http://orcid.org/0000-0003-0353-7770
Figures
Abstract
Background
Visceral leishmaniasis (VL) is a vector-borne disease that is deadly if left untreated. Understanding which factors have prognostic value may help to focus clinical management and reduce case fatality. However, information about prognostic factors is scattered and conflicting. We conducted a systematic review and meta-analysis to identify prognostic factors for mortality among VL patients in East Africa.
Methodology/Principal findings
The review protocol was registered in PROSPERO (CRD42016043112). We included studies published in English after 1970 describing VL patients treated in East African health facilities. To be included, studies had to report on associations between clinical or laboratory factors and mortality during admission or during VL treatment, with a minimal study size of ten patients. Conference abstracts and evaluations of genetic or immunological prognostic factors were excluded. We searched for studies in MEDLINE and four other databases in December 2018. To assess the risk of bias in observational studies and clinical trials, we used the Quality in Prognostic Studies (QUIPS) tool. We included 48 studies in the systematic review, describing 150,072 VL patients of whom 7,847 (5.2%) died. Twelve prognostic factors were evaluated in five or more studies and these results were submitted to meta-analysis producing one pooled crude odds ratio (OR) per prognostic factor. The following factors were strongly (OR>3) and significantly (P-value<0.05) associated with mortality: jaundice (OR = 8.27), HIV (OR = 4.60), tuberculosis (OR = 4.06), age >45 years (OR = 3.69), oedema (OR = 3.52), bleeding (OR = 3.37), and haemoglobin ≤6.5 g/dl (OR = 3.26). Factors significantly and moderately (OR between one and three) associated with death were severe malnutrition, long duration of illness, young age (<5 years), and large spleen size.
Author summary
Visceral leishmaniasis (VL) is a severe disease caused by a micro-organism called Leishmania. This disease occurs in South America, East Africa, the Mediterranean, and Asia. There are several treatments with differing success rates, side effects, and costs. Some of the VL patients who seek care in health facilities die, and knowing which patients are at risk of dying is useful for patient management. Indeed, it makes sense to reserve the best treatment and closest follow-up for those patients with the highest risk of dying. We searched the medical literature in December 2018 and found 48 studies evaluating associations between clinical or laboratory factors and risk of dying among VL patients treated in East Africa. These studies described 150,072 VL patients of whom 7,847 (5.2%) died. Factors that were evaluated in five or more studies were analysed further. VL patients were at increased risk of dying if they had jaundice, HIV, tuberculosis, old (more than 45 years) or young (below five years) age, oedema, bleeding, anaemia, malnutrition, long duration of illness, and large spleen size. We recommend to include these factors, which can easily be detected by health professionals, in future studies that aim at improving the prognosis of VL patients.
Citation: Abongomera C, van Henten S, Vogt F, Buyze J, Verdonck K, van Griensven J (2020) Prognostic factors for mortality among patients with visceral leishmaniasis in East Africa: Systematic review and meta-analysis. PLoS Negl Trop Dis 14(5): e0008319. https://doi.org/10.1371/journal.pntd.0008319
Editor: Guilherme L. Werneck, Universidade do Estado do Rio de Janeiro, BRAZIL
Received: September 8, 2019; Accepted: April 23, 2020; Published: May 15, 2020
Copyright: © 2020 Abongomera et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: CA has received a PhD scholarship granted from the European Union Seventh Framework Program (FP7/2007‐2013) under grant agreement n° 305178 via AfriCoLeish project. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Visceral leishmaniasis (VL), or kala-azar, is a neglected tropical disease which is deadly if untreated [1,2]. In East Africa and the Indian subcontinent, it is caused by Leishmania donovani and in Latin America and the Mediterranean by L. infantum. Leishmania parasites are transmitted through the bite of infected female phlebotomine sandflies [1,2]. The cycle of L. donovani is predominantly anthroponotic and that of L. infantum usually includes dogs or other mammals [1,2]. VL is endemic in 75 countries, and the estimated annual global incidence is 50,000–90,000 new cases [3–5]. In 2015, East Africa contributed the highest proportion of new VL cases world-wide– 40% [3]. South Sudan and Sudan have the highest case load, followed by Ethiopia [6–8]. VL has also been reported in Kenya, Uganda, Eritrea and Somalia [3,9–11].
VL patients usually present with prolonged fever, weight loss and splenomegaly [1,2,12]. Sodium stibogluconate (SSG), the main VL treatment in East Africa, may cause potentially lethal adverse events such as cardiac arrhythmia [13,14]. Liposomal amphotericin B (AmBisome), the alternative drug, is safer but expensive–therefore it should be administered to patients who need it most, such as those with severe VL or those at risk of adverse events with SSG [15]. The average VL case fatality rate in East Africa was 2% in 2015, but this proportion can be markedly higher in specific subgroups, such as in VL and HIV co-infected patients among whom the case fatality rate can be as high as 39% [16,17].
A critical step to improve patient outcomes is a deeper understanding of the factors that determine the prognosis among patients with VL. Evidence-based clinical decision tools based on key prognostic factors are increasingly used in other clinical domains to identify high-risk patients requiring specific care such as more intensive monitoring, additional investigations or specific treatments [18,19]. On the other hand, patients with a better prognosis might be treated on ambulatory basis or at a decentralised level [18,19]. Similarly, the stratification of patients with VL according to their risk of death could help to focus clinical management and reduce case fatality.
A systematic review reporting on prognostic factors for mortality has been reported from Latin America, where VL is caused by L. infantum [20]. A wide range of prognostic factors were found to be associated with mortality ranging from clinical signs and symptoms (jaundice, oedema, ascites, bleeding, pronounced splenomegaly, vomiting, and diarrhoea) to laboratory abnormalities (anaemia, thrombocytopenia) and coinfections (HIV and tuberculosis) [20]. However, there are important differences in reports from different geographical regions regarding many aspects related to VL, such as clinical presentation, disease severity, case fatality and treatment response [12,15]. Whether this relates to differences in Leishmania species, host population, health-seeking behaviour, or style of reporting remains largely unexplored. Consequently, whether prognostic factors identified in Latin America also apply to East Africa where L. donovani is prevalent is unknown.
