SEARCH

SEARCH BY CITATION

Keywords:

  • severe malaria;
  • bacteraemia;
  • Africa;
  • risk factors;
  • co-morbidity
  • malaria sévère;
  • bactériémie;
  • Afrique;
  • facteurs de risque;
  • co-morbidité
  • malaria severa;
  • bacteremia;
  • África;
  • factores de riesgo;
  • co-morbilidad

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Objectives  To describe the prevalence, aetiology and prognostic implications of coexisting invasive bacterial disease in children admitted with severe malaria in a rural Mozambican Hospital.

Methods  Retrospective study of data systematically collected from June 2003 to May 2007 in a rural Mozambican hospital, from all children younger than 5 years admitted with severe malaria.

Results  Seven thousand and forty-three children were admitted with a diagnosis of malaria. 25.2% fulfilled the criteria for severe malaria. 5.4% of the children with severe malaria and valid blood culture results had a concomitant bacteraemia. Case fatality rates of severe malaria cases rose steeply when bacteraemia was also present (from 4.0% to 22.0%, P < 0.0001), and bacteraemia was an independent risk factor for death among severe malaria patients (adjusted OR 6.2, 95% CI 2.8–13.7, P = 0.0001). Streptococcus pneumoniae, Gram-negative bacteria, Staphilococcus aureus and non-typhoid Salmonella (NTS) were the most frequently isolated microorganisms among severe malaria cases. Their frequency and associated case fatality rates (CFR) varied according to age and to syndromic presentation. Streptococcus pneumoniae had a relatively low CFR, but was consistently associated with severe malaria syndromes, or anaemia severity groups. No clear-cut relationship between malarial anaemia and NTS bacteraemia was found.

Conclusions  The coexistence of malaria and invasive bacterial infections is a frequent and life-threatening condition in many endemic African settings. In Mozambique, S. pneumoniae is the leading pathogen in this interaction, possibly as a consequence of the high HIV prevalence in the area. Measures directed at reducing the burden of both those infections are urgently needed to reduce child mortality in Africa.

Malaria sévère et bactériémie concomitante chez les enfants admis dans un hôpital rural mozambicain

Objectifs:  Décrire la prévalence, l’étiologie et les implications pronostiques de la coexistence de maladies bactériennes invasives chez les enfants admis avec une malaria sévère dans un hôpital rural mozambicain.

Méthodes:  Etude rétrospective de données collectées systématiquement de juin 2003 à mai 2007 dans un hôpital rural mozambicain, sur tous les enfants de moins de 5 ans admis avec une malaria sévère.

Résultats:  7043 enfants ont été admis avec un diagnostic de malaria. 25,2% répondaient aux critères de malaria sévère. 5,4% des enfants avec une malaria sévère et des résultats valides de culture de sang avaient une bactériémie concomitante. Les taux de létalité des cas de malaria sévère augmentaient fortement lorsqu’une bactériémie était également présente (de 4,0%à 22,0%; p < 0,0001) et la bactériémie était un facteur de risque indépendant de mortalité chez les patients atteints de malaria sévère (OR ajusté = 6,2; IC95%: 2,8-13,7; p = 0,0001). Streptococcus pneumoniae, bactéries à Gram négatif, Staphylococcus aureus et Salmonella non-typhoïdien (NTS) étaient les micro-organismes les plus fréquemment isolés chez les cas de malaria sévère. Leur fréquence et les taux de létalité associée variaient en fonction de l’âge et de la présentation syndromique. S. pneumoniae avait un taux de létalité relativement faible, mais était associé de façon consistante à des syndromes de malaria sévère ou à des groupes d’anémie sévère. Une relation claire entre l’anémie malarique et la bactériémie NTS n’a pas été trouvée.

Conclusions:  La coexistence de malaria et des infections bactériennes invasives est une condition fréquente et de danger la vie dans de nombreuses régions endémiques africaines. Au Mozambique, S. pneumoniae est le principal agent pathogène dans cette interaction, peut-être comme une conséquence de la forte prévalence du VIH dans la région. Mesures visant à réduire la charge de ces deux infections sont urgemment nécessaires pour réduire la mortalité infantile en Afrique.

Malaria severa y bacteremia concomitantes en niños ingresados en un hospital rural en Mozambique

Objetivos:  Describir la prevalencia, la etiología y las implicaciones del pronóstico de una enfermedad bacteriana invasiva coexistiendo en niños admitidos con malaria severa en un hospital rural de Mozambique.

