In the present study, the proportion of patients who received IV antibiotics for MABC-LD was around 38.2%, and 70.8% of this group received them within 4 weeks of treatment commencement. We found that early IV antibiotics use was associated with modified microbiological cure, whereas radiological score and treatment duration were negatively associated with it.
Only 38.2% of our patients received IV antibiotics, which was inconsistent with guideline recommendations (14, 15) and lower than reported in previous studies (53–85%) (18, 21, 22). The low rate of IV antibiotic use might be attributed to the inconvenience of admission to hospitals (27), patients’ inability to tolerate adverse effects associated with the IV antibiotics (16), or doctors’ judgments that oral therapy might be appropriate after their assessment of the disease severity (19).
In addition, the role of IV antibiotics lacks strong evidence. For example, IV agents were not specified in the 2020 IDSA/ATS guidelines (15), while a duration of over 1 month for IV initial therapy was recommended in the 2017 BTS guidelines (14). Lyu et al. posited that long-term IV duration could probably be associated with treatment success (27). Our data documented that ever IV antibiotics use for MABC-LD had a crude OR of 3.21 for modified microbiological cure. Jeon et al. found that the combination of oral antibiotics with an initial 4 weeks of IV amikacin and cefoxitin was moderately effective (31). Therefore, the use of IV agents for an adequate duration could be beneficial in curing MABC-LD.
However, few studies published to date have discussed early or late initiation of IV agents. In the present study, early initiation of IV antibiotics was found to have a strong association with better treatment outcomes (aOR 8.58) with disease severity adjusted for, whereas late initiation of IV agents was not. Furthermore, early initiation of IV agents for MABC-LD led to higher cure rates than did late initiation of IV agents (Fig. 2). To the best of our knowledge, no studies have addressed the importance of the timing of IV initiation. The results of our study suggest that IV antibiotics should be initiated within the first month of the treatment course, strengthening the rationale of the guidelines for initial-IV therapy.(14, 15) As a result, once a decision to commence MABC-LD treatment has been made, early administration of IV antibiotics might improve treatment outcomes.
The modified microbiological cure rate was 47.2%, lower than those reported in a previous study (69.2%)(32) and for the U.S. and Korea (51–80.5%)(16, 27, 31, 33), and similar to that in a study in National Jewish Health (48%)(25). However, different studies adopted different treatment outcomes, so the results are not completely comparable. Nevertheless, all results revealed that the outcomes were unsatisfactory. Individual evaluations of IV effective agents at the intensive phase for MABC-LD along with frequent treatment assessments should be encouraged to improve treatment outcomes (14, 15).
Radiological score had an aOR of 0.83 for modified microbiological cure, indicating that the severity upon diagnosis would affect treatment outcomes. By contrast, treatment duration had negative correlation for the cure, possibly due to a reverse causal relationship. In other words, those with longer treatment durations might have had poor responses to MABC-LD treatment. The AUROC value of the final model considering the radiological score, the timing of IV initiation, and treatment duration was 0.7961 (Fig. 3); this value is good and higher than the AUROC values of any other single factor. In clinical treatment for MABC-LD, we should carefully consider the disease severity, use IV drugs early, and monitor the treatment response within an adequate course.
The strength of the present study was that the study sites were six hospitals in northern and southern Taiwan, implying that the recruited participants might represent the true phenomenon from a real-world experience. However, there were several limitations. First, the sample size was small, and the study sites did not include nonurban areas. Second, due to the retrospective design, we were unable to search for possible factors that prompted physicians to prescribe IV or oral agents. Third, drug susceptibility tests for all and mycobacterial subspecies identification for some participants were not performed regularly in the medical practices, so the drug resistance profiles were unknown, and we could not compare treatment outcomes among different subspecies. Fourth, the antibiotic regimens and durations did not follow international guidelines, which is why the treatment outcomes were not comparable with those of other studies. Lastly, the outcome assessment was conducted at the end of treatment, so the true timing of culture conversion may have been missed.
In conclusion, the ratio of IV antibiotic use was as low as 38%. Low radiological score at presentation and early initiation of IV antibiotics at initial treatment might be associated with the probability of an outcome of microbiological cure. The findings of the current study not only support the use of IV antibiotics for MABC-LD but also suggest earlier use within 4 weeks after treatment commencement. In the future, larger and prospective studies are needed to validate the association.