Tuberculosis and HIV co-infection in Congolese children: risk factors of death

Introduction Human immunodeficiency virus (HIV) and tuberculosis (TB) are the leading causes of death from infectious disease worldwide. The prevalence of HIV among children with TB in moderate to high prevalence countries ranges between 10% and 60%. This study aimed to determine the prevalence of HIV infection among children treated for TB in Directly Observed Treatment Short-Course (DOTS) clinics in Lubumbashi and to identify risk of death during this co-infection. Methods This is a cross-sectional study of children under-15, treated for tuberculosis from January 1, 2013 to December 31, 2015. Clinical, paraclinical and outcome data were collected in 22 DOTS of Lubumbashi. A statistical comparison was made between dead and survived HIV-infected TB children. We performed the multivariate analyzes and the significance level set at p-value <0.05. Results A total of 840 children with TB were included. The prevalence of HIV infection was 20.95% (95% CI: 18.34-23.83%). The mortality rate was higher for HIV-infected children (47.73%) compared to HIV-uninfected children (17.02%) (p<0.00001). Age <5 years (aOR=6.50 [1.96-21.50]), a poor nutritional status (aOR=23.55 [8.20-67.64]), and a negative acid-fast bacilli testing (aOR=4.51 [1.08-18.70]) were associated with death during anti-TB treatment. Conclusion TB and HIV co-infection is a reality in pediatric settings in Lubumbashi. High mortality highlights the importance of early management.


Introduction
The most important chronic diseases affecting children in sub-Saharan Africa are tuberculosis (TB), Human Immunodeficiency Virus (HIV) infection and malnutrition. These three diseases often have similar clinical aspects or are frequently associated with the same patient. Globally, the epidemic of HIV infection has been accompanied by a serious epidemic of TB [1,2]. HIV infection is the main risk factor for tuberculosis because HIV promotes the progression of latent or recent infections of Mycobacterium tuberculosis into active disease and also increases the frequency of TB [2,3]. Both pathologies are the two leading causes of infectious disease deaths worldwide [4].
There are well-established epidemiological and biological synergies between HIV and TB, which influence the distribution, progression and outcomes of both infections. The HIV epidemic is a key factor in the resurgence of TB incidence worldwide and HIV is the pre-incident risk factor for TB development [2,4]. According to the latest global estimates, 36.9 million people are living with HIV, 70% of whom live in sub-Saharan Africa [5] and 1.7 billion people are affected by latent TB [6,7]. In 2017, UNAIDS estimated 1.8 million children under-15 living with HIV [5]. World Health Organization (WHO) estimated that 10% (1 million) of the 10.4 million new TB cases recorded in 2016 were children and 11% (1.2 million) of those with TB were HIV positive [6]. But specifically for children, there is limited data on the incidence of HIV infection among children with TB, and the information available is difficult to interpret because of problems with diagnosis, under-reporting and selection of the studied populations (hospitals rather than community) [4]. WHO estimates that HIV prevalence among children with TB in medium and high prevalence countries is between 10% and 60% [8,9] and varies with background rates of HIV infection [10].

HIV-TB co-infection has been well explored in adults in sub-Saharan
Africa, but has so far been insufficiently studied in children because of the difficulty of establishing an accurate diagnosis of TB in children in particular by microscopy. It is considered a major public health problem, particularly in resource constrained settings like Democratic Republic of Congo (DRC). The DRC is among the 30 countries heavily affected by TB; it ranks eleventh in the world and third in Africa [6].
In 2016, estimates report 32,000 cases of all forms of TB in children under-15 [6]. HIV infection and tuberculosis are two major burdens whose control or elimination is a huge challenge for African states and their health services. The co-infection tuberculosis and HIV in children is distinguished by its particular aspects. Epidemiologically, it is a direct consequence of adult tuberculosis and reflects the transmission of the disease in the community. The often atypical clinical manifestations (extrapulmonary tuberculosis) and the bacteriological and anatomo-pathological diagnosis remain difficult [11,12].
In the DRC, the fight against TB is the task entrusted to the National Tuberculosis Program (NTP), it is part of the direction of the fight against the diseases. In order to achieve its various objectives, the NTP recommends an integrated approach to TB control activities in

