), whether standard HRZE combination was used during the initial DHR, concomitant BTR with H or E. Interval between treatment initiation and DHR and concomitant BTR with E were found to have increased the risk for BTR caused by R (p ¼ 0.017 OR ¼ 1.055, 95%CI ¼ 1010-1102 and p ¼ 0.010 OR ¼ 8036 95%CI ¼ 1640-39 373 respectively) (Table 6).
In 40 of 76 patients who had BTRs, the desensitization procedure was not continued due to the primary physician's choice of alternative therapy.Twenty-seven of these patients had BTRs more than 2 h after completion of the procedure.Desensitization was successfully completed after 33 of 42 BTRs in 29 of 36 (80.6%) patients who continued the procedure after a BTR.The desensitization success after a BTR for H, R, Z, and E was 80.0% (4 of 5), 91.7% (11of 12), 71.4% (15 of 21), and 75.0%(3 of 4) respectively.The desensitization procedure was successfully completed in 132 of 139 (95%) patients; 103 patients with no BTR, and 29 of 36 patients who successfully completed the procedure after a BTR.

DISCUSSION
In this study, we retrospectively reported the characteristics of DHRs, the outcomes of rapid desensitizations, and the characteristics and risk factors for BTRs with first-line antituberculosis  To the best of our knowledge, this study has the largest patient population desensitized to antituberculosis drugs in the literature.
There was female dominance in the study group with a female-to-male ratio of 2.3:1.We previously reported that female gender is a risk factor for immediate-type DHR with first-line antituberculosis drugs. 23Some other studies reported that the female gender was a risk of antituberculosis DHR. 3,4,26mediate-type DHRs with antituberculosis drugs were reported to be developed in the early treatment period within days to a few weeks. 15,26,27Compatible with previous reports, in this study initial DHR was developed within the first week in 132 (73.7%) cases; we also realized that it was on the first day of therapy in 86 (48.04%) cases.A previous study also reported that 14.7% of the antituberculosis DHRs were developed on the first day of therapy. 17These results suggest that some of the immediate reactions develop without requiring a sensitization period and may be non-IgE mediated.][30][31] We did not perform skin tests in this study because our aim was not to identify the culprit drug but to provide rapid and effective management of patients and also, as a parenteral drug required for performing intradermal tests only the parenteral form of R is available in our country.Further studies are required to elucidate the mechanisms of immediate-type DHRs with antituberculosis drugs.
Presentation of DHRs may vary.They can be classified based on chronology, mechanism, and clinical phenotypes. 10Chronologically, reactions can be classified as immedite and delayed (nonimmediate) reactions. 10Immediate reactions typically occur within 1 h but, in some cases up to 6 h after the last drug administration and phenotypically present with cutaneous symptoms (eg, flushing, pruritus, urticaria, angioedema), rhinitis, conjunctivitis, respiratory symptoms (dyspnea, bronchospasm), gastrointestinal symptoms (nausea, vomiting, diarrhea, abdominal pain), or anaphylaxis. 10,11,32Delayed (nonimmediate) DHRs may occur at any time from 1 h after the initial drug administration, commonly after many days or, in some cases, weeks of treatment. 10,11,32The most common clinical presentations are maculopapular exanthemas and delayed urticaria. 11Although classification based on chronology and clinical phenotypes is important in clinical practice for workup planning, it may not always reflect the underlying mechanism.Several factors influence the elicitation period of reactions, including unreliable history, the administration route of drug, the role of drug metabolites, and the presence of co-factors or co-prescribed drugs may accelerate or slow down the onset or progression of a reaction.In our previous study, we demonstrated the effectiveness of our protocols in patients who had immediate type DHR to first-line antituberculosis drugs within 1 h after drug intake.One of our aims in this study was to evaluate the efficacy of this protocols in a larger group of patients phenotypically compatible with immediate-type DHR regardless of chronology.Maculopapular eruption and other forms of delayed-type DHRs were excluded from the study.In this study 26 (14.5%) of the patients had DHR chronologically 6-12 h after last drug intake; 16 of these patients had urticaria, and 10 had pruritus with erythema.There was no difference in terms of BTR development in patients who had a reaction before or after 6 h chronologically (p ¼ 0.156).
There are a limited number of studies focusing on immediate-type DHRs with antituberculosis drugs.In most studies, antituberculosis DHRs were reported as cutaneous adverse drug reactions and most of these are associated with nonimmediatetype DHRs.5,35,36 Considering the studies in which the majority of the study group consisted of patients compatible with immediatetype DHR, Smadhi et al reported Z as the most common culprit after reintroduction, and Katran et al reported Z as the most common culprit of BTRs during desensitizations. 17,29However in a study reporting adverse drug reactions to firstline antituberculosis drugs in 17 843 cases, R was the most common culprit of common adverse reactions and also DHRs such as rash, pruritus, urticaria, fever, eosinophilia, and dyspnea. 27In our study Z was found to be the most common culprit of BTRs with a BTR development rate of 32.4% and had the lowest desensitization success after a BTR (71.4%).However BTRs caused by Z were significantly seen as single drug BTR (p < 0.001).
