Serum ferritin levels in West Nile encephalitis

RESEARCH NOTE

The reproducibility of an enzyme-linked immunosorbent assay for detection of Chlamydia pneumoniae-specific antibodies J. Ngeh1, S. Gupta2 and C. Goodbourn3 1

Department of Medicine for Elderly People, Department of Cardiology and 3Department of Microbiology, Whipps Cross University Hospital, London, UK

2

ABSTRACT With use of an enzyme-linked immunosorbent assay (ELISA), Chlamydia pneumoniae immunoglobulins were detected in a consecutive series of patients’ sera (n ¼ 122 for IgA and IgG; n ¼ 138 for IgM). When the ELISA tests were repeated, the percentage disagreements were 12%, 16% and 10% for C. pneumoniae IgA, IgG and IgM, respectively. The reproducibility of the ELISA, expressed as kappa values, for IgA, IgG and IgM was 0.73, 0.60 and 0.53, respectively (p < 0.001). It was concluded that the ELISA had good reproducibility for detecting C. pneumoniae IgA, and moderately good reproducibility for detecting C. pneumoniae IgG and IgM. Keywords

Chlamydia pneumoniae, ELISA, repro-

ducibility Original Submission: 22 September 2002; Revised Submission: 31 December 2002; Accepted: 11

February 2003 Clin Microbiol Infect 2004; 10: 171–174

The association between Chlamydia pneumoniae and atherosclerotic vascular diseases has been investigated extensively by means of seroepidemiological observations, pathological specimen examinations, animal models, immunological and molecular studies, and antibiotic intervention studies [1], but seroepidemiology was the first and most commonly used method. Most seroepidemiological studies have used microimmunofluorescence (MIF), although a few have Corresponding author and reprint requests: J. Ngeh, Department of Clinical Geratology, Radcliffe Infirmary, Oxford OX2 6HE, UK Tel ⁄ Fax: + 44 1235 811637 E-mail: [email protected]

employed enzyme-linked immunosorbent assays (ELISAs). In this study, the reproducibility of ELISA tests for detection of C. pneumoniae antibodies was investigated in a case-control study on the seroprevalence of C. pneumoniae in elderly stroke patients and general medical patients. Sera were obtained from a consecutive series of 187 elderly acute stroke patients and control general medical patients (median age 80 years) recruited prospectively [2] at a large district general hospital in East London, UK. Sera were stored at ) 20
C before analysis. Commercial ELISA kits (SeroCP; Savyon Diagnostics, Ashdod, Israel) were used. The antigens in these kits originated from purified C. pneumoniae elementary bodies (TW 183), to enable specific and sensitive detection of C. pneumoniae IgA, IgG and IgM. Sera for IgM testing were pretreated with the manufacturer’s serum diluent, containing anti-IgG, to remove rheumatoid factor and reduce IgG interference. Each serum was diluted 1:105 with the supplied diluent, as recommended by the manufacturer. The case or control status of the sera was blinded during analyses performed by two experienced investigators. Only the first 122 consecutive patients’ sera were examined for C. pneumoniae IgA and IgG, and 138 for IgM. The serological tests were then repeated, and the results analysed to establish the reproducibility. Determination and validation of test results were in accordance with the manufacturer’s instructions. Absorbance or optical density (OD), proportional to the amount of specific antibodies bound to the coated antigens in each plate, was read at 450 nm. For a valid test, two criteria were met: (1) OD450 of the positive control ‡ 0.8; and (2) mean OD450 of the negative control 0.1–0.4. In order to normalise the results from different tests, a cut-off index (COI) was calculated, in which COI ¼ serum sample OD450 ‚ cut-off value (COV), and where COV ¼ twice the mean of two negative control OD450 readings. As defined by the manufacturer, a COI of < 1.0 was negative, a COI of 1–1.1 was borderline, and a COI of > 1.1 was positive. The manufacturer suggests that if borderline results are obtained from two specimens taken from the same patient 2–4 weeks apart, the specimens should be considered negative. However, in the present study, second serum samples were not collected, and

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases

172 Clinical Microbiology and Infection, Volume 10 Number 2, February 2004

IgA

IgG

Table 1a. Cut-off index (COI): samples with borderline COI in the first run with repeat results

IgM

Sample 1st run 2nd run Sample 1st run 2nd run Sample 1st run 2nd run 37 39 56 74 99 104

1.05 1.09 1.04 1.09 1.10 1.05

0.46 0.83 0.69 1.17 0.50 1.17

Mean

1.07

0.80

107 115

IgA

Sample 10 78 110 Mean Overall Mean

1.08 1.08

2.12 1.70

1.08

1.91

17 25 30 45 54 60 83 114 119

1.00 1.10 1.08 1.09 1.04 1.04 1.06 1.03 1.00 1.05

IgG

2nd run

1st run

1.02 1.08 1.03 1.04

1.85 1.27 1.17 1.43

1.06

1.01

Table 1b. Cut-off index (COI): samples with borderline COI in the second run with first-run results

IgM

Sample

2nd run

1st run

80

1.07

1.85

1.08

1.89

IgA

0.99 0.45 0.53 0.63 0.92 0.97 1.04 1.23 1.57 0.93

Sample 83 136 138

IgG

2nd run

1st run

1.04 1.02 1.01 1.02

1.06 0.71 0.54 0.77

1.04

0.89

Table 1c. Cut-off index: samples giving discrepant results in the first and second runs

IgM

Sample 1st run 2nd run Sample 1st run 2nd run Sample 1st run 2nd run 10 17 28 44 74 78 80 91 95 96 98 104 110 117 122

1.85 1.37 1.51 1.15 1.09 1.27 1.67 1.15 1.40 2.64 1.63 1.05 1.17 0.96 0.89

1.02 0.71 0.75 0.40 1.17 1.08 0.98 0.66 0.89 0.95 0.71 1.17 1.03 1.16 1.81

Mean

1.39

0.97

5 17 21 47 48 50 56 65 66 80 90 95 97 102 106 107 113 115 117 122

0.93 1.47 1.44 2.04 0.83 3.56 0.80 1.13 4.13 1.85 2.71 0.71 0.60 0.80 0.76 1.08 0.96 1.08 0.92 0.89 1.43

1.29 0.79 0.90 0.68 1.28 0.80 1.96 0.59 0.39 1.07 0.78 1.43 1.44 1.17 1.25 2.12 1.101 1.70 1.38 2.12 1.21

