Incidence and Predictors of Pulmonary Tuberculosis among Children Who Received Antiretroviral Therapy (ART), Northwest Ethiopia: A Multicenter Historical Cohorts Study 2009–2019

The human immune deficiency virus (HIV) is the strongest risk factor for endogenous reactivation of pulmonary tuberculosis (PTB) through target reduction of CD4, T-lymphocytes, and cellular immune function. Almost one-third of deaths among people living with HIV are attributed to tuberculosis. Despite this evidence, in Ethiopia, information is scarce and meager regarding PTB incidence after ART initiated for seropositive children. Methods. Facility-based multicenter historical cohort was conducted among 721 seropositive children after initiating ART from January 1, 2009, to December 31, 2019. Data from the records of children were extracted using a standardized checklist. The collected data were entered using Epi-Data version 4.2 and exported to STATA (SE) R-14 version statistical soft wares for further analysis. Bivariable and multivariable Cox regression analyses were conducted to identify predictors of PTB incidence. Results. Seven hundred twenty-one (N = 721) seropositive children were included with a mean (±SD) age of 118.4 ± 38.24 months. During the follow-up periods, 63 (15.2%) participants developed new cases of TB; majority (61/63, 96.8%) of them were PTB. The overall incidence rate and the median (±IQR) time of PTB reported were determined as 5.86 per 100 child years (95% CI: 4.58, 7.5) and 17.8 (±11) months, respectively. At baseline, children being severely stunted (AHR = 2.9 : 95% CI, 1.2–7.8, P=0.03), with Hgb ≤10 mg/dl (AHR = 4.0; 95% CI, 2.1–8.1, P=0.001), and not given isoniazid and cotrimoxazole preventive therapy (AHR = 2.4; 95% CI: 1.2; 5.1, P=0.001) (AHR = 2.5; 95% CI, 1.4–4.7, P=0.021) were significantly associated with PTB incidence. Conclusion. A high incidence rate of PTB was observed in our study as compared with the previous finding in Ethiopia. Cases at baseline not taking IPT and CPT, being severely stunted, and having low hemoglobin (≤10 mg/dl) levels were found to be at higher risk of developing PTB.


Introduction
e human immune deficiency virus (HIV) is the strongest risk factor for latent (PTB) or new infection of tuberculosis (TB) through reduction of CD4 T-lymphocytes and cellular immune function [1]. Despite this fact, both (TB/HIV) coinfections are bidirectional and comrade each other, in which HIV sustains the progression of latent tuberculosis bacilli into active TB, while tuberculosis accelerates the progression of HIV disease stages [2].
is coexistence of PTB and HIV infection increased the risk of morbidity and mortality. ART has decreased TB incidence in HIV-infected patients [3]. e waning of the immune system increased Mycobacterium tuberculosis susceptibility and progression of dormant tuberculosis bacilli to endogenous reactivation of latent incidence in the lung [4]. Pulmonary complications have been one of the commonest causes of morbidity and mortality since the advent of the AIDS pandemic and this further led to the parallel pandemic of tuberculosis in some sub-Saharan African populations where 10%-15% suffered from multiple infections [2,5]. According to the World Health Organization (WHO) report, in 2019, there were an estimated 10.1 million new cases of TB and 1.7 million new deaths, making TB the leading cause of death from a single infectious agent (ranking above HIV/AIDS) [6]. e reactivation of latent pulmonary TB is higher among people living with HIV. Africa is the second TB burden region (25%) after Southeast Asia (44%) [6,7] and is responsible for one-third of TB/HIV-associated deaths of children living with HIV [8,9].
Globally, about 36.7 million patients were living with HIV/AIDS and 2.1 million people became newly infected in 2015. Sub-Saharan Africa countries account for the largest proportion, with 25.6 million people living with HIV [1,6]. Likewise, in 2018, there were about 251,000 deaths from TB among PLWHIV, which accounts for 33% of total deaths associated with HIV, which is much higher than the case fatality rate expected ≤5% by WHO [1,10].
People living with HIV (PLWHIV) who have active TB disease were prone to several adverse outcomes. In particular, seropositive children have the risk of increased mortality, developing AIDS-defining event, and loss to follow uprate [7]. e Ethiopian Federal HIV and AIDS Prevention and Control Office estimated that the single National HIV/AIDS is among the top ten high burden counties, in which 31% of TB patients are living with HIV [9,10]. Physiologically, lack of productive cough and sputum for children living with HIV deters early diagnosing and treatment of HIV-associated PTB [11,12]. Epidemiological data on the magnitude of PTB comorbidity is important to control and prevent both diseases. In this regard, facility-based studies conducted in Ethiopia have indicated that HIV-associated PTB has an increasing trend from 5.0% in Northern Ethiopia [2] to 7.2% in Southwest Ethiopia [5]. Additional studies are required in different regions of Ethiopia to identify highrisk children's cases management. erefore, this study was intended to estimate magnitude and predictors for the occurrence of pulmonary TB among children who received antiretroviral therapy in 2020.