Tuberculosis therapy requires multiple drug administration concomitantly, and hypersensitivity to multiple drugs may develop simultaneously.15][16][17][18]29 In our study, BTRs with multiple drugs were detected in 16 of 76 (21.1%) patients, and one patient had BTRs with all first-line antituberculosis drugs.Smadhi et al. reported multiple DHR in 22% of patients after reintroduction, and two of these patients developed DHR with all first-line antituberculosis drugs. 17ta about antituberculosis DHR and its management is limited.Although drug provocation testing is the gold standard method to detect and remove the culprit drug from therapy, it carries a risk of reintroduction reactions.Lehloenya et al investigated antituberculosis drug reintroduction in patients who developed cutaneous reactions.Half the patients had reintroduction reactions, and 52% of reactions were moderate to severe.The rate of DHR with antituberculosis drugs may actually be higher in their study because 39 patients were on cotrimoxazole during the initial DHR but not reintroduced in 29 patients. 35Yee et al rechallenged antituberculosis drugs in 11 of 21 patients with rash or drug fever only 1 successfully. 4 case of a reintroduction reaction, before a new drug provocation, it is necessary to wait for the complete resolution of symptoms and also eliminate the drugs used to treat the reaction.This strategy causes a delay in starting standard therapy, especially in patients with multiple drug hypersensitivity, owing to the time needed for treatment of the repeated reactions.In addition, even if the culprit drug(s) are detected, it is not possible to easily remove them from the treatment.Therefore, the priority in this patient group must be to ensure the reuse of drugs rather than finding the culprit.Aberer and Kranke suggested desensitization rather than detailed allergy testing as the initial method in patients who have cystic fibrosis or tuberculosis in whom the drug concerned is more effective than the alternatives. 21The recommendations of Bermingham et al in the management of type-1 DHR with antituberculosis drugs are as follows: i) desensitization with culprit drug and sequencial rechallenge with other drugs if the culprit is known, ii) desensitization with all drugs if the culprit is not known. 22ere is no standardized instutional rapid desensitization protocol for immediate type DHRs with antituberculosis drugs.8][39][40] To the best of our knowledge, our protocols are currently the only protocols that have been planned to reuse the antituberculosis drugs without determining the offending drug.
With the protocols used in this study, 57.3% (103 of 179) of the patients successfully completed desensitization without any BTRs.In 36 patients who had BTRs, desensitization was continued after appropriate treatment of the reaction, and 29 of them (80.6%)successfully completed desensitization.The overall success rate of the desensitizations were found to be 95% (132 of 139).The success of our protocols has also been verified with a study conducted in a different center.Katran et al reported that HRZE was started without any problems in 49 of 81 (60.5%) patients with mild immediate reaction. 29In addition to be successful, these protocols are also advantageous because of their short implementation time.
Desensitization procedures that start from lower doses, involve more steps, or have shorter intervals between doses may improve safety, but these adaptations will also prolong the implementation time. 41Short implementation time is important in these groups of patients.Tuberculosis is a disease with a risk of transmission, so the desensitization procedure of these patients is carried out in an isolated room during daytime working hours in our allergy clinic.At the end of the procedure, the patients are monitored in our clinic for 2 h in case of a possible reaction, and then they are hospitalized in the isolated tuberculosis patient service in our hospital.
In this study, risk factors for BTR development according to each drug were also evaluated, and interval between treatment initiation and DHR and concomitant BTR with E were found to have increased the risk for BTR caused by R. To the best of our knowledge, our study is the only study in the literature that indicates the risk factors for BTR seen during antituberculosis drug desensitization.Further studies are needed to understand the contribution of these statistically significant data to clinical practice.However, there are several limitations of this study.The first limitation is the retrospective design of the study.The second limitation is in some patients who had BTR during the procedure the primary physician decided to replace the offending drug with an alternative and the protocol was not continued after a BTR.We could not evaluate the overall success of the protocols in these patients.