Métodos:  Estudio retrospectivo de datos recolectados sistemáticamente entre el 2003 y Mayo del 2007, en un hospital rural Mozambicano, provenientes de niños menores de 5 años admitidos con malaria severa.

Resultados:  Se admitieron 7043 niños con un diagnóstico de malaria. 25.2% cumplían con los criterios de malaria severa. 5.4% de los niños con malaria severa y un resultado válido en el hemocultivo, tenían una bacteremia concomitante. La tasa de letalidad (TL) de la malaria severa aumentaba considerablemente cuando la bacteremia estaba también presente (del 4.0% al 22.0%, p < 0.0001), y la bacteremia era un factor de riesgo independiente para muerte entre pacientes con malaria severa (OR ajustado 6.2, 95% IC 2.8-13.7, p = 0.0001). Streptococcus pneumoniae, bacterias Gram negativas, Staphilococcus Aureus y Salmonella no tifoidea (NTS) eran los microorganismos más frecuentemente aislados entre los casos de malaria severa. Su frecuencia y las TL asociadas variaban de acuerdo con la edad y la presentación del síndrome. S. pneumoniae tenía una TL relativamente baja, pero estaba consistentemente asociada con grupos de síndrome de malaria severa, o anemia severa. No se halló una relación clara entre la anemia por malaria y bacteremia por NTS.

Conclusiones:  La coexistencia de la malaria con infecciones bacterianas invasivas es una condición frecuente y potencialmente mortal en muchos lugares endémicos de África. En Mozambique, S.pneumoniae es el principal patógeno en esta interacción, posiblemente como una consecuencia de la alta prevalencia de VIH en el área. Se requieren urgentemente medidas dirigidas a reducir la carga de estas dos infecciones para reducir la mortalidad infantil en África.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Invasive bacterial disease and malaria remain the two leading causes of paediatric mortality and morbidity in Africa (Berkley et al. 2005; Bryce et al. 2005). Malaria causes one of the estimated 4.4 million annual child deaths in this continent (Snow et al. 2005; WHO 2005), and invasive bacterial disease is nowadays, despite the difficulties in confirming its diagnosis (Mulholland & Adegbola 2005), clearly recognised as one of the principal killers of children per se. Streptococcus pneumoniae and Salmonella species, have been consistently reported (Walsh et al. 2000; Enwere et al. 2006; Roca et al. 2006; Sigauque et al. 2009) as the bacterial isolates causing major morbi-mortality in African children.

Both malaria and bacterial infections constitute an enormous burden for the under-resourced African health facilities. They represent the principal causes of admission and together account for more than half of all paediatric in-hospital deaths (Berkley et al. 2005; Bassat et al. 2008). Both infections predominantly affect young children, which also have the highest risk of dying.

The extent to which those two diseases overlap may vary according to several factors, including the endemicity of malaria and the prevalence and aetiology of bacteraemia in different countries, and this interaction has been addressed in various ways. Firstly, data on selected series of severe malaria patients (Prada et al. 1993; Enwere et al. 1998; Berkley et al. 1999; Bronzan et al. 2007) illustrate the possible association of malaria or of certain severe malaria syndromes with specific bacterial isolates. Conversely, series of children with bacteraemia have been also used to describe potential associations with malaria infection (Graham et al. 2000; Berkley et al. 2005). Although the literature suggests that malaria may predispose to certain specific bacterial infections (Mabey et al. 1987), the physiopathological mechanisms, and the temporality of this causal relationship remains not yet clearly understood. Moreover, concomitant bacteraemia may trigger in malaria patients a more severe form of the disease (Berkley et al. 1999). This study aims to describe the relationship between severe malaria and bacterial invasive disease in children younger than 5 years admitted to a district hospital in Mozambique.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study site and population

The study is a retrospective analysis of data systematically collected over 4 years from 1st of June 2003 to 31st of May 2007, from all children younger than 5 years admitted to the Manhiça District Hospital (MDH), in southern Mozambique. Most of the analysis includes only children with severe malaria.

The area has been described in detail elsewhere (Alonso et al. 2002). The Manhiça Health Research Centre (CISM) runs a Demographic Surveillance System in the area (Nhacolo et al. 2006) and a morbidity surveillance system at MDH (Guinovart et al. 2008). Malaria transmission, mainly caused by Plasmodium falciparum, is perennial with substantial seasonality and of moderate intensity.

