JAC
Journal of Antimicrobial Chemotherapy (2005) 56, 232–235 doi:10.1093/jac/dki145 Advance Access publication 19 May 2005
Invasive Gram-negative bacilli are frequently resistant to standard antibiotics for children admitted to hospital in Kilifi, Kenya Philip Bejon1,2, Isaiah Mwangi1, Caroline Ngetsa1, Salim Mwarumba1, James A. Berkley1, Brett S. Lowe1, Kathryn Maitland1,3, Kevin Marsh1,2, Mike English1,4 and J. Anthony G. Scott1,2* 1
Wellcome Trust/Kenya Medical Research Institute (KEMRI), Centre for Geographic Medicine Research — Coast, PO Box 230, Kilifi, Kenya; 2Nuffield Department of Clinical Medicine, Oxford University, John Radcliffe Hospital, Oxford, UK; 3Department of Academic Paediatrics, Imperial College, London, UK; 4Department of Paediatrics, University of Oxford, John Radcliffe Hospital, Oxford, UK
Objectives: To determine the pattern of resistance among Gram-negative bacilli causing invasive bacterial disease for the antibiotics that are already in common use in Kilifi, Kenya and for two potential alternatives, ciprofloxacin and cefotaxime. Also, to determine whether prevalence and severity of resistance was increasing over time, to identify patients who are particularly at risk of resistant infections, and to explore which factors are associated with the development of resistance in our setting. Methods: We used Etest to study antibiotic susceptibility patterns of 90 Gram-negative bacilli cultured in blood or CSF from paediatric inpatients over 8 years. Results: Susceptibility to amoxicillin 28%, cefotaxime 95% and ciprofloxacin 99% did not vary significantly with age. Susceptibilities for isolates from children aged less than 14 days were: chloramphenicol, 81%; trimethoprim/sulfamethoxazole, 71%; and gentamicin, 91%. From older children, susceptibilities were: chloramphenicol, 62%; trimethoprim/sulfamethoxazole, 39%; and gentamicin, 73%. Chloramphenicol susceptibility was significantly more common among non-typhi salmonellae than other species (79% versus 53%, P < 0.0005). The combination of gentamicin and chloramphenicol covered 91% of all isolates. The prevalence of resistance did not increase over time and was not more common in patients with HIV or malnutrition. Age was the only clinical feature that predicted resistance. Conclusions: Gentamicin or chloramphenicol alone was suboptimal therapy for Gram-negative sepsis, although in this retrospective study, there was no association between resistance and mortality. Keywords: antibiotic resistance, Gram-negative infections, developing countries, mortality, children
Introduction Invasive Gram-negative bacilli are associated with high mortality in Kenya and throughout the world.1 Early treatment with an antibiotic to which the organism is susceptible by international standards2 is the mainstay of therapy. In developing countries, blood and cerebrospinal fluid (CSF) cultures are infrequent, reducing the information available to guide treatment at an individual and at a public health level.
In sub-Saharan Africa, the data are patchy. Across five sites in sub-Saharan Africa, resistance to ampicillin and trimethoprim/sulfamethoxazole was almost universal, but susceptibility to chloramphenicol and sometimes gentamicin appeared preserved.3 – 9 The objective of this study was to determine the pattern of resistance among Gram-negative bacilli causing invasive bacterial disease for the antibiotics in common use in Kilifi, Kenya and for two potential alternatives; ciprofloxacin and cefotaxime.
..........................................................................................................................................................................................................................................................................................................................................................................................................................
*Corresponding author. Tel/Fax: + 254-415-25453/22390; E-mail:
[email protected] ..........................................................................................................................................................................................................................................................................................................................................................................................................................
232 q The Author 2005. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail:
[email protected]
Downloaded from http://jac.oxfordjournals.org/ by guest on June 7, 2013
Received 14 February 2005; returned 23 March 2005; revised 25 March 2005; accepted 5 April 2005
Gram-negative bacteraemia in Kenya
Patients and methods Study sample The study population comprised stored frozen isolates of Gramnegative bacilli cultured from blood or cerebrospinal fluid taken from paediatric inpatients at Kilifi District Hospital (KDH), Kenya, isolated between 1 January 1994 and 31 December 2001. KDH is a rural, government-funded hospital with 42 paediatric beds and an average of 4400 admissions per year. Before August 1998, cultures of CSF and blood were initiated on clinical suspicion of meningitis or sepsis; thereafter all acute admissions, except trauma and burns, were investigated with blood cultures. Treatment of inpatients at KDH is based on WHO guidelines. _ 60 days are treated with benzylpenicillin and gentaYoung infants < micin. Older children with suspected meningitis or shock are given
benzylpenicillin and chloramphenicol. Amoxicillin or benzylpenicillin is prescribed alone to treat mild or severe pneumonia, respectively; in those with prostration or pulse oximetry readings below 90%, chloramphenicol is substituted. Gentamicin and amoxicillin are used in malnutrition. Trimethoprim/sulfamethoxazole is the commonest outpatient antibiotic in Kilifi District, but is not used for inpatients.
