With an annual incidence of 260 tuberculosis (TB) cases per 100,000 population [1], TB continues to threaten the lives of many people in Malawi. In the last 10 years, the number of previously-treated TB cases increased from 5% to 11% (Malawi National TB Programme,, unpublished data). This is of concern as these patients may acquire drug resistance and, therefore, have poor treatment outcomes [2][3]. The Malawi National TB Programme (NTP) diagnoses, registers and treats all forms of TB following international guidelines [4][5]. Previously-treated TB patients are classified into four categories: relapse, failure, re-treatment after default (RAD) and “other”. According to Malawi National TB guidelines, “other” includes chronic re-treatment TB patients, patients with recurrent smear-negative pulmonary TB and extra-pulmonary TB [4].

The categorization of TB patients as “other” may indicate poor patient management. According to the World Health Organization's (WHO) Global Tuberculosis Control report on relapse, failure, RAD and “other” for Africa, South-East Asia and Western Pacific regions in 2007, about 35% of 581,000 TB re-treatment cases were registered as “other” [1]. Although the WHO report provides information on the number of patients categorised as “other”, there are limited data on treatment regimens or treatment outcomes for this group. A study conducted in India found that TB patients categorized as “other” were treated with a category II regimen and had better treatment outcomes compared to relapse, failure and RAD [6]. In a routine programme setting, information on treatment regimen and outcomes is important as inappropriate management of TB patients can lead to patients remaining infectious and becoming a potential source of drug resistance [7].

We therefore carried out a study at the largest TB registration centre in Malawi to determine the numbers and proportions of adult TB re-treatment patients categorised as “other”. We also documented what treatment regimens were used and patient treatment outcomes for “other” TB patients, comparing them to outcomes reported for previously-treated patients registered as relapse, failure and RAD.

This is the first study from a routine TB programme setting in sub-Saharan Africa to explore characteristics and treatment outcomes in previously-treated adult TB patients registered as “other”. In this setting, 65% of the re-treatment cases were classified as “other”. Among “other”, 75% had smear-negative PTB, 75% were HIV-positive and most were men older than 34 years. The study also showed a high proportion of all TB cases had smear microscopy results, indicating the TB programme performed well. According to Malawi national TB guidelines, a re-treatment regimen (2SRHZE/1RHZE/5RHE) should have been used for all re-treatment cases. More than half of the patients classified as “other” were inappropriately treated with the first-line TB regimen (2RHZE/4RH). However, overall treatment success rate was low among re-treatment TB patients receiving a category II regimen, suggesting that this regimen may also not be the most appropriate.

There are several possible explanations for the large number of “other” patients in this study. Many of these patients had smear-negative TB and were HIV-positive, raising questions about whether these were truly recurrent smear-negative PTB as a result of re-infection or reactivation [8], or whether they had HIV-related disease that was misdiagnosed as TB [9][10]. It is also possible that these patients had other pulmonary or cardiac disease that was misdiagnosed as TB [11]. Similar misdiagnosis might have also occurred among EPTB cases because of their atypical presentations. A previous study conducted in Malawi found that 16% of patients with an EPTB diagnoses did not have TB [12]. Adoption of new accurate diagnostic tools such as GeneXpert MTB/RIF might help to reduce misdiagnosis in this group of patients. We also noted uncertainty among some TB clinical officers in other TB clinics as to how the “other” category should be defined. Interestingly, WHO TB treatment guidelines considers the “other” category as a variable group: “other” describes patients who returned with smear-negative PTB or EPTB, previously treated but with unknown outcome, or unknown whether they were ever treated or not [13]. With such a sizeable and heterogeneous group of “other” cases, more effort is needed to classify these cases into clear, exhaustive, mutually-exclusive categories. It may be worthwhile creating additional disease categories from the “other” group such as “presumed re-treatment TB” for patients with unknown previous treatment status; “smear-negative re-treatment TB” for patients who return with smear-negative PTB or EPTB; and “other re-treatment” for previously-treated patients with unknown outcomes. Overall, better diagnostic algorithms, improved access to mycobacterial culture facilities or rapid and accurate TB diagnostic tests are needed to reduce misdiagnosis and poor management.

