Acquired Drug Resistance During the Turnaround Time for Drug Susceptibility Testing Impacts Outcome of Tuberculosis

The impacts of acquired resistance to rst-line drugs (FLDs) during turnaround time (TAT) for drug susceptibility testing (DST) are still unclear. Thus, we aimed to investigate the impacts of acquired resistance to FLDs during TAT for DST on tuberculosis (TB) standard treatment. Method We performed a prospective cohort study between 2013 and 2018 in China, including sputum culture-positive TB patients with a baseline DST result for a Mycobacterium tuberculosis (Mtb) isolate collected at TB diagnosis and a follow-up DST result for a Mtb isolate collected when baseline DST result became available. Mtb isolates with acquired drug resistance were identied by the comparison between baseline and follow-up DST. Treatment outcome were evaluated by sputum culture conversion and World Health Organization (WHO) treatment outcome denitions. Data were presented by numbers (%), unless otherwise indicated. IQR, Inter-quartile range; TB, Tuberculosis; TAT for DST, turnaround time for drug-susceptibility testing; Diseases severity on chest X-ray was dened by Timika score; Severe disease, Timika score > 71; P a were calculated by Mann-Whitney test.

Drug-resistant TB, especially multidrug-resistant or rifampicin-resistant TB (MDR/RR-TB) threatens global TB care and prevention and remains a major public health concern. Globally in 2019, 3.3% of new TB cases and 17.7% of previously treated TB were MDR/RR-TB(1). For decades, diagnosis of drug-resistance has relied entirely on phenotypic drug susceptibility testing (DST), which routinely takes up to [6][7][8] weeks (2). During the two month-long turnaround time (TAT) for DST, drug susceptibility pro les of the Mycobacterium tuberculosis (Mtb) isolates might change due to suboptimal drug regimens, inadequate drug supply, low drug exposure or transmission of drug-resistant Mtb strains (3,4). It's well known that acquired RIF-resistance leads to unfavorable outcome in TB patients, but little is known how the acquired resistance to isoniazid (INH), pyrazinamide (PZA) and ethambutol (EMB) impacts the treatment outcome (5,6). Thus, we performed a prospective cohort study including three provinces in China to investigate the impacts of acquired resistance of the three rst-line drugs (FLDs), INH, PZA and EMB, during the TAT for DST on TB treatment.

Method Study population
We performed a prospective cohort study including three study provinces (Sichuan, Guizhou, and Zhejiang) in China. Midwestern Sichuan and its neighbor Guizhou to the southwest have two of the highest TB incidences in China, with 63 TB cases per 100,000 inhabitants and 122/100,000 respectively in 2018. The eastern province of Zhejiang has a lower TB incidence (33.52/100,000 in 2018). All the provinces have a similar three-tier health system consisting of village health stations, township, and county hospitals, as well as at least one TB-specialized health facilities in county level for TB diagnosis, treatment and case management. Due to limited local laboratory resources, bacterial cultures are sent to a prefectural TB reference laboratory for drug susceptibility testing (DST). All prefectural TB reference laboratories and hospital TB laboratories are quality assured in TB microbiological examinations at least twice yearly by the provincial TB reference laboratory. The study enrolled culture-con rmed pulmonary TB patients with a baseline DST result of a Mtb collected at TB diagnosis and a follow-up DST result of a Mtb isolate collected when baseline DST results were available in 2013-2018. All subjects were recruited in TB designated hospitals and received rst-line anti-TB drugs standard treatment.

Treatment and follow-up
All TB patients received rst-line drugs standard treatment after TB diagnosis that was made by smear microscopy, culture or WHO-recommended rapid diagnostic test. Standard rst-line treatment for newly treated TB was 6 months of 2HRZE/4HR according to WHO recommendations (7). The former WHO recommendations of a routinely prolonged 8 months retreatment regimen was still in practice in China at the time of the study(8), including 2-month intensive phase, 1-month continuation period of INH, RIF, PZA and EMB, 5-month continuation period of INH, RIF and EMB. Hospitalization with directly observed therapy was recommended for the rst two weeks of treatment. After discharge, drugs were supplied monthly basis and supervised by health care worker or family member. Follow-up clinical visits with collection of sputum samples were scheduled monthly during the 6-month or 8-month treatment duration.

