1. Introduction
Ranked only behind India and Russia, South Africa occupies the third position because the country has a high number of patients diagnosed with drug-resistant tuberculosis (DR-TB) in the world [
1]. The rise of DR-TB cases in South Africa and other parts of the world threatens the efforts to end TB [
1]. Globally, studies have shown that one of the major public health challenges today is the fact that numerous countries, particularly countries with limited resources, have high rates of multidrug-resistant tuberculosis (MDR-TB) and extensive drug-resistant tuberculosis (XDR-TB) [
2,
3]. It is evident through various published reports that most of those MDR-TB and XDR-TB cases are driven by the HIV epidemic [
2,
3]. In addition, the successful response to anti-TB treatment has been also challenged. For example, in 2014, reports underlined the fact that more than half of the total estimated cases of DR-TB were receiving anti-TB therapy and only half achieved treatment success [
2,
3].
Among strategies able to reduce the transmission of tuberculosis (including MDR-TB and XDR-TB), early detection of cases and their effective management are proven to be the cornerstone approach for substantially preventing the occurrence of new TB cases [
4,
5]. In patients diagnosed with TB, the conversion of smear microscopy is achieved when the initial positive result becomes negative following an appropriate anti-TB therapy [
6]. In order to achieve such a conversion of smear microscopy during the first two months of the intensive care phase, patient compliance with anti-TB drugs is mandatory, hence representing one of the most successful outcomes of any TB control program [
7,
8]. In settings with limited resources, smear conversion is, therefore, an important indicator of patient response to therapy because the use of culture techniques for monitoring patient response is costly [
4]. In addition to patient adherence to TB treatment, studies have shown that demographic factors (age and gender), HIV status, and initial level of infectiousness (measured by smear grade) are determinants for sputum conversion rate. HIV-infected patients have a 5–15% probability of developing active TB disease, making HIV coinfection one of the most significant risk factors for developing TB [
2,
3]. It has been indicated that HIV prevalence in a setting is proportionally correlated with TB incidence. About 8% of all new TB cases can be attributed to HIV [
9]. In 2018, in South Africa, HIV prevalence was estimated to be 13% at the population level, while 6 out of 10 TB patients were found to be coinfected with HIV [
10]. Moreover, the probability of death was two times more likely to occur among TB and HIV coinfected patients compared to HIV-infected patients without TB, even after adjusting for the levels of CD4+ T-cells and the initiation of antiretroviral therapy [
11].
In South Africa, the Xpert MTB/RIF (Cepheid, Sunnyvale, USA) has largely replaced direct smear microscopy for the routine detection of
Mycobacterium tuberculosis even at primary care levels. However, the microscopy smear test remains useful in patients who are confirmed to have tuberculosis for monitoring changes in loads of
M. tuberculosis observed during microscopy testing using smear grading, hence indicating treatment response in patients whose initial smear-positivity results were determined [
7,
8]. Although finding acid-fast bacilli (AFB) is highly specific in confirming the diagnosis of smear-positive tuberculosis, the viability of AFB seen can only be confirmed by cultures. Nevertheless, in many developing countries, microscopy conversion (from positive to negative results) is still considered a predictor of survival and quality care because TB transmission is highly likely to occur when a TB patient has positive sputum microscopy. However, information from the literature pertaining to the time to sputum microscopy conversion among MDR-TB patients coinfected with HIV is scanty with enormous controversies. This study determined the impacts of HIV disease progression, baseline TB infectiousness rate, and demographic factors (age and gender) on rates and time to sputum microscopy conversion among MDR-TB patients following the initiation of an appropriate anti-TB therapy. It was intended that in resource-limited settings, the use of baseline microscopy smear grading might also play a key role in risk stratification before treatment initiation of MDR-TB patients with respect to TB transmission and rational use of basic IPC supplies (i.e., surgical and N95 masks).