Methods
We conducted a cross-sectional study. Our team collected data from  15,16]. We also included patients who were unable to produce appropriate sputum specimens but whose symptoms and clinical history were consistent with active TB. Patients with both PTB and EPTB have been classified as PTB in accordance with the definition of the WHO [15].
Demographic and clinical information was collected from medical records using a standardized data collection form. The following data were extracted from the records of each patient: age, sex, contagion with a known TB case, history of TB, presence or absence of BCG vaccine scar, and time to onset of symptoms before admission to the service. The following general and physical clinical signs were systematically recorded: asthenia, weight loss, anorexia, unexplained fever (>2 weeks) greater than or equal to 38°C, persistent cough (>2 weeks), weight-for-age z-score (WAZ) ≤-2 SD and peripheral lymphadenopathy. The results of the direct tuberculosis bacillus test, as well as those of the hemogram and HIV serology of each patient at admission were also recorded. AFB was performed on sputum or gastric tube fluid from patients with pulmonary or mixed forms of tuberculosis on admission: a direct examination by the Zielh-Neelsen technique and culture of mycobacteria on the Löwenstein-Jensen medium.
The results of tuberculosis treatment have been classified into six categories according to WHO definitions: cured, completed treatment, died, defaulted treatment, failure and transfer-out [15].
Children who were cured and those who had completed treatment were classified as having successful treatment (or having a favorable outcome) [13, 15,17]. All information collected was analyzed using STATA 12 software. For each factor studied, the proportions were compared between groups of HIV-positive and HIV-negative TB children and between dead and survived HIV-infected TB children.
Chi square test was used for comparison of percentages in univariate and multivariate analysis. The significance level of all observed differences was set at p<0.05. This study was authorized by the medical ethic committee of the

University of Lubumbashi and the Health authority of Haut-Katanga
Province before data collection. Patient records/information was anonymized and de-identified prior to analysis to ensure confidentiality of individual patient information.

Discussion
This study found that the prevalence of HIV-TB co-infection among TB patients was 20.95%, which is higher as compared with 11.8% prevalence that was reported in Abidjan (Côte-d'Ivoire) by Sassan-Murokro et al. [18]. This rate is lower than the 25.7% found in Bangui This study found that HIV-infected children were more likely to have extrapulmonary forms that could also explain this high mortality in these children. This high mortality in HIV-infected children could be attributed to late diagnosis and delayed antiretroviral treatment (ART). In sub-Saharan Africa, HIV infection in the pediatric population is often diagnosed very late and especially in the course of the disease, and when ART starts late, the child´s immune system may already be seriously compromised [24]. Early pediatric HIV testing is not done because some parents / guardians rely only on their child's apparent good health, and others believe that children infected during the perinatal period do not survive until the end of their childhood [24,25]. Nearly half of HIV-infected TB children died during anti-TB treatment. HIV-TB co-infection was significantly associated with death and this is consistent with previous studies elsewhere [26][27][28]. Lolekha found that HIV-infected TB children were 6.9 times more likely to die than HIV-negative children [29]. HIV contributes substantially to the burden of childhood TB. HIV-TB coinfection is frequently associated with multiple infections complicating diagnosis and treatment, leading to increased morbidity and mortality [26]. Biruk et al. [30] explains that this would be related to pill burden, increase in adverse effect, drug-to-drug interaction, and immune reconstitution inflammatory syndrome.
In this study, a number of factors associated with death in HIVinfected TB children have been reported. The mortality rate among children under-5 was significantly higher than those over 5 years old.
Our results corroborate those of studies elsewhere [26,31]. In line with this finding, a study in Malawi found a decline in death rates with advanced age [32]. Younger children, especially those younger than two years of age, are at increased risk of death from infectious diseases, including tuberculosis due to the immaturity of the immune system. In addition, these young children are frequently affected by disseminated tuberculosis and tuberculous meningitis, which are associated with high mortality [12,33,34].

Limitations of the Study
Therefore, this study is not without limits and the results of this study should be interpreted with the following limitations. The major limitation of this work is the use of retrospective data, which is totally restricted to what is documented in the TB registers. Predictor variables which might affect outcome like treatment adherence level, comorbidities, modalities of HIV care and ART usage, other opportunistic infections as well as behavioral and social factors, were not available in the TB national system.

Conclusion
High proportion of death in HIV-infected TB children was documented in this study. Moreover, the following risk factors were identified as predictors of death: age ≤ 5 years, poor nutritional status, and negative AFB testing. Based on the findings of this study, we recommend that emphasis has to be given for children with high risk of death and targeted interventions should be carried out.
What is known about this topic  In the DRC, infectious disease are a major public health problem with an important rate of morbidity and mortality among children;  Mortality due to TB and HIV on young children is separately higher.

Competing interests
The authors declare no competing interests.