Association of HIV infection and antiretroviral therapy with the occurrence of an unfavorable TB treatment outcome in a rural district hospital in Eastern Cape, South Africa: A retrospective cohort study.

BACKGROUND: Our objective was to assess differences in TB treatment outcomes between individuals who were HIV negative, HIV positive on anti-retroviral treatment (ART) and HIV positive not on ART, at TB treatment initiation at a rural district hospital in Eastern Cape, South Africa.
METHODS: This was a retrospective cohort study of individuals diagnosed with TB between January 2017 and April 2020 at a district hospital. Adults 15 years and over with reported HIV status and treatment outcome were included (N = 711). A categorical outcome with three levels was considered: unfavorable, down referral, and success. We report descriptive statistics for the association between HIV and ART status and treatment outcome using Chi-square and Fisher's exact tests. A multinomial baseline logit model was used to estimate odds ratios for treatment outcomes.
RESULTS: Overall, 59% of included patients were HIV positive with 75% on ART. Eighty-eight patients 12% had an unfavorable outcome. Half of all patients were down referred with an additional 37% having a successful outcome. Individuals without HIV were more likely to be down referred (versus unfavorable) compared to individuals with untreated HIV (2.90 OR, 1.36, 6.17 95% CI). There was a greater likelihood for individuals without HIV having a successful TB treatment outcome compared to individuals with untreated HIV (4.98 OR, 2.07, 11.25 95% CI).
CONCLUSION: The majority of individuals had positive TB treatment outcomes (down referred or success). However, people without HIV had nearly five times greater odds of having successful outcomes than those with untreated HIV.

Background

Tuberculosis (TB) is the leading cause of infectious disease death worldwide, and in South Africa (WHO, 2019) [1]. Nearly one in every four deaths among people living with HIV is attributable to TB [1]. South Africa has an HIV prevalence rate of 20.4% among adults, 8.7% among youth 15 to 24 years, and 2.7% among children < 2 years and has more cases of HIV than any other country, with 7.7 million known HIV infections in 2018 [2, 3]. In a recent country-wide prevalence survey, it is estimated that the prevalence of bacteriologically confirmed pulmonary TB in South Africa is 852 (95% 679–1,026) per 100,000 population among individuals 15 years and older [4]. Therefore, South Africa has a dual burden of both HIV and TB, with nearly 60% of individuals diagnosed with TB disease being co-infected with HIV [3]. Yet, despite the incredibly high prevalence of HIV in South Africa, TB (and TB/HIV) continues to kill more South Africans annually than HIV alone [1, 5].

TB disease is the most serious and common opportunistic infection affecting people living with HIV [6-8]. However, since the era of antiretroviral therapy (ART), individuals dually infected with HIV and TB have had increased survival rates [6, 9, 10]. There is currently a lack of evidence regarding HIV status, ART status (or suppressed viral loads) and TB treatment outcomes [11-13]. However, ART is known to modify the effect of HIV on TB treatment outcomes [14]. A systematic review and meta-analysis of early ART initiation among patients with TB found that early ART initiation during TB treatment was associated with reduced all-cause mortality; however, it also had an increased rate of TB-associated immune reconstitution inflammatory syndrome (IRIS) and death related to TB-IRIS especially if started within the first two weeks of initiating TB treatment [6, 15, 16].

The advent of ART has changed the landscape of HIV [17]. However, among people living with HIV, the risk of developing TB is still 26 times higher than the general population [1]. Challenges for co-treatment remain; however, recent literature describes advances in effective drugs, more sensitive and rapid molecular testing for TB, and use of shorter regimens have improved outcomes among individuals co-infected with TB and HIV [18].

The objective of this study was to assess the relationship between HIV and ART status with TB treatment outcomes among a cohort diagnosed with TB in a rural, high HIV prevalence setting.

Methods

Study design and study population

Data for this study were taken from TB medical charts of a cohort of individuals 15 years or older who were diagnosed with drug-sensitive TB and who should have completed treatment at the district hospital between January 4, 2017 and April 27, 2020 (N = 1006) (Fig 1). There were 295 individuals excluded due to no documented treatment outcome (n = 246) or no HIV status reported (n = 49) for a final sample size of 711. We excluded children < 15 years due to United Nations age group definitions, differing treatment guidelines and outcomes among children, and because individuals < 15 years often are seen in pediatric settings [19-21]. Additionally, the vast majority of adults ≥15 years with drug sensitive TB in South Africa are managed with the same four drug regimen for the first two months of the “intensive phase” of treatment: rifampicin, isoniazid, pyrazinamide and ethambutol (RHZE), followed by four months of rifampicin and isoniazid (RH) for the “continuation phase” as per the South African TB guidelines [5].

Fig 1

Population flowchart.

Setting

This study took place in the King Sabata Dalindyebo (KSD) sub-district of the OR Tambo District, rural Eastern Cape, South Africa. This is one of the poorest districts in the country and one of ten districts countrywide with the highest dual burden of HIV and TB [5, 22]. The Eastern Cape has a TB incidence of 839/100,000 in 2017, higher than the country prevalence estimated at 737/100,000 [4]. Additionally, the Eastern Cape has a high overall HIV prevalence rate (25.2%) as of 2017 [22-24]. This district hospital and its surrounding 12 clinics serve a deeply rural area of the Eastern Cape with a catchment population of approximately 127,500 people [24]. Access to clinics, and particularly to the district hospital, is challenging due to the hilly geography of the area, a lack of road infrastructure, and household financial constraints. As of the latest Census data from 2011, the annual median household income is R14600 (US$1000), with only 9.7% of adults having employment, 9.6% unemployed, 10.5% discouraged work seekers, and 70.2% who are not economically active [25].