Study Design, Area, and Populations.
Metekel zone is one of the three administrative zones found in the Benishangul Gumuz region in Northwest Ethiopia, geographically located within 34°10′N and 37°40′E and latitude 09°17′N and 12°06′ N.
is zone had seven administrative district Woredas with two primary and one-referral hospitals. Gilgel Belles is the capital of Metekel zone, which is located at 565 and 394 km away from Addis Ababa and Assosa, on the way to GERD projects (Ethiopian Great Renaissance Damp at Abay River). Apart from other health service, all these health institutions have been providing ART care services since 2007 [4,13].

Study Design.
A facility-based historical cohort study design was employed among 721 seropositive children in two hospitals and two health centers from January 1, 2009, to December 31, 2019. All these health facilities provide health care services for an estimated 2968 total catchment populations as depicted in [4,14] Figure 1.

Study Populations.
All recoded pediatrics patients under the age of 15 years who received ART initiated and started care in four selected public health facilities were considered as source population. All HIV-infected children who had at least one month of ART follow-up from January 1, 2009, to December 31, 2019, were included. Patients taking anti-TB treatment at the time of HIV/AIDS enrolment were excluded from the study.

Sample Size Determination and Sampling Techniques.
e sample size was calculated by using the formula for survival analysis considering the following parameters, twosided significance level (α � 5%), Za/2 � Z value at 95% confidence interval � 1.96, power (Z B ) � 80%, and AHR � adjusted hazard ration � 1.8 [14]. e final sample size where alpha (a) � 0.05, beta (β) � 0.2, AHR is hazard ratio, E is number of events, N (sample size) � E/P (E), where P (E) � probability of event, and P � cumulative occurrence of pulmonary TB, 7.2% from this reference [5]. e final sample size was determined as 512.5 after adding 15% contingency for data incompleteness. However, from 1, 2009, to December 31, 2019. Totally, 738 charts were found within the four institutions. Since this data is manageable in resources and could have been assumed to the increased power of inference, we included all charts without sampling procedure.

Outcome Ascertainment.
e PTB incidence is considered as an event of interest after HIV/AIDS care clinic started. Children who were lost, died, or transferred out or did not develop the events until the last visit were considered censored, whereas variables including sociodemographic such as age, sex, residence, family size, parental history of TB contact, plus clinical factors like WHO clinical stage, baseline cluster of differentiation (CD4 count), Hgb, functional status, and nutritional status like stunting, wasting, and being underweight were considered as independent variables.
2.6. Operational Words. Smear-positive pulmonary tuberculosis: at least one sputum smear examination is positive for AFB on direct-microscopy. Smear-negative pulmonary tuberculosis: sputum specimens negative for AFB and radiographical abnormalities were consistent with active TB and the decision by a clinician to treat with a full course of antituberculosis chemotherapy. Extrapulmonary tuberculosis is clinically consistent with active extra-pulmonary TB and bacteriologically confirmed by AFB of one specimen from an extrapulmonary site for Mycobacterium tuberculosis and a decision by a clinician to treat with a full course of anti-TB chemotherapy [1].

Operational
Words. SAM DefinitionAccording to the WHO, weight-for-height ≤-3 Z-score, or mid-upper-arm circumference <115 mm, or presence of bilateral edema with failed appetite test should be classified for inpatient care [16]. Undernutrition was defined as a child having one of the description H/Age Z-score < −2 or W/Age Z-score < −2 or W/H Z-score < −2 SD [17,18]. Moreover, the details description is here below as moderate underweight was defined as children having W/Age Z-score < −2 SD or severe was defined as children; moderate stunting was defined as children having H/Age Z-score < −2 SD. Severe stunting was defined as children having H/Age Z-score < −3 SD [18,19]. Moderate wasting was defined as children having W/H Z-score < −2 SD. Severe wasting was defined as children having W/H Z-score < −3 SD n having W/Age Z-score < −3 Z score [17,18].
A CD4 Count CD4 below the threshold level was classified based on the age of the child's (i.e., infants CD4<1500/mm3, 12-35 months < 750/mm3, 36-59 months < 350/mm3, and ≥5 years < 200/mm3) [10]. ART Adherence for pediatrics is classified based on the percentage of drug dosage calculated from the total monthly doses of ART drugs; Good >95%, fair 85-94%, and poor <85%. Seropositive children: According to FMOH Continuum of HIV services refers to a comprehensive package of HIV prevention, diagnostic, treatment, care, and support services provided for people at risk of HIV infection or living with HIV and their families" children aged <15 years were classified and treated seropositive children [10,20].