MDH, with its 110-bed inpatient ward, is the main health facility in the area, used for primary health care by the nearby population, and one of the two referral health centres for Manhiça District (Loscertales et al. 2002).

Morbidity surveillance system

A standardised admission questionnaire is completed for all paediatric admissions. Upon arrival a finger prick blood sample is collected into heparinised capillaries to measure packed cell volume (PCV) and thick and thin blood films are prepared to quantify P. falciparum parasitaemia. A single venous blood culture is performed routinely for all children younger than 2 years of age and for children older than 2 years with confirmed hyperpyrexia (axillary temperature ≥ 39 °C), neurological signs or signs/symptoms of severity, including respiratory distress, circulatory collapse, or the clinician’s suspicion of sepsis (Sigauque et al. 2009).

Upon discharge or death, a physician revises the clinical history and records the final diagnoses, using the coding system proposed by the International Classification of Diseases.

Antiretroviral therapy only became available at MDH at the end of 2005, and before that, screening for HIV was limited to a selected number of suspected AIDS cases. For this reason, no HIV data are available for the patients included in this analysis.

Case management

Malaria patients are managed according to Mozambican national guidelines, which at the time of the study included parenteral treatment with quinine combined with Sulphadoxine–Pyrimethamine. Specific treatments for the different malaria syndromes have been described elsewhere (Bassat et al. 2008). Following national recommendations, the first line antibiotic used presumptively for suspected invasive bacterial disease is parenteral cloramphenicol or penicillin with or without gentamicin, or ampicillin plus gentamicin for severely malnourished children. Antibiotic therapy is reassessed according to culture results and sensitivity patterns and parenteral ceftriaxone is available as second-line therapy. Children requiring specialised care are transferred to Maputo Central Hospital.

Laboratory methods

PCV is measured using a microcentrifuge and a Hawksley haematocrit reader (Hawksley & Sons Ltd, Lancing, UK). Thick and thin blood films are processed and examined according to standard methods (Guinovart et al. 2008).

For blood cultures, a volume of 1–3 ml of whole blood is inoculated in a paediatric blood culture bottle (Pedibact®; Becton-Dickinson, Franklin Lakes, NJ, USA) and incubated in an automated system (BACTEC® 9050; Becton-Dickinson) for 4 days. Positive samples are examined by GRAM stain and subcultured on blood, chocolate or MacConkey agar plates as required. Bacterial isolates are identified following standard microbiologic procedures (Roca et al. 2006; Valles et al. 2006) and tested by disk-diffusion for antibiotic susceptibility (CLSI 2006).

Definitions

Bacteraemia was defined as the isolation of at least one non-contaminant bacteria from the admission blood culture. Coagulase-negative staphylococci, Bacillus species or micrococci were classified as contaminants. S. viridans isolates were not included in either the bacteraemia or the contamination group, due to the difficulty in attributing pathogenia. Non-typhoid Salmonellae (NTS) bacteraemia included all Salmonellae species, excluding Salmonella typhi cases. All other remaining gram negative bacteria were grouped for the analysis.

A severe malaria case was defined as a child admitted with a clinical diagnosis of malaria with P. falciparum asexual parasitaemia > 0 parasites/μl, and at least one of the following criteria: PCV < 15%, deep coma (Blantyre coma score ≤ 2), prostration (inability to sit unaided or to look for mother’s breast/feed in children who cannot yet sit), hypoglycaemia (<2.2 mmol/l), convulsions (≥2 reported episodes in the 24 hours prior to admission) or respiratory distress (deep breathing or indrawing). Patients meeting this definition but with a clinical diagnosis of meningitis were excluded.

Unconsciousness was defined as a Blantyre coma score on admission ≤3 in children below 8 months of age or ≤4 in older children. This age-specific difference is related to the difficulties in assessing the ability to locate a painful stimulus in younger children (Schellenberg et al. 1999). Mild anaemia was defined as a PCV 25 to <33%, moderate anaemia as a PCV 15 to <25% and severe anaemia as a PCV < 15% on admission. Malarial anaemia was defined as anaemia of any degree present in a child with severe malaria.

Nutritional status was assessed using weight for age Z-scores (WAZ), calculated using the LMS method and the 2000 CDC Growth Reference.

Data management and statistical methods

All admission questionnaires were double entered in FoxPro version 5.0 (Microsoft Corp., Seattle, WA, USA). Statistical analyses were done with Stata 9.0 (Stata Corp., College Station, TX, USA).