CONCLUSION
In conclusion, data about immediate type DHR due to antituberculosis drugs and their management are limited.Multiple drug hypersensitivity is substantial and one of the major problems in the management of patients.Due to limitations of alternative therapies, it is important to ensure the reuse of first-line drugs.Desensitization with all drugs used during DHR is an important option for the rapid and effective management of patients.We have demonstrated that our rapid desensitization protocols are effective and safe with the advantage of having short implementation time in patients phenotypically compatible with immediate type DHR with first-line antituberculosis drugs regardless of chronologically and may guide rapid and effective management of the patients.In patients for whom our rapid desensitization protocols failed, using different desensitization protocols that start from lower doses, involve more steps, or have shorter intervals between doses may be considered.Another option is switching to secondline therapy.

Table 1 .
Rapid desensitization protocols used in the study.a Solution A: Suspension was prepared by crushing one tablet of isoniazid (300 mg) into a volume of 40 ml 0.9% NaCl to a final concentration of 7.5 mg/ml.b Solution B: Suspension was prepared by mixing the content of one capsule of rifampicin (600 mg) into a volume of 30 ml 0.9% NaCl to a final concentration of 20 mg/ml.c Solution C: Suspension was made by 1:2 dilution of solution B to a final concentration of 10 mg/ml.d Solution D: Suspension was prepared by crushing one tablet of pyrazinamide (500 mg) into a volume of 10 ml 0.9% NaCl to a final concentration of 50 mg/ml.e Solution E: Suspension was made by 1:10 dilution of solution D to a final concentration of 5 mg/ml.f Solution F: Suspension was prepared by crushing one tablet of ethambutol (500 mg) into a volume of 10 ml 0.9% NaCl to a final concentration of 50 mg/ml.g Solution G: Suspension was made by 1:10 dilution of solution F to a final concentration of 5 mg/ml

Table 2
. (Continued) Demographics, disease and DHR characteristics of patients.Abbreviations: BTR, breakthrough reaction; DHR, drug hypersensitivity reaction; E, ethambutol; H, Isoniazid; R, Rifampicine; Z, pyrazinamide.a percentages shown in the table are column based.b p-values reflect comparisons between: BTR positive vs BTR negative patient groups.c Independent samples t-test.d Pearson chi-squared test.e Fisher exact test

Table 3 .
BTR characteristics of each drug a .
Abbreviations: BTR, breakthrough reaction, ND, not defined.a Percentages shown in the table are column

Table 4 .
Single and multiple drug BTR distrubitions of each drug.Abbreviations: BTR, breakthrough reaction (continued)

Table 5 .
(Continued) Statistically significant characteristics in patients with and without BTR during desensitization according to each drug.
Abbreviations: BTR, breakthrough reaction; DHR, drug hypersensitivity reaction.a Percentages shown in the table are column based.b p-values reflect comparisons between: BTR positive vs BTR negative patient groups.c Fischer exact test.d Mann whitney u test.e Pearson chi-squared test drugs in patients with active tuberculosis infection.