This study was designed on retrospective cohort data analysis, using follow-up data of the patients from registration to treatment completion in the TB control program; from January to December 2016. Ethical approval of the study was obtained from the Institutional Committee for Ethics and Review of Research, IIHMR University, Jaipur (FWA#: FWA00018806, OMB No.0990-0279) and Kandahar Medical faculty research committee. Informed consent of patients was not sought, as we collected and analysed routine program data. Unique identifiers were removed from data analysis to ensure confidentiality of the study participants.

This study was conducted in 13 public health facilities and TB treatment centers in Kandahar city. These facilities provide services including testing, treating and counseling for TB to approximately 1 million urban populations. They are collaborative units of Ministry of Public Health and funded by international non-governmental organisations.

Study participants were newly diagnosed pulmonary TB cases, those who were registered at the selected public health facilities and TB treatment centres in Kandahar city. Newly diagnosed smear-positive pulmonary TB cases were defined according to WHO criteria [18].

Inclusion and exclusion criteria: Patients registered as category I TB (smear-positive) were included. Sputum negative, relapse, and drug resistant TB cases, or those died were excluded from the cohort.

Data was collected during 2017-2018. The TB registers (National TB registers) of the health facilities were scrutinised starting from January to December 2016. Socio-demographic, clinical, microbiological, radiological, and treatment initiation and follow-up data were collected manually in the extracted sheet.

In the health facilities, suspected cases of TB were subjected to sputum microscopy. Pulmonary TB cases were classified according to bacteriological results. Sputum smear examination was done through microscopic observation of M. tuberculosis after staining with Ziehl-Neelsen’s technique. The cases are classified as smear-negative pulmonary TB (smear contains no acid-fast bacilli (AFB) in 100 fields) and smear-positive pulmonary TB. Smear grading of the cases is done as follows: “1+ (10–99 AFB in 100 fields), 2+ (1–9 AFB/field in at least 50 fields), and 3+ (>10 AFB/field in at least 20 fields)” [18].

Collected data were coded and entered into Microsoft Excel (2016). Data quality was checked for consistency, completeness, and accuracy.

The data were analysed using IBM SPSS version 21 [19]. Sputum conversion rates and delayed conversion were presented as percentages. Bivariate analysis was used for the possible factors associated with delayed sputum conversion. To determine the strength of association, Odds Ratio (Chi-square test, Mantel-Haenszel statistics) was estimated. Stepwise multiple logistic regressions (forward LR method) were carried out to identify independent predictors for delayed sputum conversion. The factors, which had an association with delayed sputum smear conversion on bivariate analysis, were included in the multiple regression analysis.

A total of 345 newly diagnosed smear-positive pulmonary TB patients who received category I treatment and had minimum five months follow-up data, were included in the analysis. The mean age of TB patients was 38 years (SD±17.6). Of the total, 52.2% were female, and 74% belonged to urban residence. All patients were HIV negative at initial visit. Major symptoms documented were fever (100%), cough (99%), chest pain (30%), night sweats (57.4%), weight loss (25%), and haemoptysis (24.3%). About 29% cases had primary bacilli density AFB1+, 43% AFB2+, and 27.8% AFB3+ before starting the treatment. [Table/Fig-1] shows socio-demographic factors, clinical, microbiological and other diagnostic findings at baseline.

The sputum conversion rate was 82.6% at the end of the 2^nd month (intensive phase), and sputum smear conversion rate was delayed in 17.4% at the end of intensive phase of treatment.

Bivariate analysis of the factors associated with the delayed sputum conversion at two months of treatment is shown in [Table/Fig-2]. Patients who showed cavitary lesion and bilateral lesion on radiography had delayed sputum conversion, although it was not statistically significant. Age 50 years and above, female sex, haemoptysis at the first visit, and bacterial density of sputum AFB 3+ was significantly associated the delayed sputum conversion. These factors were included in the multiple logistic regression models. It was found that higher bacterial load (AFB3+) was significantly associated with delayed sputum conversion.

This study was done to determine the sputum conversion rate at the intensive phase of anti-TB treatment and identify the factors associated with delayed sputum conversion. Monitoring of sputum smear conversion provides inputs for planning of treatment, diagnostic, and counseling services for TB patients and opportunity to address the emerging epidemic of Multi-drug Resistant Tuberculosis (MDR TB) [19]. It has special importance in countries like Afghanistan affected from high disease burden and conflict affected weak health systems.