3. Results
A total of 200 patients were enrolled in this study, of which 103 (51.5%) were males and 97 (48.5%) were females, with a sex ratio of males to females of 1.06:1. Out of these 200 patients, 55 (27.5%) were aged ≥35 years, while 145 (72.5%) were aged <35 years. The mean age of the study population was 34.8 (±11.6) years ranging from 18 to 67 years. HIV test results were obtained from the 200 participants, of which 114 (57%) tested HIV-positive and 86 (43%) tested HIV-negative (
Table 1). Of the 200 participants, 100 (50%) were confirmed as having multidrug-resistant tuberculosis (MDR-TB), while the remaining 100 (50%) patients were diagnosed with drug-susceptible tuberculosis (DS-TB).
3.1. Sputum Microscopy Smear Conversion Rates and Time to Sputum Conversion among HIV-Infected and -Uninfected Patients
Table 2 below depicts the number of cases of microscopy smear grades at baseline stratified by HIV serostatus of the participants. HIV-infected and -uninfected patients were matched by their microscopy smear grades at enrolment (
p = 0.794).
After 4 and 8 weeks of appropriate anti-TB treatment, there was no statistically significant difference in the microscopy smear conversion rates of HIV-infected patients as compared to HIV-uninfected patients. However,
Table 3 below shows that after 12 weeks of appropriate anti-TB treatment, there was a significant microscopy conversion rate among DS-TB patients as compared to MDR-TB patients, irrespective of their HIV status (
p = 0.0013).
Using the Wilcoxon rank sum test, there was no significant difference in time to microscopy conversion when comparing survival curves of HIV-positive patients aged ≥35 years (5.96 ± 0.387 weeks), HIV-positive patients below 35 years (6.78 ± 0.40 weeks), HIV-negative patients aged ≥35 years (5.83 ± 0.47 weeks), and HIV-negative patients below 35 years (7.06 ± 0.48 weeks) (Chi-square = 4.981; p = 0.173).
Of the 200 patients, 114 (57%) were HIV-infected of which we only obtained CD
4+ T-cell count results from 50 patients. The mean CD
4+ T-cells for the 50 patients was 179.84 cells/mm
3, ranging from 3 to 487 cells/mm
3, with a standard error of the mean (SEM) of 19.59 cells/mm
3 and standard deviation (SD) of 139.9 cells/mm
3. Of the 50 HIV-infected patients, 28 (56%) had a CD
4+ T-cell count ≤ 200 cells/mm
3 (severe immune suppression), while 17/50 (34%) and 5/50 (10%) HIV-infected patients had a CD
4+ T-cell count of 200–349 cells/mm
3 (advanced immune suppression) and 350–499 cells/mm
3 (mild immune suppression), respectively. A linear trend between levels of immune suppression by means of CD4 T-cell count (cell/mm
3) and sputum smear conversion status among MDR-TB patients was determined, and the findings show that the value of the Chi-squared test for Trend statistic was 1.356 and the
p-value was 0.244 as depicted in
Table 4. These results show that there is no significant upward trend in the proportion of MDR-TB patients who converted negative after 12 weeks of appropriate anti-TB therapy classified by immune suppression levels on treatment initiation.
However, there was a significant difference between the above levels of immuno-suppression (CD4+ T-cell count at the beginning of anti-TB therapy) and time to microscopy smear conversion. The latter was shorter (4.25 ± 2.43 weeks) among HIV-positive patients whose CD4+ T-cell counts were <200 cells/mm3 as compared to patients who had CD4+ T-cell counts between 200 and 349 cells/mm3 (6.77 ± 3.99 weeks) and 350 and 499 cells/mm3 (7.40 ± 2.97 weeks) (ANOVA; p = 0.015).
Furthermore, when the above 50 HIV-positive patients coinfected with MDR TB were categorized into two groups composed of Group 1: 28 (56%) patients with CD4+ T-cell counts <200 cells/mm
3 and Group 2: 22 (44%) patients with CD4+ T-cell counts ≥200 cells/mm
3, patients with CD4+ T-cell counts <200 cells/mm
3 converted faster (5.379 ± 0.456 weeks; 95% CI: 4.486–6.273) than those with CD4+ T-cell counts ≥200 cells/mm
3 (8.000 ± 0.746 weeks; 95% CI: 6.538–9.462), and this difference reached statistical significance using the Wilcoxon rank sum test (
p = 0.004) as depicted in
Figure 1 below.