Case fatality rates (CFRs), calculated as the number of patients who died having a specific clinical presentation divided by the total number of patients with known outcome admitted with that clinical presentation, represent in-hospital mortality and do not include patients who absconded or were transferred.

Qualitative variables were compared using a chi-square test or Fisher’s exact test. Means of normally distributed variables were compared using the Student’s t-test or ANOVA. Geometric means of parasitaemia were compared using the Wilcoxon rank sum test.

A multivariate logistic regression analysis was performed to assess whether bacteraemia was a risk factor for death among severe malaria cases, using automated backward and forward stepwise estimations (both gave the same results). Given that the dependent variable was the final outcome (dead/alive), only children with a known outcome were included in the analysis (absconded or transferred excluded). All variables that were associated with death at a significance level of P < 0.10 in the univariate analysis were included in the model. The significance level for removal from the model was set at P = 0.06 and that for addition to the model at P = 0.05. Newborns were excluded from these analyses.

Plasmodium falciparum parasitaemia results were missing in 6.3% of admitted children and PCV in 7.1%. Blood cultures were not performed on admission or missing for 12.1% of the children. 2.0% of the children were transferred and 5.5% absconded. A potential bias could have been introduced when attempting to perform a pooled analysis of children younger and older than 2 years of age, as blood cultures are not routinely performed in children older than two and having one performed implies the presence of a certain degree of clinical severity. Thus, in patients older than two years of age, many malaria cases not showing signs of severity may not have had their blood cultured, and therefore some bacteraemia cases may have remained undetected. To limit this potential methodological problem, we restricted the analysis to severe malaria cases only, assuming therefore the presence of severe disease on admission, in both age groups.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

During the 4-year study period, 14 337 children younger than 5 years were admitted to MDH. 7043 (49.1%) children had malaria, and a quarter of these (n = 1780) fulfilled the severe malaria definition. A flow chart (Figure 1) summarises the number of study children according to diagnosis, age group (<2 years and >2 years), availability and validity of blood culture results. 12.8% (207/1616) of severe malaria patients with a blood culture performed grew contaminant microorganisms, and 0.3% (5/1616) grew S. viridans. 5.4% (76/1404) of severe malaria patients with a valid culture result (excluding contaminations or S. viridans), had bacteraemia.

image

Figure 1.  Flow chart summarizing the number of study children according to diagnosis age group (<2 years and ≥2 years) availability and validity of blood culture results.

Download figure to PowerPoint

Among severe malaria patients with bacteraemia, S. pneumoniae was the most frequent isolate (26.3%, n = 20), followed by Gram negative bacteria (21.1%, n = 16), S. aureus (17.1%, n = 13) and NTS (15.8%, n = 12).

Figure 2 shows the aetiology of bacteraemia in patients with severe malaria, as compared to the aetiology of bacteraemia for all other non-malaria admissions. Distribution of microorganisms varied according to admission diagnosis, but in both cases S. pneumoniae was the most frequent isolate, although differences were non-significant. Figure 3 shows the prevalence of bacteraemia according to the different severe malaria syndromes. While the prevalence in the severe malaria cases with hypoglycaemia was very high (12.5%), it decreased substantially in the severe malaria cases with convulsions (1.7%). Figure 4 shows the relative contribution of the main bacteria within each syndromic group. S. pneumoniae was consistently the most important isolate in each of these groups, except for patients with convulsions or prostration.

image

Figure 2.  Relative contribution (in %) of the main 4 bacterial isolates as causes of bacteraemia among severe malaria cases compared to all other non-malaria diagnoses.

Download figure to PowerPoint

image

Figure 3.  Prevalence of bacteraemia among different severe malaria syndromes in children with valid blood culture results.

Download figure to PowerPoint

image

Figure 4.  Relative contribution (in %) of different bacterial isolates according to different severe malaria syndromes/symptoms.

Download figure to PowerPoint

Table 1 shows clinical and laboratory parameters of children with severe malaria according to blood culture result, and Table 2 refers them to different bacterial isolates. Bacteraemic children with severe malaria had a higher CFR, were more severely malnourished and had lower mean parasitaemias than severe malaria patients with a negative blood culture.

Table 1.   Clinical and laboratory parameters of children with severe malaria according to blood culture result
 Negative (N = 1328)Positive (N = 76) P (+ vs.−)
  1. 95% CI, 95% confidence interval; IQR, inter-quartile range.

  2. †Negative: N = 1244; positive: N = 59.

  3. ‡Negative: N = 1304; positive: N = 76.

  4. §Negative: N = 1307; positive: N = 73.

