Bacteremia among Kenyan Children
http://www.100md.com
《新英格兰医药杂志》
To the Editor: Berkley et al. (Jan. 6 issue)1 studied the prevalence and outcome of bacteremia among 19,339 children admitted to a rural hospital in Kenya. Most of the children were less than five years of age. Applying rigorous methods despite difficult conditions, the investigators found that the major causes of death were common bacterial infections, not the often-touted culprits — malaria, AIDS, and tuberculosis.
This article leaves unanswered the question of the effectiveness and availability of antibiotics for treatment. To which antibiotics were the isolates susceptible or resistant? Antimicrobial activity in blood samples was associated with increased mortality. Did the antimicrobial agent provide appropriate treatment for the organism causing the bacteremia?
A high frequency of resistance to first-line drugs has been documented among common bacterial pathogens in the recent Alliance for the Prudent Use of Antibiotics report entitled "Shadow Epidemic."2 The prevalence of resistance to penicillin among pneumococci ranges from 20 percent in the United States to more than 50 percent in Asia. Ampicillin resistance among Haemophilus influenzae isolates approaches 25 to 35 percent globally.
The study emphasizes the need to pay more attention to common bacterial causes of life-threatening infections in the developing world. Knowing the organisms' susceptibility to antibiotics will help one choose the appropriate treatment for these infections.
Stuart B. Levy, M.D.
Tufts University School of Medicine
Boston, MA 02111
stuart.levy@tufts.edu
References
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.
Shadow epidemic: the growing menace of drug resistance: 2005 GAARD report. Boston: Alliance for the Prudent Use of Antibiotics, 2004. (Accessed March 10, 2005, at http://www.apua.org.)
To the Editor: Bacteremia is an underestimated health problem for most African societies, as has been shown by Berkley et al. To draw conclusions about therapy and prevention, additional knowledge of rates of resistance in various geographic settings is necessary. Therefore, we investigated the cause of fever of unknown origin in 409 consecutive patients in three rural hospitals in Ghana after excluding those with malaria. The median age was 15 years (range, 2 weeks to 86 years). Of these patients, 205 (50.1 percent) had positive blood cultures. In 141 patients, a putative agent of septicemia was identified. In contrast to the situation in Kenya, salmonella was the most abundant bacterial genus (100 isolates) in Ghana. Of these, 59 were Salmonella typhi, a finding that has also been made in other tropical countries.1,2 More than 80 percent of all isolates of salmonella were resistant to chloramphenicol, an antibiotic that is used widely in Ghana. However, most bacteria were fully susceptible to ciprofloxacin, late-generation cephalosporins, or both (Table 1). Multidrug-resistant S. typhi has been reported from several tropical countries but not from Ghana.3 We hope that our findings help influence health planners to alter current therapeutic strategies in view of the local patterns of resistance.
Table 1. Percentage of Resistant Isolates of the Most Prevalent Bacteria Causing Fever of Unknown Origin in Ghana.
Ortrud Zimmermann
University of G?ttingen
D-37075 G?ttingen, Germany
Ring de Ciman, M.D.
St. Francis Xavier Hospital
Assin Foso, Ghana
Uwe Gross, M.D.
University of G?ttingen
D-37075 G?ttingen, Germany
ugross@gwdg.de
References
Shears P. Antibiotic resistance in the tropics: epidemiology and surveillance of antimicrobial resistance in the tropics. Trans R Soc Trop Med Hyg 2001;95:127-130.
Sinha A, Sazawal S, Kumar R, et al. Typhoid fever in children aged less than 5 years. Lancet 1999;354:734-737.
Parry CM, Hien TT, Dougan G, White NJ, Farrar JJ. Typhoid fever. N Engl J Med 2002;347:1770-1782.
The authors reply: We thank Dr. Levy and also Zimmermann and colleagues for their correspondence. The principal objective of our report in the Journal was to estimate the hospital burden of disease and death due to childhood bacteremia in a rural African setting — an area where few data exist. There is also a considerable lack of data on antimicrobial susceptibilities in sub-Saharan Africa. Because most deaths from bacteremic illnesses occur very early after admission, it is important to define appropriate empirical treatment for the first medical contact. In our study, evidence of recent use of antibiotics (antibiotic activity in plasma) was correlated with disease severity and death; we interpret this to indicate that sicker patients are more likely to have received antibiotics recently, rather than to indicate that antibiotics are failing.
The Ghanaian study investigated a clinical subgroup with fever of unknown origin and was not confined to children. This group of patients is not directly comparable to our study population. Furthermore, in Kenya, we found that axillary temperature at admission was a very poor predictor of bacteremia, even among patients without malaria. The receiver-operating-characteristic curve shows detection of bacteremia at different thresholds of axillary temperature among 16,570 unselected pediatric patients (Figure 1). Had we targeted children with an axillary temperature of 39°C or more, then 769 of 1094 children with bacteremia (70 percent) would have been missed.
