• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Lancet Infect Dis. Author manuscript; available in PMC Sep 8, 2011.
Published in final edited form as:
PMCID: PMC3168734
NIHMSID: NIHMS318140

Community-acquired bloodstream infections in Africa: a systematic review and meta-analysis

Abstract

Data on the prevalence and causes of community-acquired bloodstream infections in Africa are scarce. We searched three databases for studies that prospectively studied patients admitted to hospital with at least a blood culture, and found 22 eligible studies describing 58 296 patients, of whom 2051 (13.5%) of 15 166 adults and 3527 (8.2%) of 43 130 children had bloodstream infections. 1643 (29.1%) non-malaria bloodstream infections were due to Salmonella enterica (58.4% of these non-typhoidal Salmonella), the most prevalent isolate overall and in adults, and 1031 (18.3% overall) were due to Streptococcus pneumoniae, the most common isolate in children. Other common isolates included Staphylococcus aureus (531 infections; 9.5%) and Escherichia coli (412; 7.3%). Mycobacterium tuberculosis complex accounted for 166 (30.7%) of 539 isolates in seven studies that used mycobacterial culture techniques. HIV infection was associated with any bloodstream infection, particularly with S enterica and M tuberculosis complex bacteraemia. Where recorded, patients with bloodstream infections had an in-hospital case fatality of 18.1%. Our results show that bloodstream infections are common and associated with high mortality. Improved clinical microbiology services and reassessment of empirical treatment guidelines that account for the epidemiology of bloodstream infections might contribute to better outcomes.

Introduction

Febrile illness is a leading reason for admission to hospital in Africa,14 and rates of febrile illness are fuelled by the HIV epidemic.59 Despite the major contribution of infectious diseases to hospital admission, the availability of diagnostic microbiology services for bloodstream infections other than malaria is often limited by cost, infrastructure, and personnel constraints.10,11 Consequently, health-care workers must often rely on syndrome-oriented empirical approaches to treatment and might underestimate or overestimate the likelihood of certain diseases, risking poor clinical outcomes and the promotion of antimicrobial resistance.1214 Understanding the causes and prevalence of community-acquired bloodstream infection, which is associated with high risk of death, can inform efforts to improve health outcomes in Africa and promote the meeting of millennium development goals for the reduction of child mortality and HIV/AIDS, malaria, and tuberculosis.

Early studies of bloodstream infections in children admitted to African hospitals suggest that the prevalence of bacterial bloodstream infections among inpatients with fever or clinical sepsis exceeds that described in wealthier regions2,1518 and that bacteraemia is a common cause of illness both in areas of high and low malaria prevalence. Gram-negative organisms, particularly Salmonella enterica, rival or exceed Gram-positive organisms in importance in several published reports on bloodstream infections in both adults and children from African countries.1922 In recent years, use of blood culture to assess seriously ill patients infected with HIV has led to a growing understanding of their increased risk of a range of invasive bacterial and fungal diseases, including Streptococcus pneumoniae, disseminated tuberculosis, cryptococcosis, and Salmonella bacteraemia caused by non-typhoidal Salmonella.4,2327

We sought to review studies that used blood culture to identify non-malaria bloodstream infections among prospectively sampled adults and children with predefined, replicable, inclusion criteria admitted to hospitals in Africa, and then to aggregate these data to better quantify the prevalence of bloodstream infections and document the most commonly isolated organisms overall and among different subgroups. We postulated that bloodstream infections would be identified among many patients admitted to hospital in Africa, S enterica would be among the most commonly isolated pathogens, and that age, presence of HIV infection, and features of illness would affect the prevalence of bloodstream infections and predominant organisms isolated.

Methods

Search strategy and selection criteria

We searched two major scientific databases (PubMed and Embase) and one topical database (African Healthline) with terms defined with the assistance of a library science technologist (Megan Von Isenburg). PubMed was searched with the search string: “Africa and (fever or fevers or bacteremia or bacteremias or septicemia or septicemias) limit humans”. Embase was searched by use of the terms: “Africa” (exploded to all subheadings) and “fever/” or “fever.mp” or “fevers.mp” or “bacteremia/” or “bacteremia.mp” or “bacteraemia.mp” or “bacteremias. mp” or “septicemia/” or “septicemia.mp” or “septicaemia. mp” or “septicemias.mp”. Results were limited to humans. African Healthline was searched by use of the string: “Africa and (fever* or bacteremia* or septicemia*) limit humans and scholarly (peer reviewed) journals”. Abstracts and titles from all years and in all languages— translated though services provided by the search engines as needed—were compiled in Endnote (Thomson Reuters) and reviewed individually by two investigators (EAR and AVS) with a goal to locate articles that seemed to report on the use of bacterial blood culture to assess patients admitted to hospital. All articles flagged by either investigator as possibilities for inclusion were retrieved in full text and their references were again independently assessed by the same two investigators by use of PubMed to obtain abstracts as needed. Each article identified by this process as a candidate for inclusion was retrieved as full text. A checklist of predetermined inclusion and exclusion criteria was developed by all three authors and was used for independent assessment each of the full-text articles by two of the three authors (EAR and AVS). After both of these investigators had made a decision on inclusion versus exclusion, any disagreement was resolved through independent review by the third investigator. Full-text articles in French were read by two of the three investigators (JAC and EAR). The names of the authors of candidate articles were not masked. Conference abstracts were not included in this review, because during the selection process it was noted that inclusion criteria (subjective vs predefined, retrospective vs prospective) were not often precisely defined in abstracts and therefore determination of important selection biases would not be possible. A systematic means by which to search for reliable so-called grey literature or unpublished reports on this topic was not available.

Included studies were required to be prospective, to recruit systematically or consecutively sampled paediatric or adult hospital admissions, and to evaluate all admissions, all febrile admissions, or all febrile admissions without a focus of infection using, at least, an aerobic blood culture. Studies specifying use of predefined criteria that included afebrile patients with suspected infections in addition to febrile patients were also included.

During abstract review, we excluded articles that investigated a single or narrow cause of febrile illness; studies with the primary goal of investigating diagnostics or treatments; and review articles, editorials, policy statements, and behavioural research. We also excluded articles describing illness in people living outside the African continent (eg, returned travellers), articles that only assessed bloodstream infections in specific risk groups (eg, sickle-cell anaemia), non hospital-based studies, retrospective analyses, case reports, studies focused on nosocomial infections, and studies during an epidemic or outbreak.

During full-text review, we excluded studies that examined specific risk populations or used retrospective inclusion techniques (eg, bacteraemia as the trigger for study entry), and those that used subjective or poorly defined inclusion criteria. Articles in which separation of outpatient and inpatient cohorts was not possible, and those that reported on patient cohorts already included in another more comprehensive study were excluded. We also excluded articles that did not distinguish between community-acquired and nosocomial infections. All articles were required to quantify the total number of blood cultures obtained and the total number of pathogenic isolates, as well as to identify the three most commonly isolated pathogens and detail the numbers of patients from whom these pathogens were isolated. If incomplete data or unclear methods precluded article inclusion, authors were contacted through the contact information provided by the article or obtained though internet search engines. Only studies meeting our minimum requirements for data completeness described above were included.

Validity assessment

Study validity was established by use of the selection criteria described above, thereby excluding studies that were thought likely to have invalid results or whose results could not be compared with studies included in the analysis. We expected variability to exist between microbiological techniques and interpretation of culture results, and that such variability probably reflects realities of studies of this type done in resource-limited settings. Therefore studies were not excluded on the basis of having used media and transport techniques that might have limited the isolation of fastidious pathogens. Studies were also not excluded on the basis of lack of detailed reporting of contaminants; however, the prevalence of possible contaminants among the organisms described as pathogenic was recorded and its possible effect on the prevalence of bloodstream infections is discussed. Heterogeneity was controlled primarily by specifying the use of predefined inclusion criteria. However, since the goal of the review was to broadly assess prevalence and type of community-acquired bloodstream infections across Africa, it was important that included articles encompassed expected diversity in different hospital communities (eg, academic, government, rural, urban) without focusing too narrowly on specific patient groups at risk for bloodstream infections (eg, patients in whom treatment for malaria had already failed).28 Subgroup analyses were done to explore the effect of heterogeneity on the overall prevalence or predominant types of bloodstream infections. Publication bias was not systematically calculated, since negative studies would be extremely unlikely and no standard for measuring expected prevalence of bloodstream infections in Africa exists. However, the possibility of publication bias in exaggerating the prevalence of bloodstream infections is discussed qualitatively.29

Data extraction

Descriptive and quantitative data from each paper were extracted individually by two investigators (EAR and AVS) and entered into a Microsoft Excel 2007 spreadsheet; this included hospital setting and location, region of Africa according to regional groupings detailed within the UNAIDS epidemic update 2007,30 study time frame, specific inclusion or exclusion criteria, and culture techniques. Quantitative data collected included number of patients, age range, pathogens and contaminants isolated, antimicrobial susceptibilities, and use of additional tests (eg, for HIV or malaria). Paediatric studies were defined as those in which all included patients were younger than 15 years; studies with mixed populations of adult and children were analysed as adult studies. Inconsistencies between investigators after data extraction were resolved by return to the original papers.