3.2. Factors Associated with Prolongation of Sputum Smear Positivity in All TB Patients at the End of 12 Weeks Post-Anti-TB Initiation
Using bivariate analysis,
Table 5 shows that at the end of 12 weeks of appropriate anti-TB therapy, patients aged <35 years (
p = 0.045) who had a baseline sputum smear grade from 2+ to 3+ on microscopy (
p < 0.0001) and were diagnosed with MDR-TB (
p = 0.003) were significantly associated with a prolongation of their sputum smear positivity.
At the end of 12 weeks of anti-TB therapy, 9% of all DS-TB patients remained positive versus 25% of all their MDR-TB counterparts (
p = 0.003).
Figure 2 shows that time to sputum smear conversion of DS-TB patients was shorter (5.92 ± 0.26 weeks; CI: 5.403–6.437) than MDR-TB patients (7.00 ± 0.34 weeks; CI: 6.328–7.672) using the Wilcoxon rank sum test (Chi-square = 3.921;
p = 0.048).
In addition, 87.9% of MDR-TB patients who were aged ≥35 years converted negative at the end of 12 weeks of anti-TB therapy as compared to 78.2% of MDR-TB patients aged <35 years (
p = 0.045).
Figure 3 shows that time to sputum smear conversion for patients aged ≥35 years was shorter (5.91 ± 0.29 weeks; CI: 5.327–6.495) than the time to conversion for patients aged below 35 years (6.91 ± 0.31 weeks; CI: 6.302–7.516) (Chi-square = 4.875;
p = 0.027).
As displayed in
Table 6 and
Figure 4 below, there was a significant difference between baseline sputum smear grades and time to negative conversion (ANOVA: F = 4.467;
p = 0.005). Using Bonferroni multiple comparisons, TB patients whose baseline sputum microscopy grade was +1 converted significantly faster than patients with a microscopy grade of +3 on treatment initiation (
p = 0.009). In addition, when comparing time to negative conversion between TB patients whose sputum microscopy grades ranged from scanty to +1 (5.20 ± 2.36 weeks) and those whose sputum microscopy grades ranged from +2 to +3 (6.70 ± 3.07 weeks), a T-test for independent samples showed that the difference was statistically significant (
p = 0.0003). For the analysis, data on eight patients were missing.
However, when comparing HIV-positive and HIV-negative patients coinfected with TB with respect to whether they converted negative or remained positive at 4, 8, and 12 weeks, the difference was not statistically significant.
3.3. Prolongation of Smear Positivity at the End of 12 Weeks Post-Anti-TB Initiation in Patients Diagnosed with MDR-TB as Compared to Patients Diagnosed with Drug-Susceptible TB
In a bivariate analysis,
Table 7 below shows that the proportions of MDR-TB patients with prolonged sputum smear positivity were significantly associated with the baseline smear microscopy grade (
p = 0.014) and initial CD4 T-cell categories (
p = 0.010). These differences were not observed among DS-TB patients at the end of 12 weeks post-anti-TB initiation.
In addition, time to microscopy conversion in patients diagnosed with MDR-TB was significantly prolonged as compared to patients diagnosed with DS-TB irrespective of their HIV status using ANOVA (
p = 0.035) as shown in
Table 8. However, these findings were not supported by Bonferroni multiple comparisons (
p > 0.05) and the Wilcoxon rank sum test (
p = 0.204). The latter test indicated that there was no significant difference in time to microscopy conversion when comparing survival curves of MDR-TB patients coinfected with HIV (6.78 ± 0.399), MDR-TB patients without HIV infection (7.48 ± 0.663), DS-TB patients coinfected with HIV (5.78 ± 0.350), and DS-TB patients who are HIV-negative (6.036 ± 0.387) (Chi-square = 4.592;
p = 0.204).
Figure 5 below indicates that the mean time to sputum smear conversion from positive to negative was inversely associated with age among MDR-TB patients. Using a T-test for independent samples involving MDR-TB patients, time to microscopy conversion in patients aged <35 years was significantly prolonged (6.67 ± 3.12 weeks) as compared to patients aged >35 years (5.70 ± 2.63 weeks) (
p = 0.021).