Data collection
Information was collected regarding i) socio-demographic information: age, sex, place of residency; ii) clinical information: TB history (previously treated TB), disease severity on chest X-ray using the Timika score(9); iii) bacteriological diagnosis: sputum smear microscopy (positive or negative), sputum culture (positive or negative), drug susceptibility to the rst-line drugs RIF, INH, PZA and EMB (susceptible or resistant). As the aim of the study was to study the impact of acquired drug resistance during TB treatment with rst-line drugs, all cases of RIF-resistant Mtb was excluded from the analysis, since MDR-TB treatment was provided.

Drug susceptibility testing
In the study, sputum samples were sent for Mtb culture using BACTEC MGIT™ (Becton, Dickinson and company, USA), according to the manufacture recommendations. Baseline sputum samples were sent for smear microscopy for TB diagnosis. At the same day, the duplicates of baseline sputum samples were sent to the regional TB reference labs for culture and phenotypic DST for four rst-line drugs, INH, RIF, PZA and EMB (baseline DST). When baseline DST results became available, a sputum sample from each patient were collected for culture and follow-up DST to divide patients into groups: patients with acquired drug resistant Mtb isolates and patients with Mtb isolates with unchanged drug susceptibility pro le. According to WHO recommendations(10), the critical concentrations of INH 0.1mg/L, RIF 1.0 mg/L, PZA 100 mg/L and EMB 4 mg/L were used to classify the strains as susceptible or resistant. For quality assurance, the DST was repeated for 10 % of the isolates by an external technician from Shanghai Municipal Center for Disease Control and Prevention who had passed the WHO's external quality control assurance with a consistency of over 95%. No discrepancies were recorded.

De nitions
Disease severity of the patient's pulmonary infection was based on chest X-ray, summarized by the Timika score scale, where a score > 70 is considered severe TB disease (9). New cases were de ned as patients without history of prior TB treatment or who received less than one month of TB drugs. TAT for DST was de ned as the duration from baseline sputum collection for culture up until phenotypic DST availability of baseline isolate. Acquired drug resistance was de ned as new drug resistance (during the for DST to the rst time point when sputum culture negative for Mtb isolates, followed by three consecutive negative cultures. Successful treatment outcome was de ned as "cure" or "treatment completed. Cure was de ned as smear-or culture-negativity in the last month of treatment (and at least one previous occasion) and treatment completion if the patient had completed the 6-or 8-standard treatment, without ful lling the aforementioned de nitions of smear/culture-negativity(11). According to WHO recommendations (4th ed), if a patient whose sputum smear or culture was positive at 5 months or later during standard treatment, the treatment outcome was denoted as failure(11).

Statistical Analysis
Data were collected using standardized forms, computerized and analyzed with R software (Version 1.2.1335). Log-rank test was used to evaluate the difference of the time to sputum culture conversion between patients with acquired drug resistance and those without. A Cox proportional hazard regression model was used to evaluate the difference of time from the intermediate TAT for DST to sputum culture conversion between patients with acquired drug-resistant Mtb isolates and patients with unchanged drug susceptibility pro le by calculating hazard ratios (HRs) and 95% con dence intervals (95%CI). Logistic regression model was used to analyze the association between acquired drug resistance during TAT for DST and treatment failure. Age, sex and living areas were included as covariates in the adjusted model.
Statistical tests were two-tailed with a signi cance level set at p < 0.05.

Characteristics of patients
A total of 1,479 patients were diagnosed with cultured-con rmed pulmonary TB patients who had Mtb isolates with baseline DST results during the study period in the study sites. Of these, 144 patients with MDR-TB were excluded, 42 patients lost to follow-up, and 1,094 patients were sputum culture negative when baseline DST results were available.  (Table 1).   Data were presented by numbers (%), unless otherwise indicated. IQR, Inter-quartile range; TB, Tuberculosis; TAT for DST, turnaround time for drug-susceptibility testing; Diseases severity on chest X-ray was de ned by Timika score; Severe disease, Timika score > 71; P a were calculated by Mann-Whitney test.
Impact of acquired drug resistance on time to sputum culture conversion Acquired drug resistance during TAT for DST was associated with a signi cantly lower chance of sputum culture conversion, compared to patients with unchanged drug-susceptibility pro le (78.5% vs 94.5%, p < 0.001). Moreover, the Kaplan-Meier analysis revealed the median time to sputum culture conversion in patients whom Mtb isolates with acquired PZA-resistance was 15 days, which was close to that of patients whom Mtb isolates with unchanged susceptibility to PZA (Fig. 2). After adjustment for age, sex, TB history, disease severity on chest x-ray and baseline drug susceptibility pro le; acquired resistance to PZA (aHR 0.54, 95%CI: 0.36-0.81, p = 0.01) and acquired resistance to INH (aHR 0.50, 95%CI 0.29-0.85, p = 0.01) were still signi cantly associated with prolonged time to sputum culture conversion ( Table 2).  (Table 3).