Statistical analysis

Data on individual patients were compiled and analysed in aggregate to compare prevalence of bloodstream infections across studies and among different subpopulations. Analyses of associations between types of patients or clinical conditions (eg, HIV) and specific bloodstream infections were only done for studies in which full data were available for both the pathogens and factors being assessed; no data were imputed for any analyses. The χ2 test or Fisher's exact test for frequencies less than five was used to establish the significance of associations made to bloodstream infections or specific causative organisms; values were expressed as odds ratios (ORs) calculated with JMP statistical software version 7[bullet]0.

Results

The online database search done on June 11, 2009, yielded 10 412 articles, 7596 of which were unique articles located in at least one of the three databases (figure 1): 3366 were unique to PubMed, 1244 unique to Embase, and 180 unique to African Healthline. Most of these articles were excluded on the basis of a primary topical focus other than assessment of suspected infection in patients admitted to hospital; others were excluded because of the type of study or population. 87 full-text articles (two in French, the rest in English) were obtained for more detailed evaluation;24,14,15,17,1924,26,27,31103 one of these had been identified through a review of references.85

Figure 1
Selection of eligible articles

65 articles were excluded during detailed screening. Among the excluded articles 20 were excluded because they described analysis of patient cohorts with specific risks for bloodstream infections,25,27,31,35,36,44,45,47,52,54,59,61,62,65,68,74,75,94,98,101 12 for retrospective data collection,33,37,41,49,53,55,64,69,73,84,89,96 and 11 for subjective, non-replicable, inclusion criteria.15,17,35,55,66,71,77,82,84,91,95 22 selected articles, representing 23 unique patient cohorts, were eligible for systemic review.2,4,14,19,2123,26,32,38,40,43,46,48,57,58,60,67,79,86,88,92 One article described results from two distinct cohorts,88 and one article was included after provision of additional data from the authors (Enwere G, WHO, Geneva, Switzerland, personal communication).58 For other articles excluded because of incomplete data, attempts to contact the authors were unsuccessful or data were not available.

The 22 eligible studies were done in 34 locations between 1984 and 2006 (table 1), and included 61 327 patients from southern (23 893 patients; 39[bullet]0%), east (21 317; 34.8%), north (10 230; 16.7%), and west and central (5887; 9.6%) Africa. 45 899 patients (74.6%) were from exclusively paediatric cohorts. Figure 2 shows the study locations and the three leading causes of non-mycobacterial bloodstream infection. Of the 61 327 patients, 43 741 (71[bullet]3%) were recruited from rural district or missionary hospitals, 7456 (12[bullet]2%) from urban hospitals and referral centres, and 10 130 (16[bullet]5%) from a public infectious disease hospital system in rural and urban areas. Basic inclusion criterion was hospital or ward admission for 40 228 patients (65[bullet]6%), fever without apparent focus of infection for 11 249 (18[bullet]3%), and fever or specific signs of infection for 9850 (16[bullet]1%).

Figure 2
Study locations and summary findings including prevalent pathogens
Table 1
Summary of 22 studies of community-acquired bloodstream infection (BSI) among patients admitted to hospital in Africa, 1984–2006

All studies described the microbiological techniques used, although culture media and methods of identification of organisms varied between studies. Minimum acceptable culture volumes ranged from 1 mL to 3 mL in paediatric studies and from 5 mL to 18 mL in adult studies that provided this information.4,14,22,23,26,32,46,48,57,67,79,86,92 Only two paediatric studies detailed the actual volume of specimens received from patients.14,48

Blood culture results were available for 58 296 (95[bullet]1%) of the 61 327 patients across all studies, and 5578 (9[bullet]6%, range 4[bullet]2–38[bullet]2%) had bacterial or fungal bloodstream infections. Of patients from seven studies that used both aerobic and mycobacterial culture techniques 513 (25[bullet]3%) of 2025 patients had bacterial or fungal bloodstream infections.4,21,23,26,38,40,46 A total of 5647 non-malaria pathogic isolates were recovered across all studies: 3286 (58[bullet]2%) were Gram negative and 1885 (33[bullet]4%) were Gram positive (table 2). S enterica serotypes were the most commonly isolated pathogens, accounting for 1643 (29[bullet]1%) of isolates recovered overall, and 878 (42.3%) of pathogenic isolates among adults. S enterica serotypes were the second most prevalent organisms cultured from children, constituting 765 (21.4%) of 3569 pathogenic isolates in children. Across the whole sample, non-typhoidal serotypes (960 isolates; 17[bullet]0%) were more commonly isolated than were S enterica serotype Typhi (560; 9[bullet]9%), and 507 (90[bullet]5%) of the S enterica serotype Typhi isolates were from patients in the two north African studies.32,67 Among 960 non-typhoidal Salmonella isolates, 706 (73[bullet]5%) were serotyped; S enterica serotype Typhimurium was the most prevalent non-typhoidal Salmonella accounting for 460 (65[bullet]2%) of serotyped isolates, then S enterica serotype Enteritidis with 234 (33[bullet]1%). 688 (12[bullet]2%) of the isolates were non-Salmonella Enterobacteriaceae, 412 (59[bullet]9%) of which were Escherichia coli. Haemophilus influenzae constituted 287 (5[bullet]1%) of isolated organisms, 124 of which (82[bullet]1% of 151 documented to have been subtyped) were H influenzae type B. Brucella spp were identified in 275 (4[bullet]9%) of the positive blood cultures, all from one north African study.32

Table 2
Blood culture isolates from 22 studies, Africa, 1984–2006

Among prevalent Gram-positive organisms, Streptococcus pneumoniae accounted for 1031 (18[bullet]3%) of the isolates, and was the most common isolate in children, constituting 833 (23.3%) of the paediatric pathogens. Staphylococcus aureus constituted 537 (9[bullet]5%) of the total isolates across the age spectrum. Fungi accounted for 40 (0[bullet]7%) of the isolates. Most fungal isolates were Cryptococcus spp (29 Cryptococcus neoformans and one Cryptococcus laurentiae).

Of all included children, 3527 (8[bullet]2%) of 43 130 had bloodstream infections, compared with 2051 (13[bullet]5%) of 15 166 adults (OR 0[bullet]60, p<0[bullet]0001). HIV infection was diagnosed in 499 (18[bullet]5%) of 2695 tested patients in paediatric studies compared with 1217 (53[bullet]5%) of 2273 tested patients in adult studies (OR 0[bullet]20, p<0[bullet]0001). Compared with adults, bacteraemia in children was more likely to be caused by Gram-positive organisms (OR 1[bullet]6, p<0[bullet]0001), including S pneumoniae (OR 1[bullet]5, p<0[bullet]0001), S aureus (OR 1[bullet]4, p<0[bullet]0001), and group A streptococci (OR 10[bullet]8, p<0[bullet]0001), or by non-Salmonella Enterobacteriaceae (OR 1[bullet]4, p<0[bullet]0001). H influenzae infection was found almost exclusively in children (OR 101[bullet]2, p<0[bullet]0001). S enterica (OR 3[bullet]4, p<0[bullet]0001), mycobacteria (OR 34[bullet]6, p<0[bullet]0001), and yeasts (OR 111[bullet]2, p<0[bullet]0001) predominated among adults (table 2).