Data sources and study measures

Baseline demographic and medical history were collected through self-report and an interview with a healthcare provider as the TB file was started, GeneXpert® Ultra was performed on most patients and when required, drug-susceptibility testing (DST) results for first- and second-line TB drugs were obtained through sputum collection sent to the South African National Health Laboratory Services (NHLS). TB disease was diagnosed by clinical examination, chest x-ray, GeneXpert® Ultra or sputum smear microscopy. HIV was self-reported or confirmed with two rapid cartridge HIV antibody tests, with ELISA blood test if discrepancy between rapid tests [26]. All data were entered into and downloaded from REDCap, an electronic data management system [27].

Treatment outcome definitions are in accordance with the WHO 2013 revised definitions and reporting framework for tuberculosis [28]. Composite categorical outcomes with three levels were created: unfavorable, down referral, and success. Unfavorable outcome was defined as loss to follow-up, death, or treatment failure. Loss to follow-up (LTFU) is defined as a TB patient who did not start treatment or whose treatment was interrupted for two consecutive months or more and thus considered an unfavorable outcome, aligning with existing TB literature [29, 30]. Success was defined as cure or completed treatment. Down referral refers to patients who are down referred to be treated at their nearest primary care clinic where we could no longer follow them. The outcomes for these patients were therefore uncertain; however, down-referred patients are typically stable and doing well on treatment, such that their clinicians believe they can successfully be treated at their local clinic. Down referral is strongly encouraged and ensures that patients are treated closer to their homes, enabling better adherence. It is seen as a positive outcome and helps TB care to be more patient-centered and decentralized [5, 31]. The exposure of interest was categorical HIV and ART status: HIV infection on ART at the initiation of TB treatment (treated HIV), HIV infection but not on ART at initiation of TB treatment (untreated HIV), and no HIV infection. Information on covariates was also collected from medical histories in the medical chart including: age, sex, smoking status, alcohol use, prior/current employment in a mine, and presence of comorbidities (hypertension, diabetes, epilepsy, mental illness, liver disease, and renal insufficiency). Missing values for covariates were given their own category of “Missing”, often when a clinician did not ask or did not tick “yes” or “no” in the medical chart. For missing HIV status, often it was due to tests not being available on the day a patient was seen, and then no HIV status was recorded upon testing on a subsequent day after the initial baseline data collection. Finally, a final treatment outcome is not always recorded by clinicians in the medical chart. Therefore, we believe these missing data points are missing at random.

Statistical analysis

We first report descriptive statistics for the association between HIV and ART status and the measured demographic and clinical covariates, including treatment outcome, using Wilcoxon signed rank and Fisher’s exact tests. Demographic covariates include age, sex, past employment in a mine, tobacco use, and alcohol use. Clinical covariates include GeneXpert® Ultra for diagnosis, TB type, and multiple comorbidities.

We use a multinomial baseline logit model to estimate the odds ratios for treatment outcome (reference level: unfavorable) and HIV and ART status (reference level: no HIV infection). We adjust for age, sex, TB type, tobacco use, and alcohol use as these are known confounders of the relationship between HIV and ART status and TB treatment outcomes [22, 32, 33]. We categorized age as 15–24, 25–54, 55–65, and 65+ to capture the non-linear relationship between age and TB treatment outcomes and in standardization with other reports of TB treatment outcomes (Fig 2) [4, 34]. Missing values for TB type, alcohol, and tobacco use were made into their own “Unanswered” category.

Fig 2

TB treatment outcomes by age group (N = 936).

To test our assumption that the treatment outcomes and HIV status were missing completely at random, we provide supplementary analyses (S1 File). First, we compared the distribution of demographic characteristics between those included and excluded from the main analysis. In addition, we used multiple imputation by chained equations to create 20 datasets with imputed values for missing variables and pool the results from the main analyses. This will provide alternate odds ratio estimates assuming the data are missing conditionally at random. R V3.6.0 was used for data cleaning and statistical analyses with the tidyverse, nnet, and mice software packages.

Ethics approval

Ethics approval was provided by Harvard Medical School (IRB19-0605) and Walter Sisulu University (049/2019) as well as the Eastern Cape Department of Health (EC_201908_011). Data used in this retrospective study were identifiable upon collection; however, were fully deidentified and anonymized before statistical analysis. Both IRBs waived informed consent.

Results

Overall, 418/711 (58.8%) of included patients were HIV positive with 75% of these patients on ART (Table 1). The median age of patients was 38 years (interquartile range 30–57) with those who were HIV negative being more likely to be older (p<0.001). Half (50.0%) of patients were male, with males less likely to be HIV positive and on ART compared to females, but also more likely to not be on ART if HIV positive (p<0.001). The majority (62.0%) of patients had pulmonary TB, while 30.0% had extra pulmonary TB, 4.4% had both pulmonary and extra pulmonary, and 3.7% had TB type not recorded. Most (70.2%) patients were diagnosed using GeneXpert® Ultra, while 66.8% had a chest x-ray to assist with or confirm diagnosis. The tobacco use, alcohol use, and prior or current mining work variables were unanswered (missing) for about one third of the cohort, though the rates of missing data did not vary significantly by HIV and ART status. Among those who answered these questions, 20.0% reported tobacco use, 22.9% reported alcohol use, and 9.9% reported mining work. HIV negative persons were more likely to have been involved in mining work (p<0.001); there were no differences by alcohol or tobacco use. Additionally, about one third of patients did not have recorded non-HIV comorbidity status. Of those with known comorbidity status, 15.4% had hypertension, 1.6% diabetes, 3.7% epilepsy, 0.6% mental illness, 0.2% liver disease, and 1.0% renal insufficiency.

Table 1

Demographics of patients diagnosed with TB at a district hospital based on HIV and ART status (N = 711).