Similar to other studies [14], [15], we also observed low treatment success rate among re-treatment cases. The most important reasons for low treatment success rates were loss to follow-up (20%) and death (11%). The high proportion of patients lost to follow-up might indicate a need for closer follow-up of this group during TB treatment as loss to follow up may include hidden deaths. Also, despite the overall low treatment outcome rates, the treatment success rate for smear-positive relapse was significantly higher than the treatment success rate in “other” TB patients after adjusting for sex and age. There are various possible explanations. Relapse patients have microbiologically-proven TB in contrast to “other” cases, which, as discussed above, may be a hybrid of true TB, HIV-related disease and other pulmonary or cardiac disease. Given high HIV infection rates in “other” cases and the inclusion of conditions that may not respond to TB treatment, only a moderate proportion might improve clinically (possibly due to other concomitant management). This is consistent with the 63% TB treatment success rate in the “other” group, and the inappropriate use of first-line TB regimen would not have adversely affected clinical outcomes in such patients. It is also likely that many HIV-infected smear-negative “other” TB category patients were more immune-compromised than relapse smear-positive PTB patients, resulting in poorer outcomes [16]. As this study was retrospective, we have no means of further investigating these issues. Generally, the re-treatment group had low re-treatment failure rates which may indicate a low prevalence of MDR-TB, although this does require more in depth study. Our results also showed that males had lower treatment success rates and were more likely to default or die while on TB treatment, and this may reflect low compliance with TB treatment therapy [17], [18].

Although neither the WHO nor The International Union Against Tuberculosis and Lung Disease TB treatment guidelines provide clear guidance on what regimen should be given to patients classified as “other”, the current Malawi National TB treatment guidelines recommend re-treatment drug regimens for all “other” TB patients. However, less than half of “other” TB patients in our study were treated with a category II regimen, suggesting clinician's uncertainty about this category of TB patients, rather than a general failure to follow guidelines. Other countries treat “other” TB patients with a first-line regimen if their smear is negative because they are considered to be at lower risk of failure and developing resistance. However, some previous studies observed generally low treatment success rates among smear positive re-treatment TB patients treated with the category II regimen [15], [19]. Given the poor treatment outcomes also observed in these “other” re-treatment groups, it is certainly worthwhile and important to review the current category II regimen for re-treatment TB patients. These patients may be at high risk of drug resistance and of developing MDR, as was seen in South Africa [20].

Our findings should be viewed with the following limitations. First, 52% of treatment cards for re-treatment patients were missing although all treatment cards should have been returned to the MPC TB registration centre at the time of this study. Some cards might have been lost at the peripheral sites or within the TB registry, introducing bias if patients without treatment cards had different outcomes than those with treatment cards. However, we observed no differences in terms of demographics, TB classification and place of treatment between patients with and without treatment cards. Second, although TB outcome data were supposed to be recorded on treatment cards and in the TB register, a considerable proportion of adverse outcomes were not recorded in the TB registers. As treatment cards are the primary source of outcome information, we based the analysis on outcomes documented on the cards and not the register. Exclusion of patients with missing cards may have resulted in a biased selection of patients with better treatment outcomes, assuming that cards of defaulters were more likely to be lost. Despite these limitations, the study findings are useful to inform policy and programs in Malawi and other comparable settings to improve care for TB patients.

In conclusion, our study shows that the “other” TB category constitutes a sizeable population of re-treatment patients and is poorly managed. Accurate definitions and categories from the current “other” category should be created to avoid poor patient management. The high proportion of missing treatment cards suggests that less priority is given to these re-treatment patients. Treatment success may be improved by adoption of more accurate diagnostic tools such as Xpert MTB/RIF and more rigorous training and regular supervision of TB staff. There is also a need to reappraise the category II regimen as our results suggest that use of this regimen is not associated with successful outcomes. Some of these issues are currently under discussion within the Malawi NTP with the aim of improving routine programme management.