Real-world experience of adverse reactions-necessitated rifampicin-sparing treatment for drug-susceptible pulmonary tuberculosis

Rifampicin is an important agent for tuberculosis treatment; however, it is often discontinued because of adverse reactions. The treatment regimen then can be administered as that for rifampicin-resistant tuberculosis, which can be toxic. We retrospectively reviewed 114 patients with drug-susceptible pulmonary tuberculosis who discontinued rifampicin due to adverse reactions during an 18 year period at a tertiary referral center, of which 92 (80.7%) exhibited favorable response. Hepatotoxicity was the leading cause of intolerance. Patients with a favorable response were younger and less likely to have comorbidities. The majority of patients were administered four medications during the intensive phase and three to four during the consolidative phase. For those with a favorable response, the median duration of treatment was 10.2 months and the most common intensive regimen was a combination of isoniazid, ethambutol, pyrazinamide, and fluoroquinolone (25%). The most common consolidation regimen was a combination of isoniazid, ethambutol, and fluoroquinolone (22.8%). Among the patients with a favorable response, two (2.2%) experienced recurrence after a follow-up of 3.4 (interquartile range 1.8–6.8) years. For patients with drug-susceptible pulmonary tuberculosis who do not tolerate rifampicin owing to its toxicity, a shorter regimen may be a useful alternative.

www.nature.com/scientificreports/ be considered based on clinical experience 14 . However, such a regimen lacks solid scientific background; whereas recent studies suggest that a shorter regimen may be sufficient for drug-resistant tuberculosis 15,16 , and it may be advantageous over a longer regimen in terms of patient compliance.
In the real-world practice, rifampicin is usually included in the initial regimen unless otherwise contraindicated. Even if rifampicin is omitted during the treatment course owing to its adverse effects, short administration of this drug may be effective for patients with drug-susceptible tuberculosis. This differs from patients with rifampicin-resistant tuberculosis, in whom rifampicin is ineffective during the entire treatment course. In addition, while patients with rifampicin-resistant tuberculosis often have resistance to other first-line drugs, this is not necessarily the case for patients who cannot tolerate rifampicin. In this regard, we aimed to share the real-world experience of adverse reactions-necessitated rifampicin-sparing treatment among patients with drug-susceptible pulmonary tuberculosis.

Methods
Patient selection and baseline information. From October 1, 2003, to August 31, 2020, patients diagnosed with pulmonary tuberculosis were screened at a tertiary referral center in South Korea based on reported cases to the Korea Disease Control and Prevention Agency. Following the exclusion of MDR tuberculosis, we identified patients with pulmonary tuberculosis who exhibited intolerance to rifampicin due to adverse reactions through a review of electronic medical records. As our study was retrospective in nature, adverse reactions were not pre-defined. However, they were determined by the attending physician and the information was collected via medical records review. Adverse reactions included, but were not limited to, hepatotoxicity, skin reactions, and flu-like fever 17,18 . Patients with tuberculosis resistant to any anti-tuberculosis medications, those treated with rifampicin for > 6 months, those who were transferred out, those who were still under ongoing treatment, and those who had an unclear treatment history were excluded. Drug susceptibility was tested conventionally on Lowenstein-Jensen media using the absolute concentration method (Supplementary Appendix 1), except for pyrazinamide, for which the pyrazinamidase test was performed. We selected patients confirmed to be susceptible for all anti-tuberculosis medications.
Baseline patient data, including baseline demographics, underlying diseases, treatment regimens, and treatment outcomes, were extracted retrospectively from the electronic medical records system. The Institutional Review Board of Seoul National University Bundang Hospital (B-2204-750-106) approved this study and waived the requirement for informed consent due to the use of anonymized data and the study's retrospective nature. This study was conducted in accordance with the principles of the Declaration of Helsinki.
Interpretation of the treatment regimen. Patients were treated in accordance with the local treatment recommendations, 12,19-21 which remained mostly consistent throughout the study period. The treatment regimen could be modified with the liberty of the attending physician depending on the patient's response to treatment. Anti-tuberculosis medications used during the treatment period were screened. The types of medications and their duration of use were inspected, and the duration of each drug was described in detail. Duration of each drug was calculated as actual dosage days prescribed, while the overall treatment duration was calculated by counting the calendar days from the first day of treatment to the last. Furthermore, a representative regimen was chosen, which was characterized as regimens that last for a minimum of 2 months 22 . The patients were predominantly subjected to two distinct treatment regimens. The initial regimen was deemed as the intensive regimen, while the subsequent regimen was considered as the consolidative regimen.
Definition of initial treatment response. As the study was retrospective in nature, there were no preestablished protocols for conducting microbiological follow-ups subsequent to treatment. Therefore, patients were divided into two groups according to their treatment response: favorable and unfavorable. A favorable response was defined as being cured or completing the treatment according to the recent 2021 World Health Organization definition 23 , apart from the possibility of changing treatment regimen due to adverse reactions. An unfavorable response was defined as death during treatment, treatment failure regarding clinical and/or bacteriological response, or treatment abandonment including a loss to follow-up 23,24 . Follow-up after treatment completion. For patients who initially achieved a favorable response, we reviewed the electronic medical records using their unique hospital ID to determine if any recurrence of tuberculosis or death occurred after treatment. within addition to the medical records, sputum culture studies and chest X-ray results were also inspected. The date and primary cause of death were acquired from the death certificates prepared by the attending physicians and were confirmed using the Statistics Korea database with the Korean Standard Classification of Diseases, 7th edition, as of 31 December 2020. For those whose status could not be assessed using the above methods, we contacted them by telephone for their status as of May 10, 2022.