Finally, at the end of 8 weeks of anti-TB therapy, all MDR-TB patients who had a baseline smear microscopy ranging from scanty to +1 converted negative, while 25% of all MDR-TB patients with smear grade ranging from +2 to +3 remained positive until the end of 12 weeks following treatment initiation (p = 0.014).
As depicted in
Figure 6 below, all MDR-TB patients who had a baseline smear microscopy from scanty to +1 converted faster (5.20 ± 0.305 weeks; CI: 4.602–5.798) than MDR-TB patients who had a baseline smear microscopy graded from +2 to +3 (7.00 ± 0.272 weeks; CI: 6.466–7.534) using the Wilcoxon rank sum test (Chi-square = 15.543;
p = 0.0001).
In a multivariate analysis, Cox regression showed that only the baseline sputum smear microscopy grade was independently associated with prolonged smear positivity in MDR-TB patients at the end of 12 weeks following anti-TB initiation (B = −0.405; SE = 0.159; Wald = 6.464; Exp (B) = 0.667; p = 0.011) after adjusting for HIV status, CD4+ T-cell count, and age.
4. Discussion
In South Africa and other countries with limited resources, despite numerous efforts and progress made, DR-TB remains not only a public health challenge but also an economic challenge because the management of DR-TB cases is extremely costly as compared to drug-sensitive TB [
5,
16]. This situation is worsening by the presence of a large HIV epidemic in these settings and also by the astronomic rise of cases of MDR-TB and XDR-TB [
5,
16]. Because regular cultures of patient specimens are not cost-effective in order to monitor patient response to therapy, smear conversion is still effectively used as an indicator in many countries with limited resources [
17]. Healthcare workers in settings with limited resources rely on findings from smear microscopy grading to make clinical decisions. The present study was conducted in order to gain insight into delayed time and rate of sputum smear conversion, particularly in the face of dual MDR-TB/XDR-TB and HIV epidemics, as a proxy indicator of the persistence level of TB infectiousness. Factors associated with delays in smear conversion from positive to negative were also explored.
Anecdotal evidence has been suggesting that DR-TB patients, when coinfected with HIV, become more infectious with less response to treatment as compared to DR-TB patients who are HIV-negative. Although previous studies conducted elsewhere have refuted this assumption, it was necessary to test the hypothesis in settings where dual DR-TB and HIV epidemics are prevalent and resources are limited. The underlying assumption was that in resource-limited settings, the use of baseline smear microscopy grading might play a key role in risk stratification before treatment initiation of MDR-TB patients, hence supporting the rational use of basic IPC supplies (i.e., surgical and N95 masks) due to their regular shortage to effectively prevent the transmission of DR-TB.
Our results show that after 8 weeks of appropriate anti-TB treatment, there was no significant difference in the smear microscopy conversion rates of HIV-infected patients as compared to HIV-uninfected patients. Furthermore, it is shown that after 8 weeks of treatment, one patient out of every six treated patients remained potentially infectious, irrespective of the HIV status. This is contrary to the general belief that patients become noninfectious after two weeks of the standard treatment regimen. This finding is in line with other results, which have also shown that the conversion to a negative test, hence the loss of infectiousness of pulmonary tuberculosis patients during therapy, does not occur rapidly in all patients [
18,
19]. In Tanzania, it was found that after 2 weeks of treatment conversion was significantly higher in HIV-positive (72.8%) than HIV-negative (63.3%) patients [
19]. From our results, it is also shown that after 2 weeks of treatment, the conversion rate was significantly higher in HIV-positive (56.6%) than HIV-negative (43.4%) patients. Previous studies have shown that HIV patients do not show delayed sputum conversion [
20,
21]. On the contrary, it has been argued that HIV patients may have early sputum conversion [
21]. A study that was performed in the Karonga district in Malawi showed that the presence of HIV infection was associated with a shorter time to sputum smear conversion [
22]. In this study, HIV-negative TB patients were at higher risk for 8-week sputum smear nonconversion than their HIV-positive counterparts. It could be that the specialist-driven treatment-adherence training received by HIV-positive patients also had a positive effect on these patients when they initiated their TB treatment, consequently influencing their sputum conversion.
