Table 1 Summary of 22 observational studies recording invasive bacterial infection (IBI) in children with Plasmodium falciparum malaria in sSA

SEVERE MALARIA

Angyo [13]

Jos, Nigeria

Aug 92 – Oct 93

PCS;

SPN 34.3%

SM definition included slide negative cases (n = 214). SM mortality = 3.2%.

Severe malaria. Urban TH

S. aureus 20%

Salm spp 14.3%

Bassat [23]

Manhica, Mozambique

Jun 03 – May 07

RCS;

All EGN 36.8% NTS 15.8%

High mortality SMA and IBI (31%).

Severe malaria. Rural DH

SPN 26.3%.

Valid BC data excludes contaminants

S. aureus 17%,

Berkley [36]

Kilifi, Kenya

Apr 93 – May 96

RCS

SPN 30%

Bacteremia less prevalent in >30 m (2.5%).

S. aureus 16%

Severe malaria. Rural DH

NTS 14%

Berkley [24]

Kilifi, Kenya

Aug 98 – Jul 02

PCS

EGN at high parasite counts >50,000

Mortality strongly correlated with IBI (but not with SPN and HIb).

Severe malaria. Rural DH

Case fatality inversely related to parasitemia.

Bronzan [38]

Blantyre, Malawi

1996 - 2005

PCS;

NTS 58%

Invasive NTS assoc with severe anemia (73% of NTS cases). Not associated with mortality.

(Rainy seasons)

Severe malaria. Urban RH

SPN 11%

Dorndorp [2]

Multisite

Oct 05 –

RCT;

NR

60% children received antibiotics.

Mozambique

Jul 10

Severe malaria.

Gambia

Multi center

Kenya DRC

Ghana

Tanzania

Uganda

Nigeria

Rwanda

Enwere [39]

Banjul, Gambia

1992-1994

RCT; Cerebral malaria

S. aureus 42.9%

Possible contaminants. High survival rate without antibiotic treatment.

(BCS ≤2) Urban TH

EGN 28.6%

Evans [22]

Kumasi, Ghana

nr

PCS;

NTS 43%

OR for IBI 3.5 (95% CI 1.4-8.2). No positive association of SM and bacteremia.

Pub 2004

Severe malaria. Urban TH

S.aureus 39%

E.coli 4%

Kremsner [42]

Lambarene, Gabon

nr

PCS; sub-analysis of Abx trial; Severe malaria

S. aureus (3) 42.9%

Addition of clindamycin to quinine shortened disease course

Pub 1995

Rural DH

EGN (2) 28.5%

Prada [43]

Lagos, Nigeria

May 92 – Aug 92

PCS; Cerebral malaria

All EGN 75%

Hemolytic activity detected in 50% of bacterial strains analyzed

Urban TH

Salmonella 25%

ALL SEVERITY MALARIA

Akinyemi [44]

Lagos, Nigeria

Oct 04 – Sept 05

PCS; Febrile illness.

S. tyhpi 45.2%

Only salmonella isolates reported. Included adults Child results reported here (unless indicated).

S. enteritidis 24.2%

2 urban RHs & 2 urban clinics

(Incl adults)

Akpede [40]

Benin, Nigeria

Oct 88 – Oct 89

PCS;

S. aureus 39.5%

Reanalysis of data. Retracted original suggestion of association of IBI and malaria. Note case fraction of IBI in malaria inaccurate (non-malarial febrile illness not differentiated)

Signs of febrile illness

EGN 23.2%

Urban TH

Bahwere [25]

Lwiro, DRC

Jan 89 – Dec 90

RCS;

All patients^b

Positive malaria film and anemia (Hb < 8) associated with higher IBI case fraction

All hospital admissions. Rural DH

NTS 36.5%,

E. coli 15.1%, Citrobacter 6.3%

Berkley [52]

Kilifi, Kenya

Feb 99 – Dec 01

RCS;

NR

Source data (Berkley 2009) [24]

All hospital admissions. Rural DH

1/9 with syndrome indicating Abx Rx had IBI.

Mabey [26]

Fajara, Gambia

Jan 79 – Oct 84

RCS;

NTS 69.8%

Seasonal association of iNTS with malaria. NTS rare (3%) in kids >4y. iNTS more common in malaria season.

All hospital admissions. Rural DH

S. typhi 11.6%

Mtove [45]

Muheza, Tanzania

Mar 08 – Feb 09

PCS;

All patients^b

Invasive NTS more likely to have malaria, recent malaria, anemia, low glucose

Febrile illness + ≥1 severity criteria.

NTS 29%

E. coli 17%

Rural DH

S. typhi 9%

Nadjm [5]

Muheza, Tanzania

June 06 - May 07

PCS;

NTS 52%

>50% organisms not susceptible to 1st line Abx. Case fatality non-malaria IBI = 19%

Fever in last 48 h.

Rural DH

Okwara [20]

Nairobi, Kenya

Jan 01 – Mar 01

CSS;

Any GNR 62.5%

Also examined urinary isolates. 60% malaria dx in non-endemic area

Febrile Illness.

NTS 34.4%

Urban TH

Sigauque [71]

Manhica, Mozambique

May 01- April 06

RCS;

All patients^b

Almost half of the community acquired bacteremias had concomitant malaria

All hospital admissions. Rural DH

NTS 26%

SPN 25%

S. aureus 12%

E. coli 10%

Were [41]

Siaya, Kenya

Mar 04 – Jan 06

PCS; Malaria admissions;

NTS 42.3%

Proportion of IBI decreased with increasing parasitemia. Excluded previous hospitalization and CM.

S. typhi 37%;

Rural DH

S. aureus 35.6%

NON-SEVERE MALARIA

Ayoola [27]

Ibadan, Nigeria

Jun – Nov 98

PCS;

Gram pos 56%

Excluded SM and if history of antibiotic use in preceding 7 days

Fever. Urban TH

Gram neg44%

Brent [28]

Kilifi, Kenya

May 03 – Oct 03

PCS; Unselected outpatient attendees.

All patients^b

Recruitment outside of malaria season. Excluded children with recent admission (10 days)

Rural DH

SPN 50%