Results
Patient characteristics. During the study period, 6,667 patients were diagnosed with pulmonary tuberculosis and 267 patients were diagnosed with MDR tuberculosis. A total of 158 patients were unable to tolerate rifampicin due to adverse effects, of which 114 were included in the final analysis based on the selection criteria. Among 114 patients, 92 (80.7%) showed a favorable response to treatment. Among the patients in the favorable response group, 79 and 13 achieved cure and treatment completion, respectively. Among those in the unfavorable response group, treatment was abandoned in 12 patients either by the attending physician (n = 9) or the patient (n = 3), and the other 10 died during the treatment. Seven patients died due to infectious diseases possibly related to tuberculosis, while the other three of extrapulmonary malignancy (Fig. 1).
Patients with a favorable response were younger (median 61.0 [IQR 48.6-74.0] vs. 78.0 [IQR 70.0-80.0] years, P = 0.001), and were less likely to have comorbidities, such as diabetes mellitus (16.3% vs. 40.9%, P = 0.024) and chronic kidney disease (1.1% vs. 13.6%, P = 0.026). Cavities (25.0% vs. 59.1%, P = 0.002) and bilateral involvement (44.6% vs. 77.3%, P = 0.006) on chest radiographs were less common in patients with favorable responses. There Figure 1. Flowchart of the patient selection process. a Because some patients cannot expectorate further sputum after treatment, those who achieved clinical improvement and completed treatment according to the attending physician were included in the favorable response group. b Includes patients who cannot continue treatment owing to underlying conditions (n = 9) and those who are lost to follow-up (n = 3). c The duration of follow-up was a median of 3.4 (interquartile range 1.8-6.8) years. Abbreviation: MDR, multidrug-resistant. www.nature.com/scientificreports/ were no significant differences with respect to sex (proportion of females: 40.2% vs. 27.3%, P = 0.379). Based on a multivariable analysis, it was determined that a younger age (adjusted odds ratio 0.94, 95% CI 0.90-0.99), the absence of chronic liver disease (adjusted odds ratio 0.05, 95% CI 0.01-0.60), and the absence of cavities (adjusted odds ratio 0.23, 95% CI 0.07-0.75) were significantly associated with a favorable response. Please refer to Supplementary Table 1 for further details.
Of the 114 patients, rifampicin was reintroduced once in 56 patients and twice in three patients. However, it was eventually excluded from the treatment regimen due to adverse reactions. Hepatotoxicity (59.6%) was the most common reason for discontinuation of rifampicin, followed by skin rash and/or pruritus (54.4%), gastrointestinal discomfort (50.9%), fever (28.9%), abnormal blood cell count (14.9%), headache and/or dizziness (13.2%), general weakness (9.6%), and renal injury (3.5%). Three patients did not use rifampicin in the first place due to expected hepatotoxicity regarding the underlying liver condition. The types of adverse reactions did not differ according to the initial treatment response (P > 0.050) ( Table 1).