Among the five adult studies that included mycobacterial blood-culture techniques, M tuberculosis complex was the most common isolate, accounting for 166 (33[bullet]8%) of 491 isolates from 1716 patients.4,23,26,38,46 Four of these studies detailed data on polymicrobial bloodstream infections; 13 (8[bullet]5%) of 153 patients with M tuberculosis bacteraemia also had bacteraemia with another organism, most commonly non-typhoidal Salmonella (5; 39%) or S pneumoniae (4; 31%).4,23,26,38 In the two paediatric studies which assessed for mycobacteraemia, one child had Mycobacterium avium complex bacteraemia, but M tuberculosis was not recovered from any child, despite bloodstream infections with other pathogens being identified in 12 (15.0%) of 80 and 35 (15.3%) of 229 included patients.21,40 All of the children in these studies had previously received a BCG vaccination; cultures were obtained by drawing 3 mL of blood into BACTEC MYCO/F LYTIC blood culture bottles (Becton Dickinson Inc).

Organisms thought to be skin-flora contaminants were explicitly reported to have been excluded from the analyses in 17 studies;2,14,19,21,23,26,32,38,40,48,57,58,60,67,79,86,92 in 11 studies providing full data, contaminants were isolated from 5448 (13[bullet]1%) of 41 443 paediatric blood cultures versus 47 (3[bullet]5%) of 1355 adult blood cultures (OR 3[bullet]8, p<0[bullet]0001).2,14,19,23,26,38,48,57,79,86,92 In addition to the organisms reported to have been excluded, 137 (7[bullet]3%) of 1885 Gram-positive and 21 (0[bullet]6%) of 3286 of Gram-negative organisms were reported as pathogens justified by clinical scenarios, but could have been classified as contaminants in other studies. These included 27 Staphylococcus epidermidis isolates, 109 Streptococcus viridans or unspecified streptococci, one Bacillus cereus, and 21 Gram-negative rods of the genera Burkholderia, Aeromonas, Flavimonas, Flavibacterium, and Xanthomonas. 132 (83[bullet]5%) of 158 possible contaminants were identified in paediatric studies. Additionally, Falade and colleagues60 reported that 18 of 78 isolates of S aureus isolated from their cohort were sent to a reference laboratory for confirmation (the same laboratory used by Enwere and colleagues58 in their analysis included in this study), but only one of the 18 was confirmed to be S aureus; the remaining 17 isolates were coagulase negative staphylococci. Confirmation for the remaining 60 isolates was not available.

13 studies reported the prevalence of malaria parasitaemia identified by thick or thin smear.2,4,19,21,38,40,43,46,57,67,86,88,92 Of 21 131 patients for whom malaria film results were reported 11 914 (56[bullet]4%, range 5[bullet]7–61[bullet]7%) were parasitaemic; in five adult and two paediatric cohorts, prevalence of bacterial or fungal bloodstream infections was greater than the prevalence of malaria parasitaemia.4,21,38,40,57,67,88 Among two adult and seven paediatric studies, 769 (6[bullet]5%) of 11 814 patients with parasitaemia also had fungal or bacterial bloodstream infections.2,14,19,21,38,43,57,86,92

Archibald and colleagues38 in Dar es Salaam, Tanzania, and Bell and colleagues46 in Lilongwe, Malawi, assessed the prevalence of malaria among febrile study patients compared with asymptomatic outpatients or people admitted to hospital with non-febrile disorders. Among patients from Lilongwe, parasitaemia was significantly more common (relative risk 2[bullet]3, p=0[bullet]0021) among study patients (72 [31%] of 231) compared with randomly selected afebrile outpatients (10 [14%] of 73), but there was no difference (p=0[bullet]575) in prevalence of parasitaemia between study patients (49 [9[bullet]5%] of 517) and afebrile patients admitted to the orthopaedic ward in Dar es Salaam (12 [8[bullet]0%] of 150). In a rural western Kenya cohort described by Dougle,57 197 (86.0%) of 229 study patients were smear negative for malaria, yet 137 (59.8%) of all 229 study patients had a primary diagnosis of malaria. Bloodstream infections were identified in 51 (22.3%) of the patients in this study, 35 (69%) of whom were diagnosed with malaria on admission.

In addition to blood culture, Afifi and colleagues32 tested serum samples for Brucella spp by use of slide or tube agglutination with Brucella abortus antigens. Agglutination testing had a specificity of 79% and a sensitivity of 89% compared with growth in biphasic blood-culture media (PML Microbiologicals, Wilsonville, OR, USA). A positive serum agglutination test for Brucella spp in the absence of isolation in blood culture was identified in one of 336 patients in Petit and colleagues' 1990 cohort.88 Only one article reported viral serology, and identified dengue as the most common cause of laboratory-identified febrile illness, accounting for 18 of 100 cases of fever without a focus in Port Sudan.67 Okwara and colleagues reported urine culture results and found that ten (26[bullet]3%) of 38 children with urinary tract infection also had bacteraemia with the same organisms, most commonly with Citrobacter spp (five children; 50[bullet]0%) and Enterococcus spp (two children; 20[bullet]0%); these patients accounted for ten of 32 patients with bacteraemia in the study.86

Of 4968 patients from 11 studies that included HIV antibody testing, 1738 (35[bullet]0%) were HIV seropositive.4,21,23,26,38,40,46,48,57,79,88 In seven studies with sufficient data for analysis, 486 (34[bullet]5%) of 1407 patients infected with HIV had bloodstream infections versus 110 (15[bullet]2%) of 725 people seronegative for HIV (OR 3[bullet]4, p<0[bullet]0001; table 3).4,21,26,38,40,46,79 Infection with non-typhoidal Salmonella (OR 8[bullet]2, p<0[bullet]0001)4,23,38,79 and M tuberculosis (OR 23[bullet]4, p<0[bullet]0001)4,23,26,38,46 were significantly more likely among people infected with HIV than among uninfected patients in studies reporting sufficient data for analysis, whereas S enterica serotype Typhi bacteraemia was identified in fewer patients infected with HIV than uninfected patients (OR 0[bullet]07, p<0[bullet]0001).23,26,38,79 C neoformans (21 patients) and Histoplasma capsulatum (two patients) were found exclusively in patients infected with HIV in studies where pathogen-specific HIV serostatus was provided.23,26,38,46 Infection with S pneumoniae,4,23,26,38,79 S aureus,23,26,38,79 and E coli4,23,26,38,57,79 were not associated with HIV infection among the studies in this Review.

Table 3
Causes of bloodstream infection by HIV serostatus* in seven African studies, 1993–2004

Antimicrobial use before admission to hospital was common, ranging from 3% to 56% of patients across eight studies reporting such data.2,21,23,26,43,46,48,86 In four studies,21,23,43,86 receipt of antimicrobial drugs as an outpatient was not associated with blood culture results. One study2 showed an increased risk of bacteraemia and one study identified fewer bloodstream infections in patients previously treated with antimicrobial drugs than in those who were not.26

13 studies, which included 3916 (69[bullet]3%) of the total pathogens isolated, reported results of antimicrobial susceptibility testing, all of which used disc diffusion or Etest methods.2,22,23,26,32,38,46,57,58,60,79,88,92 Five of these studies specified use of interpretive criteria according to National Committee on Clinical Laboratory Standards guidelines (now the Clinical Laboratories Standards Institute).38,58,60,88,92 Table 4 presents the major findings for prevalent isolates. Although 90[bullet]3% of S pneumoniae and 89[bullet]1% of S enteria serotype Typhi were classified as ampicillin susceptible, only 8[bullet]9% of E coli isolates were ampicillin susceptible. Susceptibility of S enterica serotype Enteritidis and S enterica serotype Typhimurium were 33[bullet]6% and 3[bullet]0% for ampicillin, 93[bullet]2% and 15[bullet]0% for co-trimoxazole, and 38[bullet]9% and 100% for chloramphenicol, respectively. More than 95% of Gram-negative organisms were susceptible to third-generation cephalosporins and ciprofloxacin.

