Characteristics of the subjects
A flow diagram summarizing the identified eligible subjects and the study participants was shown in Fig. 1. The baseline characteristics of the study population were listed in Table 2.
We retrospectively studied 1,836 subjects with PTBH in Hangzhou from 2005 to 2018. Participants in the training set (n = 1719) and the external validation set (n = 117) were analyzed respectively. There was not a significant difference between the two sets (Table 2). The mean age was 48.90 ± 20.95 and 49.41 ± 21.84, and the ratio of males to females was 2.42 to 1 and 3.03 to 1 in the training and validation set, respectively. Notably, most of the subjects [1, 357 (73.91%)] were with the education level of high school and below (Table 2).
Predictor's Selection
Table 3 summarizes the results of the univariate analyses of the association between an individual covariate and the risk of incident MDR-TB in individuals with PTBH. Twenty of the 43 tested covariates were associated with a high risk of incident MDR-TB from this study population in the training set (P < 0.10). The significant covariates were (a) sociodemographic characteristics, including age < 60 years, a history of direct contact, family income of low level, high-risk occupation, high school and below, and rural areas, (b) clinical characteristics, including passive mode of TB case finding, HIV infection, RTH, unsuccessful treatment, TIOFMV, FDC-2HRZE/4HR, 2HRZES/6HRE, 3HRZES/6HRE, frequencies of chest X-ray examination, duration of pulmonary cavities (DPC), and duration of abnormal X-ray findings, and (c) microbiological characteristics, including frequencies of sputum culture, duration of positive sputum culture (DPSC), and duration of negative sputum culture. The remaining 23 covariates, including gender, nationality, a history of direct contact (e.g., unknown), registered household of patients, comorbidities, patients with severe infection, mode of TB case management, TB treatment time, TIOLC, 2H3R3Z3/4H3R3, 2H3R3Z3E3/4H3R3, 2HREZ/4H3R3, 2HRZE/4HR, 3HRZE/6HRE, 2H3R3Z3E3S3/6H3R3E3, individualized TRs, duration of pulmonary miliary tubercles, duration without radiological findings, duration without sputum culture, frequencies of sputum smear, duration of positive sputum smear, duration of negative sputum smear, and duration without sputum smear, were not associated with incident MDR-TB among individuals with PTBH (P > 0.10).
Table 3
Univariate Cox regression model showing risk factors associated with incident MDR-TB in the training and validation sets (N = 1836)
Variables | Training set (n = 1719)a | Validation set (n = 117)a |
HR (95% CI) | P Value* | HR (95% CI) | P Value* |
Sociodemographic characteristics | | | |
Age (years) | | | | |
< 60 | 1.90 (1.57–2.31) | ༜0.001 | 2.74 (1.26-6.00) | 0.011 |
≥ 60 | Reference | | | |
Gender | | | | |