2 3 6 8 24 27 29 31 37 41 42 98 114 119

therefore a single borderline result was defined as negative. The COI of positive samples ranged from 1.15 to 6.16 for C. pneumoniae IgA, from 1.12 to 4.91 for

1.22 1.21 1.58 1.23 1.12 1.29 2.52 1.93 1.25 1.19 1.25 0.64 1.03 1.00

0.46 0.55 0.60 0.64 0.78 0.61 0.91 0.70 0.65 0.66 0.66 1.20 1.23 1.57

1.32

0.80

IgG, and from 1.12 to 5 for IgM. In the repeat tests, the COI of positive samples ranged from 1.14 to 5.23 for IgA, from 1.101 to 3.45 for IgG, and from 1.20 to 3.49 for IgM. Table 1 shows the COI values

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

Research Note 173

of samples giving borderline or discrepant results. The ELISA reproducibility results are summarised in Table 2. Serological testing is considered the most useful means of determining the prevalence of C. pneumoniae infection [3]. MIF is currently the standard in C. pneumoniae serology, but is subjective and requires an expert microscopist to interpret the results [3,4]. Inter-laboratory variation of MIF shows an overall agreement with reference standard titres of c. 80% [5]. In comparison, ELISA is more objective than MIF [3]. It can be automated, with the advantages of high throughput and electronic records. ELISA is therefore easier to standardise and is the preferred diagnostic method [4]. However, although there is good agreement between serological tests and detection of C. pneumoniae by PCR or culture [6], others have reported a poor correlation between C. pneumoniae serology and culture [7]. In the clinical setting, interpretation of serological results should be based on a combination of C. pneumoniae IgA, IgG and IgM profiles. Positive IgM, negative or positive IgG, and negative or positive IgA results indicate current infection. Positive IgG, and negative IgM and IgA results indicate past or current infection. Positive IgA, with positive or negative IgG and negative IgM results indicate current or chronic infection. However, samples obtained early during primary infection may not contain detectable antibodies. If C. pneumoniae infection is suspected, a second sample should be obtained 2–4 weeks later and tested in parallel with the original sample. However, all clinical and laboratory data should be considered when making a diagnosis. The SeroCP ELISA result is qualitative, and is reported on the basis of a single 1:105 dilution of a

serum sample, and not as an endpoint titre. According to the manufacturer, overall agreement between in-house MIF and SeroCP ELISA is 98% for IgA, 95% for IgG, and 89% for IgM. Others have also reported a moderately good correlation between MIF results and ELISA tests (r0.8, p0.001) [4,8]. Similarly, as in other studies [9,10], the sensitivity and specificity compared to MIF were reported to be c. 90% by the manufacturer. However, ELISA reproducibility needs proper evaluation, and this was the focus of the current study. A previous study reported good reproducibility for the Labsystem ELISA kit for detection of C. pneumoniae IgG, with mean and median coefficients of variation of 10.2% and 8.6%, respectively [8]. The present study was the first to evaluate the reproducibility of the SeroCP ELISA, and the results in Table 2 demonstrated a good sample agreement of > 80%, similar to that reported for MIF [5]. The kappa value quantifies the agreement between the first and repeated ELISA tests, after allowing for random variation and the number of observations used [11]. A kappa value of 0 indicates no agreement, whereas 1.0 indicates perfect agreement. If 0.61 < kappa < 0.80, good concordance is indicated, while 0.41 < kappa < 0.60 indicates moderate concordance. The kappa values for the SeroCP ELISA were 0.73 for IgA, 0.60 for IgG, and 0.53 for IgM (p < 0.001). It was concluded that the SeroCP ELISA has good reproducibility for IgA, and moderately good reproducibility for IgG and IgM. ELISA may become a preferred objective test in the seroepidemiological study of C pneumoniae infection and its link with atherosclerotic vascular disease. Whether the results obtained by ELISA correlate with endovascular chronic infec-

Table 2. ELISA reproducibility Chlamydia pneumoniae antibody IgA (n ¼ 122) IgG (n ¼ 122) IgM (n ¼ 138)

Both tests positive

One test negative or borderline, one positive

36

71

15

25

77

20

114

10

14

Both tests negative

% sample disagreement (95% CI)

Kappa value

12% (6–18%) 16% (9–23%) 10% (5–15%)

0.73 (p < 0.001) 0.60 (p < 0.001) 0.53 (p < 0.001)

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

174 Clinical Microbiology and Infection, Volume 10 Number 2, February 2004

tion remains controversial. However, further efforts to standardise new commercially available ELISA kits against MIF will be required [3,12]. Although the SeroCP ELISA is comparable to MIF, its level of reproducibility is not perfect, and this fact should be considered when such a kit is used in further epidemiological or diagnostic studies.

11. Siegel S, Castellan NJ. Nonparametric statistics for the behavioural sciences. New York: McGraw-Hill International Editions, 1988. 12. Tuuminen T, Palomaki P, Paavonen J. The use of serologic tests for the diagnosis of chlamydial infections. J Microbiol Methods 2000; 42: 265–279.

RESEARCH NOTE

ACKNOWLEDGEMENTS We thank Savyon Diagnostics Ltd, Israel, for supplying the SeroCP ELISA kits used in this study. We also thank the hospital’s Chief Medical Laboratory Scientific Officer, Mrs Joyce Honeycombe, for C. pneumoniae serology, and Mr Allan Hackshaw of London University for statistical analyses.

Similar inflammatory response in human whole blood to live Streptococcus pneumoniae of different serotypes P. Kragsbjerg1, M. Jurstrand2 and H. Fredlund2 1