Age (mean (SD)) (months)22.0 (14.7)19.1 (13.9)0.09
Case fatality rate† (% (n))4.0 (50)22.0 (13)<0.0001
Respiratory distress (% (n))40.8 (542)48.7 (37)0.175
Haematocrit‡ (mean (SD))23.4 (8.4)22.7 (8.4)0.440
Mean parasitaemia (parasite/μl) (geometric mean (95% CI))16 639 (14 914–18 562)6537 (3718–11 494)0.002
Weight for age Z-score§ (median (IQR))−1.6 (−2.7/−0.6)−2.2 (−3.5/−1.2)0.001
Table 2.   Clinical and laboratory parameters of children with severe malaria and bacteraemia according to blood culture isolate
  S. pneumoniae (n = 20)Non-typhoid Salmonella NTS (n = 12) Gram negative bacteria (n = 16) S. aureus (n = 13)
  1. 95% CI, 95% confidence interval; IQR, inter-quartile range.

  2. †Negative: n = 1244; positive: n = 59.

Age (mean (SD)) (months)17.2 (10.8)16.4 (6.1)23.3 (19.9)22.0 (14.8)
Case fatality rate (% (n/N))13.3 (2/15)25.0 (2/8)25.0 (3/12)30.0 (3/10)
Respiratory distress (% (n))60.0 (12)33 (4)37.5 (6)30.8 (4)
Haematocrit† (mean (SD))20.8 (6.9)22.1 (9.3)22.3 (8.8)23.2 (7.6)
Mean parasitaemia (parasite/μl) (geometric mean (95% CI))1753 (564–5449)9687 (3233–29 030)13 338 (3079–57 778)11 883 (2326–60 717)
Weight for age Z-score (median (IQR))−2.7 (−3.4/−1.6)−2.6 (−3.3/−2.1)−2.2 (−4.1/−1.1)−2.0 (−3.1/−0.4)

The prevalence of NTS bacteraemia was not significantly different between severe malaria and non-malaria cases (data not shown). Respiratory bacteria (S. pneumoniae and Hib) were isolated more frequently from patients with severe malaria who had respiratory distress (43.2%) than from patients without it (20.5%; P = 0.03).

When assessing specific signs and symptoms among severe malaria patients with valid blood culture results, bacteraemia was significantly more frequent in patients who presented with hypoglycaemia on admission (12.5%vs. 5.1% in those that did not, P = 0.01), hepatomegaly (16.1%vs. 5.0%, P < 0.0001) or oral candidiasis (30.8%vs. 5.2%, P < 0.0001), and significantly less frequent among those with convulsions (1.7%vs. 6.0%, P = 0.02). No significant differences were observed between children presenting and not presenting other signs or symptoms.

Anaemia and bacteraemia in severe malaria patients

The prevalence of anaemia (PCV < 33%) among severe malaria patients was high (83.0%). The frequency of bacteraemia seemed to be independent of the presence of anaemia and this remained true for the different degrees of anaemia severity, being 6.3% (26/416) among children with mild anaemia, 4.6% (22/474) among those with moderate anaemia and 7.0% (18/256) among severe anaemia cases (p = 0.4). Figure 5 illustrates the aetiology of bacteraemia cases among severe malaria patients according to admission haematocrit. S. pneumoniae was the most frequent bacterial isolate among all malarial anaemia groups (mild, moderate and severe).

image

Figure 5.  Aetiology (% (n)) of bacteraemia cases in severe malaria patients according to haematocrit on admission. PCV, packed cell volume.

Download figure to PowerPoint

Case fatality rates

Severe malaria accounted for 12% (81/669) of all in-hospital deaths in children <5 years of age. Thirteen children with bacteraemia and severe malaria died. CFR for severe malaria was overall 4.9% (81/1648), rising significantly to 22.0% (13/59) when associated with concomitant bacteraemia.

The impact of having a concomitant bacteraemia was important in some but not all severe malaria syndromic presentations. Among patients with severe malaria and severe anaemia which also had bacteraemia, CFR rose when compared to those without bacteraemia [31.3% (5/16) vs. 5.2% (11/211), P = 0.002]. Similarly, severe malaria patients with respiratory distress and superimposed bacteraemia had a much higher CFR than those without bacteraemia [32.1% (9/28) vs. 5.2% (26/498), P < 0.0001]. This increased CFR related to the concomitant bacteraemia could not be confirmed in those severe malaria cases with deep coma, although numbers were very small.