Table 4
Antimicrobial susceptibilities of common bloodstream isolates in seven reporting studies, Africa, 1993–2006

Table 5 displays the clinical features associated with non-malaria bloodstream infections in at least two included cohorts. Clinical features significantly associated with bloodstream infections included lethargy, restlessness, oral candidiasis, and jaundice. Hepatomegaly and splenomegaly showed variable association with bloodstream infections across studies. Malnutrition, which was defined differently by different authors, correlated in some studies but not in others; four paediatric studies found an association between malnutrition and bloodstream infection,14,21,48,92 whereas two other paediatric and one adult study found no association or a negative association with malnutrition.2,23,79

Table 5
Clinical and laboratory associations with non-malaria bloodstream infection from 12 studies, Africa, 1991-2006

The proportion of patients with bloodstream infections was greater among cohorts that restricted enrolment to patients with fever or other signs of infection versus all admissions (13[bullet]4% vs 7[bullet]4%; OR 1[bullet]9, p<0[bullet]0001) and in cohorts that included fever with or without a focus over those that included patients with non-focal febrile illness alone (16[bullet]6% vs 10[bullet]5%; OR 1[bullet]7, p<0[bullet]0001). Fever higher than 38[bullet]9°C was associated with bloodstream infections in four studies4,14,19,46 but showed no relation in one.43

Mycobacteraemia was associated with chronic cough in two studies (OR 3[bullet]8–16[bullet]4, p<0[bullet]01).4,26 However, 23 (38%) of 60 patients and 13 (23%) of 57 patients with mycobacteraemia from two studies had no associated respiratory signs or symptoms.38,46 Three studies identified anaemia more often in patients with mycobacteraemia compared with all patients (OR 9[bullet]5, p<0[bullet]001),4 all patients infected with HIV (p<0[bullet]001),23 or all patients with bloodstream infections (p<0[bullet]05).26

16 studies reported in-hospital fatality for all enrolled patients with and without bloodstream infections; 2,4,14,19,2123,26,32,43,46,48,67,79,88,92 2703 (5[bullet]3%) of 51 346 died in hospital. In 13 reporting studies, 2,4,19,23,26,32,43,46,48,67,79,88,92 700 (18[bullet]1%) of 3873 patients with bloodstream infections died in hospital compared with 1966 (4[bullet]4%) of 45 030 patients without bloodstream infections (OR 4[bullet]8, p<0[bullet]0001). 47 (42%) of 111 patients with mycobacteraemia died in hospital in four reporting studies.23,26,38,46

Associations between seasons and rates of bloodstream infection and prevalent organisms were described in seven studies.14,32,19,40,46,58,92 Data from Malawian cohorts in both Blantyre and Lilongwe showed a shift from a predominance of non-typhoidal Salmonella during the wet season to a predominance of S pneumoniae during the dry season.40,46 Enwere and colleagues58 also noted an increased prevalence of S pneumoniae during the hot dry season in the Gambia, and Afifi and colleagues32 showed a relation between how often S enterica serotype Typhi was isolated with the onset of the rainy season in Egypt from November through January. Factors other than rainfall might also have a role. In Egypt, Brucella spp were isolated most often from March to June, the period of livestock parturition.32 Sigauque and colleagues92 found no association between season and bloodstream infections.

Discussion

Bacterial or fungal bloodstream infections are common among prospectively sampled adults and children admitted to hospital in Africa, with a mean prevalence of 13[bullet]4% (range 8[bullet]5–38[bullet]2%) among those with fever and 7[bullet]4% (4[bullet]2–16[bullet]9%) among all admissions irrespective of fever history. These data underscore the importance of considering bacterial or fungal bloodstream infections in the differential diagnosis of all patients admitted to hospital, particularly those who are febrile. Our findings suggest that health outcomes in Africa could be improved by scale-up of laboratory facilities to allow for accurate diagnosis and directed treatment of patients with bloodstream infections and to inform local policy on likely prevalence, pathogens, and antimicrobial susceptibility patterns.

WHO guidelines for management of acute illness in children (Integrated Management of Childhood Illness) recommend use of an appropriate antibacterial drug in addition to an antimalarial drug in children with certain signs of severe illness.104 However, the algorithm remains structured around malaria treatment even in areas of low malaria risk, recommending use of an antimalarial drug before clinical follow-up in febrile illness without focal signs. Our Review indicates that the presence of malaria parasitaemia does not exclude bloodstream infection with other organisms, particularly in areas with high malaria prevalence. In the large paediatric cohort from Mozambique investigated by Sigauque and colleagues,92 621 (5.8%) of 10 699 children admitted to hospital with both tests available were found to have concomitant bacterial bloodstream infection with malaria parasitaemia.

The common practice of diagnosing and treating fever as malaria was described in Kenyan57 and Malawian46 cohorts included in this Review, and was also shown by Reyburn and colleagues105 in Tanzania, who found that most (2412; 53[bullet]9%) patients treated for malaria in hospital were slide-negative, and that 1571 (66%) of these patients did not receive antibacterial drugs.105 In the study by Reyburn and colleagues,105 malaria slide-negative patients had a higher mortality than did slide-positive patients (6[bullet]9 vs 12[bullet]1%, p<0[bullet]001), suggesting that empirical treatment of malaria without subsequent treatment for other causes of febrile illness might have played a part in this excess mortality. Our Review suggests that S enterica is one of the pathogens most commonly isolated throughout various settings in Africa and, like malaria, S enterica bacteraemia can cause undifferentiated fever and splenomegaly. We have also shown that M tuberculosis bacteraemia can present without focal signs, although this has only been proven in adults. Taken together, this information presents a strong case that empirical treatment of malaria alone in regions of both high and low prevalence will result in inadequate management of a substantial portion of patients with non-malaria bloodstream infections.

HIV infection is linked to the distribution, prevalence, and cause of community-acquired bloodstream infections in patients admitted to hospital in Africa. Non-malarial bloodstream infection—particularly with non-typhoidal Salmonella, M tuberculosis, or fungi—was more prevalent in patients infected with HIV than in those without HIV, and might have also influenced the prevalence and distribution of types of bloodstream infections between adults and children, given that HIV infection was present much more commonly among adults. Among the five studies in adults that used mycobacterial blood-culture techniques,4,23,26,38,46 M tuberculosis was the most common isolate, comprising just over one third of all bacterial and fungal bloodstream isolates. The vast majority of these patients were infected with HIV, and a substantial proportion of them did not present typical chest signs associated with tuberculosis disease.38,46 In these patients, chronic fever and wasting were common and many had abnormal chest radiographs even in the absence of chest symptoms.4,26,38 Knowledge of local HIV prevalence and routine HIV counselling and testing of patients admitted to hospital whose HIV status is unknown will assist health-care workers in formulating a differential diagnosis. Evidence suggests that even low levels of immune suppression can be associated with significantly increased risks of invasive bacterial disease.106 The influence of expanded access to antiretroviral therapy programmes on the epidemiology of bloodstream infections in Africa is yet to be established.

Of note, bloodstream infection with M tuberculosis was not identified among the Malawian children described by Archibald and colleagues,21,40 despite a high prevalence of HIV and other bloodstream pathogens. Blood-culture volume, effects of the BCG vaccine, or unique disease pathogenesis in adults and children might lead to this lack of association; this merits further investigation in other studies. Additionally, the inverse association between S enterica serotype Typhi bacteraemia and HIV is striking and further research to substantiate and explore this might reveal important immunological or epidemiological differences between different serotypes of S enterica.

With the exception of E coli and non-typhoidal Salmonella, pathogens included in this Review tended to be susceptible to one or more locally available antimicrobial drugs. However, the studies in which antimicrobial susceptibility was assessed were done between 3 years and 16 years ago and might not reflect current local patterns. Few (<30%) non-typhoidal Salmonella isolated among studies in this Review were susceptible to ampicillin, most S enterica serotype Enteritidis were resistant to chloramphenicol, and most S enterica serotype Typhimurium were resistant to co-trimoxazole. Frequent isolation of non-typhoidal Salmonella resistant to commonly used antimicrobial drugs has been described by other investigators.107110 Additionally, meticillin-resistant S aureus,32,111,112 penicillin-resistant S pneumoniae,38,111 and multidrug-resistant Enterobacteriaceae and M tuberculosis50,107,113,114 have been reported in Africa and present a threat to individual and public health. Most invasive bacterial isolates across all studies in this Review were susceptible to third-generation cephalosporins or fluoroquinolones. However, these drugs might be beyond the financial reach of patients and institutions, and their increased use as empirical treatment might promote antimicrobial resistance. Taken together, these data emphasise the complexity and potential pitfalls of attempting to treat common illnesses without improvement in directed diagnostics.