REFERENCES 1. Ngeh J, Anand V, Gupta S. Chlamydia pneumoniae and atherosclerosis—what we know and what we don’t. Clin Microbiol Infect 2002; 8: 2–13. 2. Ngeh J, Gupta S, Goodbourn C, Panayiotou B, McElligott G. Seroprevalence of Chlamydia pneumoniae in elderly stroke and medical patients: a case-control study. Age Ageing 2001; 30(suppl 2): S60. 3. Dowell SF, Peeling RW, Boman J et al. Standardizing Chlamydia pneumoniae assays: recommendations from the Centers for Disease Control and Prevention (USA) and the Laboratory Centre for Disease Control (Canada). Clin Infect Dis 2001; 33: 492–503. 4. Ossewaarde JM, Tuuminen T, Boersma WG, Sandstrom M, Palomaki P, Boman J. A preliminary evaluation of a new enzyme immunoassay to detect Chlamydia pneumoniae-specific antibodies. J Microbiol Methods 2000; 43: 117–125. 5. Peeling RW, Wang SP, Grayston JT et al. Chlamydia pneumoniae serology: interlaboratory variation in microimmunofluorescence assay results. J Infect Dis 2000; 181(suppl 3): S426–S429. 6. Verkooyen RP, Willemse D, Hiep-van Casteren SCAM et al. Evaluation of PCR, culture, and serology for diagnosis of Chlamydia pneumoniae. J Clin Microbiol 1998; 36: 2301–2307. 7. Boman J, Hammerschlag MR. Chlamydia pneumoniae and atherosclerosis: critical assessment of diagnostic methods and relevance to treatment studies. Clin Microbiol Rev 2002; 15: 1–20. 8. Messmer TO, Martinez J, Hassouna F et al. Comparison of two commercial microimmunofluorescence kits and an enzyme immunoassay kit for detection of serum immunoglobulin G antibodies to Chlamydia pneumoniae. Clin Diagn Lab Immunol 2001; 8: 588–592. 9. Persson K, Boman J. Comparison of five serologic tests for diagnosis of acute infections by Chlamydia pneumoniae. Clin Diagn Lab Immunol 2000; 7: 739–744. 10. Numazaki K, Ikebe T, Chiba S. Detection of serum antibodies against Chlamydia pneumoniae by ELISA. FEMS Immunol Med Microbiol 1996; 14: 179–183.

Department of Internal Medicine, Division of Haematology and 2Department of Clinical ¨ rebro University Hospital, Microbiology, O ¨ rebro, Sweden O

ABSTRACT Differences in inflammatory responses in human adult whole blood to live pneumococcal serotypes 3, 7F, 9V and 23F were investigated. Using flow cytometry and ELISA, oxidative burst, expression of activation markers CD11b ⁄ CD18, and in-vitro production of tumour necrosis factor-a, interleukin-6 (IL-6) and interleukin-8 were measured. There was no significant difference between the serotypes regarding any of the variables investigated, although there was a trend towards higher concentrations of IL-6 induced by serotypes 9V and 23F. In the present experimental model, the serotypes of Streptococcus pneumoniae shown previously to cause different degrees of inflammation were found to cause a similar inflammatory response in human whole blood. Keywords Streptococcus pneumoniae, serotypes, oxidative burst, cytokine response Original Submission: 15 March 2002; Revised Submission: 5 November 2002; Accepted: 17 Decem-

ber 2002 Clin Microbiol Infect 2004; 10: 174–177 Corresponding author and reprint requests: P. Kragsbjerg, Division of Haematology, Department of Internal Medicine, ¨ rebro University Hospital, S-701 85 O ¨ rebro, Sweden O Tel: + 46 19 6021533 E-mail: [email protected]

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

Research Note 175

Clinical infection caused by Streptococcus pneumoniae is usually characterised by intense inflammation. Patients with bacteraemic pneumococcal infection have high levels of C-reactive protein and inflammatory cytokines, and patients with pneumococcal meningitis have high concentrations of pro-inflammatory cytokines in the cerebrospinal fluid [1–3]. It is believed that a balanced inflammatory response is decisive for the outcome of clinical infection. A minority of the 90 known serotypes of S. pneumoniae cause clinical infection, but the occurrence of different serotypes varies geographically and over time [4]. Different serotypes of S. pneumoniae have been shown to cause different degrees of inflammation in the cerebrospinal fluid of rabbits [5,6] and in cultures of peripheral blood mononuclear cell cultures [7,8]. In a previous study, clinical isolates of different serotypes of S. pneumoniae were shown to induce an oxidative burst and increased expression of CD18 [9], and to induce pro-inflammatory cytokine production in whole blood [10]. In the present study, 12 different live clinical isolates belonging to four different serotypes of S. pneumoniae were studied with regard to oxidative burst, CD11b ⁄ CD18 expression and pro-inflammatory cytokine production in adult human whole blood. The aim of the study was to investigate whether different live serotypes of S. pneumoniae cause different degrees of inflammation in human whole blood. All isolates were from the blood of patients with bacteraemic pneumococcal infection and were serotyped at the Statens Serum Institut, Copenhagen, Denmark. Three strains of each of the four serotypes 3, 7F, 9V and 23F were included in the study. The strains were chosen after review of previous studies in which serotypes 3 and 7F were shown to cause mild inflammation, and serotypes 9V and 23F severe inflammation [5,6]. The strains were kept frozen at ) 70
C, thawed and grown on blood agar for 24 h to check purity, and then cultured in fastidious anaerobic broth on a shaker overnight at 37
C. After centrifugation at 2500 g for 20 min, the bacteria were washed three times and suspended in phosphate-buffered saline solution (PBS) with Ca2+ (0.9 mmol ⁄ L) and Mg2+ (0.5 mmol ⁄ L) [11]. Whole blood was obtained from healthy donors and collected in sterile tubes with sodium heparin as an additive. The concentration of polymorphonuclear leukocytes (PMNLs) was determined by

counting in a haemacytometer. In each experiment, blood from three donors with the same ABO and Rhesus blood group was used. When isolated PMNLs were used, preparation was as described previously [9,11]. Pooled serum [9,11] was used in experiments with isolated PMNLs. The concentrations of serotype-specific antibodies in the pooled serum (expressed as a percentage of international standard 89-SF) were as follows: type 3, 85%; type 7F, 27%; type 9V, 25%; and type 23F, 22% (analysed at the Statens Serum Institut). Heat-inactivation of complement was done as described previously [11]. Anti-pneumococcal antibodies were absorbed by incubating live S. pneumoniae serotype 3 with pooled serum. In short, one volume of serum was mixed with one volume of bacterial suspension and kept on ice for 60 min, after which the pneumococci were centrifuged at 2500 g for 20 min. The absorbed serum was filtered through a 0.2-lm membrane filter. The haemolytic activity of the absorbed sera did not decrease by more than one two-fold titre step in comparison with the complement titre of unabsorbed serum. Flow cytometric analyses (fluorescence activated cell sorting) were performed as described previously [9]. In total, 5000 PMNLs were analysed ⁄ tube. The fluorescence distribution was displayed as a single histogram and the mean fluorescence channel was used for further calculations. The PMNL oxidative burst response was determined using hydroethidine (Molecular Probes, Leiden, The Netherlands), which becomes oxidised to ethidium bromide when reacting with oxidative metabolites within the PMNLs. Staining of PMNLs with hydroethidine was performed as described previously [9]. When expression of CD18 and CD11b was studied, 10 lL of each monoclonal antibody was added to the appropriate tubes, and these were held for 60 min on ice. The samples were then lysed and fixed with 2 mL of lysing solution (Becton Dickinson Immunocytometric Systems, San Jose, CA, USA) at room temperature for 10 min. After centrifugation (400 g, 5 min) and removal of the supernatant, the pellet was resuspended in 2 mL of PBS with formaldehyde 1% v ⁄ v, followed by centrifugation and resuspension in 0.4 mL of PBS with formaldehyde 1% v ⁄ v. The tubes were then stored at 4
C for a maximum of 18 h until analysis.