CFRs of the different bacterial isolates among severe malaria patients ranged from 12.5% for Streptococci to 33.3% for H. influenzae. Among severe malaria bacteraemic patients, CFRs did not significantly vary according to antibiotic use (25.0% for children who received antibiotics vs. 17.4% among those who did not; P = 0.49).

Risk factors for death

Independent risk factors for death among severe malaria patients are shown in Table 3. Importantly, a positive blood culture independently increased the odds of death by 6.2 (95% CI 2.8–13.7).

Table 3.   Risk factors independently associated with mortality in children <5 years (n = 1165) with severe malaria according to a multivariate logistic regression model
Risk factorAdjusted OR 95% CI P (LR χ2-test 1 d.f.)
  1. LR, likelihood ratio.

  2. †Only includes children with a valid blood culture result (positive versus negative blood culture).

  3. ‡Only includes peripheral health posts traditional healers and informal health sector.

Positive blood culture†6.22.8–13.7<0.0001
Unconsciousness2.81.3–6.10.02
Oral candidiasis7.61.4–42.80.04
Weight for age Z-score (per unit increase)0.740.62–0.880.0005
Nasal flaring3.31.8–5.90.0001
Hypoglycaemia3.01.2–7.70.04
Having been previously visited‡2.21.2–4.00.01

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Severe malaria is the predominant cause of hospital admission among children younger than 5 years of age attending Manhiça hospital (Bassat et al. 2008). This study portrays the frequent coexistence of bacteraemia among severe malaria patients, and its detrimental prognostic implications. Despite a substantial degree of overlap between the two diseases, it seems difficult to establish whether one infection predisposes to the other or whether they are both coincidental. Indeed, 5.4% of all admitted patients with severe malaria and valid blood culture results had concurrent bacteraemia, and conversely, 7.3% of bacteraemic patients had also severe malaria. Bacteraemia substantially increased, as anticipated (Berkley et al. 1999; Bronzan et al. 2007), the risk of death. Bacteraemia should therefore be considered in future revisions as one of the defining criteria for severe malaria. The remaining independent risk factors for death among malaria cases are concordant with those previously described for the area (Bassat et al. 2008), disparities being attributable to differences in sample size. CFRs of patients with both diseases were much higher than CFRs of any of the infections alone. Parasite densities among severe malaria patients with bacteraemia were significantly lower than those among non-bacteraemic, suggesting that some of those low-grade malarial infections could be incidental. This finding contrasts with previous studies (Berkley et al. 1999; Bronzan et al. 2007) in which parasitaemias in severe malaria patients with or without bacteraemia did not differ significantly, but such studies did not look at all consecutive admissions, and were only based on selected series of severe malaria cases predominantly presenting with cerebral malaria, severe anaemia or respiratory distress. It is important to note, however, that due to the age-related differences in the criteria for taking blood cultures, an important proportion of children older than two years of age may lack a blood culture in admission, as compared to younger children. The fact that only 117 severe malaria patients (15.7%) older than two years of age did not have a blood culture performed reassures us that most severe malaria patients did indeed fulfil severity criteria despite their age for blood to be cultured. The lack of microbiological data on these admitted children may, however, bias our estimates on the magnitude of the coexistence of both diseases in any direction, but should not represent, in any case, a major deviation from reality.

As all blood cultures were performed on admission in this study, both diseases, when coexisting, were simultaneously present on arrival. It is however not possible to ascertain the causal pathway leading to such co-infection, as it is problematic to distinguish whether one patient dies with bacteria or parasites in their blood or as a result of such parasites or bacteria.

A consistent relationship between malaria and NTS bacteraemia has been documented (Graham et al. 2000; Walsh et al. 2002; Bronzan et al. 2007), and several studies have suggested that malaria parasites or malarial anaemia may predispose to NTS bacteraemia (Mabey et al. 1987). Other African series (Prada et al. 1993; Bahwere et al. 2001; Evans et al. 2004; Norton et al. 2004) confirmed this and also underlined the importance of other GRAM negative bacteria among malaria cases. In contrast to these findings, our data highlight the importance of S. pneumoniae as the principal isolate related with severe malaria cases in Manhiça. Difficulties inherent to the isolation of S. pneumoniae (such as its slow growth rate and subsequent potential for contamination of the truly infected samples by other faster-growing contaminant bacteria) could partially explain why such infection was rare in previous African studies (Evans et al. 2004). The HIV infection, highly prevalent in Manhiça [23.6% positivity among pregnant women attending the antenatal clinic in 2001–2006 (Menendez et al. 2008)], has traditionally been considered a risk factor for pneumococcal invasive disease (Greenwood 1999). Unfortunately, no HIV data were available for these patients, and its relevance cannot be confirmed, potentially bringing an important bias to these results. However, the importance of HIV infection may be inferred indirectly by the high prevalence of oral candidiasis among children in the study, an unusual finding in severe malaria patients, but a good marker of immunosuppression. Similarly, the absence of sickle cell disease in Manhiça (Menéndez, personal communication), a genetic trait traditionally associated with an increase in NTS and other GRAM negative bacterial infections, may also contribute to enhance the burden of pneumococcal disease in detriment of NTS.