Treatment regimen.
Most patients started their treatment with the standard four-drug regimen (93.0%, isoniazid, rifampicin, ethambutol, and pyrazinamide) or three-drug regimen (2.6%, isoniazid, rifampicin, and ethambutol) (Supplementary Table 2). The attending physician subsequently prescribed a suitable intensive phase regimen for the patients, consisting of two to six drugs as outlined in Table 2. Out of the total of 114 patients, 66 were administered four or more medications, whereas 48 were administered three or fewer medications. In terms of achieving an initial favorable response to treatment, no statistically significant difference was observed between patients receiving four or more drugs (80.3%) and those receiving three or less drugs (81.3%) (P = 0.899).
For the 92 patients with a favorable outcome, the median duration of treatment was 10.2 (IQR 8.8-12.6) months. During the intensive treatment phase, a four-drug (53.3%) or three-drug (38.0%) regimen consisting of first-line drugs with or without fluoroquinolone, was the most common. For the consolidation phase, threedrug (44.6%) or four-drug (41.3%) regimens consisting of first-line drugs with fluoroquinolone were the most frequently used ( Table 2). The most common intensive regimen was a combination of isoniazid, ethambutol, pyrazinamide, and fluoroquinolone (25.0%), whereas the most common consolidation regimen was a combination of isoniazid, ethambutol, and fluoroquinolone ( (Fig. 2). Details are presented in Table 3.

Follow-up after favorable response.
For the 92 patients with an initially favorable response, the recent health status (within 6 months) of 47 (51.1%) patients was obtained from their electronic medical records. The remaining 45 (48.9%) patients were contacted by telephone, and 22 (23.9%) responded. Three patients (3.3%) had no follow-up visits or telephone responses. The median duration of follow-up, including the telephone survey, was 3.4 (IQR 1.8-6.8) years after treatment completion (Fig. 3). Two patients (2.2%) experienced recurrence 8.5 and 65.5 months after treatment completion, respectively. One patient experienced recurrence of drug-susceptible tuberculosis, while another one experienced tuberculous empyema without known resistance patterns. The details of the two patients are presented in Supplementary Table 5.