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

176 Clinical Microbiology and Infection, Volume 10 Number 2, February 2004

Interleukin 6 (pg/mL)

10000 7500 5000 2500 0 Lo

Hi

Neg

Pos 1ng Pos 0.1ng

Fig. 1. The in-vitro production of IL-6 in human adult whole blood after incubation for 4 h at 37
C with live Streptococcus pneumoniae. Lo ¼ serotypes 3 and 7F; Hi ¼ serotypes 9V and 23F. Neg ¼ negative control, i.e., whole blood incubated without S. pneumoniae; Pos 1 ng and 0.1 ng ¼ positive controls, i.e., whole blood incubated with lipopolysaccharide 1 ng and 0.1 ng.

%

To study the effects of complement and antibody, the pneumococci (60 lL of an appropriate concentration) were opsonised with 40 lL of either pooled serum, heat-inactivated pooled serum, or pooled serum absorbed with S. pneumoniae as described above. Opsonisation was performed at 37
C for 15 min with agitation, after which 20 lL was added to the isolated PMNLs (at a ratio of five bacteria to one PMNL), and fluorescence activated cell sorting analysis was performed as described above. In-vitro experiments were also performed as described previously [10]. Live S. pneumoniae were added to whole blood at a ratio of five bacteria ⁄ monocyte. Whole blood (900 lL) was mixed with 100 lL of bacterial suspension and incubated for 4 h at 37
C in sterile 12 · 75 mm glass tubes. The tubes were then put into icy water and immediately centrifuged; the plasma was separated and stored at – 20
C for later analysis. Two negative controls (PBS and whole blood) and four positive controls (whole blood plus either 100 pg or 1 ng of lipopolysacharide (Sigma-Aldrich, Deisenhofen, Germany)) were included. Cytokine analysis was performed using commercially available tumour necrosis factor-a (TNF-a), interleukin6 (IL-6) and interkeukin-8 (IL-8) ELISA kits (R&D Systems Europe, Abingdon, UK). One-way ANOVA with a Bonferroni post test was performed using GraphPad Prism version 3.00 for Windows (GraphPad Software, San Diego, CA, ¨ rebro USA). The study was approved by the O County Council Ethics Committee. All serotypes tested caused an increase in oxidative burst and in expression of CD11b ⁄ CD18, but there was no significant difference between the four different serotypes. All serotypes also induced production of IL-6, IL-8 and TNF-a in whole blood. Production of IL-8 and TNF-a showed no difference between the serotypes, but there was a trend towards higher concentrations of IL-6 caused by the serotypes 9V and 23F, shown previously to cause severe inflammation (p0.054) (Fig. 1). When isolated neutrophils were used, the presence of complement was shown to be a prerequisite for an oxidative burst. In the presence of complement, even low levels of serotype-specific antibodies were sufficient for a respiratory burst to take place. As shown in Fig. 2, a major reduction in oxidative burst was seen when complement was heat-inactivated. Absorption of serotype 3 anti-

100 90 80 70 60 50 40 30 20 10 0 PMNL

SP + PMNL

SP + Po

SP + Pi

Fig. 2. The neutrophil oxidative burst response to live Streptococcus pneumoniae using human pooled serum with known concentrations of complement and serotype-specific immunoglobulin. The results are shown as a mean percentage (+ SE) of the maximal oxidative response in four experiments. PMNL, polymorphonuclear leukocytes; SP, S. pneumoniae; Po, pooled serum; Pi, pooled serum, heat inactivated.

bodies caused only a minor reduction in oxidative burst from 781 to 603 (mean channel fluorescence) in the presence of normal complement activity. When complement was heat-inactivated, the oxidative burst response was reduced to 460. The experimental conditions in the present investigation were as similar as possible to the intravascular situation in patients with bacteraemic pneumococcal infection. Previous studies of the inflammatory response showed that serotypes 5 and 7F caused mild inflammation, and serotypes 6B, 14 and 23F severe inflammation, when the live pneumococcal strains were injected directly into the cisterna magna of rabbits [6]. It was suggested that different serotypes of S. pneumoniae possess individual virulence determinants causing different degrees of inflammation. Tauber et al. [5] found three levels of inflammation using the same

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

Research Note 177

experimental model, with serotype 9 causing severe, serotype 3 mild, and serotype 1 intermediate inflammation [5]. Clearly, the experimental setting of the present study differs greatly from that of the rabbit meningitis model, and it is also possible that the difference in timing of measurement may account for the differences in results. When the relative importance of complement and serotype-specific antibodies was evaluated, complement was shown to be a prerequisite for an oxidative burst, whereas even low levels of serotype-specific antibody may be sufficient to produce an oxidative burst in neutrophils. In conclusion, the inflammatory response in human adult blood caused by the four different serotypes of S. pneumoniae was similar with regard to oxidative burst, neutrophil expression of CD11b ⁄ CD18, and production of pro-inflammatory cytokines such as TNF-a, IL-6 and IL-8, although there was a trend towards higher concentrations of IL-6 induced by the serotypes shown previously to induce severe inflammation (9V and 23F). Thus, the results of the present study do not confirm previous suggestions that different serotypes of S. pneumoniae can cause different degrees of inflammatory response in humans. REFERENCES 1. Kragsbjerg P, Holmberg H, Vikerfors T. Dynamics of blood cytokine concentrations in patients with bacteremic infections. Scand J Infect Dis 1996; 28: 391–398. 2. Kragsbjerg P, Jones I, Vikerfors T, Holmberg H. Diagnostic value of blood cytokine concentrations in acute pneumonia. Thorax 1995; 50: 1253–1257. 3. Glimaker M, Kragsbjerg P, Forsgren M, Olcen P. Tumor necrosis factor-alpha (TNF alpha) in cerebrospinal fluid from patients with meningitis of different etiologies: high levels of TNF alpha indicate bacterial meningitis. J Infect Dis 1993; 167: 882–889. 4. Scott JA, Hall AJ, Dagan R et al. Serogroup-specific epidemiology of Streptococcus pneumoniae: associations with age, sex, and geography in 7,000 episodes of invasive disease. Clin Infect Dis 1996; 22: 973–981. 5. Tauber MG, Burroughs M, Niemoller UM, Kuster H, Borschberg U, Tuomanen E. Differences of pathophysiology in experimental meningitis caused by three strains of Streptococcus pneumoniae. J Infect Dis 1991; 163: 806–811. 6. Engelhard D, Pomeranz S, Gallily R, Strauss N, Tuomanen E. Serotype-related differences in inflammatory response to Streptococcus pneumoniae in experimental meningitis. J Infect Dis 1997; 175: 979–982. 7. Arva E, Andersson B. Kinetics of cytokine release and expression of lymphocyte cell-surface activation markers after in vitro stimulation of human peripheral blood mononuclear cells with Streptococcus pneumoniae. Scand J Immunol 1999; 49: 237–243.

