Case fatality rate and viral aetiologies of acute respiratory tract infections in HIV positive and negative people in Africa: The VARIAFRICA-HIV systematic review and meta-analysis.

BACKGROUND: To set priorities for efficient control of acute respiratory tract infection (ARTI) in Africa, it is necessary to have accurate estimate of its burden, especially among HIV-infected populations.
OBJECTIVES: To compare case fatality rate (CFR) and viral aetiologies of ARTI between HIV-positive and HIV-negative populations in Africa. STUDY
DESIGN: We searched PubMed, EMBASE, Web of Knowledge, Africa Journal Online, and Global Index Medicus to identify studies published from January 2000 to April 2018. Random-effect meta-analysis method was used to assess association (pooled weighted odds ratios (OR) with 95% confidence interval (CI)).
RESULTS: A total of 36 studies (126,526 participants) were included. CFR was significantly higher in patients with HIV than in HIV-negative controls (OR 4.10, 95%CI: 2.63-6.27, I²: 93.7%). The risk was significantly higher among children ≤5 years (OR 5.51, 95%CI 2.83-10.74) compared to people aged >5 years (OR 1.48, 95%CI 1.17-1.89); p = 0.0002. There was no difference between children (15 years) and adults and between regions of Africa. There was no difference for viral respiratory aetiologies (Enterovirus, Adenovirus, Bocavirus, Coronavirus, Metapneumovirus, Parainfluenza, Influenza, and Respiratory Syncytial Virus) of ARTI between HIV-positive and HIV-negative people, except for Rhinovirus where being HIV-negative was associated with Rhinovirus (OR 0.70; 95%CI 0.51-0.97, I²: 63.4%).
CONCLUSIONS: This study shows an increased risk of deaths among HIV-infected individuals with ARTI, however with no difference in viral aetiologies compared to HIV-negative individuals in Africa. ARTI deserves more attention from HIV health-care providers for efficient control. Specific strategies are needed for HIV-positive children under 5.

Background

Acute respiratory infections represent one of the major public health concerns. Among these respiratory infections, pneumonia causes about 15% (1 million) of infectious and non-infectious deaths of children under 5 every year and represents the leading infectious cause of death worldwide among this age group [1]. In general population, respiratory viruses are responsible for most of the acute respiratory infections and several risk factors including HIV infection, have been associated with severe acute respiratory infections [[2], [3], [4]]. Data from original studies have shown that irrespective of antiretroviral therapy, morbidity and mortality from acute respiratory infections remains higher in HIV positive individuals than HIV negative individuals [5,6]. Evidence suggest that the burden, in term of deaths, of respiratory viral infections like influenza is higher in Africa compared to other regions [7]. This burden can be fuelled by HIV infection in Africa where about 70% of all HIV positive individuals resided in 2016 [8]. Estimating the burden of respiratory infections in HIV-positive patients in Africa is therefore crucial for guiding health decision making for efficient control of ARTI in Africa. To date, influenza vaccine is not integrated in the immunization program of almost all African countries [9]. In addition, with the imminent availability of human respiratory syncytial virus vaccines, it is necessary to set priorities for vaccination and management of acute respiratory infections in Africa, and more particularly in people living with HIV.

Although several studies in Africa have contributed to show a greater severity of respiratory infections among HIV positive people [10], there is currently no systematic review with meta-analysis comparing the case fatality rate (CFR) between HIV-positive and -negative individuals with respiratory tract infections at the continent level. Comparing HIV-positive and -negative individuals regarding ARTI would help to know in which direction specific interventions are needed. Any identified difference would help to know in which direction specific strategies are needed for people living with HIV.

Therefore, to the best of our knowledge, we conducted the first systematic review and meta-analysis to investigate whether current data may suggest specific attention to HIV-positive individuals; with the aim for better control and preventive strategies for acute respiratory infections. Specifically, we aimed to compare the proportion of Viral Aetiologies of Acute Respiratory tract Infections in Africa between HIV-positive (VARIAFRICA-HIV) and HIV-negative populations. We also aimed to compare CFR between HIV-positive and HIV-negative populations with acute respiratory tract infections (ARTI).

Study design

Design

This systematic review and meta-analysis was reported following the PRISMA guidelines as illustrated by the checklist (Supplementary Table 1) [11]. The systematic review is registered in PROSPERO, CRD42018088261.