	HIV+, no ART (N = 104)	HIV+, ART (N = 314)	HIV- (N = 293)	Total (N = 711)	P-value1
Age					<0.001
Median (IQR)	32 (28, 44)	37 (31, 45)	53 (28, 67)	38 (30, 57)
Male Sex	59 (56.7)	130 (41.5)	166 (56.7)	355 (50.0)	<0.001
Tobacco Use					0.330
No	55 (52.9)	180 (57.3)	170 (58.0)	405 (57.0)
Yes	19 (18.3)	39 (12.4)	41 (14.0)	99 (13.9)
Unanswered	30 (28.8)	95 (30.3)	82 (28.0)	207 (29.1)
Alcohol Use					0.303
No	52 (50)	168 (53.5)	166 (56.7)	386 (54.3)
Yes	21 (20.2)	51 (16.2)	42 (14.3)	114 (16.0)
Unanswered	31 (29.8)	95 (30.3)	85 (29.0)	211 (29.7)
Mining status					0.001
No	57 (54.8)	192 (61.1)	158 (53.9)	407 (57.2)
Unanswered	41 (39.4)	113 (36.0)	105 (35.8)	259 (36.4)
Yes	6 (5.8)	9 (2.9)	30 (10.2)	45 (6.3)
Hypertension					<0.001
No	62 (59.6)	185 (58.9)	171 (58.8)	418 (58.8)
Yes	5 (4.8)	11 (3.5)	60 (20.5)	76 (10.7)
Unanswered	37 (35.6)	118 (37.6)	62 (21.2)	217 (30.5)
Diabetes					0.229
No	66 (63.5)	194 (61.8)	221 (75.4)	481 (67.7)
Yes	1 (1.0)	1 (0.3)	6 (2.0)	8 (1.1)
Unanswered	37 (35.6)	119 (37.9)	66 (22.5)	222 (31.2)
Epilepsy					0.945
No	65 (62.5)	188 (59.9)	220 (75.1)	473 (66.5)
Yes	2 (1.9)	8 (2.5)	8 (2.7)	18 (2.5)
Unanswered	37 (35.6)	118 (37.6)	65 (22.2)	220 (30.9)
Mental illness					1.000
No	66 (63.5)	195 (62.1)	225 (76.8)	486 (68.4)
Yes	0 (0)	1 (0.3)	2 (0.7)	3 (0.4)
Unanswered	38 (36.5)	118 (37.6)	66 (22.5)	222 (31.2)
Liver disease					0.135
No	65 (62.5)	195 (62.1)	228 (77.8)	488 (68.6)
Yes	1 (1.0)	0 (0)	0 (0)	1 (0.1)
Unanswered	38 (36.5)	119 (37.9)	65 (22.2)	222 (31.2)
Renal insufficiency					0.355
No	64 (61.5)	193 (61.5)	225 (76.8)	488 (68.6)
Yes	1 (1.0)	3 (1.0)	1 (0.3)	5 (0.7)
Unanswered	39 (37.5)	118 (37.6)	67 (22.9)	224 (31.5)
TB type					0.050
Pulmonary	70 (67.3)	185 (58.9)	186 (63.5)	441 (62.0)
Extra pulmonary	22 (21.2)	109 (34.7)	82 (28.0)	213 (30.0)
Both	8 (7.7)	12 (3.8)	11 (3.8)	31 (4.4)
Unanswered	4 (3.8)	8 (2.5)	14 (4.8)	26 (3.7)
GeneXpert done					0.007
No	31 (29.8)	111 (35.4)	69 (23.5)	211 (29.7)
Yes	73 (70.2)	203 (64.6)	223 (76.1)	499 (70.2)
Unanswered	0 (0)	0 (0)	1 (0.4)	1 (0.1)
Chest x-ray done					<0.001
No	39 (37.5)	129 (41.1)	67 (22.9)	235 (33.1)
Yes	65 (62.5)	184 (58.6)	226 (77.1)	475 (66.8)
Unanswered	0 (0)	1 (0.3)	0 (0)	1 (0.1)

1 Wilcoxon rank sum (age) and Fisher’s exact test were used to test for differences in distribution by HIV/ART category. Unanswered values were excluded from testing.

Down referral

Half of all patients were down referred (n = 359, 50.5%) with an additional 264 (37.1%) having a successful outcome–thus, the majority of all (87.6%) patients included in our cohort had a positive treatment outcome. Patients with HIV and on ART were more likely to be down referred (vs. unfavorable outcome) than patients with untreated HIV (1.64 OR 0.86, 3.11 95%CI) (Table 3). Individuals without HIV were much more likely to be down-referred compared to individuals with untreated HIV (2.90 OR, 1.36, 6.17 95% CI). When comparing individuals with treated HIV, individuals without HIV were also more likely to be down referred (1.77 OR, 0.93–3.35 95% CI). When imputing missing data, we found similar estimated odds ratios and 95% CIs (S2 Table in S1 File).

Table 3

Estimated odds ratios for Down referral vs. Unfavorable and Successful vs. Unfavorable TB treatment outcome (N = 711).