8. Arva E, Andersson B. Induction of phagocyte-stimulating and Th1-promoting cytokines by in vitro stimulation of human peripheral blood mononuclear cells with Streptococcus pneumoniae. Scand J Immunol 1999; 49: 417–423. 9. Kragsbjerg P, Fredlund H. The effects of live Streptococcus pneumoniae and tumor necrosis factor- alpha on neutrophil oxidative burst and beta2-integrin expression. Clin Microbiol Infect 2001; 7: 125–129. 10. Kragsbjerg P, So¨derquist B, Holmberg H, Vikerfors T, Danielsson D. Production of tumor necrosis factor-a and interleukin-6 in whole blood stimulated by live Gramnegative and Gram-positive bacteria. Clin Microbiol Infect 1998; 4: 129–134. 11. Fredlund H, Olcen P, Danielsson D. A reference procedure to study chemiluminescence induced in polymorphonuclear leukocytes by Neisseria meningitidis. APMIS 1988; 96: 941–949.

RESEARCH NOTE

Characterisation of invasive pneumococcal isolates in Catalan children up to 5 years of age, 1989–2000 C. Latorre, A. Gene´, T. Juncosa, C. Mun˜oz-Almagro and A. Gonza´lez-Cuevas Microbiology Service, Hospital Sant Joan de De´u, Passeig Sant Joan de De´u 2, 08950 Esplugues, Barcelona, Spain

ABSTRACT Ninety-six Streptococcus pneumoniae strains isolated between January 1989 and December 2000 from usually sterile sites of children aged < 5 years of age were included in the study. Resistance to penicillin (38.6% intermediate, 10.4% high-level), cefotaxime (20.8%), tetracycline (41.7%), chloramphenicol (33.3%) and erythromycin (27.1%), as well as serogroup ⁄ type, were related to age and pathology. Strains from children aged < 2 years showed the highest penicillin resistance rate. Resistance to penicillin, tetracycline, chloramphenicol and erythromycin was the most common pattern (18.8% of strains). Most isolates (80.2%) belonged to seroCorresponding author and reprint requests: C. Latorre, Microbiology Service, Hospital Sant Joan de De´u, Passeig Sant Joan de De´u 2, 08950 Esplugues, Barcelona, Spain Tel: + 34 93 253 21 07 Fax: + 34 93 280 36 E-mail: [email protected]

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

178 Clinical Microbiology and Infection, Volume 10 Number 2, February 2004 groups ⁄ types included in the heptavalent conjugate vaccine. Keywords Antibiotic resistance, invasive disease, paediatrics, serogroups, Streptococcus pneumoniae Original Submission: 7 February 2002; Revised Submission: 25 July 2002; Accepted: 28 July 2002

Clin Microbiol Infect 2004; 10: 177–181 Drug-resistant Streptococcus pneumoniae strains continue to spread worldwide [1], mainly among infants and young children. Some countries, namely South Africa, Spain and Hungary, have higher resistance rates than others [2], which may result in a therapeutic problem. The distribution of serotypes depends on the country [3], and some serotypes are more prevalent in young children than in adults [4]. The recent introduction of heptavalent conjugate pneumococcal vaccine makes it possible to prevent the disease in children aged < 2 years old, so epidemiological knowledge of both antibiotic resistance and serotypes is of interest for our country and others in which this vaccine is not included in current vaccination programmes. Ninety-six pneumococcal strains (those that survived freezing) isolated from usually sterile sites of children aged £ 5 years during the 12-year period from January 1989 to December 2000 were included in the present study. All strains were frozen at ) 70
C in skimmed milk after identification (Gram’s stain and optochin growth inhibition) and antimicrobial susceptibility testing, and thawed later for further susceptibility and typing studies. All strains were screened initially for susceptibility to penicillin (1-lg oxacillin disk), chloramphenicol and erythromycin (Kirby–Bauer technique) on Mueller–Hinton agar plates supplemented with sheep blood 5% v/v and incubated in CO2 5% v/v. Later, MICs of penicillin, cefotaxime, chloramphenicol, tetracycline and erythromycin were determined by the agar dilution technique, according to the criteria of the National Committee for Clinical Laboratory Standards (NCCLS) [5]. S. pneumoniae ATCCR 49619 was used as a control. Serotyping was carried out by the Quellung reaction, with the use of 46 antisera provided by the Statens Serum Institut (Copenhagen, Denmark). Both MICs and serotypes were determined annually at the National Pneumococcus Reference Centre.

Statistical analysis of all data was based on chisquare tests. A difference between groups was considered to have significance when p was < 0.05, with one degree of freedom and risk a was 0.05. Sixty-one (63.5%) strains were isolated from children aged < 2 years. The focus or the main site of the invasive infection was respiratory in 51 cases (31 aged < 2 years), meningeal in 20 (11 aged < 2 years), osteoarticular in ten (seven aged < 2 years), cutaneous in four (two aged < 2 years), and unknown in 11 (ten aged < 2 years). As shown in Table 1, 47 (48.9%) strains were non-susceptible to penicillin. Age was not statistically significant with regard to degree of resistance, although a high proportion of children aged < 2 years had diminished penicillin susceptibility compared to those aged 2–5 years (54.1% vs. 40%). Isolates causing respiratory pathology were those with the lowest resistance rate (38%), while osteoarticular isolates had the highest resistance rate (90%); this difference was statistically significant (p < 0.005). Nineteen strains had a cefotaxime MIC of 1 mg ⁄ L, one strain had an MIC of 2 mg ⁄ L, and the remaining 76 strains were susceptible to this b-lactam. The cefotaxime MIC was equal to or one dilution below that of penicillin, except in six strains, whose cefotaxime MIC was one dilution above that of penicillin. Table 1. Resistance patterns according to penicillin MICs of non-susceptible strains No. of strains with penicillin MIC (mg ⁄ L) No. of strains P T C E PT PC TC TE PCE PTC PTE PCTE S

14 3 2 1 1 1 4 3 1 8 4 18 36

0.125

0.25

0.5

1

2

5

6

3

1 1

2 1

2 1 2

5

1 4 1 5

2 5

P, resistant to penicillin; T, resistant to tetracycline; C, resistant to chloramphenicol; E, resistant to erythromycin; S, susceptible to all antibiotics tested.