Laboratory methods Before June 1998, brain heart infusion broth was used to culture blood for 7 days at 378C in 5% CO2 with sub-cultures at 1, 2 and 7 days. Subsequently, BacTec Peds Plus media (Becton Dickinson, NJ, USA) were incubated for 5 days and sub-cultured as indicated by the BacTec instrument.
Table 1. Resistance patterns of 690 invasive isolates of Gram-negative bacilli for four antibiotics determined by Etest
Bacterial group and antibiotic Escherichia coli amoxicillin chloramphenicol trimethoprim/sulfamethoxazole gentamicin ceftriaxone ciprofloxacin Enterobacter spp. amoxicillin chloramphenicol trimethoprim/sulfamethoxazole gentamicin ceftriaxone ciprofloxacin Klebsiella spp. chloramphenicol trimethoprim/sulfamethoxazole gentamicin ceftriaxone ciprofloxacin Pseudomonas aeruginosa chloramphenicol trimethoprim/sulfamethoxazole gentamicin ciprofloxacin Non-typhi salmonellae amoxicillin chloramphenicol trimethoprim/sulfamethoxazole gentamicin ceftriaxone ciprofloxacin All organisms amoxicillin chloramphenicol trimethoprim/sulfamethoxazole gentamicin ceftriaxone ciprofloxacin
Intermediate
Resistant
n
%
n
%
n
%
Total
31 81 21 118 141 139
22 57 15 84 100 100
0 1 0 0 0 0
0 1 0 0 0 0
110 59 120 23 0 2
78 42 85 16 0 0
141 141 141 141 141 141
1 30 29 35 44 45
2 67 64 78 98 0
1 0 0 0 0 0
2 0 0 0 0 0
43 15 16 10 1 0
96 33 36 22 2 0
45 45 45 45 45 45
32 30 41 57 62
51 48 65 90 98
1 1 0 2 1
2 2 0 3 2
30 32 22 4 0
48 51 35 6 0
63 63 63 63 63
1 17 52 56
2 30 91 98
2 11 0 0
4 20 0 0
54 28 5 1
95 50 9 2
57 56 57 57
129 234 145 197 296 296
44 79 49 67 100 100
3 2 1 10 0 0
1 1 0 3 0 0
164 60 150 89 0 0
55 20 51 30 0 0
296 296 296 296 296 296
192 441 302 526 653 683
28 64 44 76 95 99
6 9 13 10 27 3
1 1 2 1 4 0
492 240 374 154 10 4
71 35 54 22 1 1
690 690 689 690 690 690
Resistance to amoxicillin was not determined for Pseudomonas aeruginosa and Klebsiella species.
233
Downloaded from http://jac.oxfordjournals.org/ by guest on June 7, 2013
Susceptible
Bejon et al. Isolates were stored at 708C in tryptone soya broth with 15% glycerol for up to 8 years. Antimicrobial susceptibilities were determined by Etest in the laboratories of the manufacturer (AB Biodisk, Solna, Sweden).
Data analysis STATA 8.2 (StataCorp, College Station, TX, USA) was used. Susceptibility breakpoints were taken from the NCCLS guidelines 2003.2 x2 was used to test hypotheses of association. Logistic regression models were developed with backward stepwise regression, excluding variables when the likelihood ratio test P value _ 0.05. Isolate species and source of isolate were included was > throughout. Intermediate resistance (I) was classified within resistance (R).
Results
Discussion This study shows that, beyond the first 2 weeks of life, the antibiotics that are used most frequently to treat children with
Table 2. Logistic regression odds ratios and 95% CIs in final models of risk factors for resistance to amoxicillin, chloramphenicol, trimethoprim/sulfamethoxazole and gentamicin Antibiotic Amoxicillin Chloramphenicol Trimethoprim/sulfamethoxazole
Gentamicin
Risk factor
OR
Lower CI
Upper CI
P value
positive malaria slide positive malaria slide isolated after 1997 _ 14 days age > male sex positive malaria slide isolated after 1997 _ 14 days age > isolated after 1997 _ 14 days age >
0.51 0.57 0.58 2.79 1.41 0.46 0.69 4.37 0.63 4.88
0.34 0.35 0.41 1.72 1.00 0.29 0.48 2.82 0.42 2.40
0.78 0.94 0.83 4.53 1.98 0.72 0.98 6.77 0.93 9.91
0.002 0.025 0.0003