Although several studies in East Africa have aimed at identifying prognostic factors [21–26], the information is scattered and sometimes conflicting, and it is currently unclear to which extent the findings of these studies are valid and generalizable. Hence, the classification of VL severity remains poorly defined. We therefore set out to conduct a systematic review to identify key prognostic factors associated with mortality among patients with VL in East Africa.
Methods
Protocol and registration
This review was conducted in line with the ‘preferred reporting items for systematic reviews and meta-analyses’ (PRISMA) guidelines [27]. The completed PRISMA checklist is available in S1 Text. The review protocol was registered in a repository of systematic review protocols prior to starting the research (PROSPERO, protocol number CRD42016043112) [28].
Eligibility criteria
We aimed to include studies on VL patients from East African countries, namely Ethiopia, Sudan, South Sudan, Kenya, Uganda, Somalia and Eritrea, with a laboratory-confirmed or clinical VL diagnosis and treated in hospitals, health centres or other health care structures. Studies had to evaluate the association between clinical or laboratory parameters and mortality during admission or during VL treatment, with a minimal study size of ten patients. To be included, studies had to provide either a summary measure for the association with mortality [odds ratio (OR), relative risk (RR), or hazard ratio (HR)] with a P-value or confidence interval (CI), or had to give crude data that allowed the calculation of a measure of association. We excluded conference abstracts and studies on prognostic factors relating to genetic or immunological features. Only studies published after 1970 and in English language were considered.
Information sources, search strategy, and study selection
In December 2018, we searched for studies in five electronic databases. The search strategy combined terms indicating the disease (such as “kala azar” and “visceral leishmaniasis”) with terms indicating the geographical region. Details are given in Table 1.
Additionally, we reviewed the reference lists of selected publications and contacted VL specialists to check if the automatic search had missed any important studies.
Titles and abstracts of all retrieved studies were independently screened by two authors (CA and SvH). These same authors also independently assessed the full texts of the retained studies. At both steps, disagreements were resolved by a third author (FV).
Data items and data collection process
The core information was the strength of association between prognostic factors and mortality. In addition, we extracted information about study setting (place, time, type of health facility); study population (demographics, clinical characteristics, inclusion/exclusion criteria); type of treatment; and study design and methods (including information needed for the risk of bias assessment). Data from the included studies were extracted using a standardised, pre-piloted form. Two authors (CA and SvH), independently extracted the data. Disagreements were resolved by a third author (JvG).
Risk of bias in individual studies
Two authors (CA and SvH) independently assessed the risk of bias in the studies using the Quality in Prognostic Studies (QUIPS) tool [29]. Disagreements were resolved by a third author (JvG/FV). The results of the risk of bias assessment are described as part of the narrative synthesis but were not used in the meta-analysis.
Synthesis of results
All included studies are presented in a narrative synthesis and summary table. For each prognostic factor evaluated in at least three studies, we also summarised the findings in a forest plot. For factors reported in at least five studies, we did a meta-analysis. The information that was pooled was the strength of the association between a specific prognostic factor and mortality during admission or VL treatment. The summary measure was a pooled OR for each candidate prognostic factor.
As different studies reported continuous variables with different cut-offs, we defined standardised cut-offs based on information from the literature and the available data. For age we used <5, <15, 15–45 and ≥45 years, for duration of illness <2 months and ≥2 months, for malnutrition body mass index (BMI) <16 kilogrammes/metre2 (kg/m2) or Z score <-3, for haemoglobin ≤6.5 grams/decilitre (g/dl) and >6.5 g/dl, and for spleen size <10 centimetres (cm) or ≥10 cm.
Current guidelines for prognostic studies recommend to report both crude and adjusted measures of association [30]. However, adjusted measures are comparable only if they are based on multivariable models that include a comparable set of variables (key or core prognostic factors). As most of the studies included in our review either did not report adjusted measures or presented findings based on differing sets of variables, we decided to use only the crude measures of association for our meta-analyses. Findings from the few studies reporting multivariable analyses were described separately in the narrative synthesis. For these studies, we also reported on the types of models used, the procedures used for building the models, how validation were conducted and how performance of the models were assessed.
For the different predictors, we performed a fixed effects and random effects meta-analysis of the odds ratio of mortality. Since we pooled studies conducted in different settings and with high heterogeneity, the main conclusions were based on the random effects model. However, we also reported the fixed effects model for comparison. The amount of heterogeneity was quantified with the I square statistic (I2), which expressed the proportion of variation across studies that is due to heterogeneity.
We performed subgroup analyses for prognostic factors that were evaluated in at least five studies and where a variable of interest for subgroup analysis was also reported. Therefore, subgroup analyses for some prognostic factors were not conducted as they were studied in less than five studies. Subsequently, we only conducted subgroup analyses according to the HIV prevalence and the countries where the studies were conducted. The HIV prevalence was stratified as follows: HIV prevalence >90%, 10–90%, <10%, and unknown HIV prevalence.
For all prognostic factors evaluated in meta-analyses, funnel plots were made to assess the risk of publication bias [31,32]. Funnel plots visualize for all the included studies their precision against the reported strength of association. Plots taking the shape of a symmetric, inverted funnel suggest that publication bias is unlikely. All statistical analyses were performed with R version 3.6.0.
Results
Characteristics of the selected studies
Of the 1,524 studies identified, 48 were included in the systematic review [17,21–26,33–73] (Fig 1).