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

Research Note 179

Rates of resistance to tetracycline (41.7%), chloramphenicol (33.3%) and erythromycin (27.1%) were lower than those for penicillin. The erythromycin MIC for resistant strains was ‡ 128 mg ⁄ L for all except two strains, which had erythromycin MICs of 8 mg ⁄ L and 64 mg ⁄ L. There were 12 different resistance patterns, seven of them including penicillin, and 62.5% of strains were resistant to one or more antibiotics (70.2% of these had some degree of penicillin resistance). Resistance to penicillin, tetracycline, chloramphenicol and erythromycin (PTCE) was present in 18.8% of strains. Apparently, the degree of penicillin resistance, in terms of MIC, was not related to multiresistance. Nineteen serotypes distributed into 15 serogroups were present in our series (Table 2). Seventy-seven (80.2%) of the 96 strains, including 100% of those with some degree of penicillin resistance and 83.3% of those isolated from children aged < 2 years, belonged to serotypes related antigenically to those included in the heptavalent conjugate vaccine. Resistant strains were restricted to five serogroups, all included in the seven-valent vaccine. More than half (57.9%) of the 19 strains whose serotype was not included in the seven-valent vaccine belonged to serotypes 1 and 5, and 26.3% of non-vaccine strains (serotypes 5, 7, 12 and 24) had some degree of resistance to penicillin, tetracyclines, or chloramphenicol; they came from children aged < 2 years, and were agents of respiratory infections, except for one cerebrospinal fluid isolate. One of the risk factors for an increase in penicillin-resistant and multiresistant S. pneumoniae is age [6], and the high rate of antibiotic resistance of pneumococci obtained from our paediatric patients has been described previously [7–9]. The source of isolates and the type of infection are also related to the degree of antibiotic susceptibility [10], with strains producing localised infections expected to be more resistant than invasive strains. The prevalence of intermediate resistance is also recognised generally [11], as is the high percentage of resistant strains in children aged < 2 years [12]. Our finding of lower resistance rates in isolates causing respiratory pathology with respect to other invasive isolates contrasts with the idea of higher resistance in local respiratory infections [11]. The ability of serogroup 18 and other minor susceptible serotypes to produce generalised respiratory infections from a local focus could explain this situation.

Although found in the USA [13], highly cefotaxime-resistant strains with intermediate resistance to penicillin do not constitute a major problem in Spain [14]. Spain was considered to be the focus of chloramphenicol resistance [15], with rates > 40%, a percentage that diminished during the 1990s, especially in blood strains [10]. Our incidence of 33.3% is still higher than that of countries such as the USA, where chloramphenicol resistance has an incidence of 8.3% [11]. Tetracycline resistance varies over time and in different parts of the world [12,15] as a result of different patterns of drug use [16]. The data presented here show a high percentage of resistance, but this is clearly lower than reported in previous publications [7,8]. In contrast with North America [17], erythromycin resistance of pneumococci in Spain is caused mainly by ribosomal changes. This is suggested by the fact that all but one of our isolates had erythromycin MICs ‡ 64 mg ⁄ L. Some studies in Spain [10,12] quote a higher prevalence of erythromycin resistance in children than in adults, as well as a growing resistance trend, probably associated with trends in antibiotic use. The multiresistance pattern PTCE is a major cause of concern, mainly because of its continuous increase. Nearly a fifth (18.8%) of pneumococci in the series showed this resistance pattern, compared with 12.5% found in all Spanish strains received at the National Reference Centre during 1990–96 [18]. The serotype distribution of the isolates is similar to that in the USA [4] or Europe [10], but contrasts with that of oriental countries [3]. The distribution of penicillin-susceptible and -resistant strains among serotypes is also consistent with other descriptions [2,19], as is the generalised penicillin susceptibility of strains not included in the sevenvalent conjugate vaccine. The fact that all nonvaccine strains resistant to any antibiotic were isolated from children aged < 2 years suffering from respiratory infections may be a matter of concern in cases of severe respiratory infections of young infants caused by one of these serotypes, even if they were correctly vaccinated. Although our health authorities have recommended pneumococcal capsular vaccine for risk populations [20], nothing is yet stipulated regarding conjugated vaccine for infants and young children. The excellent in-vitro coverage of heptavalent conjugate vaccine suggests that its implementation in vaccination programmes could lower the incidence of invasive disease during the first 2–5 years of life in

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

1

8 4

8

1

5

5 1

8 2 1

4 7 4 1

4

6

1 6

1

4

1

1 4 1 2 2

1

2 1 4 1 1

1

2 2

1

1

2 1

3 1

3

1

2

3 1

3

1

1

1 1

1 1

1 1

1

SG ⁄ T, serogroups ⁄ serotypes; (1), serotypes studied from 1998; < 5 years, isolated from children aged < 5 years; < 2 years, strains isolated from children aged < 2 years; pen R, with penicillin MIC > 0.06 mg ⁄ L; respiratory, producing respiratory infections; articular, producing osteoarticular infections; cutaneous, producing cutaneous infections.

1

12 6

10

1

7

8 2

5 3 6 2 2 5 1 1 1 1 8 5 1 11 5 1 1

6 1 1 4 4 4

6 1 1 6 5 11 2 2 7 1 1 1 1 10 13 1 18 7 1 1

1 2 3 4 5 6A 6B 7 9 9A (1) 9V (1) 10 12 14 18 18C (1) 19 23F 24 NT

2

< 5 years < 2 years Respiratory Meningeal Osteoarticular Cutaneous < 5 years pen R < 2 years pen R Respiratory pen R Meningeal pen R Osteoarticular pen R Cutaneous pen R

SG/T

No. of strains

Table 2. Serogroup ⁄ type distribution according to age, penicillin resistance and site of infection

180 Clinical Microbiology and Infection, Volume 10 Number 2, February 2004

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

Research Note 181

our zone. According to these data, and those of other Spanish and European studies [12], a nine-valent vaccine, including serotypes 1 and 5, could be of interest in Spain, as it would increase coverage of children aged < 5 years from 80.2% to 91.7%. ACKNOWLEDGMENTS We thank Dr Asuncio´n Fenoll and all those from the Pneumococcal Reference Laboratory for their excellent work over the years.