	Estimated OR for Down referral vs. Unfavorable TB Treatment Outcome		Estimated OR for Successful vs. Unfavorable TB Treatment Outcome
Variable	Odds Ratio and 95% CI	p-value	Odds Ratio and 95% CI	p-value
HIV+, no ART	1 (Reference)		1 (Reference)
HIV+, ART	1.64 (0.86, 3.11)	0.065	2.29 (1.13, 4.63)	0.010
HIV-	2.90 (1.36, 6.17)	0.003	4.98 (2.21, 11.25)	<0.001
Age category
15–24	1 (Reference)		1 (Reference)
25–54	0.73 (0.3, 1.77)	0.244	0.83 (0.33, 2.08)	0.347
55–64	0.53 (0.19, 1.46)	0.108	0.42 (0.14, 1.21)	0.054
65+	0.21 (0.08, 0.56)	0.001	0.31 (0.11, 0.83)	0.01
Sex
Female	1 (Reference)		1 (Reference)
Male	0.89 (0.53, 1.48)	0.323	0.92 (0.54, 1.56)	0.378
TB Type
Pulmonary	1 (Reference)		1 (Reference)
Extra Pulmonary	0.71 (0.4, 1.24)	0.114	1.47 (0.83, 2.59)	0.094
Both	0.34 (0.12, 0.98)	0.023	1.01 (0.38, 2.74)	0.49
Unanswered	0.57 (0.19, 1.72)	0.158	0.5 (0.15, 1.73)	0.138
Alcohol use
Yes	1 (Reference)		1 (Reference)
No	0.34 (0.05, 2.34)	0.136	0.44 (0.06, 3.24)	0.208
Unanswered	1.42 (0.55, 3.69)	0.234	1.38 (0.51, 3.74)	0.261
Tobacco use
Yes	1 (Reference)		1 (Reference)
No	1.39 (0.2, 9.77)	0.372	1.02 (0.13, 7.65)	0.494
Unanswered	0.66 (0.26, 1.69)	0.195	0.49 (0.18, 1.31)	0.078

Success

Among 711 patients with HIV status reported, only 12.4% (n = 88) had an unfavorable outcome, 37.5% of these were LTFU (Table 2). Individuals on ART at initiation of TB treatment were more likely to have a successful outcome compared to individuals with untreated HIV (2.29 OR, 1.13, 4.63 95% CI) (Table 3). In addition, there was an even greater likelihood for individuals without HIV having a successful TB treatment outcome compared to individuals with untreated HIV (4.98 OR, 2.07, 11.25 95% CI). When comparing individuals with treated HIV, individuals without HIV were also more likely to have a successful TB treatment outcome (2.18 OR, 1.13–4.20 95% CI). When imputing missing data, we found similar estimated odds ratios and 95% CIs (S2 Table in S1 File).

Table 2

TB treatment outcome by HIV and ART status (N = 711).

	HIV+, no ART (N = 104)	HIV+, ART (N = 314)	HIV- (N = 293)	Total (N = 711)
Down referred	56 (53.8)	160 (51.0)	143 (48.8)	359 (50.5)
Success	27 (26.0)	116 (36.9)	121 (41.3)	264 (37.1)
Cure	2	8	11	21
Completed treatment	25	108	110	243
Unfavorable	21 (20.2)	38 (12.1)	29 (9.9)	88 (12.4)
Died	15	26	12	53
Treatment failure	0	0	2	2
Lost to follow up	6	12	15	33

Discussion

In this analysis, we observed high rates of positive outcomes (either success or down referral)– 88%—among all drug sensitive TB patients regardless of HIV or ART status. Compared with other studies, and South African statistics, these rates are impressive [35, 36]. South Africa estimates a nearly 54% treatment success rate among the burden of individuals with drug-sensitive TB, and a 52% treatment success rate among people co-infected with TB and HIV [35].

We found that individuals without HIV infection had nearly five times greater odds of having a successful outcome compared with individuals with untreated HIV. This disparity highlights the importance of integrating HIV and TB care–i.e. testing and treating HIV within TB programs and vice versa–and of paying special attention to patients with HIV who have not yet initiated ART when TB treatment is started [37]. Additionally, although we were unable to measure adherence in this study, other studies in South Africa have found that people receiving concurrent treatment for HIV/TB are highly adherent [38].

A recent study from Botswana found 88–91% of patients co-infected with HIV/TB were successfully treated for TB [39]. Dolutegravir-based ART was associated with favorable TB treatment outcomes and high rates of viral load suppression were found across all regimen categories. Another recent study from Nigeria found that 84% of patients co-infected with HIV/TB were successfully treated for TB [40]. Successful treatment was associated with being newly registered for TB, receiving TB treatment for the first time, and being treated at a private health facility. Thus, other studies have also found high rates of treatment success among patients co-infected with TB and HIV.

In an Italian cohort of 246 patients with HIV-associated TB, they found that ART initiation during TB treatment was associated with a substantial reduction of death compared to not starting ART; however, even patients already on ART when they developed TB disease remained at an elevated risk of death [41]. The data available to us in this study did not allow us to determine whether ART was initiated, nor the timing of ART initiation; however, clinicians at the hospital responsible for TB and ART patients report that the majority of patients who are naïve at the start of TB treatment remain off of ART until at least two weeks past initiation of TB treatment in order to avoid immune reconstitution inflammatory syndrome (IRIS) associated complications [42, 43]. Per conversations with treating clinicians at this rural district hospital, most patients diagnosed with HIV and TB at the same time initiate ART within two months post TB initiation, depending on their CD4 count.

As mentioned above, a down referral is considered a positive outcome. Typically, ‘healthier’ patients are down referred as they are seen as individuals that can be managed at the clinic level, usually by a nurse. At times, patients who are fairly unwell are also referred to their local clinic, especially when they live far away from the district hospital and transport costs to return are prohibitive, but the sickest patients are typically retained for hospital follow-up and provided with transport money by a local NGO. The decentralization of TB care has been encouraged due the preference by patients to receive care closer to home, as well as the cost-effectiveness of decentralized models of care in South Africa [44]. This cohort of TB patients had an impressive rate of down-referral, which could be seen as a positive indication of the overall health system in this setting working effectively and efficiently.