Male | 1.18 (0.85–1.42) | 0.083 | 1.40 (0.64–3.06) | 0.399 |
Female | Reference | | | |
Nationality | | | | |
Han | 1.03 (0.46–2.31) | 0.940 | 0.98 (0.13–7.17) | 0.986 |
National minority | Reference | | | |
A history of direct contact | | | | |
Yes | 3.25 (2.96–3.56) | ༜0.001 | 3.56 (2.38–5.33) | ༜0.001 |
Unknown | 1.18 (0.76–1.84) | 0.231 | 0.99 (0.46–1.69) | 0.213 |
No | Reference | | | |
Family income | | | | |
Low-income level | 1.33 (1.13–1.58) | 0.001 | 1.42 (0.76–2.68) | 0.275 |
Middle level and above | Reference | | | |
Occupational risk | | | | |
High-risk | 1.51 (1.26–1.80) | ༜0.001 | 2.22 (1.18–4.18) | 0.013 |
Non-high-risk | Reference | | | |
Education levels | | | | |
High school and below | 1.26 (1.04–1.54) | 0.022 | 1.44 (0.69–3.15) | 0.359 |
Universities and higher | Reference | | | |
Residences | | | | |
Rural areas | 1.32 (1.12–1.57) | 0.001 | 1.45 (0.76–2.76) | 0.265 |
Urban areas | Reference | | | |
Registered household | | | | |
Migrant individuals with PTBH | 1.133 (0.88–1.34) | 0.143 | 1.29 (0.68–2.45) | 0.445 |
Resident individuals with PTBH | Reference | | | |
Clinical characteristics | | | | |
Mode of TB case finding | | | | |
Passive | 2.76 1.59–4.78) | ༜0.001 | 2.13 (1.12–4.05) | 0.022 |
Active | Reference | | | |
Comorbidities | | | | |
Yes | 0.80 (0.54–1.16) | 0.240 | 0.93 (0.33–2.64) | 0.892 |
No | Reference | | | |
HIV infection | | | | |
Positive | 3.96 (2.97–5.26) | ༜0.001 | 2.96 (1.16–7.61) | 0.024 |
Negative | Reference | | | |
Patients with severe infection | | | | |
Yes | 1.07 (0.80–1.45) | 0.660 | 1.20 (0.37–3.91) | 0.763 |
No | Reference | | | |
Mode of TB case management | | | | |
Family members' management or self-management | 1.16 (0.93–1.44) | 0.197 | 1.19 (0.49–2.89) | 0.700 |
Community doctor management | Reference | | | |
Different PTBH | | | | |
NDTH | Reference | | | |
RTH | 1.67 (1.38–2.02) | ༜0.001 | 3.66 (1.86–7.21) | ༜0.001 |
Treatment outcomes | | | | |
Unsuccessful treatment | 5.14 (4.33–6.11) | ༜0.001 | 5.37 (2.54–11.34) | ༜0.001 |
Successful treatment | Reference | | | |
TB treatment time, days | 0.99 (0.99-1.00) | 0.130 | 1.05 (1.02–2.01) | 0.015 |
Time from illness onset to the first medical visit, days | 1.01 (1.00-1.02) | 0.025 | 0.99 (0.98–1.01) | 0.162 |
Time from illness onset to laboratory confirmation, days | 0.99 (0.99-1.00) | 0.945 | 0.99 (0.98–1.01) | 0.100 |
TRs of PTBH | | | | |
2H3R3Z3/4H3R3 | 0.67 (0.32–1.43) | 0.304 | 0.72 (0.22–2.36) | 0.592 |
2H3R3Z3E3/4H3R3 | 0.82 (0.62–1.10) | 0.185 | 0.74 (0.22–2.57) | 0.640 |
2HREZ/4H3R3 | 0.85 (0.40–1.79) | 0.611 | 0.97 (0.23–4.03) | 0.963 |
2HRZE/4HR | 0.87 (0.73–1.04) | 0.114 | 1.03 (0.97–2.02) | 0.596 |
FDC-2HRZE/4HR | 0.57 (0.36–0.93) | 0.025 | 0.73 (0.17–3.11) | 0.674 |
2HRZES/6HRE | 1.64 (1.12–2.40) | 0.011 | 0.48 (0.07–3.51) | 0.469 |
3HRZE/6HRE | 1.25 (0.89–1.77) | 0.196 | 1.84 (0.65–5.21) | 0.251 |