Data Collection Instruments.
A standard and pretested data extraction form was used to extract the required information from Ethiopia's Federal Ministry of Health Pediatrics antiretroviral therapy (ART) follow-up [11]. Before the actual data collection, the prepared checklists were pretested on 36 case notes of follow-up in seropositive children from Jawi Hospital. e two-day training was given for six-diploma nurse's data collectors and two-degree public health officers of supervisors on the objective of the study, variables of interest, and maintaining data confidentiality. Strict follow-up and supervision were carried out during data collection by the principal investigators and feedback was given on daily basis. e collected data were checked for inconsistencies, coding errors, completeness, accuracy, and missing values.

Data Collection Procedures and Quality Assurance.
To assure the quality of data, data collectors and supervisors were trained about how and what information they should collect from the medical records for one day. e checklist was pretested on 5% of randomly selected charts, which were not included in the actual study. After the pretest, necessary modification of the data collection tool was made. Strict follow-up and supervision were carried out during data collection by principal investigators and feedback was given daily. Individual records with incomplete data during data collection were excluded. e collected data were first being checked and cleaned for completeness.  Journal of Tropical Medicine to generate the Z score (WAZ, HAZ, WHZ/BAZ) to define the nutritional status of seropositive children. We used the Cox regression hazards, regression model, to estimate mortality incidence and predictors during risky successive follow-up. Kaplan-Meier survival analysis was used to determine the cumulative probability of death for all seropositive children and the meantime to death. Variable with P value <0.25 in bivariable Cox regression analysis was included in the multivariable Cox regression model. We tested the assumptions of Cox Proportional Hazard Models using Schoenfeld residuals. Variables with an AHR and 95% confidence interval (CI) and a P value <0.05 were claimed as significant predictors SAM associated mortality of seropositive children.

Sociodemographic Characteristics of Seropositive
Children. As depicted in Figure 1, data of seven hundred thirty-two (N � 732) seropositive children who initiated ART care were reviewed. During the data extraction process, 12 pediatrics files were excluded due to previous TB treatment and incomplete documentation. Based on this, the overall response rate found was 721/732 (99.6%). More than half (384, 53.26%) of participant cases were females in gender, and 510 (70.74%) of them were from urban areas. e largest percentage of participant cases was categorized under the age group 11-15 years, which accounted for 389 (53.96%) of the total subjects. e overall mean (±SD) age of participant children was found to be 118.4 ± 38.24 months (Table 1).

Incidence of Pulmonary TB.
About 721 participants' children were followed up for 16678.07 person per month risk of observation. During the follow-up period, 8.79% (n � 63) of new cases of TB occurred. e majority, 52.4% (N � 36/63) vs. 38.2% (n � 25/63) cases, were smear negative and smear positive pulmonary TB as documented, respectively. However, the remaining 3.4% (n � 2) of cases were extrapulmonary TB (EXPTB). Regarding clinical presentations of SNPTB cases, 27 (72.7%) and 23 (60.6%) had a history of high-grade fever and night sweating, respectively. e overall incidence of PTB in this study was determined as 5.8 per 100 child years (95% CI: 4.58, 7.5). In addition, the incidence of those two new TB cases was estimated as 0.21 per 100 child years risks of observations (Table 4). Table 5 in the multivariable Cox regression analysis, after adjustment and controlling of certain confounding in the final model, four variables were found significantly associated with PTB incidence. Based on this, the risks of developing PTB for children who were not given CPT were more than two times increased as compared with those who were addressed CPT at baseline (AHR � 2.5: 95% CI, 1.84-4.74, P < 0.021). Likewise, the risks of developing PTB infection for seropositive children being nutritionally curved height for age (HFA) ≤−3 Z score) or being severely stunted were three times (AHR � 2.9: 95% CI, 1.2-7.8, P < 0.03) higher as Both orphaned 97 13.5