HIV continues to drive the TB epidemic in South Africa, where nearly 60% of patients are co-infected [1]. Setting policy goals and initiatives to integrate TB and HIV care is important. South Africa has been a champion of HIV care—specifically increasing HIV testing and increasing initiation of ART [5]. Due to the dual high burden of TB and HIV in the country, integrated TB/HIV care has also improved substantially in the past decade [17, 45–47]. However, more must be done for individuals suffering from dual infection, especially when ART has not yet been established. In the recent country-wide TB prevalence survey, they found that 56% of HIV positive participants had not sought care for their TB symptoms compared to 69% of individuals who were HIV negative [4]. It is also important to advocate for early and regular HIV and TB screening for all South Africans–especially men, who often have delays in seeking care and barriers to healthcare access [4]. Providing more male friendly health services are needed.

Limitations

First, a major limitation in this cohort was the number of individuals we excluded due to missing treatment outcomes. This is always a risk when using retrospective data and routinely collected data. In our main analysis, we excluded 295 (29%) of individuals because of missing treatment outcome values or missing HIV status because we believe these are missing at random. Our supplemental analysis (S1 File) did not find any difference in demographic characteristics between the individuals excluded from the analysis, but did find that excluded individuals were slightly more likely to have extra pulmonary TB (37.6% vs. 30.0%). When we repeated our analysis with imputed missing values, we found no meaningful difference from the complete case analysis presented in the main text.

Second, we considered LTFU as an unfavorable outcome to align with existing TB research that find people who are LTFU have worse treatment outcomes [30]. Unfortunately, no information about the course of TB (nor HIV) treatment was available among patients in order to conduct a more granular time-to-event analysis, potentially accounting for informative censoring. Third, measures of adherence were not collected in the study, so we were unable to investigate the potential mediating role of adherence on the relationship between HIV/ART status and treatment outcomes [48].

Additionally, we did not have laboratory data to further investigate CD4 count and HIV viral loads to understand how each of those play a role in TB treatment outcomes [49]. We also were unable to assess timing of treatment initiation for either TB or HIV—both of which can have significant impact on treatment outcomes [6, 50]. Similarly, we do not know how many (if any) of the people living with HIV not on ART may have started ART during their TB treatment. It is likely that many individuals did start ART while engaged with care per South African guidelines; however, this was not captured longitudinally [26].

Conclusion

Despite positive outcomes among most TB patients, HIV positive patients who are not on ART when TB treatment is initiated, do significantly worse than patients without HIV and patients with treated HIV. Therefore, patients with untreated HIV should be seen as a high-risk group, be swiftly followed up and the biopsychosocial aspects of their care need to be carefully considered. Ensuring that patients with HIV/TB coinfection are correctly treated for both diseases is paramount and the drive for the integration of TB/ART services in South Africa needs to continue.

Supplemental analysis.

Investigating missing outcomes.

(DOCX)

Click here for additional data file.

15 Sep 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

G. K. Balasubramani

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In your ethics statement in the Methods section and in the online submission form, please provide additional information about the data used in your retrospective study. Specifically, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data from their medical records used in research, please include this information.

3. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

4. Please include a separate caption for each figure in your manuscript.

5. Please ensure that you refer to Figure 1 in your text as, if accepted, production will need this reference to link the reader to the figure.

Additional Editor Comments (if provided):

The above-referenced paper has been reviewed by two experts serving as peer reviewers for PLOS One.

After careful review, I am sending you the manuscript for significant revisions per the reviewers' request. One of the substantial concern from the reviewers are statistical analysis are not correctly handled. The reviewer's comments are appended below.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have examined the association of HIV treatment on the TB treatment outcome. The study is based on moderate sample size data from a hospital in South Africa. The authors mention that there is not much literature on HIV treatment vs TB treatment and hence this study has some insight for treatment of TB.

There are a number of issues associated with this manuscript and the presentation has to be improved.

The main concerns are:

There is no mention about the confounding effect to type of TB which could have influence on the outcome and on HIV treatment. It is not clear if this is considered in the current analysis. Also other covariates such as alcohol and tobacco use may also have an influence on the TB outcome.

Although ORs and CIs are for the main analysis are provided, some additional details on original parameter estimates, SE and the p-values for the HIV status variable and for the covariates in the model would be useful. Results in table 3 and 4 can be provided together as these are from single multinomial logit analysis.

The authors should briefly discuss the type of TB treatment which could be useful for the reader interested in TB research.

Table 1 has lot of additional information about covariates/illness but the details about the assessment of these conditions (whether self-reported or diagnosed) are missing. Separate row for p-value in this table can be omitted. Are these reported p-values for three group or two group comparisons? It is not easy to understand this as the authors discuss the results separately for HIV treatment groups.

Some additional minor comments are:

The figure 1 which is important for the reader to understand the numbers presented in the study design is not mentioned inside the text.

Does “unanswered” refer to missing data or refused to answer (if self-reported)?

It is not clear why the authors present the age results with 936 samples when all other results are based on 711 samples.

Incidence and prevalence estimates should have a time interval.

It is hard to follow the description provided in data source and study measures. The authors should describe the information about some of the self-reported measures.

Provide the reference for R software and the r-function used for the multinomial logit model.

Reviewer #2: A major limitation of this paper is the number of people excluded in this study (414 no treatment outcome, 77 no HIV status reported and 212 < 15 years). They start with 1429 subjects with TB and the analysis sample contains only 711 (48%). They claim that missing completely was at random but no results are present to substantiate this claim.

Lines 102 -106 talks about children < 15 years that are excluded from the study and not appropriate in the methods section

Line 111 - The TB incidence rate is mentioned as 838/100,000 per year. If this is the case, then wouldn't we expect much more cases during the study period (Jan 2017 to Apr 2020)? (even after excluding children). Were all TB cases in the Eastern Cape been accounted in this study?