These data also challenge a clear-cut relation between severe malaria cases with any degree of anaemia and NTS. Indeed, S. pneumoniae was again the most frequent isolate in any of the different anaemia groups, and also showed the lowest associated mean admission haematocrit. NTS bacteraemias were associated with higher geometric mean parasitaemias than S. pneumoniae bacteraemias, contrary to previous reports (Mabey et al. 1987; Brent et al. 2006a) which linked NTS infections to low-level parasite loads or recently cured malaria episodes. This study did not replicate the previously described relationship between malarial anaemia and NTS bacteraemia, although an association between them may still exist, and might not have been detected due to the relatively small number of severe malaria cases with bacteraemia. The contribution of other factors to the development of anaemia in this setting may partially explain why such an association was not clearly confirmed.

This retrospective analysis of all under-five admissions over four years has given insight on the way malaria and bacteraemia intertwine. Importantly, these data may however underestimate what really occurs in the community (Brent et al. 2006b), where an important proportion of the burden of these two diseases remains undetected. Individually, these two diseases have an enormous impact on the health of African children, and this study has shown that the coexistence of bacteraemia drastically worsens the clinical prognosis in malaria. The deployment of existing, already registered and effective antibacterial tools, such as the pneumococcus conjugate vaccine, necessary to decrease the silent but constant burden of bacteraemia, may also have an impact in reducing malaria mortality.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We are indebted to the children and mothers participating in the study. We would like to thank Madalena Ripinga and Mariano Sitaúbe. The work of the clinical officers, data managers, laboratory workers and laboratory coordinator was important for the successful collection of data. CISM receives financial support from the Spanish Agency for International Cooperation (AECI). QB is supported by the Spanish Ministry of Health (Contrato post formación sanitaria especializada, Fondo de Investigaciones Sanitarias, Instituto de Salud Carlos III, ref. CM05/00134).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Alonso PL, Saúte F, Aponte JJ et al. (2002) Manhiça DSS, Mozambique. In: Population and Health in Developing Countries. INDEPTH, Ottawa, pp. 189195.
  • Bahwere P, Levy J, Hennart P et al. (2001) Community-acquired bacteremia among hospitalized children in rural central Africa. International Journal of Infectious diseases 5, 180188.
  • Bassat Q, Guinovart C, Sigauque B et al. (2008) Malaria in rural Mozambique. Part II. Children admitted to hospital. Malaria Journal 7, 37.
  • Berkley J, Mwarumba S, Bramham K, Lowe B & Marsh K (1999) Bacteraemia complicating severe malaria in children. Transactions of the Royal Society of Tropical Medicine and Hygiene 93, 283286.
  • Berkley JA, Lowe BS, Mwangi I et al. (2005) Bacteremia among children admitted to a rural hospital in Kenya. The New England Journal of Medicine 352, 3947.
  • Brent AJ, Oundo JO, Mwangi I, Ochola L, Lowe B & Berkley JA (2006a) Salmonella bacteremia in Kenyan children. The Pediatric Infectious Disease Journal 25, 230236.
  • Brent AJ, Ahmed I, Ndiritu M et al. (2006b) Incidence of clinically significant bacteraemia in children who present to hospital in Kenya: community-based observational study. Lancet 367, 482488.
  • Bronzan RN, Taylor TE, Mwenechanya J et al. (2007) Bacteremia in Malawian children with severe malaria: prevalence etiology HIV coinfection and outcome. The Journal of Infectious Diseases 195, 895904.
  • Bryce J, Boschi-Pinto C, Shibuya K & Black RE (2005) WHO estimates of the causes of death in children. Lancet 365, 11471152.
  • Clinical Laboratory Standards Institute (CLSI) Ed. (2006) Performance Standards for Antimicrobial Disk Susceptibility Tests; Approved Standard, 9th edn. CLSI, Wayne.