Moreover, previous research in Tanzania [
19] and Uganda [
20] has shown that HIV status is not a significant predictor of two-month sputum smear nonconversion. Coinfection with HIV has not been reported to be a significant factor for the smear conversion rate or time to conversion [
18,
19]. HIV coinfection did not significantly influence the duration of infectiousness during treatment. In this present study, although there was no significant difference in the sputum smear conversion rates of HIV-infected patients versus HIV-uninfected patients, the conversion rate was, however, higher (>90%) among HIV-infected patients whose baseline microscopy grade ranged from scanty to 2+ as compared to patients who had a grade of 3+ at baseline (conversion rate of 76%).
At 8 weeks, 82.5% of all our study participants converted negative. Two separate studies in India reported two widely different conversion rates at the end of 8 weeks: one reported 58% [
23], while another reported 84% [
24]. It is estimated that more than 20% of smear-positive patients remain infectious after 2 months of treatment [
25,
26]. Studies performed in many settings have shown that the proportions of sputum smear nonconversion at the end of the intensive phase of TB treatment range from 5% to 32% [
27,
28].
In this study, acid-fast smear positivity status was also assessed in correlation with CD4+ T-cell counts. Paradoxically, HIV-positive patients with CD4+ T-cell counts < 200 cells/mm
3 converted faster than those with CD4+ T-cell counts ≥ 200 cells/mm
3. All HIV-infected MDR-TB patients who had a baseline CD
4+ T-cell count ≥ 350 cells/mm
3 at the beginning of the anti-TB treatment had a sputum microscopy conversion from positive to negative after 12 weeks. There is a clear benefit to the individual in starting HIV treatment before reaching a CD4+ T-cell count of 350 cells/mm
3. In 2012, SA raised the treatment threshold to a CD4+ T-cell count from 200 to 350 cells/mm
3 for all patients, while pregnant women and anyone with TB are given ART irrespective of the CD4+ T-cell count [
29].
Currently, the duration of MDR-TB infectiousness after the initiation of effective treatment is still a subject of discussion. Measures are to be maintained until noninfectiousness has been demonstrated. According to the Centers for Disease Control and Prevention guidelines [
30], patients with DS-TB may be considered to be noninfectious when (i) they have completed at least two weeks of standard anti-TB therapy, preferably with direct observation by a TB program-appointed treatment supervisor; (ii) they have demonstrated clinical improvement; and (iii) there is negligible chance of multidrug resistant TB (MDR-TB). However, the World Health Organization’s compendium of indicators for monitoring and evaluating national TB programs states that the majority of new smear-positive pulmonary TB patients should convert to smear-negative after 2 or 3 months of treatment [
2,
3,
6]. In addition, the demonstration of noninfectiousness is best performed by demonstrating culture conversion [
2,
3,
6]. Unfortunately, the time required to report culture results and the availability of resources are major limitations to the use of cultures for the purpose of infection prevention and control. On the other hand, sputum smear microscopy, though less sensitive than cultures according to the literature, can be reported much earlier, but the presence of positive microscopy does not necessarily indicate the presence of living mycobacteria, hence the risk of infectiousness.
For most of our patients who were diagnosed with DS-TB versus MDR-TB, the conversion rate was 91% and 75%, respectively. A study by DeRiemer et al. reported that for managing MDR-TB patients, the TB program can achieve a cure rate of 73% [
31]. The present study shows that potential TB infectiousness decreases from the baseline to 56.5%, 82.5%, and 83.0% at the end of 4 weeks, 8 weeks, and 12 weeks, respectively. Our smear conversion rate at the end of 12 weeks (83%) was shown to be higher than the reported 80.0% in Taiwan [
32] but lower than the 98.6% found during a study in Tanzania [
19] and also lower than the 92% rate reported by Bawri et al. [
20].