Vaccine-preventable diseases continue to cause a substantial burden of disease in Africa.115,116 S pneumoniae was the most commonly isolated paediatric pathogen among patients in our Review, and invasive H influenzae (the majority of which was documented to be the vaccine-preventable type B), was found in 286 (8.0%) children with bloodstream infections. Enhanced uptake of available vaccines and increased pathogen surveillance should be prioritised to optimise benefits from vaccines and to inform their further development. Given the prevalence of non-typhoidal Salmonella among patients in this Review, further research into vaccine development is crucial.117

There are several important limitations to this study. We examined hospital-based cohorts to define community-acquired bloodstream infections, and thereby might have included patients with more severe illness and included only those who have access to health-care facilities.118 However, non-malaria bloodstream infection might be as common among certain groups of outpatients as in inpatients in Africa.14,39,70 Some important features related to the risks and clinical importance of bloodstream infections in different groups of patient (eg, infants vs older children, patients with underlying clinical conditions other than HIV, patients with HIV across the spectrum of immune suppression) were not explored in this study. Enrolment and inclusion criteria varied by study, and studies spanned a period of more than 20 years; therefore overall prevalence of bloodstream infection and the types of organisms isolated should not be seen as a guide to diagnosis for individual clinical encounters, rather as an indication of the anticipated range of pathogens and to emphasise the need to consider the most locally appropriate treatment for a patient who might have a bloodstream infection. Aggregation of raw data resulted in weighting of large studies; the local importance of some pathogens isolated in smaller studies might have been underrepresented.

Some pathogens were either incompletely identified or incompletely reported (4.7% of all pathogens reported); these might have been bloodstream contaminants and might have resulted in an enhanced estimate of the true prevalence of bloodstream infections. Studies varied in how a blood-culture contaminant was defined, and as was noted by Falade and colleagues,60 the prevalence of S aureus as a pathogen and of bloodstream infection as a whole might have been overestimated by some laboratories incorrectly identifying coagulase-negative staphylococci as S aureus. Underestimation of bloodstream infections could have occurred as well, because the diagnostic sensitivity of the laboratory techniques used in various studies could not be rigorously assessed, and fastidious pathogens that need optimum culture and transport methods might not have been isolated in some laboratories. Commonly isolated organisms, such as S pneumoniae and H influenzae, might have been underestimated due to their fastidious nature. Bloodstream infection in children, in particular, might also have been underestimated as a result of blood-culture volume inadequacy. On analysis of blood volume from 16 570 cultures from children, Berkley and colleagues48 noted increased prevalence of bloodstream infection (7[bullet]9%) in 3 mL samples compared with 1 mL samples (5[bullet]6%, p=0[bullet]006).

Studies that found high prevalences of bloodstream infections might have been preferentially published over other studies. However, the exclusion of studies that enrolled patients who were subjectively defined as “septic” probably reduced the degree of bias presented in this Review. Patterns of vaccine use might have influenced the prevalence and types of pathogens isolated; the cohort presented by Enwere and colleagues58 was part of a study to assess the efficacy of the pneumococcal vaccine.

This Review presents a strong case for the importance of considering non-malarial bloodstream infections in all patients admitted to hospital in Africa, particularly those with fever, known HIV infection, signs of impaired immunity, or lethargy. Our findings underscore the importance of continuous re-evaluation of WHOs Integrated Management of Childhood Illness and other empirical treatment algorithms for febrile illness in view of changing patterns of disease. When used, empirical treatment with antimicrobial drugs should include coverage for Gram-negative and Gram-positive organisms, and be effective against S enterica and S pneumoniae. Furthermore, our Review highlights that the prevalence of bacteraemia exceeds the prevalence of malaria parasitaemia among patients admitted to hospital in some regions, and that malaria parasitaemia and bacteraemia might coexist. Therefore, clinicians should be alert to local malaria transmission patterns and maintain a high index of suspicion for bacteraemia when proceeding though a diagnostic and treatment plan for an ill patient with a blood film showing malaria in a region of high prevalence. Public health leaders should prioritise the use of available vaccines, and consider means to alleviate the direct cost of antimicrobials to patients admitted to hospital. Further resources and research to define risk factors and prevention strategies, including vaccines, for non-typhoidal Salmonella in Africa are needed. Expanded and improved clinical microbiology services in Africa might contribute substantially to the improved management of community-acquired bloodstream infections. In the meantime, regular collection and dissemination of microbiological data from sentinel hospital studies, such as those included in the Review, are needed to inform empirical guidelines for management of patients with suspected community-acquired bloodstream infection.

Acknowledgments

The authors gratefully acknowledge the Hubert-Yeargan Center for Global Health for infrastructure support for the Kilimanjaro Christian Medical Centre-Duke University Collaboration, and Megan Von Isenburg, Associate Director of Public Services and Liaison to Graduate Medical Education at Duke University, for her assistance with the search of published work. JAC received support from US National Institutes of Health awards International Studies on AIDS Associated Co-infections U01 AI062563, AIDS International Training and Research Program D43 PA-03-018, and the Duke Clinical Trials Unit and Clinical Research Sites U01 AI069484-01. EAR received support from the AIDS International Research Training Program T32 Training Grant, A1007392-18, and the Duke University Center for AIDS Research (CFAR), an NIH-funded programme (P30 AI 64518). AVS was supported by the Alpha Omega Alpha Carolyn L Kuckein Student Research Fellowship.

Footnotes

Contributors: JAC conceived the idea for this study. EAR, AVS, and JAC designed the study. AVS and EAR searched published work, reviewed published papers, and made the primary selection of eligible papers. JAC resolved disagreements regarding the eligibility of papers. EAR and AVS compiled and analysed the data. EAR prepared the first draft of the paper. All authors contributed to the writing of the report and have seen and approved the final version.

Conflicts of interest: JAC has received honoraria from Abbott laboratories. EAR and AVS declare that they have no conflicts of interest.

Search strategy and selection criteria: These are detailed in the Methods section.