Line 117 Is the adult unemployment population 91.5%? This appears to be very high. (9.9% reported mining work alone!)

Line 124 Why was the rapid cartridge HIV antibody test not done for the 77 patients with no HIV status?

Line 129-130 Who created the composite categorical outcomes? Why couldn't it be created for the 414 patients?

Line 153 - Why was the "unfavorable" outcome chosen as the reference category (which includes loss to follow-up)? Why can't the majority (88%) who had positive outcomes be the reference category?

Line 155- Why can't age be introduced as a continuous variable in the models? Dichotomizing a continuous variable arbitrarily with a cut-off as 45 years may not be appropriate. Is there any clinical evidence? Would childbearing age be appropriate for males too?

Table 1: Approximately 30% of most variables in Table 1 appear to be missing (unanswered) and are excluded from tests of association. Are these variables self-reported or taken from the clinical notes?

Tables 2 and 3 are not presented in a logical manner and need to be summarized properly

Line 206 - Why are all age reported (n=936) when the study includes only >= 15 years (n=711). They are not relevant to the study

Lines 218 - 223 The statements are ambiguous and not clear. Especially lines 222 & 223 are vague.

Line 274 - It is mentioned that 341 (34%) are excluded because of missing treatment outcomes. This is not in agreement with (n=414) presented in figure 1. Figure 2 should include only the study participants and not excluded children.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

20 Oct 2021

RE: PLOS ONE PONE-D-21-14796

Dear Dr. Balasubramani and reviewers,

Thank you very much for your thorough review and comments on our manuscript. We have revised the manuscript to be a much-improved version addressing all of your concerns.

Below we list how we have addressed each reviewer’s comment.

Ethics approval was provided by Walter Sisulu University in South Africa, the Eastern Cape Department of Health, and the Harvard Medical School Institutional Review Board.

Thank you for the opportunity to revise our manuscript. We look forward to hearing back soon.

Sincerely,

Brittney van de Water, on behalf of the authors

Editors Comments

Editor 1.1: Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming.

RESPONSE: Thank you. We have reviewed the style requirements and have named files appropriately.

Editor 1.2: In your ethics statement in the Methods section and in the online submission form, please provide additional information about the data used in your retrospective study. Specifically, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data from their medical records used in research, please include this information.

RESPONSE: Thank you. We have added into our Ethics statement “Data used in this retrospective study were identifiable upon collection; however, were fully deidentified and anonymized before statistical analysis. Both IRBs waived informed consent.”.

Editor 1.3: We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly.

RESPONSE: Thank you for this guidance. We will make a limited dataset available through Boston College’s Dataverse repository.

Editor 1.4: In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

RESPONSE: Thank you. As noted above, we have no limitations and data will be deposited into Boston College’s Dataverse repository.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see: http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. We will update your Data Availability statement on your behalf to reflect the information you provide.

RESPONSE: Many thanks.

Editor 1.4: Please include a separate caption for each figure in your manuscript.

RESPONSE: Thank you – we have added a separate caption for Figure 1 (and have omitted Figure 2 in line with reviewers’ comments).

Editor 1.5: Please ensure that you refer to Figure 1 in your text as, if accepted, production will need this reference to link the reader to the figure.

RESPONSE: Thank you, and apologies for omitting. We have included reference to Figure 1 in the first paragraph of the Methods section.

Reviewer 1 Comments

Reviewer 1.1: There is no mention about the confounding effect to type of TB which could have influence on the outcome and on HIV treatment. It is not clear if this is considered in the current analysis. Also, other covariates such as alcohol and tobacco use may also have an influence on the TB outcome.

RESPONSE: Thank you for this comment. We have now adjusted for TB type, alcohol use, and tobacco use in the analysis with minimal change, as detailed in the Methods, Statistical Analysis section.

Reviewer 1.2: Although ORs and CIs are for the main analysis are provided, some additional details on original parameter estimates, SE and the p-values for the HIV status variable and for the covariates in the model would be useful. Results in table 3 and 4 can be provided together as these are from single multinomial logit analysis.

RESPONSE: Thank you. We have added p-values to tables 3 and 4 and have combined them into just one table, Table 3. Note that we do not report standard errors as these are less interpretable on the odds ratio scale.

Reviewer 1.3: The authors should briefly discuss the type of TB treatment which could be useful for the reader interested in TB research.

RESPONSE: Thank you. We have added “The vast majority of adults > 15 years with drug sensitive TB in South Africa is managed with four drugs for the first two months of the “intensive phase“ of treatment: rifampicin, isoniazid, pyrazinamide and ethambutol (RHZE), followed by four months of rifampicin and isoniazid (RH) – the “continuation phase”, as per the South African National TB guidelines.“ into the Methods section for background on types of TB treatments used during this period in South Africa.

Reviewer 1.4: Table 1 has lot of additional information about covariates/illness but the details about the assessment of these conditions (whether self-reported or diagnosed) are missing. Separate row for p-value in this table can be omitted. Are these reported p-values for three group or two group comparisons? It is not easy to understand this as the authors discuss the results separately for HIV treatment groups.

RESPONSE: Thank you for this comment. We added in the Methods, Data sources and study measures section that “Baseline demographic and medical history were collected through self-report and an interview”. To clarify, we have added into the Methods that p-values correspond to a Fisher’s exact test that compares across all three groups simultaneously (HIV negative, HIV positive on ART, and HIV not on ART). This can be interpreted as testing for differences in the distribution of HIV/ART for each demographic characteristic.

Reviewer 1.5: Figure 1 is important for the reader to understand the numbers presented in the study design is not mentioned inside the text.

RESPONSE: Thank you for noticing this. We have now referenced Figure 1 in the text at first mention in the Methods section, Study design and study population.