3HRZES/6HRE | 4.79 (2.47–9.29) | ༜0.001 | 2.44 (1.11–5.37) | 0.027 |
2H3R3Z3E3S3/6H3R3E3 | 1.00 (0.53–1.87) | 0.988 | 1.23 (0.43–3.47) | 0.699 |
Individualized TRs | 1.11 (0.93–1.33) | 0.266 | 1.39 (0.71–2.75) | 0.334 |
Chest radiological findings | | | | |
Frequencies of X-ray examination | 0.90 (0.85–0.96) | 0.001 | 1.01 (0.83–1.23) | 0.911 |
Duration of cavities, months | 1.21 (1.14–1.28) | ༜0.001 | 1.53 (1.15–2.04) | 0.004 |
Duration of miliary tubercles, months | 1.08 (0.54–2.15) | 0.836 | 0.77 (0.19–3.21) | 0.720 |
Duration of abnormal findings, months | 1.15 (1.09–1.21) | ༜0.001 | 0.98 (0.78–1.23) | 0.852 |
Duration without radiological findings, months | 0.81 (0.53–1.25) | 0.337 | 0.72 (0.28–1.84) | 0.492 |
Microbiological characteristics | | | | |
Frequencies of sputum culture | 0.88 (0.83–0.93) | ༜0.001 | 0.77 (0.60-1.00) | 0.050 |
Duration of positive sputum culture, months | 1.14 (1.04–1.24) | 0.004 | 0.99 (0.64–1.54) | 0.968 |
Duration of negative sputum culture, months | 0.77 (0.69–0.86) | ༜0.001 | 0.46 (0.22–0.97) | 0.042 |
Duration without sputum culture, months | 0.92 (0.79–1.06) | 0.249 | 1.01 (0.33–3.11) | 0.983 |
Frequencies of sputum smear | 1.05 (0.95–1.18) | 0.434 | 0.94 (0.87–1.02) | 0.121 |
Duration of positive sputum smear, months | 1.02 (0.98–1.06) | 0.405 | 1.05 (0.86–1.28) | 0.628 |
Duration of negative sputum smear, months | 0.98 (0.81–1.20) | 0.880 | 0.89 (0.80–0.99) | 0.030 |
Duration without sputum smear, months | 0.62 (0.32–1.21) | 0.161 | 0.83 (0.09–7.71) | 0.869 |
a Data are shown as hazard ratio (95% CI), P Value |
* Bold values are those that reach statistical significance (P < 0.05) |
Abbreviations: MDR-TB multidrug-resistant tuberculosis, TB tuberculosis, HIV human immunodeficiency virus, PTBH previous TB history, NDTH newly diagnosed TB history, RTH re-treated TB history, HR hazard ratio, TRs treatment regimens, FDC fixed-dose combination, H isoniazid, R rifampicin, Z pyrazinamide, E ethambutol, S streptomycin, CI confidence interval |
To further explore independent predictors of incident MDR-TB in individuals with PTBH, we performed a multivariate Cox proportional hazard regression analysis. Table 4 listed the multivariable Cox regression results for this study population. The analysis showed that age < 60 years, a history of direct contact, passive mode of TB case finding, HIV infection, RTH, unsuccessful treatment, 3HRZES/6HRE, DPC, and DPSC were significantly linked to the MDR-TB risk among individuals with PTBH (P༜0.05). From this model, we could also see that the unsuccessful treatment (HR 2.72, 95% CI 2.20–3.37, P༜0.001) was one of the strongest predictors for incident MDR-TB in this population (Table 4). These findings were used to create a practical clinical nomogram for predicting the probability of incident MDR-TB among individuals with PTBH (Fig. 2).