Eligibility criteria

We considered cross-sectional, case control, and cohort studies including both HIV-positive and HIV-negative populations with ARTI, residing in Africa regardless of age and ethnicity. The primary outcome of interest was case fatality rate and the secondary outcome was proportion of respiratory viral aetiologies of ARTI. Respiratory viruses had to be identified with polymerase chain reaction technique (PCR). Studies conducted during outbreak periods, studies including only HIV positive people or only HIV negative people were excluded.

Search strategy for identifying relevant studies

We did a comprehensive search on electronic databases among which Medline (through PubMed), Excerpta Medica Database, African Journals Online, Web of Knowledge, and Global Index Medicus. The search strategy was adapted to suit each database as illustrated by the search on PubMed (Supplementary Table 2). We considered studies published from 1st January 2000 until 30 April 2018, without any language restriction. We arbitrary limited to 2000 to have contemporaneous and recent data since the epidemiology of HIV infection was drastically changed in the last years. To supplement the electronic search, references of all relevant studies were also screened for potential consideration.

Selection of studies for inclusion in the review

Two review authors (SK and AFM) independently screened the titles and abstracts of aggregated citations retrieved from the electronic search, and full texts of potentially eligible articles were further assessed for inclusion. Disagreements were resolved through discussion and unreached consensus was resolved by a third review author (JJB). The Kappa statistics were used to estimate the agreement on study selection between the review authors before discussion [12].

Data extraction and management

Data were extracted independently by two review authors (SK and AFM) and double checked by a third review author (JJB) for accuracy. In case of missing data in an article, the authors of this article were contacted to request the unreported details. Data extracted included:

Description of the article: name of the first author, year of publication;

Study characteristics: study design, sampling method, timing of data collection, period of participants’ recruitment, setting, location (country and regions according to United Nations Statistical Division);

Participants’ characteristics: clinical presentation, types of sample used for viral detection, mean/median age, age group, and proportion of males.

Proportions of respiratory viral aetiologies of ARTI and case fatality rate in both HIV positive and HIV negative groups or enough data to compute these estimates.

Risk of bias assessment

The methodological quality of the studies was assessed using Newcastle Ottawa Scale [13]. Scoring questions on a scale of 0 (no) and 1 (yes) was assigned to each item to have a total score for each article. According to the total score obtained, each article was classified as high risk (0–3 points), moderate risk (4–6 points), and low risk of bias (7–9 points). The assessment was done by two review authors (SK and JJB) and discrepancies were resolved through discussion and consensus. The kappa statistics were used to investigate the agreement between the review authors before discussion [12].

Data synthesis and analysis

We did data meta-analyses using the package ‘meta’ (version 4.9-2) of R (version 3.5.1, The R Foundation for Statistical Computing). Random-effect meta-analysis by Mantel-Haenszel (MH) method was used to assess the association between HIV infection and case fatality rate or viral aetiologies, and reported as pooled weighted odds ratios (OR) with both 95% confidence and 95% prediction intervals [14]. Symmetry of counter-enhanced funnel plots (for outcome with 10 studies or more) and Harbord test were used to assess reporting and publication bias [15]. Consideration of significant publication bias was at a threshold of p < 0.10. Heterogeneity across studies was assessed by χ² on Cochran’s Q test, and reported as I² statistics [16]. In case of substantial heterogeneity (I² > 50%) [17], we did subgroup-analysis (people aged ≤ 5 years versus > 5 years, adults (> 15 years) versus children, and UNSD regions) to investigate possible sources of residual heterogeneity.

Results

Study selection and characteristics

Initially, we identified 1721 records. Finally, 36 full texts were retained for the review and included in the meta-analysis [[18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53]] (Fig. 1 ). The inter-rater agreement for study selection was high (κ = 0.74). Of the 36 included studies, 18 (50%) low risk of bias, 18 (50%) moderate risk of bias, none had high risk of bias. In total, 126,526 participants were included. Data were from nine countries. No data were from western and northern Africa. One study was from central Africa, 14 from eastern Africa, and 21 from southern Africa. Characteristics of included studies are presented in the Supplementary Table 3 and detailed in the Supplementary Table 4.

Fig. 1

Flow diagram of study selection.