Reviewer 1.6: Does “unanswered” refer to missing data or refused to answer (if self-reported)?

RESPONSE: Thank you for this question. As described in the Methods, “unanswered” refers to questions that were typically not asked by the clinician. In trying to not “patient blame” we want readers to understand that often it is not because patients are non-forthcoming with their health status, rather often times it is the provider (nurse or doctor) who never asks the question and/or leaves it blank. Of note, these charts are from a low-resourced setting where clinicians are often hard pressed for time and charting is not always the priority. The sentence reads “Missing values for covariates were given their own category of “Missing”, often when a clinician did not ask or did not tick “yes” or “no” in the medical chart”.

Reviewer 1.7: It is not clear why the authors present the age results with 936 samples when all other results are based on 711 samples.

RESPONSE: Thank you for this comment. Our original analysis only focused on adults, but we chose to display results for individuals <15 years in Figure 2. We now realize that this caused confusion and have removed persons <15 years from Figure 2 and only focus on adults in this manuscript and Figure 2.

Reviewer 1.8: Incidence and prevalence estimates should have a time interval.

RESPONSE: Thank you for this concern. We have added dates for prevalence and incidence referenced in the Methods per the cited references.

Reviewer 1.9: It is hard to follow the description provided in data source and study measures. The authors should describe the information about some of the self-reported measures.

RESPONSE: Thank you for this comment. We have revised the Methods section which now reads “Baseline demographic and medical history were collected through self-report and an interview with a healthcare provider as the TB file was started…” and expands on self-reported measures.

Reviewer 1.10: Provide the reference for R software and the r-function used for the multinomial logit model.

RESPONSE: Thank you for catching this. We have now included the R software and R-function used for analysis in the Methods. We also include the R scripts necessary to run our analyses.

Reviewer 2 Comments

Reviewer 2.1: A major limitation of this paper is the number of people excluded in this study (414 no treatment outcome, 77 no HIV status reported and 212 < 15 years). They start with 1429 subjects with TB and the analysis sample contains only 711 (48%). They claim that missing completely was at random but no results are present to substantiate this claim.

RESPONSE: We appreciate your concern. We have revised the Methods to only describe individuals who were > 15 years. Therefore, it now reads “Data for this study were taken from TB medical charts of a cohort of individuals 15 years or older who were diagnosed with drug-sensitive TB and who should have completed treatment at the district hospital between January 4, 2017 and April 27, 2020 (N=1006) (Fig 1). There were 295 individuals excluded due to no documented treatment outcome (n=246) or no HIV status reported (n=49) for a final sample size of 711.” We also added into the Data Sources/Study Measures section “Missing values for covariates were given their own category of “Missing”, often when a clinician did not ask or did not tick “yes” or “no” in the medical chart, as described above, this is a busy hospital with busy clinicians and lacks administrative capacity to close loops for charting, which does not impact clinical care. For missing HIV status, often it was due to tests not being available on the day a patient was seen, and then no HIV status was recorded upon testing on a subsequent day after the initial baseline data collection. Finally, a final treatment outcome is not always recorded by clinicians in the medical chart. Therefore, we believe these missing data points are missing at random.”. Also, we have now included a supplemental analysis where we (1) investigate differences in demographic characteristics between our included study population and those that were excluded due to missingness and (2) repeated analysis with a multiple imputation procedure to impute the missing outcomes and HIV status. Importantly, we feel confident that our results were not impacted by this large source of missingness as we found minimal differences in demographic characteristics and the results did not change in our imputed version.

Reviewer 2.2: Lines 102 -106 talks about children < 15 years that are excluded from the study and not appropriate in the methods section.

RESPONSE: Thank you. We have removed the sentences regarding the rationale and only provide inclusion/exclusion criteria now. We have also cited 3 additional sources explaining why we excluded individuals < 15 years.

Reviewer 2.3: Line 111 - The TB incidence rate is mentioned as 838/100,000 per year. If this is the case, then wouldn't we expect much more cases during the study period (Jan 2017 to Apr 2020)? (even after excluding children). Were all TB cases in the Eastern Cape accounted in this study?

RESPONSE: Thank you for this observation. No, not all cases of TB were accounted for in this study – this study was only at one district hospital in a rural area with a population of nearly 130,000. There are 12 referring clinics to the hospital where many cases of TB are also diagnosed. We only provided this information as background information about the setting.

Reviewer 2.4: Line 117 Is the adult unemployment population 91.5%? This appears to be very high. (9.9% reported mining work alone!)

RESPONSE: Thank you. Reviewing the data we took from the official 2011 South African Census (as presented through Wazimap.co.za), we realize that the manner in which we have reported this data may be unclear. Although the rate of employment in the catchment area of Zithulele Hospital, at 9.7% of working-age adults is accurate and indeed exceedingly low, the remainder of the adults should not, strictly speaking, be categorized as “unemployed”. The South African Census data uses the following categories and definitions:

Employed: Persons who worked for pay, profit, or family gain, even for just one hour, in the seven days before they were interviewed.

Unemployed (official definition): Persons who did not work, but who looked for work and were available to work in the past seven days.

Discouraged work-seeker: Someone who has lost hope in finding any kind of work.

Not economically active: Persons who were neither employed nor unemployed (e.g. full-time students; retired persons; and homemakers who did not want to work).

Looking at the areas/wards making up the catchment area of Zithulele Hospital (excluding Mthatha, a major city in KSD sub-district), the percentages for each category are as follows:

Employed: 9.7%

Unemployed: 9.6%

Discouraged work seeker: 10.5%

Not economically active: 70.2%

It is therefore more accurate to report an employment rate of 9.7%, an unemployment rate of 20% (of which half are discouraged work seekers) and an extremely high proportion (70.2%) of economically inactive people - which we have now corrected in the manuscript.