REFERENCES 1. Appelbaum PC. Antimicrobial resistance in Streptococcus pneumoniae: an overview. Clin Infect Dis 1992; 15: 77–83. 2. Rahav G, Toledano Y, Engelhard D et al Invasive pneumococcal infections. A comparison between adults and children. Medicine 1997; 76: 295–303. 3. Saha SK, Rikitomi N, Biswas D et al. Serotypes of Streptococcus pneumoniae causing invasive childhood infections in Bangladesh, 1992–95. J Clin Microbiol 1997; 35: 785–787. 4. Orange M, Gray BM. Pneumococcal serotypes causing disease in children in Alabama. Pediatr Infect Dis J 1993; 12: 244–246. 5. National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial susceptibility testing. Eleventh Informational Supplement. M100-S11. Villanova, PA: NCCLS, 2001. 6. Appelbaum PC. Epidemiology and in vitro susceptibility of drug-resistant Streptococcus pneumoniae. Pediatr Infect Dis J 1996; 15: 932–939. 7. Latorre C, Juncosa T, Sanfeliu I. Antibiotic resistance and serotypes of 100 Streptococcus pneumoniae strains isolated in a children’s hospital in Barcelona, Spain. Antimicrob Agents Chemother 1985; 28: 357–359. 8. Latorre C, Juncosa T, Sanfeliu I. Antibiotic susceptibility of Streptococcus penumoniae isolates from paediatric patients. J Antimicrob Chemother 1988; 22: 659–665. 9. Latorre C. Streptococcus pneumoniae isolated from a pediatric population: changes in ten years. Acta Paediatr 1998; 87: 940–944. 10. Lin˜ares J, Tubau F, Domı´nguez MA. Antibiotic resistance in Streptococcus pneumoniae in Spain: an overview of the 1990s. In: Tomasz A, ed. Streptococcus pneumoniae. Molecular biology and mechanisms of disease—update for the 1990s. New York: Mary Ann Liebert, 2000: 399–407. 11. Doern GV, Heilmann KP, Huynh HK, Rhomberg PR, Coffman SL, Brueggemann AB. Antimicrobial resistance among clinical isolates of Streptococcus pneumoniae in the United States during 1999–2000, including a comparison of resistance rates since 1994–1995. Antimicrob Agents Chemother 2001; 45: 1721–1729. 12. Fenoll A, Jado I, Vicioso D, Berron S, Yuste JE, Casal J. Streptococcus pneumoniae in children in Spain: 1990–1999. Acta Paediatr Suppl 2000; 89: 44–50. 13. Mun˜oz R, Dowson CG, Daniels M et al. Genetics of resistance to third-generation cephalosporins in clinical isolates of Streptococcus pneumoniae. Mol Microbiol 1992; 6: 2461–2465.

14. Ruiz J, Sempere M, Simarro E, Fenoll A. Description of two new isolates of Streptococcus pneumoniae in Spain that are highly resistant to cefotaxime. Antimicrob Agents Chemother 1998; 42: 2768–2469. 15. Klugman KP. Pneumococcal resistance to antibiotics. Clin Microbiol Rev 1990; 3: 171–196. 16. Diekema DJ, Brueggemann AB, Doern GV. Antimicrobial drug use and changes in resistance in Streptococcus pneumoniae. Emerg Infect Dis 2000; 6: 605–614. 17. Johnston NJ, De Azavedo JC, Kellner JD, Low DE. Prevalence and characterization of the mechanisms of macrolide, lincosamide, and streptogramin resistance in isolates of Streptococcus pneumoniae. Antimicrob Agents Chemother 1998; 42: 2425–2426. 18. Fenoll A, Jado I, Vicioso D, Pe´rez A, Casal J. Evolution of Streptococcus pneumoniae serotypes and antibiotic resistance in Spain: update (1990–1996). J Clin Microbiol 1998; 36: 3447–3454. 19. Kaplan SL, Mason EO, Barson WJ et al. Three-year multicenter surveillance of systemic pneumococcal infections in children. Pediatrics 1998; 102: 538–545. 20. Salleras L, Vidal J, Bruguera M et al. Vacunaciones del adulto. Med Clin (Barc) 1994; 102(suppl): 42–55.

RESEARCH NOTE

Serological investigation of the prevalence of anti-dengue IgM and IgG antibodies in Attapeu Province, South Laos G. Peyerl-Hoffmann1, B. Schwo¨bel1, S. Jordan1, V. Vamisaveth2, R. Phetsouvanh2, E. M. Christophel3, S. Phompida2, F. V. Sonnenburg1 and T. Jelinek1,4 1

Department of Infectious Diseases and Tropical Medicine, University of Munich, Germany, 2 Center of Malariology, Parasitology and Entomology (CMPE), Ministry of Health, Vientiane, Lao PDR, 3WHO Regional Office, Vientiane, Lao PDR, 4Institute of Tropical Medicine, Berlin, Germany

ABSTRACT The prevalence of dengue antibodies was determined in the Attapeu region of South Laos with 225 blood samples collected from mostly febrile Corresponding author and reprint requests: Tomas Jelinek, Institute of Tropical Medicine, Spandauer Damm 130, 14050 Berlin, Germany E-mail: [email protected]

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

182 Clinical Microbiology and Infection, Volume 10 Number 2, February 2004

patients during the rainy season August – October 2001. An IgM capture ELISA was positive for one (0.4%) sample, while 177 (79%) samples were positive in an indirect IgG ELISA. Of the positive IgG samples, 20 (11.3%) were also positive on blood slides for Plasmodium falciparum. Dengue fever seems to be widespread in this area, but clinical dengue diagnosis remains difficult, especially in the first days of illness when physicians have to discriminate between dengue and other febrile illnesses. Keywords

Dengue fever, ELISA, IgG, IgM, Sero-

prevalence Original Submission: 12 December 2002; Revision Submission: 7 April 2003; Accepted: 24 April 2003