Table 4
Multivariate Cox regression model showing risk factors associated with incident MDR-TB in the training and validation sets (N = 1836)
Variables | Training set (n = 1719)a | Validation set (n = 117)a |
HR (95% CI) | P Value* | HR (95% CI) | P Value* |
Sociodemographic characteristics | | | | |
Age < 60 years | 1.25 (1.01–1.57) | 0.049 | 1.37 (0.51–3.71) | 0.531 |
Family income of low level | 1.05 (0.56–1.99) | 0.867 | NA | |
High-risk occupation | 0.98 (0.78–1.22) | 0.833 | 1.90 (0.78–4.65) | 0.160 |
High school and below | 1.22 (0.95–1.58) | 0.125 | NA | |
Rural areas | 1.13 (0.64–1.97) | 0.256 | NA | |
A history of direct contact | 2.71 (2.42–3.04) | ༜0.001 | 2.34 (1.33–4.13) | 0.003 |
Clinical characteristics | | | | |
Passive mode of TB case finding | 2.38 (1.24–4.58) | 0.009 | 1.28 (0.46–3.53) | 0.639 |
HIV infection | 2.36 (1.75–3.18) | ༜0.001 | 1.32 (0.42–4.16) | 0.640 |
Re-treated TB history | 1.36 (1.11–1.68) | 0.004 | 1.83 (0.26–12.70) | 0.540 |
Unsuccessful treatment | 2.72 (2.20–3.37) | ༜0.001 | 2.65 (1.06–6.62) | 0.037 |
TB treatment time, days | NA | | 1.03 (1.01–1.14) | 0.016 |
Time from illness onset to the first medical visit, days | 1.00 (0.99–1.01) | 0.368 | NA | |
FDC-2HRZE/4HR | 0.90 (0.52–1.54) | 0.692 | NA | |
2HRZES/6HRE | 0.71 (0.47–1.06) | 0.09 | NA | |
3HRZES/6HRE | 2.18 (1.31–3.62) | 0.003 | 0.71 (0.26–1.96) | 0.510 |
Chest radiological findings | | | | |
Frequencies of X-ray examination | 0.71 (0.65–1.77) | 0.149 | NA | |
Duration of cavities, months | 1.18 (1.10–1.27) | ༜0.001 | 1.51 (1.01–2.25) | 0.046 |
Duration of abnormal findings, months | 1.21 (0.91–1.34) | 0.253 | NA | |
Microbiological characteristics | | | | |
Frequencies of sputum culture | 1.00 (0.87–1.15) | 0.978 | 0.71 (0.45–1.13) | 0.148 |
Duration of positive sputum culture, months | 1.26 (1.10–1.44) | 0.001 | NA | |
Duration of negative sputum culture, months | 0.90 (0.75–1.08) | 0.997 | 0.95 (0.44–2.05) | 0.896 |
a Data are shown as hazard ratio (95% CI), P Value * Bold values are those that reach statistical significance (P < 0.05) Abbreviations: MDR-TB multidrug-resistant tuberculosis, TB tuberculosis, HIV human immunodeficiency virus, HR hazard ratio, FDC fixed-dose combination, H isoniazid, R rifampicin, Z pyrazinamide, E ethambutol, S streptomycin, CI confidence interval, NA not available |
Construction Of The Nomogram
A nomogram was developed to assess the risk of incident MDR-TB using significant factors from the 1719 patients’ data in the training set. With 9 independent predictors of training set, it was possible to create a nomogram to predict the probability of incident MDR-TB among individuals with PTBH (Fig. 2). The top row of the nomogram corresponds to the general score. For each predictors listed on the left (including age < 60 years, a history of direct contact, passive mode of TB case finding, HIV infection, RTH, unsuccessful treatment, 3HRZES/6HRE, DPC, and DPSC), there is a corresponding row on the right indicating possible descriptors. After characterizing the patient for each variable, a perpendicular line toward the first row should be drawn to identify the value. This action should be performed for all 9 predictors, followed by tallying the final score. This final score should be identified in a total point row and then a perpendicular line is drawn that corresponds to the probability of incident MDR-TB from individuals with PTBH.
Calibration And Validation Of The Nomogram
After internal validation using the bootstrap technique, the C-index of this nomogram is 0.833 (95% CI 0.807–0.859) and 0.871 (95% CI 0.773–0.969) for the training and validation sets, respectively, which indicates adequate discriminatory power. The calibration plots are also performed separately using the training and external validation sets. As shown in Fig. 3 (A), the calibration plots show that the predicted 1-, 5-, and 10-year probability of incident MDR-TB corresponded closely with the actual 1-, 5-, and 10-year probability of incident MDR-TB estimated in the training set. Figure 3 (B) illustrates that the nomogram appears well calibrated, and there is a strong correlation between predicted and observed outcomes across the spectrum of predictions in the external validation set.