Discussion
In this study, we found an alarmingly high rate (33.9%, 65/192) of acquired drug resistance to three rstline drugs during TAT for DST. Acquired PZA-resistance and acquired INH-resistance during TAT for DST were associated with longer time to sputum culture conversion. Furthermore, acquired INH-resistance as well as acquired PZA-resistance during TAT for DST were signi cantly associated with increased risk of treatment failure.
In this study, 33.9% (65/192) developed acquired resistance to the three rst-line drugs, INH, PZA and EMB, during TAT for DST. Compared to our study with 33.9% acquired drug resistance, another crosssectional survey from a high TB burden area found that acquired drug resistance developed in 31.0% (19/62) Mtb isolates during rst 2 months of treatment (12). A possible explanation for the acquired drug resistance is variable and sub-therapeutic drug exposure, which is commonly seen as highlighted by a previous study. Therapeutic drug monitoring is thus recommended in patients with risk factors for altered drug exposure, for instance malabsorption or diabetes (13). Our ndings suggest that a large proportion of TB patients received inadequate treatment in China, which might hamper TB control. Optimized therapy based on timely and accurate DST results, especially in high drug-resistant TB burden areas, would be bene cial for TB treatment.
We found that acquired PZA-resistance and acquired INH-resistance during TAT for DST were independent predictors of longer time for sputum culture conversion. In a previous study, PZA-resistance as well as previously treated TB was linked with prolonged time to sputum culture conversion (14)(15)(16).
Without the sterilizing activity of PZA, there is a risk of continued TB transmission and increased risk of relapse (17). Given the importance of early sputum culture conversion, individual treatment regimens coupled with interventions, such as regular surveillance to drug susceptibility pro le of Mtb isolates, are necessary.
Furthermore, acquired INH-resistance and acquired PZA-resistance were observed to be independently associated with treatment failure. The nding was in line with previous meta-analysis where treatment failure was signi cantly more common in patients with INH-resistant strains than in those infected with drug susceptible strain, when treated with rst-line tuberculosis drugs (18). Therefore, the WHO treatment recommendation for INH resistant TB from 2017 should be applied (6 months RIF, EMB, PZA and a uoroquinolone) (10). A cohort study also found that undiagnosed PZA-resistance might compromise treatment outcome of PZA containing drug regimens (19). Accordingly, if drug susceptibility pro le of Mtb isolate changes during TAT for DST, physicians should promptly adjust and individualize treatment. Additionally, the higher treatment failure rate among patients with acquired INH-and PZA-resistance indicate the need for more vigilant clinical drug-resistance monitoring and timely repeat DST results during anti-TB treatment.
Relying on the systematic clinical investigation and microbiological testing, the present study made it possible to observe how acquired drug resistance during TAT for DST impact time to sputum culture conversion and treatment outcome. The fact that DST was performed in reference TB laboratories, including repeat DST, provides quality assurance. Meanwhile, the relatively large sample size made it possible to consider covariates and confounders in the multivariate analysis. The present study also have some limitations. Firstly, phenotypic DST for PZA using BACTEC MGIT™ might be affected by some factors, such as bacterial activity and load, but WGS for genes associated with PZA-resistance was performed for strict quality control and the consistency rate reached more than 90%. Secondly, the timing of acquired drug resistance were di cult to determine and thus we used the middle points of TAT for DST as an approximation. Moreover, the studied subjects received standardized regimens and some nding might be not generalizable to a population with high proportion of individualized treatment. However, our results support the use of an individualized treatment with follow-up DST in case of continued culture positivity, as the proportion of acquired drug resistance was signi cant. Furthermore, the routine use of a prolonged 8 month treatment regimen with rst-line drugs without a renewed DST should be abolished, considering the risk of causing more drug resistance.

Conclusions
In conclusion, we found that acquired PZA-resistance during TAT for DST might prolong the time to sputum culture conversion as well as increase the treatment failure rate and acquired INH-resistance was also an independent risk factor for treatment failure. Our observation highlights the need for regular drugresistance monitoring and adjustment of treatment regimens based the changed drug susceptibility pro le during TB treatment.

Abbreviations
FLDs: First-line drugs TAT: Turnaround time