Discussion
In this real-world report, we evaluated the treatment regimen, initial treatment response, and follow-up outcomes in patients with drug-susceptible pulmonary tuberculosis who had to omit rifampicin because of adverse reactions. The proportion of patients with a favorable response was 80.7%, most of whom were younger with fewer comorbidities and had mild extent of radiographic involvement compared with those with an unfavorable response. In patients with a favorable response, the median duration of treatment was 10.2 months, with a fourdrug or three-drug regimen consisting of first-line drugs with or without fluoroquinolones. The recurrence rate was 2.2% with a median follow-up of 3.4 years. To the best of our knowledge, this is the largest study to report the real-world experience of rifampicin-sparing treatment for drug-susceptible tuberculosis with a long-term follow-up assessment. The high rate of the favorable response (80.7%) in our study is comparable to that achieved with regimens evaluated in recent studies for rifampicin-resistant tuberculosis, such as STREAM stage 2 (83%), 10 MDR-END (75.0%), 8 and the NExT study (51.0-75.0%). 11 Although data are scarce, a few studies have reported the consequences of rifampicin-sparing treatment owing to its adverse reactions. A previous study by Park et al. analyzed patients who could not tolerate rifampicin because of adverse reactions; 25 in their study, 20 of 25 patients (80.0%) achieved treatment success after a median of 371 days of treatment. In another study, Gupta and colleagues www.nature.com/scientificreports/ studied renal transplant patients who received rifampicin-sparing treatment. 26 The study showed favorable responses in 60 of 67 (93.7%) patients after a median treatment time of 12 months, and one patient experienced recurrence. Patients included in our study had a relatively shorter duration of treatment (median 10.2 months) compared to that in previous studies, but with a similar favorable response rate. These findings align with recent efforts to reduce the duration of treatment for drug-susceptible pulmonary tuberculosis 27,28 . Although examining the reason for this encouraging result is beyond the scope of this study, we suggest some explanations with caution. First, rifampicin was used at least once in all 92 patients with a favorable response, for a median duration of 1.22 (IQR 0.63-2.23) months. Rifampicin is highly bactericidal and effective against M. tuberculosis populations, not only in the metabolic state but also in the dormant state 29 . With adequate companion drugs, rifampicin is known to exert early bactericidal activity, which results in greater decrease in tuberculosis burden in the first 2 days of treatment than in the following 12 days 30 . The action of rifampicin against M. tuberculosis may have affected the treatment response, although it was only partly used during the early treatment course.
Second, other fundamental anti-tuberculosis medications have been actively utilized as substitutes for rifampicin. In patients with a favorable response, isoniazid and ethambutol were generally maintained during treatment (median duration of 8.87 and 8.93 months, respectively), and pyrazinamide was used for a median of www.nature.com/scientificreports/ 2.07 months, which is the recommended duration of use in the first-line treatment 22 . Other than the first-line drugs, fluoroquinolone was most actively used (91.3% of patients, median duration of 8.35 months), which was found to be comparable with the rifampicin-based regimen in a small group of renal transplant recipients 31 . The Curry International Tuberculosis Center and the California Department of Health also recommended a treatment regimen comprising isoniazid, ethambutol, and fluoroquinolone for 12-18 months and pyrazinamide for 2 months in cases of rifampicin-resistant tuberculosis 32 . In our study, cycloserine was also commonly used in patients with a favorable response (44.6%, median duration of 6.8 months). Cycloserine is a bacteriostatic antibiotic that inhibits cell-wall biosynthesis 33 . It is known to be efficacious against MDR tuberculosis and is currently included in category B for the treatment of MDR tuberculosis 7 .  www.nature.com/scientificreports/ Third, regimens consisting of at least three-drug combinations were favored. For those with a favorable response, 88 of 92 (95.7%) patients used ≥ 3 drug combinations during the intensive phase, while 81 of 92 (88.0%) patients used ≥ 3 drug combinations during the consolidation phase. This is in line with the principle of tuberculosis treatment, which requires preferentially three efficient drugs for a prolonged duration 34 .
Regarding the risk of recurrence, the results of our study should be interpreted with caution. There were no predefined follow-up protocols in our study. Although the median follow-up was 3.4 (IQR 1.8-6.8) years for 92 patients with a favorable response, 15 (16.3%) of them had < 1 year of follow-up after treatment completion. This can lead to some concern of possible recurrence, which usually occurs within a year of treatment completion. 35 Previous well-designed clinical trials (REMOX, OFLOTUB, and RIFAQUIN) have revealed higher rates of recurrence with shorter regiments compared to the standard regimen [36][37][38] . However, the findings of our study reflect our real-world practice, and every effort was made to assess recent health status for each patient.
This study has some limitations. First, the recurrence of tuberculosis was screened via electronic medical records or telephonic contact. This could have omitted patients with subclinical tuberculosis. Attention should be paid to the possible acquisition of rifampicin resistance, considering the short-term use of rifampicin 39,40 . Second, anti-tuberculosis regimens were frequently altered by the attending physician during the treatment course. Nonetheless, we have undertaken a thorough medical records review and provided detailed information regarding each drug used. Third, as a retrospective study, there were no pre-established protocols for defining adverse reactions, re-challenging drugs, or follow-up intervals. Nevertheless, we have incorporated long-term follow-up data on numerous patients to address some of these limitations.
In conclusion, we report a real-world experience of adverse reactions-necessitated rifampicin-sparing treatment for drug-susceptible pulmonary tuberculosis. Over 80% of the patients achieved an initial favorable response. They were treated with a four-or three-drug regimen for a median of 10.2 months, which included rifampicin for a median of 1.2 months. Considering the recurrence rate observed in this study (2.2% during a median follow-up of 3.4 years), attention should be paid to the patient health condition after treatment completion. We cannot make a specific recommendation based on our observations, but patients who cannot tolerate rifampicin due to adverse reactions may benefit from a shorter regimen, rather than a long-term treatment targeted for MDR tuberculosis. Future investigations involving immunomodulatory treatments for tuberculosis could potentially benefit patients experiencing adverse reactions to anti-tuberculosis medications 41 .

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.