For the training set, the AUCs of the nomogram predicting the 1-, 5- and 10-year incidence of MDR-TB were 0.904, 0.921, and 0.908, respectively [Fig. 4 (A)]. Regarding the external validation set, the AUCs of the nomogram for predicting the 1-, 5- and 10-year incidence of MDR-TB were 0.954, 0.970, and 0.919, respectively [Fig. 4 (B)]. As Fig. 4 shows, the nomogram demonstrated the superior prediction ability of incidence of MDR-TB.
Predicting An Individual Patient's Risk
To make it easier to interpret our results, we represented the final reduced model with a nomogram that can be used to calculate a prognostic score and estimate the risk of incident MDR-TB for an individual with PTBH (Fig. 2). The nomogram produced the following mathematical predictive model for the presence of incident MDR-TB risk in the training set, with h (t, x) denoting the probability of incident MDR-TB among individuals with PTBH [24]:
h (t, x) = h0 (t) exp (β1x1 +… + βixi+…+ βkxk) = h0 (t) exp [0.9975×(a history of direct contact) + 0.2253×(age < 60 years) + 0.2544 × RTH + 0.8685×(passive mode of TB case finding) + 0.8596×(HIV infection) + 1.0023×(unsuccessful treatment) + 0.7779 × (3HRZES/6HRE) + 0.1682 × DPC + 0.2308 × DPSC]
where h (t, x) is the hazard at time t after a defined starting point for an individual with variables x = (x1… xi … xk) is being predicted by h0 (t), the so-called underlying hazard at time t, and the predictor variables x1 to xk (recorded at time zero), each variable xi being multiplied by a corresponding regression coefficient βi. Here, exp stands for exponential function, e.g., exp (βx) = e βx, and the underlying hazard h0 (t) is the hazard at time t of an individual whose xi’s are all zero.
The predicted probabilities associated with each factor are mapped into points on a scale from 0 to 100. The presence or the level of each predictive factor is associated with a point system, allowing summing up the points for all the factors. The total points accumulated by the various covariates correspond to the predicted probability of incident MDR-TB. For example, for an individual with the characteristics of age < 60 years, a history of direct contact, RTH, unsuccessful treatment, 3HRZES/6HRE, DPC (such as 3 months), and DPSC (such as 2 months) in individuals with PTBH (see Table 5).
Table 5
Predicting an individual patient’s MDR-TB risk
Risk factor | Value | Points |
Age < 60 years | Yes | 12.5 |
A history of direct contact | Yes | 100 |
Passive mode of TB case finding | No | 0 |
HIV infection | No | 0 |
Re-treated TB history | Yes | 14.0 |
Unsuccessful treatment | Yes | 51.5 |
3HRZES/6HRE | Yes | 18.0 |
Duration of pulmonary cavities, months | 3 | 15.0 |
Duration of positive sputum culture, months | 2 | 24.0 |
Total points | 235.0 |
Estimate of MDR-TB risk, % | |
1-year probability of incident MDR-TB | 74.5 |
5-year probability of incident MDR-TB | 84.0 |
10-year probability of incident MDR-TB | 90.5 |
Abbreviations: TB tuberculosis, MDR-TB multidrug-resistant tuberculosis, HIV human immunodeficiency virus, H isoniazid, R rifampicin, Z pyrazinamide, E ethambutol, S streptomycin |
Abbreviations: MDR-TB multidrug-resistant tuberculosis, TB tuberculosis, DST drug susceptibility |
Abbreviations: MDR-TB multidrug-resistant tuberculosis, MTCF mode of TB case finding, HIV human immunodeficiency virus, PRTH previous re-treated TB history, NDTH newly diagnosed TB history, RTH re-treated TB history, TO treatment outcome, H isoniazid, R rifampicin, Z pyrazinamide, E ethambutol, S streptomycin, DPC duration of pulmonary cavities, DPSC duration of positive sputum culture |
Abbreviations: PTBH previous tuberculosis history, MDR-TB multidrug-resistant tuberculosis |