Discussion
At the end of the cohort of 721 HIV-infected children, 90 children (12.4%) died, 24 (3.3%) were lost to follow-up, 68 (15.1%) were transferred to other health institutions, and 539 (74.7%) remained in the follow-up. Moreover, the present study revealed that the crude incidence of TB among children living with HIV/AIDS was determined as 8.7% (6.8; 11.04). is report is higher than the finding from Debre Tabor hospital 5.0% [2], yet it is comparable with finding stated in Arba Minch (7.2%) [5] and Tanzania (8.5%) [21] hospitals. is might be due to similarity in the study setting and the association of protective effects of ART. In the contrary, this report is lower than the finding from Zimbabwe (12.01%) [22] and Yaoundé, Cameroon (61.8%) [23]. e possible elucidation for the difference may be the size of the study population and the higher burden of tuberculosis in resource-limited settings [1]. In addition, HIV waned the immune system and accelerated viral replication for the depletion of CD4 count [10], which precipitated the new episode of lethal opportunistic infections [24]. is can be reduced by early addressing of CPT and IPT, which is inexpensive and highly effective for reducing loads of endogenous reactivation of latent PTB [25]. Equivalent to this report, however, were seropositive children who missed CPT and had significantly associated with risks of developing PTB. is is consistent with a previous document in Northwest Ethiopia and Gondar hospitals [14,26]. Moreover, in line with research finding with Northwest Ethiopia, Zaire, and Cameroon [4,23,27], our study indicated seropositive children who did not take IPT at baseline had a risk of developing PTB. Consistent with research findings in Northwest Ethiopia [4], concomitant administration of IPT with ART after undertaking in-depth active TB signs, power to have demoted more than ninety-six percent of new TB incidence for children living with HIV. Irrationally, nearly two in five of (37.45%) participant children did not take IPT in our study.
Furthermore, result in our research revealed that seropositive children who missed CPT were significantly associated with risks of developing PTB.
is discovery is comparable with the previous description in Gondar [28] and Adam hospitals [12]. Cotrimoxazole prophylaxis is safe,  inexpensive, and highly effective in reducing morbidity and mortality among HIV-infected infants and children. e World Health Organization (WHO) recommends that all HIV-exposed infants be started on cotrimoxazole prophylaxis at four to six weeks of age, to provide adequate prevention against early opportunistic infections (OIs). is is particularly critical for HIV-infected infants. Data from a randomized clinical trial conducted in Zambia demonstrated the effectiveness of cotrimoxazole prophylaxis in reducing mortality and morbidity among HIV-infected children, despite high levels of bacterial resistance [6, 10,14]. One of the main challenges countries must address to scale up cotrimoxazole prophylaxis is the absence of mechanisms to systematically identify and follow up HIV-exposed infants at and after birth. If HIV-infected children are to be identified early in the course of their disease-a prerequisite for receiving cotrimoxazole prophylaxis and early ART-systems need to be in place to ensure that health workers consider the possibility of HIV infection in infants at birth and at all clinics or other health encounters thereafter [1,4,10,29]. Our finding also indicated that the median (±IQR) time of TB reported and the overall TB incidence (IDR) were determined as 17.8 (±11) months and 5.86 per 100 child years (95% CI: 4.58, 7.5), respectively. e human immune deficiency virus (HIV) is the strongest risk factor for endogenous reactivation of pulmonary tuberculosis (PTB) through target reduction of CD4 and T-lymphocytes, with rapid bacilli reactivation after children are already exposed to infections [14].
Our study also indicated that participant children at baseline being nutritionally severely stunted were independently associated with PTB occurrence as compared with those who have WHO curve ≥−2 Z score.
is is similar to finding s in Adama [30], Tanzania [31], and Uganda and Zimbabwe hospitals [32]. is might be due to HIV infection increasing nutrient mal-absorption due to metabolic alterations that culminate in weight loss and stunting with time leading to early exposure for opportunistic infections [11]. Comparably, the existence of rapid viral replication consumed body energy and creates an arena for developing deadly opportunistic infections [33,34]. is finding also showed that children having hemoglobin ≤10 mg/dl were independently related to HIV-associated PTB occurrence as compared with HIV-infected countergroups. is is in line with the study finding in Adama hospital [30], Gondar hospital [12], Northern Ethiopia [28], Dar es Salaam, Tanzania [31], and England [35]. is is due to hemoglobin levels having a high predictive value for tuberculosis incidence [36].

Limitation of the Study.
e retrospective nature of this study is one of the limitations of this study. Due to this, some of the clinically important predictor variables that have been independently associated with the incidence of TB occurrence in other studies, like the educational status of children and economic status of the family, were not included in these studies.

Conclusion
We found a high burden of pulmonary TB among children living with HIV/AIDS as compared with previous reports in Ethiopia. Cases at baseline not taking IPT and CPT, being severely stunted, and having low hemoglobin (≤10 mg/dl) levels were found to be at higher risk to PTB occurrence. Besides, intensified screening for provisions of isoniazid preventive therapy to children living with HIV/AIDS is highly recommended.