References

1. Petit PL, van Ginneken JK. Analysis of hospital records in four African countries, 1975–1990, with emphasis on infectious diseases. J Trop Med Hyg. 1995;98:217–27. [PubMed]
2. Bahwere P, Levy J, Hennart P, et al. Community-acquired bacteremia among hospitalized children in rural central Africa. Int J Infect Dis. 2001;5:180–88. [PubMed]
3. Campbell JD, Kotloff KL, Sow SO, et al. Invasive pneumococcal infections among hospitalized children in Bamako, Mali. Pediatr Infect Dis J. 2004;23:642–49. [PubMed]
4. Peters RP, Zijlstra EE, Schijffelen MJ, et al. A prospective study of bloodstream infections as cause of fever in Malawi: clinical predictors and implications for management. Trop Med Int Health. 2004;9:928–34. [PubMed]
5. Floyd K, Reid RA, Wilkinson D, Gilks CF. Admission trends in a rural South African hospital during the early years of the HIV epidemic. JAMA. 1999;282:1087–91. [PubMed]
6. Onankpa B, Airede L, Paul I, Dorcas I. Pattern of pediatric HIV/AIDS: a five-year experience in a tertiary hospital. J Natl Med Assoc. 2008;100:821–25. [PubMed]
7. Kengne AP, Kaze FF, Dzudie A, Awah PK, Ngu KB. HIV/AIDS occurrence in the main university teaching hospital in Cameroon: audit of the 2001 activities of the service of internal medicine. J Int Assoc Physicians AIDS Care (Chic Ill) 2007;6:61–65. [PubMed]
8. Bradshaw D, Schneider M, Dorrington R, Bourne DE, Laubscher R. South African cause-of-death profile in transition—1996 and future trends. S Afr Med J. 2002;92:618–23. [PubMed]
9. Ole-Nguyaine S, Crump JA, Kibiki GS, et al. HIV-associated morbidity, mortality and diagnostic testing opportunities among inpatients at a referral hospital in northern Tanzania. Ann Trop Med Parasitol. 2004;98:171–79. [PubMed]
10. Archibald LK, Reller LB. Clinical microbiology in developing countries. Emerg Infect Dis. 2001;7:302–05. [PMC free article] [PubMed]
11. Petti CA, Polage CR, Quinn TC, Ronald AR, Sande MA. Laboratory medicine in Africa: a barrier to effective health care. Clin Infect Dis. 2006;42:377–82. [PubMed]
12. Perkins BA, Zucker JR, Otieno J, et al. Evaluation of an algorithm for integrated management of childhood illness in an area of Kenya with high malaria transmission. Bull World Health Organ. 1997;75(suppl 1):33–42. [PMC free article] [PubMed]
13. English M, Berkley J, Mwangi I, et al. Hypothetical performance of syndrome-based management of acute paediatric admissions of children aged more than 60 days in a Kenyan district hospital. Bull World Health Organ. 2003;81:166–73. [PMC free article] [PubMed]
14. Brent AJ, Ahmed I, Ndiritu M, et al. Incidence of clinically significant bacteraemia in children who present to hospital in Kenya: community-based observational study. Lancet. 2006;367:482–88. [PubMed]
15. Idoko JA, Lawande RV, Mohammed I. A prospective study of septicaemia in Zaria, northern Nigeria. East Afr Med J. 1986;63:515–21. [PubMed]
16. Haddon RA, Barnett PL, Grimwood K, Hogg GG. Bacteraemia in febrile children presenting to a paediatric emergency department. Med J Aust. 1999;170:475–78. [PubMed]
17. Alausa KO, Montefiore D, Sogbetun AO, Ashiru JO, Onile BA, Sobayo E. Septicaemia in the tropics: a prospective epidemiological study of 146 patients with a high case fatality rate. Scand J Infect Dis. 1977;9:181–85. [PubMed]
18. Kuppermann N. Occult bacteremia in young febrile children. Pediatr Clin North Am. 1999;46:1073–109. [PubMed]
19. Akpede GO, Abiodun PO, Sykes RM. Relative contribution of bacteraemia and malaria to acute fever without localizing signs of infection in under-five children. J Trop Pediatr. 1991;38:295–98. [PubMed]
20. Ayoola OO, Adeyemo AA, Osinusi K. Aetiological agents, clinical features and outcome of septicaemia in infants in Ibadan. West Afr J Med. 2003;22:30–34. [PubMed]
21. Archibald LK, Kazembe PN, Nwanyanwu O, Mwansambo C, Reller LB, Jarvis WR. Epidemiology of bloodstream infections in a bacille Calmette-Guerin-vaccinated pediatric population in Malawi. J Infect Dis. 2003;188:202–08. [PubMed]
22. Gordon MA, Walsh AL, Chaponda M, et al. Bacteraemia and mortality among adult medical admissions in Malawi— predominance of non-typhi salmonellae and Streptococcus pneumoniae. J Infect. 2001;42:44–49. [PubMed]
23. Vugia DJ, Kiehlbauch JA, Yeboue K, et al. Pathogens and predictors of fatal septicemia associated with human immunodeficiency virus infection in Ivory Coast, west Africa. J Infect Dis. 1993;168:564–70. [PubMed]
24. Norrgren H, Cardoso AN, da Silva ZJ, et al. Increased prevalence of HIV-2 infection in hospitalized patients with severe bacterial diseases in Guinea-Bissau. Scand J Infect Dis. 1997;29:453–59. [PubMed]
25. Grant AD, Djomand G, De Cock KM. Natural history and spectrum of disease in adults with HIV/AIDS in Africa. AIDS. 1997;11(suppl B):S43–54. [PubMed]
26. Ssali FN, Kamya MR, Wabwire-Mangen F, et al. A prospective study of community-acquired bloodstream infections among febrile adults admitted to Mulago Hospital in Kampala, Uganda. J Acquir Immune Defic Syndr Hum Retrovirol. 1998;19:484–89. [PubMed]
27. Arthur G, Nduba VN, Kariuki SM, Kimari J, Bhatt SM, Gilks CF. Trends in bloodstream infections among human immunodeficiency virus-infected adults admitted to a hospital in Nairobi, Kenya, during the last decade. Clin Infect Dis. 2001;33:248–56. [PubMed]
28. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283:2008–12. [PubMed]
29. Thornton A, Lee P. Publication bias in meta-analysis: its causes and consequences. J Clin Epidemiol. 2000;53:207–16. [PubMed]
30. UNAIDS/WHO. AIDS epidemic update. [accessed 4 Feb, 2008]; http://www.unaids.org.
31. Adejuyigbe EA, Ako-Nai AK, Adisa B. Bacterial isolates in the sick young infant in Ile-Ife, Nigeria. J Trop Pediatr. 2004;50:323–27. [PubMed]
32. Afifi S, Earhart K, Azab MA, et al. Hospital-based surveillance for acute febrile illness in Egypt: a focus on community-acquired bloodstream infections. Am J Trop Med Hyg. 2005;73:392–99. [PubMed]
33. Ako-Nai AK, Taiwo O, Ebri A, Adeniran MO. Bacterial isolates involved in cases of septicaemia in a Nigerian hospital. East Afr Med J. 1990;67:407–12. [PubMed]
34. Akpede GO, Abiodun PO, Sykes RM. Acute fevers of unknown origin in young children in the tropics. J Pediatr. 1993;122:79–81. [PubMed]
35. Akpede GO, Abiodun PO, Sykes RM. Pattern of infections in children under-six years old presenting with convulsions associated with fever of acute onset in a children's emergency room in Benin City, Nigeria. J Trop Pediatr. 1993;39:11–15. [PubMed]
36. Anglaret X, Messou E, Ouassa T, et al. Pattern of bacterial diseases in a cohort of HIV-1 infected adults receiving cotrimoxazole prophylaxis in Abidjan, Côte d'Ivoire. AIDS. 2003;17:575–84. [PubMed]
37. Angyo IA, Opkeh ES, Opajobi SO. Predominant bacterial agents of childhood septicaemia in Jos. Niger J Med. 2001;10:75–77. [PubMed]
38. Archibald LK, den Dulk MO, Pallangyo KJ, Reller LB. Fatal Mycobacterium tuberculosis bloodstream infections in febrile hospitalized adults in Dar es Salaam, Tanzania. Clin Infect Dis. 1998;26:290–96. [PubMed]
39. Archibald LK, McDonald LC, Nwanyanwu O, et al. A hospital-based prevalence survey of bloodstream infections in febrile patients in Malawi: implications for diagnosis and therapy. J Infect Dis. 2000;181:1414–20. [PubMed]
40. Archibald LK, Nwanyanwu O, Kazembe PN, et al. Detection of bloodstream pathogens in a bacille Calmette-Guerin (BCG)-vaccinated pediatric population in Malawi: a pilot study. Clin Microbiol Infect. 2003;9:234–38. [PubMed]