Most studies included in the systematic review were observational [77.0%, (37/48)] and the rest were clinical trials [23.0%, (11/48)] (S1 Table and S2 Table). Of the 11 trials, eight were randomised and three non-randomised. Most studies were conducted in Ethiopia [33.3%, (16/48)], followed by Sudan [31.3%, (15/48)], South Sudan [12.6%, (6/48)], Uganda [6.3%, (3/48)], Kenya [6.3%, (3/48)], and the rest were multicentre studies [10.4%, (5/48)]. Approximately one third (17/48, 35.4%) of the studies were conducted through a collaboration of non-governmental and governmental institutions, while 14 (29.2%) were conducted by non-governmental institutions and 13 (27.1%) by governmental institutions. Most studies [32/48 (66.6%)] were conducted in hospitals (S1 Table and S2 Table).
Patient inclusion criteria varied across studies: some studies included all VL patients, whereas others had strict enrolment criteria. For example, some studies focused on specific VL populations, such as HIV coinfected patients [36,45,57,72], people with severe VL or those failing first line treatment [45,62], patients on specific VL treatment regimens [23,36,40,46,52,56,72], or pregnant women [33]. There were also studies that excluded patients with the above criteria (e.g. HIV coinfected patients [35,39,56,69], or pregnancy [35,37,39,64,69]).
The total number of patients in the 48 studies [17,21–26,33–73] included in the systematic review was 150,072 and 7,847 (5.2%) of them died (S1 Table and S2 Table). In the observational studies [17,21–26,33–59,71–73], there were 146,698 patients and 7,703 (5.3%) died (S1 Table), whereas in the trials [60–70], there were 3,374 patients and 144 (4.3%) died (S2 Table). Patients from Sudan made up 72.2% of the total number, since most of the studies conducted there were large, while only 4.7% of all patients were from Ethiopia, where many studies were of smaller size. There were no studies included from Eritrea or Somalia (S1 Table and S2 Table). Out of the 48 studies [17,21–26,33–73], 27 were included in the meta-analysis [17,21–26,33–47,60,65,71–73] (Fig 1).
Pentavalent antimonials were the most commonly used treatment regimen. They were used in 39/48 studies and the death rate ranged from 0 to 55.6% (mean 6.8%) [17,21–26,34–42,47–55,57–66,68,70,71,73]. Liposomal amphotericin based regimens were used in 19/48 studies and the death rate ranged from 0 to 12.7% (mean 7.9%) [21,24,34,38,39,44–46,48,53,54,56,57,62,67,69,71–73].
Risk of bias
The risk of bias assessment for all the observational studies included in the systematic review (n = 37) is given in S1 Fig, and the assessment for the observational studies included in the meta-analysis (n = 25) in Fig 2.
The majority of these 25 studies [17,21–26,33–47,71–73] scored well (low risk of bias) in three of the six domains of the QUIPS tool, i.e. study participation, study attrition, and outcome measurement. The other three domains, i.e. prognostic factor measurement, study confounding, and statistical analysis and reporting were more problematic. Here, the majority of studies were classified as having moderate or high risk of bias (Fig 2).
The bias assessment for the 11 trials [60–70] included in the systematic review is given in S2 Fig. The two trials [60,65] included in the meta-analysis scored relatively well. In one trial [65], the risk of bias was low in all the domains except in study confounding, where the risk of bias was moderate. In the other trial [60], the risk of bias was low in the outcome measurement and study confounding domains, but it was moderate in all the other domains (S2 Fig).
Prognostic factors for mortality
A wide variety of potential prognostic factors were evaluated in the included studies (Table 2).
The prognostic factors that were assessed most frequently (in at least ten studies) were age, malnutrition, HIV status, duration of illness and treatment regimen. On the other hand, many factors were assessed in only one or two studies (Table 2).
The funnel plots for 12 prognostic factors are shown in supplementary information (S2 Text). The number of observations on the plots (i.e. the number of studies per prognostic factor) was relatively low (ranging between 5 and 11). The funnel plots that were least symmetrical were those for age and relapse VL, but overall, there were no indications of substantial publication bias.
Meta-analysis
Twelve prognostic factors (reported in 27 studies) were evaluated in at least five studies and were submitted to meta-analysis. The overview of the meta-analyses of the crude ORs can be found in Table 3.
The individual forest plots, subgroup analyses, and funnel plots for each prognostic factor are given in the supplementary information (S2 Text). There were ten factors for which we found a statistically significant association with mortality: HIV, tuberculosis, young or old age, jaundice, oedema, bleeding, anaemia, severe malnutrition, long duration of illness, and pronounced splenomegaly. Jaundice was the strongest predictor of mortality, with a pooled OR of 8.27, followed by HIV infection (OR 4.60) and tuberculosis (OR 4.06) (Table 3). For factors such as age below five years, pronounced splenomegaly, and longer duration of illness, the association was less pronounced, with an OR below two (Table 3).
Heterogeneity across studies was high for most prognostic factors, with the exception of jaundice, HIV status, bleeding, malnutrition, age below five years, and spleen size (Table 3). However, despite the large heterogeneity, for all prognostic factors except relapse vs. primary VL, the associations were found to be in the same direction. For relapse vs. primary VL, the heterogeneity could be explained by the proportion of HIV-infected patients in the studies. Indeed, a subgroup analysis showed that in studies with >90% HIV-infected patients the odds of dying was more than three times lower in relapse than in primary VL (OR 0.29, 95% CI 0.13–0.61) while for the other studies with a lower HIV prevalence, there was no clear association (S2 Text). In subgroup analysis by country (S2 Text), for most variables there were no clear patterns between countries because of wide and overlapping confidence intervals. However, some associations were different for Sudan compared to the other countries. Children <15 years had lower odds of death in all countries except Sudan, while for Sudan the odds ratio of dying after long duration of illness was higher than in Ethiopia. In Sudan, females have lower odds of dying than males, in Ethiopia and South Sudan there was no significant difference, while in Uganda, females had a higher chance of death.