Figure 1. Receiver-Operating-Characteristic Curves for Axillary Temperature for the Prediction of Bacteremia among Patients with and without Malarial Parasites at a Kenyan Hospital.
Sensitivity and 1 minus specificity are shown at various thresholds of axillary temperature (39°C, 38°C, and 37°C). The blue line denotes admissions of patients without parasitemia (9378 patients; area under the curve, 0.58 [95 percent confidence interval, 0.56 to 0.60). The red line denotes admissions of patients with parasitemia (7051 patients; area under the curve, 0.48 [95 percent confidence interval, 0.44 to 0.51]). The area under the curve for all admissions (16,570 pediatric patients) was 0.52 (95 percent confidence interval, 0.50 to 0.54; not shown). A reference line is also shown.
Different age groups and different inclusion criteria will yield different bacterial pathogens. Salmonella species are very common among adults in Kenya, as they are in Ghana, and antimicrobial susceptibilities have recently been reported.1,2 The use of a conjugate Haemophilus influenzae type b vaccine since 2002 may explain an absence of H. influenzae, but the failure of the study by Zimmermann et al. to isolate Streptococcus pneumoniae is remarkable, since this fastidious organism is prevalent in Ghana3 and elsewhere in West Africa. It is possible that the undefined clinical selection criteria may have especially favored salmonella. However, without details of microbiologic methods or quality standards, interpretation of the findings is not possible. Standardization of clinical definitions4 and laboratory procedures5 through networks and use of longitudinal data sets is essential to the development of appropriate antimicrobial policies.
James A. Berkley, M.D.
Brett S. Lowe, M.Phil.
J. Anthony G. Scott, M.R.C.P.
Centre for Geographic Medicine Research (Coast)
Kilifi, Kenya
jberkley@kilifi.mimcom.net
References
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.
Kariuki S, Revanthi G, Muyodi J, et al. Characterization of multidrug-resistant typhoid outbreaks in Kenya. J Clin Microbiol 2004;42:1477-1482.
Denno DM, Frimpong E, Gregory M, Steele RW. Nasopharyngeal carriage and susceptibility patterns of Streptococcus pneumoniae in Kumasi, Ghana. West Afr J Med 2002;21:233-236.
Management of the child with a serious infection or severe malnutrition: guidelines for care at the first referral level in developing countries. Geneva: World Health Organization, 2000.
Yeo SF, Akalin E, Arikan S, et al. Susceptibility testing of Haemophilus influenzae: an international collaborative study in quality assessment. J Antimicrob Chemother 1996;38:363-386.
This article leaves unanswered the question of the effectiveness and availability of antibiotics for treatment. To which antibiotics were the isolates susceptible or resistant? Antimicrobial activity in blood samples was associated with increased mortality. Did the antimicrobial agent provide appropriate treatment for the organism causing the bacteremia?
A high frequency of resistance to first-line drugs has been documented among common bacterial pathogens in the recent Alliance for the Prudent Use of Antibiotics report entitled "Shadow Epidemic."2 The prevalence of resistance to penicillin among pneumococci ranges from 20 percent in the United States to more than 50 percent in Asia. Ampicillin resistance among Haemophilus influenzae isolates approaches 25 to 35 percent globally.
The study emphasizes the need to pay more attention to common bacterial causes of life-threatening infections in the developing world. Knowing the organisms' susceptibility to antibiotics will help one choose the appropriate treatment for these infections.
Stuart B. Levy, M.D.
Tufts University School of Medicine
Boston, MA 02111
stuart.levy@tufts.edu
References
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.
Shadow epidemic: the growing menace of drug resistance: 2005 GAARD report. Boston: Alliance for the Prudent Use of Antibiotics, 2004. (Accessed March 10, 2005, at http://www.apua.org.)
To the Editor: Bacteremia is an underestimated health problem for most African societies, as has been shown by Berkley et al. To draw conclusions about therapy and prevention, additional knowledge of rates of resistance in various geographic settings is necessary. Therefore, we investigated the cause of fever of unknown origin in 409 consecutive patients in three rural hospitals in Ghana after excluding those with malaria. The median age was 15 years (range, 2 weeks to 86 years). Of these patients, 205 (50.1 percent) had positive blood cultures. In 141 patients, a putative agent of septicemia was identified. In contrast to the situation in Kenya, salmonella was the most abundant bacterial genus (100 isolates) in Ghana. Of these, 59 were Salmonella typhi, a finding that has also been made in other tropical countries.1,2 More than 80 percent of all isolates of salmonella were resistant to chloramphenicol, an antibiotic that is used widely in Ghana. However, most bacteria were fully susceptible to ciprofloxacin, late-generation cephalosporins, or both (Table 1). Multidrug-resistant S. typhi has been reported from several tropical countries but not from Ghana.3 We hope that our findings help influence health planners to alter current therapeutic strategies in view of the local patterns of resistance.