Case fatality rate

In total, 60,224 participants were included in the analysis for case fatality rate. Case fatality rate was significantly higher among HIV-positive (1514/18159, 8.3%) than HIV-negative people (1555/38996, 4.0%) (pooled OR 4.10, 95% CI 2.63–6.27, I²: 93.7%, 25 studies; Fig. 2 ). There was most probably missing data on the bottom left-hand-side of the counter-enhanced funnel plot (Supplementary Fig. 1), as confirmed by the Harbord test (p = 0.0009; Supplementary Table 5). However, since the majority of this area contains regions of high statistical significance, this reduces the plausibility that publication bias is the underlying cause of this funnel asymmetry. However, after adjustment with Trim-and-Fill method (Supplementary Fig. 2), the pooled OR was 1.28 (95%CI 0.84–1.95, 43 (18 added) studies, I² = 95.7%) (Supplementary Table 6). The sensitivity analysis including only low risk of bias studies yielded a finding close to that of crude analysis: pooled OR 4.33, 95% CI 2.02–9.29, 11 studies.

Fig. 2

Comparison of case fatality rate between HIV-positive and HIV-negative populations with acute respiratory tract infections in Africa.

In subgroup analysis, the strength of the association between HIV-infection and CFR was significantly higher among children ≤ 5 years (pooled OR 5.51, 95%CI 2.83–10.74) compared to people aged > 5 years (pooled OR 1.48, 95%CI 1.17–1.89), p = 0.0002 (Supplementary Fig. 3 ). There was no difference between children (<15 years) and adults (Supplementary Figure 4) and between regions (Supplementary Figure 5).

Fig. 3

Comparison of viral aetiologies of acute respiratory tract infections between HIV-positive and HIV-negative populations with in Africa.

Viral aetiologies of acute respiratory tract infections

In total, 76,174 participants were included in the analysis for viral aetiologies. There was no difference for all the viral respiratory aetiologies (Enterovirus, Human Adenovirus, Human Bocavirus, Human Coronavirus, Human Parainfluenza Virus, Influenza Virus, Human Metapneumovirus, and Respiratory Syncytial Virus) of ARTI between HIV positive and HIV negative people, except for Rhinovirus where HIV-positive persons were less likely to be infected with Rhinovirus than HIV-negative (pooled OR 0.70; 95%CI 0.51-0.97) (Fig. 3). The figure with detailed studies is in the Appendix (Supplementary Figure 6).

Funnel plots for Human Adenovirus, Respiratory Syncytial Virus, Human Parainfluenza Virus, Influenza Virus, and Human Metapneumovirus were presented in Supplementary Figures 7–11, respectively. There was no publication bias for all viruses except for Human Parainfluenza Virus and Human Metapneumovirus (Supplementary Table 5). For both viruses with significant publication bias, there was most probably missing data on the bottom left-hand-side of the counter-enhanced funnel plot reducing the plausibility that publication bias is the underlying cause of this funnel asymmetry. After adjustment with Trim-and-Fill method, HIV positive individuals were less likely to be infected by Human Parainfluenza virus (OR: 0.52; 95%CI: 0.31-0.87; I² = 86.4%) and Human Metapneumovirus (OR: 0.46; 95%CI 0.29-0.73, 19; I² = 69.2%) (Supplementary Table 5).

In subgroup analysis, there was no difference between children and adults for all respiratory viruses (Supplementary Table 6). The association of HIV infection with the following respiratory viral infections was different between people aged > 5 years and those aged ≤5 years: Human Adenovirus (≤5 years: OR 1.00 [95%CI 0.61–1.64] vs >5 years: OR 3.91 [95%CI 2.01–7.61]; p = 0.0013), Respiratory Syncytial Virus (≤5 years: OR 0.62 [95%CI 0.31–1.26] vs >5 years: OR 4.30 [95%CI 2.09–8.87]; p = 0.0002), Human Parainfluenza Virus (≤5 years: OR 1.38 [95%CI 0.73–1.26] vs >5 years: OR 4.30 [95%CI 2.09–8.87]; p = 0.0002), and Human Metapneumovirus (≤5 years: OR 0.46 [95%CI 0.35-0.60] vs >5 years: OR 3.00 [95%CI 1.27–7.04]; p < 0.0001) (Supplementary Table 6). The association of HIV infection with the following respiratory viral infections was different between different regions: Respiratory Syncytial Virus (Eastern region: OR 1.50 [95%CI 0.58–3.89] vs Southern: OR 0.40 [95%CI 0.18-0.90]; p = 0.039) and Human Parainfluenza Virus (Eastern region: OR 4.01 [95%CI 2.23–7.24] vs Southern: OR 0.69 [95%CI 0.50-0.96]; p < 0.0001), and Human Metapneumovirus (Eastern region: OR 1.50 [95%CI 0.53–4.22] vs Southern: OR 0.48 [95%CI 0.38-0.61]; p = 0.037) (Supplementary Table 6).