Six factors (diarrhoea, vomiting, weakness, malaria, tissue parasite load and treatment regimen) were studied in at least three but less than five studies (S2 Text). For diarrhoea and vomiting, all available studies showed an association with higher mortality, but these findings were not always significant. For weakness, malaria, and tissue parasite loads, the estimates went in opposite directions (S2 Text). Although many different studies incorporated treatment as a prognostic factor, a variety of regimens were compared in each study, allowing for only three comparisons. These showed that amphotericin deoxycholate is associated with higher mortality than antimonials, pentostam was associated with higher mortality compared to generic SSG (non-significant) and SSG alone with higher mortality than a combination of SSG and paromomycin (non-significant) (S2 Text).
There were only nine studies reporting multivariable analyses [17,23–26,38,60,71,72]. Table 4 gives an overview of the prognostic factors included in the models and indicates which of them remained significantly associated with death after adjusting for other factors.
All multivariable analyses were based on logistic regression (S1 Table and S2 Table). Most predictors analysed were based on findings from previous studies. Predictor selection during modelling was described in only 4 [23,71,72,75] of the 9 studies reported in Table 4. Two studies developed and externally validated clinical prediction tools [38,71]. The procedures for validation and assessment of performance of the clinical prediction tools were fully described. As a measure of discrimination, areas under the receiver operating curve were reported: 0.83 for the Ethiopian study (adults only), 0.74 for the model including adults in Sudan and 0.83 for children/adolescents in Sudan. Calibration was not done.
The prognostic factors included in multivariable analyses differed widely across studies, precluding the conduct of a meta-analysis of adjusted measures of effect. Some factors were fairly consistently identified as prognostic factors in adjusted analyses, such as HIV, tuberculosis, jaundice, bleeding, weakness, and anaemia (Table 4). Other factors such as malnutrition, oedema/ascites, young/old age, vomiting, and diarrhoea were identified in some, but not in others. For factors such as spleen size and duration of illness, most studies did not find a significant association in adjusted analyses (Table 4).
There were only two studies that developed and validated a clinical prediction tool and reported absolute risks of mortality (Abongomera et. al. [71] and Kamink et. al. [38] in Table 4). One of these studies was conducted in Ethiopia in an area with a high HIV prevalence where VL is particularly common amongst young male HIV infected migrant workers [71]. The other study in South Sudan excluded HIV infected patients and developed a separate tool for children/adolescents and adults [38]. Even though a few factors such as bleeding and tuberculosis were not assessed in the South Sudanese study, independent prognostic factors were very similar in both studies, such as jaundice, anaemia, weakness, and oedema/ascites [38,71]. Factors such as spleen size and duration of symptoms were not retained in both studies. In a setting with high HIV prevalence in Ethiopia, HIV coinfection and tuberculosis were also retained as prognostic factors [71].
Discussion
We conducted a systematic review and meta-analysis of prognostic factors for mortality in VL patients in East Africa. A total of 48 studies [17,21–26,33–73] were included, out of which 12 prognostic factors could be submitted to meta-analysis (sourced from 27 studies [17,21–26,33–47,60,65,71–73]). Out of these, ten factors were found to be significantly associated with mortality. HIV and tuberculosis are concurrent infections which are prevalent in East Africa. Other markers such as jaundice, oedema, pronounced splenomegaly, anaemia, bleeding, duration of illness, and malnutrition likely indicate severity and/or progression of the disease. Additionally, old and young age were also associated with mortality. Most of these prognostic factors can be easily identified by health professionals in resource-constrained settings. All of these were also identified as prognostic factors in a meta-analysis from Latin America [20], suggesting they apply across continents. Future prognostic studies within East Africa evaluating novel or other prognostic markers should consider these as “core” prognostic factors for inclusion in multivariable analysis.
While, ideally, adjusted estimates would have been calculated in meta-analysis, this was not possible due to the wide variability of prognostic factors included in multivariable analysis in the different studies. Only one study included all core factors [71]. Nevertheless, several of the core factors were found to retain their prognostic value in the studies performing multivariable analysis. The effect of age varied across the few studies [17,23–26,38,71].
We found two studies which developed clinical prediction tools [38,71]. The factors included in these two tools are very similar to those recently proposed in a clinical prediction tool in Brazil [20]. It remains to be defined whether the two tools also perform well in neighbouring countries within East Africa. We note that we could not include any studies on prognostic factors from Eritrea or Somalia.
Meanwhile, these tools could be used in routine practice for early detection of VL cases at high risk of death. This would require systematic assessment of several clinical factors and some laboratory testing (HIV, tuberculosis, haemoglobin). Patients with a predicted high risk of death could be triaged to the appropriate medical department or services providing more intensive care or closer patient observation. Those with a very low risk could be managed at the decentralised and/or ambulatory level. Nevertheless, impact studies are required to quantify the effectiveness of the implementation of such tools to reduce VL mortality.
Future studies could also aim to further improve the current prediction tools by adding a number of factors (especially laboratory tests) which were found predictive in other continents, but which are not yet (fully) assessed in East Africa. This includes renal impairment, elevated liver function tests, serum albumin levels, thrombocytopenia, and leukocytopenia [20]. However, some of these tests might not be available at the lower health care level in East-African countries. Diarrhoea and vomiting were retained as independent prognostic factors in the Latin American meta-analysis [20], but were not frequently studied in East Africa (and therefore not included in the present meta-analysis).
Bacterial coinfection was also found to be associated with mortality in several studies in Latin America [20]. No studies assessed this in East Africa. There was however one study which found sepsis to be associated with adverse outcomes, defined as death or treatment failure (composite outcome) [21]. Of interest, bacterial sepsis has been found in one in five patients in a study in Ethiopia [76]. Particularly given the increasing burden of multi-drug resistant bacterial infections globally, detailed studies on the prevalence, pattern, and prognostic significance of bacterial infections in VL in East Africa are merited.