41. Ashiru JO, Osoba AO. Gram negative septicaemia in Ibadan, Nigeria. East Afr Med J. 1986;63:471–76. [PubMed]
42. Ayoola OO, Adeyemo AA, Osinusi K. Concurrent bacteraemia and malaria in febrile Nigerian infants. Trop Doct. 2005;35:34–36. [PubMed]
43. Ayoola OO, Adeyemo AA, Osinusi K. Predictors of bacteraemia among febrile infants in Ibadan, Nigeria. J Health Popul Nutr. 2002;20:223–29. [PubMed]
44. Babirekere-Iriso E, Musoke P, Kekitiinwa A. Bacteraemia in severely malnourished children in an HIV-endemic setting. Ann Trop Paediatr. 2006;26:319–28. [PubMed]
45. Bachou H, Tylleskar T, Kaddu-Mulindwa DH, Tumwine JK. Bacteraemia among severely malnourished children infected and uninfected with the human immunodeficiency virus-1 in Kampala, Uganda. BMC Infect Dis. 2006;6:160. [PMC free article] [PubMed]
46. Bell M, Archibald LK, Nwanyanwu O, et al. Seasonal variation in the etiology of bloodstream infections in a febrile inpatient population in a developing country. Int J Infect Dis. 2001;5:63–69. [PubMed]
47. Berkley J, Mwarumba S, Bramham K, Lowe B, Marsh K. Bacteraemia complicating severe malaria in children. Trans R Soc Trop Med Hyg. 1999;93:283–86. [PubMed]
48. Berkley JA, Lowe BS, Mwangi I, et al. Bacteremia among children admitted to a rural hospital in Kenya. N Engl J Med. 2005;352:39–47. [PubMed]
49. Berkowitz FE. Bacteremia in hospitalized black South African children: a one-year study emphasizing nosocomial bacteremia and bacteremia in severely malnourished children. Am J Dis Child. 1984;138:551–56. [PubMed]
50. Blomberg B, Jureen R, Manji KP, et al. High rate of fatal cases of pediatric septicemia caused by Gram-negative bacteria with extended-spectrum beta-lactamases in Dar es Salaam, Tanzania. J Clin Microbiol. 2005;43:745–49. [PMC free article] [PubMed]
51. Blomberg B, Manji KP, Urassa WK, et al. Antimicrobial resistance predicts death in Tanzanian children with bloodstream infections: a prospective cohort study. BMC Infect Dis. 2007;7:43. [PMC free article] [PubMed]
52. Bronzan RN, Taylor TE, Mwenechanya J, et al. Bacteremia in Malawian children with severe malaria: prevalence, etiology, HIV coinfection, and outcome. J Infect Dis. 2007;195:895–904. [PubMed]
53. Campbell JD, Sow SO, Levine MM, Kotloff KL. The causes of hospital admission and death among children in Bamako, Mali. J Trop Pediatr. 2004;50:158–63. [PubMed]
54. Commey J, Quarm-Goka B, Agyepong I. Persistent fever in severe malaria in children. Cent Afr J Med. 1994;40:257–60. [PubMed]
55. Cotton MF, Burger PJ, Bodenstein WJ. Bacteraemia in children in the south-western Cape: a hospital-based survey. S Afr Med J. 1992;81:87–90. [PubMed]
56. Cotton MF, Wasserman E, Smit J, Whitelaw A, Zar HJ. High incidence of antimicrobial resistant organisms including extended spectrum beta-lactamase producing Enterobacteriaceae and methicillin-resistant Staphylococcus aureus in nasopharyngeal and blood isolates of HIV-infected children from Cape Town, South Africa. BMC Infect Dis. 2008;8:40. [PMC free article] [PubMed]
57. Dougle ML, Hendriks ER, Sanders EJ, Dorigo-Zetsma JW. Laboratory investigations in the diagnosis of septicaemia and malaria. East Afr Med J. 1997;74:353–56. [PubMed]
58. Enwere G, Biney E, Cheung Y, et al. Epidemiologic and clinical characteristics of community-acquired invasive bacterial infections in children aged 2–29 months in The Gambia. Pediatr Infect Dis J. 2006;25:700–05. [PubMed]
59. Evans JA, Adusei A, Timmann C, et al. High mortality of infant bacteraemia clinically indistinguishable from severe malaria. QJM. 2004;97:591–97. [PubMed]
60. Falade AG, Lagunju IA, Bakare RA, Odekanmi AA, Adegbola RA. Invasive pneumococcal disease in children aged <5 years admitted to 3 urban hospitals in Ibadan, Nigeria. Clin Infect Dis. 2009;48(suppl 2):S190–96. [PubMed]
61. Gilks CF, Brindle RJ, Otieno LS, et al. Life-threatening bacteraemia in HIV-1 seropositive adults admitted to hospital in Nairobi, Kenya. Lancet. 1990;336:545–49. [PubMed]
62. Grant AD, Djomand G, Smets P, et al. Profound immunosuppression across the spectrum of opportunistic disease among hospitalized HIV-infected adults in Abidjan, Côte d'Ivoire. AIDS. 1997;11:1357–64. [PubMed]
63. Grant AD, Sidibe K, Domoua K, et al. Spectrum of disease among HIV-infected adults hospitalised in a respiratory medicine unit in Abidjan, Côte d'Ivoire. Int J Tuberc Lung Dis. 1998;2:926–34. [PubMed]
64. Green SD, Cheesbrough JS. Salmonella bacteraemia among young children at a rural hospital in western Zaire. Ann Trop Paediatr. 1993;13:45–53. [PubMed]
65. Hendrickse RG, Hasan AH, Olumide LO, Akinkunmi A. Malaria in early childhood: an investigation of five hundred seriously ill children in whom a “clinical” diagnosis of malaria was made on admission to the children's emergency room at University College Hospital, Ibadan. Ann Trop Med Parasitol. 1971;65:1–20. [PubMed]
66. Hill PC, Onyeama CO, Ikumapayi UN, et al. Bacteraemia in patients admitted to an urban hospital in west Africa. BMC Infect Dis. 2007;7:2. [PMC free article] [PubMed]
67. Hyams KC, Oldfield EC, Scott RM, et al. Evaluation of febrile patients in Port Sudan, Sudan: isolation of dengue virus. Am J Trop Med Hyg. 1986;35:860–65. [PubMed]
68. Kassa-Kelembho E, Mbolidi CD, Service YB, Morvan J, Minssart P. Bacteremia in adults admitted to the Department of Medicine of Bangui Community Hospital (Central African Republic) Acta Trop. 2003;89:67–72. [PubMed]
69. Kezzal K, Ali-Yahia O, Rahal K, Toudji A, Zerhouni A. Importance of blood culture in septicemia, particularly bacterial endocarditis. Arch Inst Pasteur Alger. 1986;55:41–60. in French. [PubMed]
70. Lepage P, Bogaerts J, Van Goethem C, et al. Community-acquired bacteraemia in African children. Lancet. 1987;1:1458–61. [PubMed]
71. Lepage P, Van de Perre P, Nsengumuremyi F, Van Goethem C, Bogaerts J, Hitimana DG. Bacteraemia as predictor of HIV infection in African children. Acta Paediatr Scand. 1989;78:763–66. [PubMed]
72. Lewis DK, Peters RP, Schijffelen MJ, et al. Clinical indicators of mycobacteraemia in adults admitted to hospital in Blantyre, Malawi. Int J Tuberc Lung Dis. 2002;6:1067–74. [PubMed]
73. Maher D, Kumwenda J. Bacteraemia in Blantyre. Trop Doct. 1994;24:82–83. [PubMed]
74. McCarthy MC, Haberberger RL, Salib AW, et al. Evaluation of arthropod-borne viruses and other infectious disease pathogens as the causes of febrile illnesses in the Khartoum Province of Sudan. J Med Virol. 1996;48:141–46. [PubMed]
75. McDonald LC, Archibald LK, Rheanpumikankit S, et al. Unrecognised Mycobacterium tuberculosis bacteraemia among hospital inpatients in less developed countries. Lancet. 1999;354:1159–63. [PubMed]
76. McDonald LC, Jarvis WR. Community-acquired bacteremia in Zimbabwe and the global cost of contaminated blood cultures. Pediatr Infect Dis J. 1997;16:537–38. [PubMed]
77. Meremikwu MM, Nwachukwu CE, Asuquo AE, Okebe JU, Utsalo SJ. Bacterial isolates from blood cultures of children with suspected septicaemia in Calabar, Nigeria. BMC Infect Dis. 2005;5:110. [PMC free article] [PubMed]
78. Montasser MF, Ibrahim FH, Abdel Wahab MF. Rapid diagnosis of non-prolonged febrile illnesses necessitating fever hospital admission. J Egypt Public Health Assoc. 1991;66:675–92. [PubMed]
79. Nathoo KJ, Chigonde S, Nhembe M, Ali MH, Mason PR. Community-acquired bacteremia in human immunodefi ciency virus-infected children in Harare, Zimbabwe. Pediatr Infect Dis J. 1996;15:1092–97. [PubMed]