In this study, the sputum conversion rate was 82.6% at the end of intensive phase of treatment in smear-positive pulmonary TB, which was remarkably similar in the studies conducted in Pakistan, India, West Africa, and Iran (80.5%-84%) [10,11,14,16]. It may be an overestimation of real figure in this study as this study population consisted of only newly diagnosed persons in affected area which is dominated by those aged below 50 years, and the absence of HIV positives in the study population. According to published studies, the poor conversion is associated with resistant and transferred cases, and HIV positive patients [9-13,15,17]. Although the conversion rate, in this study, was higher than other study conducted in Portugal which was 74.6% [12]; it was lower than Taiwan (89.9%) [20], Gambia (90%) [21], and Cameroon (92.4%) [22]. Variation in sputum conversion rates in different studies may also be attributed to differences in geographic locations, baseline bacillary load, study methods, and other non-explored factors.

This study revealed the higher bacilli density (AFB3+) as a single independent predictor of delayed sputum conversion at the end of two months of treatment. Heavy bacterial load at initial testing was documented as an important risk factor of delay in sputum smear conversion at the end of intensive phase of TB treatment in the literature. Several studies have found high bacterial load as an important predictor of delayed conversion (Iran and Portugal – AFB2+ and AFB3+ [11,12], and Pakistan, Burkina Faso, and Rwanda – AFB3+ [10,13,14]). The patients with heavy bacterial load are at higher risk of delayed sputum conversion, development of MDR TB, and associated poor outcomes of treatment [11,23,24]. Hence, all TB patients should be identified and treated at the earliest to avoid delayed conversion. Further research is required to detect the determinants of heavy initial bacterial load among smear-positive pulmonary TB patients.

In addition to heavy bacterial load, an association between delayed sputum smear conversion and old age, female sex, and haemoptysis was also identified. Although, these factors did not achieve significance level in the multivariate analysis; it may be due to smaller sample size. Many studies in the past have reported that sputum conversion was higher in young patients [9-13,15,17]. Behnaz F et al., in Iran, and Caetano Mota P et al., in Portugal have also reported the association between the old age and the delayed conversion [11,12]. Poor conversion in old age might be due their increased incidence of physical disabilities, lack of efficacious bacilli clearance due to low immune response, and delay in seeking diagnosis and care. The co-morbidities associated with old age affecting anti-TB drug absorption and metabolism might be interfering with smear conversion in these patients [11].

In this study, TB was slightly preponderant among females (52.2%). This finding is similar to the WHO estimates (2016) for Afghanistan, where 63% of all newly diagnosed TB cases were women [5]. In contrast to this findings, studies in Iran, India, West Africa, and Portugal found that male sex was significantly associated with the delayed conversion [9,11,12,14]. The exact reason is not known but the severity of disease, illiteracy, and delayed presentation may play an important role. Cultural and financial barriers can act as major obstacles for women in seeking care, which results in delayed presentation and more severe illness especially in the conflict zones [5]. Women and children are the main victims of conflicts as their access to health care are limited [25].

Similar to these findings, a retrospective study conducted in Thailand have also reported that sputum smear conversion failure was associated with haemoptysis [26]. However, a study conducted in Cameroon found contradictory findings [27]. There is limited research which explored the association of symptoms with sputum conversion, thus more research and prospective studies are required to confirm these findings.

In this study, the patients with cavitary and bilateral lesion showed delayed conversion, but it was not statistically significant. Previous studies from India found that cavitary lesion and bilateral involvement was significantly associated with delayed conversion [9]. Small sample size and retrospective design of our study limit us on making conclusions on this finding in our study.

This study was a secondary analysis of TB patients’ treatment follow-up data. Therefore, it is difficult to comment on the causality of identified determinants. Being routine data of health services, this data set suffered from incompleteness, inaccuracy, and missing data such as chest radiograph findings, however, the authors carefully examined all registers to minimise these limitations. In addition, the potential risk factors not found in the registers were not assessed such as vitamin D deficiency, diabetes, nutritional status, etc. Although, this study has a small sample size which restricts its scope to generalise the findings; it is the first study of its kind from Afghanistan populations.

In this study, the delayed sputum conversion rate of newly diagnosed smear-positive pulmonary TB patients was 17.4% at the end of the intensive phase. The sputum microscopy finding AFB3+ was the single independent factor that could significantly predict sputum conversion. Hence, the policymakers in Afghanistan should take a note of the results of this study to accelerate the TB control in the country and the need of prospective studies to identify other risk factors associated with delayed conversion and unfavorable treatment outcomes.