Many of the studies used routine data and were retrospective in nature. Consequently, only a limited number of prognostic factors were assessed in most studies. Moreover, differences in definitions and cut-offs of prognostic factors or in the reliability of the clinical assessments likely exist across studies. This could explain the considerable heterogeneity found in the present review. However, despite this heterogeneity, the associations with mortality generally went in the same direction. Standardised prospective data collection within VL programs (or in selected sites) would improve the quality and consistency of the available data, and such an enhanced routine data collection system should include the core indicators of mortality. Furthermore, a recent initiative for data sharing for infectious diseases–including VL—could make the meta-analysis of individual patient data possible [77]. We did not have data on drug resistance. However this is probably of a lesser concern in the East African region, compared to the Indian subcontinent where high resistance rates have been reported [15].
Conclusion
A total of 48 studies [17,21–26,33–73] were included with a total population of 150,072 VL cases from five countries. Ten prognostic factors were identified in the meta-analyses. These factors should be considered “core” prognostic factors in future prognostic studies. Several of these such as HIV, tuberculosis, jaundice, bleeding, and weakness remained significant after adjustment for other factors in the few studies conducting multivariable analysis [17,23–26,38,71]. One clinical prediction tool was developed in Ethiopia [71], another in South Sudan [38]. As both performed relatively well, future prognostic factor studies should evaluate whether these tools also have a good performance in neighbouring countries, and whether they can be further improved by adding additional prognostic factors not yet (fully) explored within East Africa.
Supporting information
S2 Text. Forest plots, subgroup analyses, and funnel plots for each prognostic factor.
1. Forest plots displaying the crude odds ratios of the association between the prognostic factors and mortality for factors included in meta-analysis (at least five estimates); 2. Subgroup analysis according to HIV prevalence and Country; 3. Forest plots displaying the crude odds ratios of the association between the prognostic factors and mortality for factors with at least three but less than five estimates and mortality; 4. Funnel plots.
https://doi.org/10.1371/journal.pntd.0008319.s002
(DOCX)
S1 Table. Overview and summary of the observational studies included.
https://doi.org/10.1371/journal.pntd.0008319.s003
(DOCX)
S2 Table. Overview and summary of the trials included.
https://doi.org/10.1371/journal.pntd.0008319.s004
(DOCX)
S1 Fig. Assessment of risk of bias for 37 observational studies using the Quality in Prognostic Studies (QUIPS) tool.
(Figure structure: on top “studies included in the meta-analysis”–below: “studies included in narrative synthesis only”).
https://doi.org/10.1371/journal.pntd.0008319.s005
(JPG)
S2 Fig. Assessment of risk of bias for 11 trials using the Quality in Prognostic Studies (QUIPS) tool.
(Figure structure: on top “studies included in the meta-analysis”–below: “studies included in narrative synthesis only”).
https://doi.org/10.1371/journal.pntd.0008319.s006
(JPG)
References
- 1. Murray HW, Berman JD, Davies CR, Saravia NG. Advances in leishmaniasis. Lancet. 2005;366:1561–77. pmid:16257344
- 2. Chappuis F, Sundar S, Hailu A, Ghalib H, Rijal S, Peeling RW, et al. Visceral leishmaniasis: what are the needs for diagnosis, treatment and control? Nat Rev Microbiol. 2007;5:873–82. pmid:17938629
- 3. World Health Organization. Global leishmaniasis update, 2006–2015: a turning point in leishmaniasis surveillance. Wkly Epidemiol Rec. 2017;38:557–72.
- 4.
World Health Organization. Leishmaniasis [Internet]. 2019. Available from: https://www.who.int/en/news-room/fact-sheets/detail/leishmaniasis
- 5.
Drugs for Neglected Diseases Initiative. Leishmaniasis—Disease fact sheet [Internet]. 2018. Available from: https://www.dndi.org/wp-content/uploads/2018/12/Factsheet2018_Leishmaniasis.pdf
- 6. World Health Organization. Leishmaniasis country profile—2015: South Sudan. 2017.
- 7. World Health Organization. Leishmaniasis country profile—2015: Sudan. 2017.
- 8. World Health Organization. Leishmaniasis country profile—2015: Ethiopia. 2017.
- 9. World Health Organization. Leishmaniasis country profile—2015: Kenya. 2017.
- 10. World Health Organization. Leishmaniasis country profile—2015: Uganda. 2017.
- 11. World Health Organization. Leishmaniasis country profile—2015: Somalia. 2017.
- 12. van Griensven J, Diro E. Visceral leishmaniasis. Infect Dis Clin North Am. 2012;26:309–22. pmid:22632641
- 13. Laguna F, López-vélez R, Pulido F, Salas A, Torre-cisneros J, Torres E, et al. Treatment of visceral leishmaniasis in HIV-infected patients: a randomized trial comparing meglumine antimoniate with amphotericin B. AIDS. 1999;13:1063–9. pmid:10397536
- 14. Delgado J, Macías J, Pineda JA, Corzo JE, González-Moreno MP, de la Rosa R, et al. High frequency of serious side effects from meglumine antimoniate given without an upper limit dose for the treatment of visceral leishmaniasis in human immunodeficiency virus type-1-infected patients. Am J Trop Med Hyg. 1999;61:766–9. pmid:10586909
- 15.
World Health Organization. Control of the leishmaniases. [Internet]. WHO Technical Report Series. 2010. Available from: http://apps.who.int/iris/bitstream/10665/44412/1/WHO_TRS_949_eng.pdf
- 16. Diro E, Lynen L, Ritmeijer K, Boelaert M, Hailu A, van Griensven J. Visceral Leishmaniasis and HIV coinfection in East Africa. PLoS Negl Trop Dis. 2014;8:e2869. pmid:24968313
- 17. Mengistu G, Ayele B. Visceral Leishmaniasis and HIV co-infection in patients admitted to Gondar university hospital, northwest Ethiopia. Ethiop J Heal Dev. 2007;21:53–60.