80. Nesbitt A, Mirza NB. Salmonella septicaemias in Kenyan children. J Trop Pediatr. 1989;35:35–39. [PubMed]
81. Norton EB, Archibald LK, Nwanyanwu OC, et al. Clinical predictors of bloodstream infections and mortality in hospitalized Malawian children. Pediatr Infect Dis J. 2004;23:145–151. [PubMed]
82. Obi JO, Ileli CI. Childhood hyperpyrexia in Benin City, Nigeria. J Trop Med Hyg. 1982;85:115–17. [PubMed]
83. O'Dempsey TJ, McArdle TF, Lloyd-Evans N, et al. Importance of enteric bacteria as a cause of pneumonia, meningitis and septicemia among children in a rural community in The Gambia, West Africa. Pediatr Infect Dis J. 1994;13:122–28. [PubMed]
84. Odhiambo FA, Wamola IA, Ndinya-Achola JO. Aerobic and facultative bacterial isolates from blood cultures of children with clinically diagnosed septicaemia. East Afr Med J. 1991;68:869–74. [PubMed]
85. Odio WT, Mangalaboyi LJ, M'Belepe MR, Ditu MS. Fever origins in the tropics. Scand J Infect Dis. 1982;14:79–80. [PubMed]
86. Okwara FN, Obimbo EM, Wafula EM, Murila FV. Bacteraemia, urinary tract infection and malaria in hospitalised febrile children in Nairobi: is there an association? East Afr Med J. 2004;81:47–51. [PubMed]
87. Parker TM, Murray CK, Richards AL, et al. Concurrent infections in acute febrile illness patients in Egypt. Am J Trop Med Hyg. 2007;77:390–92. [PubMed]
88. Petit PL, Haarlem JV, Poelman M, Haverkamp MC, Wamola IA. Bacteraemia in patients presenting with fever. East Afr Med J. 1995;72:116–20. [PubMed]
89. Rayner BL, Willcox PA. Community-acquired bacteraemia; a prospective survey of 239 cases. Q J Med. 1988;69:907–19. [PubMed]
90. Rosen EU. The disease profile of hospitalized Third World urban black adolescents. J Adolesc Health Care. 1985;6:448–52. [PubMed]
91. Samuel SO, Fadeyi A, Akanbi AA, 2nd, Ameen NB, Nwabuisi C, Onile BA. Bacterial isolates of blood cultures in patients with suspected septicaemia in Ilorin, Nigeria. Afr J Med Med Sci. 2006;35:137–41. [PubMed]
92. Sigauque B, Roca A, Mandomando I, et al. Community-acquired bacteremia among children admitted to a rural hospital in Mozambique. Pediatr Infect Dis J. 2009;28:108–13. [PubMed]
93. Slusher T, Gbadero D, Howard C, et al. Randomized, placebo-controlled, double blinded trial of dexamethasone in African children with sepsis. Pediatr Infect Dis J. 1996;15:579–83. [PubMed]
94. Smith C, Arregui LM, Promnitz DA, Feldman C. Septic shock in the Intensive Care Unit, Hillbrow Hospital, Johannesburg. S Afr Med J. 1991;80:181–84. [PubMed]
95. Sow SO, Diallo S, Campbell JD, et al. Burden of invasive disease caused by Haemophilus influenzae type b in Bamako, Mali: impetus for routine infant immunization with conjugate vaccine. Pediatr Infect Dis J. 2005;24:531–37. [PubMed]
96. Sow-Goerger MT, Sall B, Ka S, et al. Hospital septicemias in Dakar: epidemiologic study. Dakar Med. 1989;34:49–52. in French. [PubMed]
97. van Deventer JD, Carter CL, Schoeman CJ, Joubert G. The impact of HIV on the profi le of paediatric admissions and deaths at Pelonomi Hospital, Bloemfontein, South Africa. J Trop Pediatr. 2005;51:391–92. [PubMed]
98. Waddell RD, Lishimpi K, von Reyn CF, et al. Bacteremia due to Mycobacterium tuberculosis or M bovis, Bacille Calmette-Guerin (BCG) among HIV-positive children and adults in Zambia. AIDS. 2001;15:55–60. [PubMed]
99. Walsh AL, Phiri AJ, Graham SM, Molyneux EM, Molyneux ME. Bacteremia in febrile Malawian children: clinical and microbiologic features. Pediatr Infect Dis J. 2000;19:312–18. [PubMed]
100. Wesley AG, Scragg JN, Rubidge CJ, Wallace HL. The racial incidence of disease in hospital children in Durban. S Afr Med J. 1967;41:332–35. [PubMed]
101. Whitelaw DA, Rayner BL, Willcox PA. Community-acquired bacteremia in the elderly: a prospective study of 121 cases. J Am Geriatr Soc. 1992;40:996–1000. [PubMed]
102. Woodruff PW, Morrill JC, Burans JP, Hyams KC, Woody JN. A study of viral and rickettsial exposure and causes of fever in Juba, southern Sudan. Trans R Soc Trop Med Hyg. 1988;82:761–66. [PubMed]
103. Zimmermann O, de Ciman R, Gross U. Bacteremia among Kenyan Children. N Engl J Med. 2005;352:1379–81. [PubMed]
104. WHO, UNICEF. Integrated management of childhood illness chart booklet—standard. Geneva: World Health Organization; 2008. [accessed April 15, 2010]. http://www.who.int/child_adolescent_health/documents/IMCI_chartbooklet/en/index.html.
105. Reyburn H, Mbatia R, Drakeley C, et al. Overdiagnosis of malaria in patients with severe febrile illness in Tanzania: a prospective study. BMJ. 2004;329:1212. [PMC free article] [PubMed]
106. Danel C, Moh R, Minga A, et al. CD4-guided structured antiretroviral treatment interruption strategy in HIV-infected adults in west Africa (Trivacan ANRS 1269 trial): a randomised trial. Lancet. 2006;367:1981–89. [PubMed]
107. Gordon MA, Graham SM, Walsh AL, et al. Epidemics of invasive Salmonella enterica serovar enteritidis and S enterica serovar typhimurium infection associated with multidrug resistance among adults and children in Malawi. Clin Infect Dis. 2008;46:963–69. [PubMed]
108. Kariuki S, Revathi G, Kariuki N, et al. Invasive multidrug-resistant non-typhoidal Salmonella infections in Africa: zoonotic or anthroponotic transmission? J Med Microbiol. 2006;55:585–91. [PubMed]
109. Kariuki S, Revathi G, Kariuki N, et al. Increasing prevalence of multidrug-resistant non-typhoidal salmonellae, Kenya, 1994–2003. Int J Antimicrob Agents. 2005;25:38–43. [PubMed]
110. Kariuki S, Revathi G, Kiiru J, Lowe B, Berkley JA, Hart CA. Decreasing prevalence of antimicrobial resistance in non-typhoidal Salmonella isolated from children with bacteraemia in a rural district hospital, Kenya. Int J Antimicrob Agents. 2006;28:166–71. [PubMed]
111. Cotton MF, Wasserman E, Smit J, Whitelaw A, Zar HJ. High incidence of antimicrobial resistant organisms including extended spectrum beta-lactamase producing Enterobacteriaceae and methicillin-resistant Staphylococcus aureus in nasopharyngeal and blood isolates of HIV-infected children from Cape Town, South Africa. BMC Infect Dis. 2008;8:40. [PMC free article] [PubMed]
112. Perovic O, Koornhof H, Black V, Moodley I, Duse A, Galpin J. Staphylococcus aureus bacteraemia at two academic hospitals in Johannesburg. S Afr Med J. 2006;96:714–17. [PubMed]
113. Akinyemi KO, Smith SI, Oyefolu AO, Coker AO. Multidrug resistance in Salmonella enterica serovar typhi isolated from patients with typhoid fever complications in Lagos, Nigeria. Public Health. 2005;119:321–27. [PubMed]
114. Wasfy MO, Frenck R, Ismail TF, Mansour H, Malone JL, Mahoney FJ. Trends of multiple-drug resistance among Salmonella serotype typhi isolates during a 14-year period in Egypt. Clin Infect Dis. 2002;35:1265–68. [PubMed]
115. O'Brien KL, Wolfson LJ, Watt LP, et al. Burden of disease caused by Streptococcus pneumoniae in children younger than 5 years: global estimates. Lancet. 2009;374:893–902. [PubMed]
116. Watt JP, Wolfson LJ, O'Brien KL, et al. Burden of disease caused by Haemophilus influenzae in children younger than 5 years: global estimates. Lancet. 2009;374:903–11. [PubMed]
117. Morpeth SC, Ramadhani HO, Crump JA. Invasive non-typhi Salmonella disease in Africa. Clin Infect Dis. 2009;49:606–11. [PMC free article] [PubMed]
118. Crump JA, Youssef FG, Luby SP, et al. Estimating the incidence of typhoid fever and other febrile illnesses in developing countries. Emerg Infect Dis. 2003;9:539–44. [PMC free article] [PubMed]
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...