We have also changed the Setting section to now read “The median annual household income is R14600 (US$1000)”.

Reviewer 2.5: Line 124 Why was the rapid cartridge HIV antibody test not done for the 77 patients with no HIV status?

RESPONSE: Thank you for this astute observation. As with all types of clinical care in low resource settings, not all evaluations and tests are done for every individual. At times the hospital runs out of cartridges, clinicians are too rushed to do an examination at the initial visit and plan to do it later, or in most cases the rapid test was completed, and just not documented in the source document. Three co-authors are clinicians in the hospital and confirm that due to extenuating circumstances, things do get missed from time to time.

Reviewer 2.6: Line 129-130 Who created the composite categorical outcomes? Why couldn't it be created for the 414 patients?

RESPONSE: Thank you. We (BvdW, IF, and KlR) created the composite categorical outcomes. These were created from source document outcomes including: loss to follow-up, death, treatment failure, cure, completed treatment, and down referral. Those 414 individuals did not have a source document outcome documented. Similar to the above response regarding missing HIV status, clinicians do not always document a final outcome in the TB Medical Record. We have now included a Supplement with an investigation into the missing data. Please see our response to Comment 2.1.

Reviewer 2.7: Line 153 - Why was the "unfavorable" outcome chosen as the reference category (which includes loss to follow-up)? Why can't the majority (88%) who had positive outcomes be the reference category?

RESPONSE: Thank you. We chose “unfavorable” as the reference category because both “success” and “down referral” are considered “positive” outcomes, so it did not make sense to do a comparison of “success” to “down referral” but rather compare each of these to “unfavorable”. Importantly, this was the reference category preferred by our collaborators at the hospital, and we agree that it is easier to interpret “positive” outcomes vs. “negative” outcomes in this manner. Lastly, the reference category does not change the results and an interested reader has all the available information to “flip” the reference category if they so wished.

Reviewer 2.8: Line 155- Why can't age be introduced as a continuous variable in the models? Dichotomizing a continuous variable arbitrarily with a cut-off as 45 years may not be appropriate. Is there any clinical evidence? Would childbearing age be appropriate for males too?

RESPONSE: Thank you for this question. We chose not to include age as a continuous (linear) term because this would force the relationship between age and treatment outcome to be linear, which it is not. However, we agree with the Reviewer that this binary cutoff at 45 years is likely not sufficient. Our remaining options were to model age: (1) using a spline or polynomial terms or (2) with more categories informed from prior literature or our data. We chose the latter as this is more common in the literature, and we believe it accurately captures the non-linear relationship between age and the treatment outcome and is similar to what the recent South African Prevalence Survey had with age brackets. (Figure 2).

Reviewer 2.9: Table 1: Approximately 30% of variables in Table 1 appear to be missing (unanswered) and are excluded from tests of association. Are these variables self-reported or taken from the clinical notes?

RESPONSE: Thank you. Yes, we agree that many of the variables have high rates of missingness, hence, why we reported the amount “unanswered” in the table for transparency. This study was a retrospective chart review, and therefore all variables were taken from clinical charts (where they are primarily self-reported as described in the Methods).

Reviewer 2.10: Tables 2 and 3 are not presented in a logical manner and need to be summarized properly.

RESPONSE: Thank you for this helpful comment. We have combined Tables 3 and 4 into one Table, per other Reviewer comments, and additionally, to address your concern here, we have rearranged our Results to discuss down referral prior to successful outcomes.

Reviewer 2.11: Line 206 - Why are all age reported (n=936) when the study includes only >= 15 years (n=711). They are not relevant to the study.

RESPONSE: Thank you. We have omitted this paragraph to help streamline the paper and now only focus on adults (along with Figure 2).

Reviewer 2.12: Lines 218 - 223 The statements are ambiguous and not clear. Especially lines 222 & 223 are vague.

RESPONSE: Thank you for this comment. We have deleted those three sentences for clarity.

Reviewer 2.13: Line 274 - It is mentioned that 341 (34%) are excluded because of missing treatment outcomes. This is not in agreement with (n=414) presented in figure 1.

RESPONSE: Thank you. Apologies, it was a mistake that we mentioned n=341. We have edited the Methods to say “Data for this study were taken from TB medical charts of a cohort of individuals 15 years or older who were diagnosed with drug-sensitive TB and who should have completed treatment at the district hospital between January 4, 2017 and April 27, 2020 (N=1006) (Fig 1). There were 295 individuals excluded due to no documented treatment outcome (n=246) or no HIV status reported (n=49) for a final sample size of 711.”

Reviewer 2.14: Figure 2 should include only the study participants and not excluded children.

RESPONSE: Thank you – we have revised Figure 2.

Submitted filename: Response to reviewers_PLOSONE_TBtxoutcomes_10192021.docx

Click here for additional data file.

15 Mar 2022

Association of HIV infection and antiretroviral therapy with the occurrence of an unfavorable TB treatment outcome in a rural district hospital in Eastern Cape, South Africa: A retrospective cohort study

PONE-D-21-14796R1

Dear Dr. van de Water,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Petros Isaakidis MD, PhD

Academic Editor

PLOS ONE

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have substantially revised the manuscript. Some minor points.

It is important to acknowledge the authors of the R packages. Please include references for R packages.

Table 2 could be mentioned before table 3.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Anbupalam Thalamuthu

23 Mar 2022

PONE-D-21-14796R1

Association of HIV infection and antiretroviral therapy with the occurrence of an unfavorable TB treatment outcome in a rural district hospital in Eastern Cape, South Africa: A retrospective cohort study

Dear Dr. van de Water:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Petros Isaakidis

Academic Editor

PLOS ONE