Table 1. Percentage of Resistant Isolates of the Most Prevalent Bacteria Causing Fever of Unknown Origin in Ghana.
Ortrud Zimmermann
University of G?ttingen
D-37075 G?ttingen, Germany
Ring de Ciman, M.D.
St. Francis Xavier Hospital
Assin Foso, Ghana
Uwe Gross, M.D.
University of G?ttingen
D-37075 G?ttingen, Germany
ugross@gwdg.de
References
Shears P. Antibiotic resistance in the tropics: epidemiology and surveillance of antimicrobial resistance in the tropics. Trans R Soc Trop Med Hyg 2001;95:127-130.
Sinha A, Sazawal S, Kumar R, et al. Typhoid fever in children aged less than 5 years. Lancet 1999;354:734-737.
Parry CM, Hien TT, Dougan G, White NJ, Farrar JJ. Typhoid fever. N Engl J Med 2002;347:1770-1782.
The authors reply: We thank Dr. Levy and also Zimmermann and colleagues for their correspondence. The principal objective of our report in the Journal was to estimate the hospital burden of disease and death due to childhood bacteremia in a rural African setting — an area where few data exist. There is also a considerable lack of data on antimicrobial susceptibilities in sub-Saharan Africa. Because most deaths from bacteremic illnesses occur very early after admission, it is important to define appropriate empirical treatment for the first medical contact. In our study, evidence of recent use of antibiotics (antibiotic activity in plasma) was correlated with disease severity and death; we interpret this to indicate that sicker patients are more likely to have received antibiotics recently, rather than to indicate that antibiotics are failing.
The Ghanaian study investigated a clinical subgroup with fever of unknown origin and was not confined to children. This group of patients is not directly comparable to our study population. Furthermore, in Kenya, we found that axillary temperature at admission was a very poor predictor of bacteremia, even among patients without malaria. The receiver-operating-characteristic curve shows detection of bacteremia at different thresholds of axillary temperature among 16,570 unselected pediatric patients (Figure 1). Had we targeted children with an axillary temperature of 39°C or more, then 769 of 1094 children with bacteremia (70 percent) would have been missed.
Figure 1. Receiver-Operating-Characteristic Curves for Axillary Temperature for the Prediction of Bacteremia among Patients with and without Malarial Parasites at a Kenyan Hospital.
Sensitivity and 1 minus specificity are shown at various thresholds of axillary temperature (39°C, 38°C, and 37°C). The blue line denotes admissions of patients without parasitemia (9378 patients; area under the curve, 0.58 [95 percent confidence interval, 0.56 to 0.60). The red line denotes admissions of patients with parasitemia (7051 patients; area under the curve, 0.48 [95 percent confidence interval, 0.44 to 0.51]). The area under the curve for all admissions (16,570 pediatric patients) was 0.52 (95 percent confidence interval, 0.50 to 0.54; not shown). A reference line is also shown.
Different age groups and different inclusion criteria will yield different bacterial pathogens. Salmonella species are very common among adults in Kenya, as they are in Ghana, and antimicrobial susceptibilities have recently been reported.1,2 The use of a conjugate Haemophilus influenzae type b vaccine since 2002 may explain an absence of H. influenzae, but the failure of the study by Zimmermann et al. to isolate Streptococcus pneumoniae is remarkable, since this fastidious organism is prevalent in Ghana3 and elsewhere in West Africa. It is possible that the undefined clinical selection criteria may have especially favored salmonella. However, without details of microbiologic methods or quality standards, interpretation of the findings is not possible. Standardization of clinical definitions4 and laboratory procedures5 through networks and use of longitudinal data sets is essential to the development of appropriate antimicrobial policies.
James A. Berkley, M.D.
Brett S. Lowe, M.Phil.
J. Anthony G. Scott, M.R.C.P.
Centre for Geographic Medicine Research (Coast)
Kilifi, Kenya
jberkley@kilifi.mimcom.net
References
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.
Kariuki S, Revanthi G, Muyodi J, et al. Characterization of multidrug-resistant typhoid outbreaks in Kenya. J Clin Microbiol 2004;42:1477-1482.
Denno DM, Frimpong E, Gregory M, Steele RW. Nasopharyngeal carriage and susceptibility patterns of Streptococcus pneumoniae in Kumasi, Ghana. West Afr J Med 2002;21:233-236.
Management of the child with a serious infection or severe malnutrition: guidelines for care at the first referral level in developing countries. Geneva: World Health Organization, 2000.
Yeo SF, Akalin E, Arikan S, et al. Susceptibility testing of Haemophilus influenzae: an international collaborative study in quality assessment. J Antimicrob Chemother 1996;38:363-386.