- 18. Knaus WA, Wagner DP, Draper EA, Zimmerman JE, Bergner M, Bastos PG, et al. The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest. 1991;100:1619–36. pmid:1959406
- 19. Canet J, Gallart L, Gomar C, Paluzie G, Vallès J, Castillo J, et al. Prediction of postoperative pulmonary complications in a population-based surgical cohort. Anesthesiology. 2010;113:1338–50. pmid:21045639
- 20. Belo VS, Struchiner CJ, Barbosa DS, Nascimento BWL, Horta MAP, da Silva ES, et al. Risk Factors for Adverse Prognosis and Death in American Visceral Leishmaniasis: A Meta-analysis. PLoS Negl Trop Dis. 2014;8:e2982. pmid:25058582
- 21. Hurissa Z, Gebre-Silassie S, Hailu W, Tefera T, Lalloo DG, Cuevas LE, et al. Clinical characteristics and treatment outcome of patients with visceral leishmaniasis and HIV co-infection in northwest Ethiopia. Trop Med Int Health. 2010;15:848–55. pmid:20487426
- 22. Collin S, Davidson R, Ritmeijer K, Keus K, Melaku Y, Kipngetich S, et al. Conflict and Kala-Azar: Determinants of Adverse Outcomes of Kala-Azar among Patients in Southern Sudan. Clin Infect Dis. 2004;38:612–9. pmid:14986243
- 23. Mueller Y, Mbulamberi DB, Odermatt P, Hoffmann A, Loutan L, Chappuis F. Risk factors for in-hospital mortality of visceral leishmaniasis patients in eastern Uganda. Trop Med Int Health. 2009;14:910–7. pmid:19552645
- 24. Herrero M, Orfanos G, Argaw D, Mulugeta A, Aparicio P, Parreño F, et al. Natural history of a visceral leishmaniasis outbreak in highland Ethiopia. Am J Trop Med Hyg. 2009;81:373–7. pmid:19706898
- 25. Lyons S, Veeken H, Long J. Visceral leishmaniasis and HIV in Tigray, Ethiopia. Trop Med Int Heal. 2003;8:733–9.
- 26. Seaman J, Mercer AJ, Sondorp HE, Herwaldt BL. Epidemic visceral leishmaniasis in southern Sudan: Treatment of severely debilitated patients under wartime conditions and with limited resources. Ann Intern Med. 1996;124:664–72. pmid:8607595
- 27. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JPA, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6:e1000100. pmid:19621070
- 28. Abongomera Charles, Mohammed Rezika, Vogt Florian, Buyze Jozefien, Diro Ermias, Ritmeijer Koert, Verdonck Tine, van Griensven Johan. International prospective register of systematic reviews A systematic review of risk factors for mortality in visceral leishmaniasis patients in East Africa [Internet]. 2016. Available from: http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42016043112
- 29. Hayden JA, Windt DA Van Der, Cartwright JL, Côté P. Research and Reporting Methods Annals of Internal Medicine Assessing Bias in Studies of Prognostic Factors. Ann Intern Med. 2013;144:427–37.
- 30. Peat G, Riley RD, Croft P, Morley KI, Kyzas PA, Moons KGM, et al. Improving the Transparency of Prognosis Research: The Role of Reporting, Data Sharing, Registration, and Protocols. PLoS Med. 2014;11:e1001671. pmid:25003600
- 31.
The Cochrane Collaboration. Cochrane Handbook for Systematic Reviews of Interventions. 2011.
- 32. Sterne JAC, Sutton AJ, Ioannidis JPA, Terrin N, Jones DR, Lau J. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised. Br Med J. 2011;342:d4002.
- 33. Adam GK, Omar SM, Ahmed MA, Abdallah TM, Ali AA. Cross-sectional study of the case–fatality rate among patients with visceral leishmaniasis infections during pregnancy in Sudan. Int J Gynecol Obstet. 2017;140:119–20.
- 34. Aderie EM, Diro E, Zachariah R, Abongomera C, Dolamo BL, Ritmeijer K. Does timing of antiretroviral treatment influence treatment outcomes of visceral leishmaniasis in Northwest Ethiopia? Trans R Soc Trop Med Hyg. 2017;111:107–16. pmid:28633331
- 35. Atia AM, Mumina A, Tayler-smith K, Boulle P, Alcoba G, Siddig M. Sodium stibogluconate and paromomycin for treating visceral leishmaniasis under routine conditions in eastern Sudan. Trop Med Int Heal. 2015;20:1674–84.
- 36. Diro E, Lynen L, Mohammed R, Boelaert M, Hailu A, van Griensven J. High parasitological failure rate of visceral leishmaniasis to sodium stibogluconate among HIV co-infected adults in Ethiopia. PLoS Negl Trop Dis. 2014;8:e2875. pmid:24854196
- 37. Hailu W, Weldegebreal T, Hurissa Z, Tafes H, Omollo R, Yifru S, et al. Safety and effectiveness of meglumine antimoniate in the treatment of Ethiopian visceral leishmaniasis patients with and without HIV co-infection. Trans R Soc Trop Med Hyg. 2010;104:706–12. pmid:20870258
- 38. Kämink SS, Collin SM, Harrison T, Gatluak F, Mullahzada AW, Ritmeijer K. A clinical severity scoring system for visceral leishmaniasis in immunocompetent patients in South Sudan. PLoS Negl Trop Dis. 2017;11:e0005921. pmid:28968400
- 39. Khalil EAG, Hassan AM El, Zijlstra EE, Hashim FA, Ibrahim ME, Ali MS, et al. Treatment of visceral leishmaniasis with sodium stibogluconate in Sudan: management of those who do not respond. Ann Trop Med Parasitol. 1998;92:151–8. pmid:9625910
- 40. Kimutai R, Musa AM, Njoroge S, Omollo R, Alves F, Hailu A, et al. Safety and Effectiveness of Sodium Stibogluconate and Paromomycin Combination for the Treatment of Visceral Leishmaniasis in Eastern Africa: Results from a Pharmacovigilance Programme. Clin Drug Investig. 2017;37:259–72. pmid:28066878
- 41. Maru M. Clinical and laboratory features and treatment of visceral leishmaniasis in hospitalized patients in Northwestern Ethiopia. Am J Trop Med Hyg. 1979;28:15–8. pmid:219723
- 42. Mengesha B, Abuhoy M. Kala-azar among labour migrants in metema-humera region of ethiopia. Trop Geogr Med. 1978;30:199–206. pmid:726033
- 43. Mohammed HB, Ali A, Idriss MI, Gasmelseid M, Yousif MM, Ibrahem AA. Prevalence of Hepatitis B, Hepatitis C, HIV and Malaria Co Infection among Patients Infected with Visceral Leishmaniasis in Gedarif, Eastern Sudan. Glob J Infect Dis Clin Res. 2016;2:21–4.