Discussion

To the very best of our knowledge this is the first systematic review and meta-analysis to compare the CFR between HIV-positive and -negative people with ARTI in Africa and also to investigate the association between HIV infection and viral aetiologies of ARTI in Africa. Key findings of this review include higher rate of mortality in HIV positive people compared to HIV negative with the burden higher among those aged < 5 years. Another key finding is that in general there was no difference in the viral aetiologies of ARTI between HIV positive and HIV negative people, except for Rhinovirus which was less associated with HIV infection.

As for other communicable and for non-communicable diseases, HIV positive individuals with ARTI have a higher risk of death compared to HIV negative individuals. The explanation lies globally in the immune response and on how work antiretroviral therapy. Immune response in HIV positive patients with ARI may include low levels of antibodies, low level of CD4 + T cell, and other potential immunological dysfunctions that may promote increased mortality in this group of patients [[54], [55], [56]]. The specific pathogenesis that may explain the association between HIV infection and the increased risk of death in patients with acute respiratory infections is not well understood. Unlike our review, studies conducted outside Africa suggested that the presence of HIV infection did not influence the clinical outcomes of people with ARTI [[57], [58], [59], [60]]. This difference can be explained by the fact that people living with HIV in Africa compared to developed countries can experience more limited access to healthcare for advanced HIV disease, malnutrition, parasitic diseases, and low antiretroviral treatment coverage [57,58,60]. Therefore, in an African context, the higher risk of death in people with HIV can not only be explained by HIV infection itself. We found an increased risk of death among children < 5 years. This finding is close to that of general population of children. Indeed, ARTI have a disproportionate effect on children younger than 5 years with increased risk of death and disability-adjusted life-years [61]. The increased mortality among this population may be related to the immaturity of their immune response compared to adults, especially in the case of HIV infection [61]. In addition, evidence support that the adherence to antiretroviral therapy is lower in children compared to adults [62,63]. As antiretroviral therapy is a key component to curb the mortality in HIV positive people, children under five with suboptimal observance to antiretroviral therapy, in addition to a weak immune response in the case of ARTI, can therefore present higher risk of death compared to adults.

The lack of difference in the respiratory virus for ARTI between HIV-positive and negative individuals suggests that the difference in the germs leading to ARTI may come from others such as bacteria and fungi including tuberculosis and pneumocystis jirovecii [[64], [65], [66], [67]]. Indeed, HIV-positive and negative individuals live in the same ecosystems and can therefore be contaminated in the community through air transmission with same microorganisms. In the strategies for curbing the burden of disease related to ARTI, efforts to control respiratory viral infection in terms of viral aetiologies should be the same between HIV positive and negative individuals.