  • Enwere G, Van Hensbroek MB, Adegbola R et al. (1998) Bacteraemia in cerebral malaria. Annals of Tropical Paediatrics 18, 275278.
  • Enwere G, Biney E, Cheung YB et al. (2006) Epidemiologic and clinical characteristics of community-acquired invasive bacterial infections in children aged 2–29 months in The Gambia. The Pediatric Infectious Disease Journal 25, 700705.
  • Evans JA, Adusei A, Timmann C et al. (2004) High mortality of infant bacteraemia clinically indistinguishable from severe malaria. The Quarterly Journal of Medicine 97, 591597.
  • Graham SM, Walsh AL, Molyneux EM, Phiri AJ & Molyneux ME (2000) Clinical presentation of non-typhoidal Salmonella bacteraemia in Malawian children. Transactions of the Royal Society of Tropical Medicine and Hygiene 94, 310314.
  • Greenwood B (1999) The epidemiology of pneumococcal infection in children in the developing world. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 354, 777785.
  • Guinovart C, Bassat Q, Sigauque B et al. (2008) Malaria in rural Mozambique. Part I. Children attending the outpatient clinic. Malaria Journal 7, 36.
  • Loscertales MP, Roca A, Ventura P et al. (2002) Epidemiology and clinical presentation of respiratory syncytial virus infection in a rural area of southern Mozambique. The Pediatric Infectious Disease Journal 21, 148155.
  • Mabey DC, Brown A & Greenwood BM (1987) Plasmodium falciparum malaria and Salmonella infections in Gambian children. The Journal of Infectious Diseases 155, 13191321.
  • Menendez C, Bardaji A, Sigauque B et al. (2008) A randomized placebo-controlled trial of intermittent preventive treatment in pregnant women in the context of insecticide treated nets delivered through the antenatal clinic. PLoS ONE 3, e1934.
  • Mulholland EK & Adegbola RA (2005) Bacterial infections – a major cause of death among children in Africa. The New England Journal of Medicine 352, 7577.
  • Nhacolo AQ, Nhalungo DA, Sacoor CN, Aponte JJ, Thompson R & Alonso P (2006) Levels and trends of demographic indices in southern rural Mozambique: evidence from demographic surveillance in Manhica district. BMC Public Health 6, 291.
  • Norton EB, Archibald LK, Nwanyanwu OC et al. (2004) Clinical predictors of bloodstream infections and mortality in hospitalized Malawian children. The Pediatric Infectious Disease Journal 23, 145151.
  • Prada J, Alabi SA & Bienzle U (1993) Bacterial strains isolated from blood cultures of Nigerian children with cerebral malaria. Lancet 342, 1114.
  • Roca A, Sigauque B, Quinto L et al. (2006) Invasive pneumococcal disease in children <5 years of age in rural Mozambique. Tropical Medicine & International Health 11, 14221431.
  • Schellenberg D, Menendez C, Kahigwa E et al. (1999) African children with malaria in an area of intense Plasmodium falciparum transmission: features on admission to the hospital and risk factors for death. The American Journal of Tropical Medicine and Hygiene 61, 431438.
  • Sigauque B, Roca A, Mandomando I et al. (2009) Community-acquired bacteremia among children admitted to a rural hospital in Mozambique. The Pediatric Infectious Disease Journal 28, [Epub ahead of print].
  • Snow RW, Guerra CA, Noor AM, Myint HY & Hay SI (2005) The global distribution of clinical episodes of Plasmodium falciparum malaria. Nature 434, 214217.
  • Valles X, Flannery B, Roca A et al. (2006) Serotype distribution and antibiotic susceptibility of invasive and nasopharyngeal isolates of Streptococcus pneumoniae among children in rural Mozambique. Tropical Medicine and Interactional Health 11, 358366.
  • Walsh AL, Phiri AJ, Graham SM, Molyneux EM & Molyneux ME (2000) Bacteremia in febrile Malawian children: clinical and microbiologic features. The Pediatric Infectious Disease Journal 19, 312318.
  • Walsh AL, Molyneux EM, Kabudula M, Phiri AJ, Molyneux ME & Graham SM (2002) Bacteraemia following blood transfusion in Malawian children: predominance of Salmonella. Transactions of the Royal Society of Tropical Medicine and Hygiene 96, 276277.
  • WHO (2005) World Malaria Report. WHO, Geneva.