- 44. Omar SM, Ma K, Ahmed MAA, Fadul Y, Adam GK, Abdallah TM, et al. Current Research on HIV/AIDS Seroprevalence of HIV Infection among Visceral Leishmaniasis Patients in Gadarif, Eastern Sudan. Curr Res HIV/AIDS. 2016;2016:1–4.
- 45. Ritmeijer K, ter Horst R, Chane S, Aderie EM, Piening T, Collin SM, et al. Limited effectiveness of high-dose liposomal amphotericin B (AmBisome) for treatment of visceral leishmaniasis in an Ethiopian population with high HIV prevalence. Clin Infect Dis. 2011;53:e152–8. pmid:22016502
- 46. Salih NA, van Griensven J, Chappuis F, Antierens A, Mumina A, Hammam O, et al. Liposomal amphotericin B for complicated visceral leishmaniasis (kala-azar) in eastern Sudan: How effective is treatment for this neglected disease? Trop Med Int Heal. 2014;19:146–52.
- 47. Zijlstra EE, Ali MS, el-Hassan AM, el-Toum IA, Satti M, Ghalib H. Clinical aspects of kala-azar in children from the Sudan: a comparison with the disease in adults. J Trop Pediatr. 1992;38:17–21. pmid:1315397
- 48. Boateng K, Usman A, Pita J, Mpairwe A, Bimpa D-D, Lyosi E, et al. The Emerging Questions from the Current Epidemiology of Visceral Leishmaniasis in South Sudan: A Cross-sectional Study Design. J Emerg Infect Dis. 2017;2:1–7.
- 49. Chulay JD, Spencer HC, Mugambi M. Electrocardiographic changes during treatment of leishmaniasis with pentavalent antimony (sodium stibogluconate). Am J Trop Med Hyg. 1985;34:702–9. pmid:2992303
- 50. Gorski S, Collin SM, Ritmeijer K, Keus K, Gatluak F, Mueller M, et al. Visceral leishmaniasis relapse in Southern Sudan (1999–2007): a retrospective study of risk factors and trends. PLoS Negl Trop Dis. 2010;4:e705. pmid:20544032
- 51. Kager PA, Rees PH, Manguyu FM, Bhatt KM, Wellde BT, Hockmeyer WT LWJ. Clinical, haematological and parasitological response to treatment of visceral leishmaniasis in Kenya. A study of 64 patients. Trop Geogr Med. 1984;36:21–35. pmid:6328709
- 52. Melaku Y, Collin SM, Keus K, Gatluak F, Ritmeijer K, Davidson RN. Treatment of kala-azar in southern Sudan using a 17-day regimen of sodium stibogluconate combined with paromomycin: a retrospective comparison with 30-day sodium stibogluconate monotherapy. Am J Trop Med Hyg. 2007;77:89–94. pmid:17620635
- 53. Mueller Y, Nguimfack A, Cavailler P, Couffignal S, Rwakimari JB, Loutan L, et al. Safety and effectiveness of amphotericin B deoxycholate for the treatment of visceral leishmaniasis in Uganda. Ann Trop Med Parasitol. 2008;102:11–9. pmid:18186974
- 54. Mueller YK, Kolaczinski JH, Koech T, Lokwang P, Riongoita M, Velilla E, et al. Clinical epidemiology, diagnosis and treatment of visceral leishmaniasis in the Pokot endemic area of Uganda and Kenya. Am J Trop Med Hyg. 2014;90:33–9. pmid:24218406
- 55. Ritmeijer K, Davidson RN. Royal Society of Tropical Medicine and Hygiene joint meeting with Médecins Sans Frontières at Manson House, London, 20 March 2003: field research in humanitarian medical programmes. Médecins Sans Frontières interventions against kala-azar in the Sudan, 19. Trans R Soc Trop Med Hyg. 2003;97:609–13. pmid:16134257
- 56. Tamiru A, Tigabu B, Yifru S, Diro E, Hailu A. Safety and efficacy of liposomal amphotericin B for treatment of complicated visceral leishmaniasis in patients without HIV, North-West Ethiopia. BMC Infect Dis. 2016;16:548. pmid:27724891
- 57. van Griensven J, Simegn T, Endris M, Diro E. Visceral leishmaniasis and HIV co-infection in Northwest Ethiopia: Antiretroviral treatment and burden of disease amongst patients enrolled in HIV care. Am J Trop Med Hyg. 2017;1–10.