Our findings demonstrated that ARTI deserves more attention from HIV health-care providers, researchers, policy makers, and stakeholders for improved detection, overall proper management, and efficient control of ARTI in people with HIV in Africa, especially for children under 5. Although strategies to curb the burden of ARTI coordinated by the WHO made some advances from 1980 to 1990 toward reducing childhood mortality from LRTI, the HIV epidemic had reduced the efforts made so far [68]. The reduction of morbidity and mortality of ARTI in Africa should not only focus on HIV infection and also requires a multifaceted approach that includes addressing risk factors like smoking, second hand smoking, indoor pollution, undernutrition especially among children under five years old [69]. Strategies to curb the burden among HIV positive individuals can include using viral and conjugate bacterial vaccines to reduce colonization and consequently the microbiological load in the environment of HIV positive individuals [70]. When indicated among HIV-positive individuals, healthcare providers should adhere to guidelines for chemoprophylaxis with cotrimoxazole to prevent among others, respiratory infections among people living with HIV [71,72]. To improve the immunity of HIV positive individuals, routine micronutrient supplements, especially multivitamins can be used. Multivitamins have shown to have benefit in reducing the prevalence of ARTI in HIV positive people [73]. One of the most important point to curb the burden of ARTI is to early initiate antiretroviral treatment among all HIV positive people regardless of CD4 cells count and age as recommended by World Health Organization (WHO) [74]. Antiretroviral treatment has been shown to reduce the frequency of most opportunistic infections and significantly reduce HIV-related mortality, including among people with respiratory tract infections [[57], [58], [59], [60]]. However, implementing this WHO guidelines remain challenging in Africa, a continent with poor human and financial resources and weak healthcare infrastructures [75]. Surveillance for antimicrobial resistance resulting from repeated and prophylactic antibiotic use among HIV positive individuals and its clinical impact should be investigated in future studies. Research are also needed to investigate the effectiveness of vaccine for respiratory pathogens among HIV positive individuals.

This review should be interpreted in consideration of certain limitations. First and common to most reviews of this type, we found a high heterogeneity between studies explained by age groups. However, we were unable to investigate other sources due to inconsistency in reporting of covariates throughout included studies. Second, the various UNSD sub-regions in Africa were not proportionally represented in the review, which may hinder the translatability of our findings to the entire continent. Third, all included studies were observational, which could lead to bias because of unmeasured confounders between HIV positive and negative. Nonetheless and to the best of our knowledge, this is the first systematic review and meta-analysis which has given a clear and accurate estimate of CFR among HIV positive individuals with ARTI compared to HIV-negative with ARTI. It is also the first to investigate association between HIV infection and specific respiratory viral aetiology of ARTI. We searched the biggest electronic databases and used rigorous methodological and statistical procedures to generate our estimates. All studies were assessed as having low to moderate risk of bias in their methodological quality, suggesting that we can be confident in the quality of these estimates. Indeed, the sensitivity analysis including only low risk of bias studies yielded an estimate close to that of crude analysis in CFR analysis. We also considered only studies using polymerase chain reaction as diagnostic method of viral aetiologies of ARTI.

This systematic review and meta-analysis shows an increased risk of deaths among HIV-positive individuals with ARTI compared to HIV negative with ARTI control counterparts, especially among children < 5 years. However, there is no difference in viral aetiologies of ARTI infections between HIV positive and HIV negative individuals in Africa. As such, ARTI deserves more attention from HIV health-care providers, researchers, policy makers, and stakeholders for overall proper management, and efficient control of ARTI in people with HIV in Africa. Specific strategies are needed for HIV-positive children under 5.

Data sharing and data accessibility

All data generated or analysed during this study are included in this published article and its supplementary information files.

Authors’ contributions

Conception: SK, JJB, RN. Design: SK, JJB, ET, MSN, RN. Literature search: JJB. Selection of studies: SK, AFM, JJB. Full texts search: SK. Data extraction: SK, AFM, JJB. Data synthesis and analysis: JJB, SK. Data interpretation: JJB, SK, MSN, ET. Manuscript drafting: SK, JJB. Manuscript editing and revision: SK, JJB, AFM, FBNS, PAN, MSN, ST, ET, RN. Manuscript final version approval: SK, JJB, AFM, FBNS, PTN, MSN, ST, ET, RN. Guarantor of the review: RN.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Ethical approval

Not required.

Declaration of Competing Interest

The authors declare that they have no competing interests.

CRediT authorship contribution statement

Sebastien Kenmoe: Conceptualization, Data curation, Methodology, Project administration, Writing - original draft, Writing - review & editing. Jean Joel Bigna: Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing - original draft, Writing - review & editing. Abdou Fatawou Modiyingi: Data curation, Writing - review & editing. Marie S. Ndangang: Writing - review & editing. Paul Alain Ngoupo: Writing - review & editing. Fredy Brice N. Simo: Writing - review & editing. Serges Tchatchouang: Writing - review & editing. Elvis Temfack: Data curation, Methodology, Writing - review & editing. Richard Njouom: Conceptualization, Methodology, Project administration, Supervision, Validation, Writing - review & editing.