The present study shows that the age group <35 years was a significant predictor of nonconversion of sputum smears among our smear-positive TB patients after the initial anti-TB treatment. This is contrary to numerous other studies that reported that the elderly were the least likely to have documented sputum conversion [
25,
33,
34,
35]. The authors have identified older age as a significant risk factor for sputum smear nonconversion at the end of the intensive treatment phase and also showed that age ≥ 40 years was an independent predictor of nonconversion of sputum smears [
25,
33,
34,
35]. Singla et al. observed in a similar study that patients aged over 60 years had an almost six times greater risk of remaining sputum-positive after two months of treatment than patients aged 21–40 years, while patients aged 41–60 years were twice as likely to remain sputum-positive [
25]. The reasons for these differences are not clear.
An independent risk factor for nonconversion identified in this present study was a higher smear grade (≥2+) at baseline, and this was irrespective of HIV status. A higher smear grade at baseline has also been identified in other studies as a predictor for nonconversion after 2 months of treatment [
4,
17,
25,
26,
36]. According to the authors, a heavy initial bacillary load has been documented as an important risk factor for delay in sputum smear conversion at the end of the intensive phase of TB treatment. They argued that delayed treatment onset is associated with a higher bacillary load at diagnosis, which in turn is related to higher sputum smear nonconversion [
4,
17,
25,
26,
36].
Many other reasons can also explain the nonconversion of sputum smear at the end of 12 weeks of treatment. First of all, nonviable bacteria can remain visible under the microscope. Ideally, culture is the best method to evaluate the viability of
M. tuberculosis [
37]. Unfortunately, this method is not practical in resource-limited settings, such as in developing countries where TB is prevalent. Moreover, a study has shown a good correlation between cultures and sputum acid-fast bacilli smears [
22].
Other potential explanations of nonconversion of sputum smears at the end of 12 weeks of treatment are poor supervision of the initial phase therapy, poor treatment adherence by patients, doses of anti-TB drugs below the recommended range, comorbid conditions, drug-resistant
M. tuberculosis that is not responding to first-line treatment, and heavy initial bacillary load [
4,
17,
25,
26,
36]. The presence of lung cavitation in a TB patient is a well-recognized risk factor for delayed conversion and treatment failure in TB due to decreased penetration and antibacterial activity of drugs [
21,
38,
39].
The intensive phase completion conversion rate of 83% demonstrated in this study compares well with the expected rate, ranging between 80% and 90%, among MDR-TB patients who are put on appropriate therapy and comply fully with treatment instructions. With full patient compliance and strict adherence to therapeutic instructions, it is anticipated that 80 to 90% of infected patients will respond to therapy and become smear-negative within 2–3 months of treatment [
40].
5. Conclusions
This study demonstrates that a reasonable number of MDR-TB patients did not convert to a negative smear result at the end of 8–12 weeks and that HIV-seropositive status is not a principal factor in delaying sputum conversion among patients receiving intensive phase tuberculosis treatment. We found that a heavy initial bacillary load is associated with sputum smear nonconversion at the end of the intensive phase of TB treatment. As a result, MDR-TB patients with a heavy initial bacillary load should thus be closely monitored, and healthcare professionals should maintain infection prevention measures until conversion occurs.
The adoption therefore in developing countries of modern rapid methods for identifying and testing the susceptibility of M. tuberculosis to at least rifampicin should be encouraged. These will help considerably in the risk stratification of MDR-TB patients, hence improving clinical and infection prevention decision making. In addition, strict monitoring and evaluation of drug adherence especially among MDR-TB patients as well as periodic surveillance of regional drug-resistant status among TB patients are warranted.
Infection prevention and control measures are recommended for all sputum smear-positive patients to minimize the spread of infection. Measures are to be maintained until noninfectiousness has been demonstrated. Patients with a heavy initial bacillary load should thus be closely monitored.
Limitations of the study included the presence of missing data, mainly data on CD4+ T-cell count and HIV plasma viral load results. Further, culture is the best method to evaluate the viability of M. tuberculosis and, if performed, could have helped in drawing more accurate conclusions. However, we compared patients with DS-TB versus MDR-TB smear conversion rates and time to conversion and used multivariate Cox regression models to rule out confounders.