- 58. van den Bogaart E, Berkhout MMZ, Adams ER, Mens PF, Sentongo E, Mbulamberi DB, et al. Prevalence, features and risk factors for malaria co-infections amongst visceral leishmaniasis patients from Amudat Hospital, Uganda. PLoS Negl Trop Dis. 2012;6:e1617. pmid:22506087
- 59. van den Bogaart E, Berkhout MMZ, Nour ABYM, Mens PF, Talha A-BA, Adams ER, et al. Concomitant malaria among visceral leishmaniasis in-patients from Gedarif and Sennar States, Sudan: a retrospective case-control study. BMC Public Health. 2013;13:332. pmid:23577673
- 60. Ritmeijer K, Dejenie A, Assefa Y, Hundie TB, Mesure J, Boots G, et al. A Comparison of Miltefosine and Sodium Stibogluconate for Treatment of Visceral Leishmaniasis in an Ethiopian Population with High Prevalence of HIV Infection. Clin Infect Dis. 2006;43:357–64. pmid:16804852
- 61. Hailu A, Musa A, Wasunna M, Balasegaram M, Yifru S, Mengistu G, et al. Geographical variation in the response of visceral leishmaniasis to paromomycin in East Africa: a multicentre, open-label, randomized trial. PLoS Negl Trop Dis. 2010;4:e709. pmid:21049059
- 62. Khalil E.A.G., Hashim F. A., Ali M. S., Ghalib H.W., Zijlstra E.E; EL. Hag I. A., Mustafa M.D., Satti M. H., Kordafani A., El-Safi S., Ali Hag M., El Hassan A. M. Comparative study of liposomal amphotericin B (ambisome) and sodium stibogluconate in the treatment of kala-azar in the sudan. East Afr Med J. 1998;75:481–4.
- 63. Moore E, O’Flaherty D, Heuvelmans H, Seaman J, Veeken H, de Wit S, et al. Comparison of generic and proprietary sodium stibogluconate for the treatment of visceral leishmaniasis in Kenya. Bull World Health Organ. 2001;79:388–93. pmid:11417033
- 64. Musa A, Khalil E, Hailu A, Olobo J, Balasegaram M, Omollo R, et al. Sodium stibogluconate (SSG) & paromomycin combination compared to SSG for visceral leishmaniasis in East Africa: a randomised controlled trial. PLoS Negl Trop Dis. 2012;6:e1674. pmid:22724029
- 65. Ritmeijer K, Veeken H, Melaku Y, Leal G, Amsalu R, Seaman J, et al. Ethiopian visceral leishmaniasis: generic and proprietary sodium stibogluconate are equivalent; HIV co-infected patients have a poor outcome. Trans R Soc Trop Med Hyg. 2001;95:668–72. pmid:11816442
- 66. Seaman J., Pryce D., Sondorp H., Moody A., Bryceson A., Davidson RN. Epidemic visceral leishmaniasis in Sudan: a randomized trial of aminosidine plus sodium stibogluconate versus sodium stibogluconate alone. J Infect Dis. 1993;168:715–20. pmid:8394861
- 67. Seaman J, Boer C, Wilkinson R, de Jong J, de Wilde E, Sondorp E, et al. Liposomal amphotericin B (AmBisome) in the treatment of complicated kala-azar under field conditions. Clin Infect Dis. 1995;21:188–93. pmid:7578729
- 68. Veeken H, Ritmeijer K, Seaman J, Davidson R. A randomized comparison of branded sodium stibogluconate and generic sodium stibogluconate for the treatment of visceral leishmaniasis under field conditions in Sudan. Trop Med Int Heal. 2000;5:312–7.
- 69. Wasunna M, Njenga S, Balasegaram M, Alexander N, Omollo R, Edwards T, et al. Efficacy and Safety of AmBisome in Combination with Sodium Stibogluconate or Miltefosine and Miltefosine Monotherapy for African Visceral Leishmaniasis: Phase II Randomized Trial. PLoS Negl Trop Dis. 2016;10:e0004880. pmid:27627654
- 70. Zijlstra EE, Ali MS, El-Hassan AM, El-Toum I, Satti M, Ghalib HW, et al. The treatment of kala-azar in the Sudan with sodium stibogluconate: a randomized trial of three dosage regimens. Trans R Soc Trop Med Hyg. 1993;87:307–9. pmid:8236402
- 71. Abongomera C, Ritmeijer K, Vogt F, Buyze J, Mekonnen Z, Admassu H, et al. Development and external validation of a clinical prognostic score for death in visceral leishmaniasis patients in a high HIV co-infection burden area in Ethiopia. PLoS One. 2017;12:e0178996. pmid:28582440
- 72. Abongomera C, Diro E, de Lima Pereira A, Buyze J, Stille K, Ahmed F, et al. The initial effectiveness of liposomal amphotericin B (AmBisome) and miltefosine combination for treatment of visceral leishmaniasis in HIV co-infected patients in Ethiopia: A retrospective cohort study. PLoS Negl Trop Dis. 2018;12:e0006527. pmid:29799869
- 73. Adam GK, Ali KM, Abdella YH, Omar SM, Ahmed MAA, Abdalla TM, et al. Trend in cumulative cases and mortality rate among visceral leishmaniasis patients in Eastern Sudan: a 14-year registry, 2002–2015. Int J Infect Dis. 2016;51:81–4. pmid:27596686
- 74. Seaman J, Mercer AJ, Sondorp E. The epidemic of visceral leishmaniasis in western Upper Nile, southern Sudan: course and impact from 1984 to 1994. Int J Epidemiol. 1996;25:862–71. pmid:8921468
- 75. Kämink SS, Collin SM, Harrison T, Gatluak F, Mullahzada AW, Ritmeijer K. A clinical severity scoring system for visceral leishmaniasis in immunocompetent patients in South Sudan. PLoS Negl Trop Dis. 2017;11:e0005921. pmid:28968400
- 76. Endris M, Takele Y, Woldeyohannes D, Tiruneh M, Mohammed R, Moges F, et al. Bacterial Sepsis in Patients with Visceral Leishmaniasis in Northwest Ethiopia. Biomed Res Int. 2014;2014:1–7.
- 77.
Infectious Diseases Data Observatory. Infectious Diseases Data Observatory (IDDO): Visceral leishmaniasis [Internet]. online. 2019 [cited 2019 Jul 2]. Available from: https://www.iddo.org/research-themes/visceral-leishmaniasis