Clin Microbiol Infect 2004; 10: 181–184 Dengue is a mosquito-transmitted arboviral disease, with millions of cases occurring each year [1]. Dengue virus causes a spectrum of clinical manifestations such as dengue fever (DF), a self-limiting flu-like illness with very low mortality, or dengue hemorrhagic fever (DHF). Severe cases with signs of circulatory failure may develop a hypovolaemic shock, termed dengue shock syndrome (DSS), which is associated with a case-fatality rate of > 10% [2]. The dengue virus belongs to the family Flaviviridae, with infection caused by four serotypes (DEN-1–DEN-4). Infection in humans with one of the serotypes (primary infection) produces life-long immunity against reinfection with the same serotype, but does not protect against a different serotype [3]. Secondary infection with a different serotype following primary infection is associated with an increased risk of DHF. A challenging problem with regard to patient management is the rapid differential diagnosis of early symptoms in order to distinguish dengue from other diseases such as malaria. Detection of IgM and IgG antibodies is a useful tool to distinguish between primary and secondary infection. Primary infection is characterised by the presence of significant or rising levels of IgM antibodies in the period 3–5 days after onset of infection, and can persist for 3–5 months. Anti-dengue IgG levels are comparatively low during primary infection, but

secondary infection often results in the appearance of high levels of IgG before the IgM reponse. IgG levels rise quickly, peak about 2 weeks after the onset of symptoms, and then decline slowly over 3–6 months. Anti-dengue IgM levels are comparatively low during a secondary infection [3–5]. The objective of the present study was to examine serum samples for anti-dengue IgM and IgG antibodies in the Province of Attapeu in South Laos to provide basic epidemiological data on the situation of dengue infections in this rural area. From August–October 2001, sera from 225 patients, mostly with a history of fever (n ¼ 129), were collected in the Province of Attapeu in South Laos. This region is situated in the south-east part of Lao PDR, close to the border with Vietnam and Cambodia. The study was approved by the ethical committee of the University of Munich, Germany and by the Ministry of Health, Vientiane, Lao PDR. The patients were of ethnic Lao Thung or Lao Loum origin. Inclusion criteria were an age of > 1 year, fever or history of fever, and informed consent by the patient or the parents. Initially, samples from all patients with febrile symptoms were screened with a Plasmodium falciparum dipstick test (Paracheck; Orchid, Goa, India). Following a positive dipstick test, thin and thick blood smears for malaria were made, stained with Giemsa by standard procedures, and later read by experienced technicians in the hospitals. The blood samples were collected either in the District Hospital of Xaysettha or in Attapeu Provincial Hospital, allowed to coagulate for some hours at 4
C, aliquotted into tubes, frozen at ) 20
C and later transported to Germany (Department of Infectious Diseases and Tropical Medicine, University of Munich) for further testing. All sera were tested by two different ELISAs (PanBio, Brisbane, Australia) for detection of specific IgM antibodies to dengues–flaviviruses, and for detection of IgG antibodies to the four dengue serotypes. Data were stored in a Microsoft Access database, crosschecked, and then transferred into SPSS v.10 (SPSS Inc, Chicago, IL, USA) for all statistical analyses. Table 1 summarises the results obtained. The IgM capture ELISA was positive for one (0.4%)

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

Research Note 183

Table 1. Summary of results obtained for 225 serum samples with the ELISAs for dengue IgM and IgG antibodies Dengue IgG-positive (n = 177) Females (n ¼ 106) Male (n ¼ 114) Age (years, n ¼ 224) < 10 10–19 29–29 30–39 40–49 > 50 Hospital Province (Attapeu, n ¼ 130) District (Xaysettha, n ¼ 95) History of fever (n ¼ 129) No fever (< 37.5
C) Fever (> 37.5
C) P. falciparum-positive P. falciparum-negative Dengue IgM ELISA-positive Dengue IgM ELISA-negative

Dengue IgG-negative (n = 48)

78 (73.6%) 94 (82.4%)

28 (26.4%) 20 (17.6%)

3 32 47 40 29 25

3 19 11 7 6 2

(50.0%) (62.7%) (81.0%) (85.1%) (82.9%) (92.6%)

(50.0%) (37.3%) (10.0%) (14.9%) (17.1%) (7.4%)

107 (82.3%)

23 (17.7%)

70 (73.7%)

25 (26.3%)

101 (78.3%)

28 (21.7%)

120 27 20 156 1

34 12 6 42

(76.9%) (69.2%) (76.9%) (78.8%) (100.0%)

176 (78.6%)

(23.1%) (30.8%) (23.1%) (21.2%)

48 (21.4%)

sample from the group aged 10–19 years, while 177 (79%) samples were positive with the indirect IgG ELISA, which indicates a past or recent dengue infection. Of the 130 (57.8%) samples from outpatients in the Province Hospital, 107 were positive in the IgG ELISA. Of the 95 (42.2%) samples from the District Hospital, one was positive for IgM, and 70 were positive for IgG. There was no significant difference in IgG prevalence at the province (82.3%) and district (73.7%) levels. The age distribution was 3–73 years (mean 31 years), with an equivalent gender distribution (female, 48.2%; male, 51.8%). There was no significant difference in the male: female ratio between the seropositive (54.7%: 45.3%) and the seronegative (41.9%: 58.1%) samples. The prevalence of dengue IgG antibodies within the different age groups increased from 50% in the group (n ¼ 6) aged < 10 years to 92.6% in the group (n ¼ 25) aged > 50 years. There was a

strong linear association of increasing prevalence with age (Spearman Correlation 0.233, p < 0.001; ChiQuadrat 15.4, d.f. ¼ 5, p < 0.009). Unfortunately, no paired samples were available to distinguish between primary and secondary dengue infection. In the IgG-seropositive group, 101 patients reported a recent febrile illness (with a fever history of 1–7 days), but only 27 patients had fever of > 37.5
C at presentation (Table 1). Among all 177 IgG-seropositive patients, 20 were also positive for P. falciparum, compared to six of the 48 IgG-seronegative group. In this group, 28 patients reported a history of fever, while 12 patients had a temperature > 37.5
C at presentation. This seroprevalence study is, to our knowledge, the first conducted in South Laos, and demonstrated a seroprevalence of 0.4% for anti-dengue IgM and of 79% for anti-dengue IgG. High transmission rates for flaviviruses have been described in South-east Asia. Therefore serological cross-reactivity across the flavivirus group (e.g., Japanese encephalitis virus) is common at the IgG level, and the results must be treated with caution. No previous data on the dengue prevalence situation in the Lao PDR are available, but a similar study from Thailand among children aged £ 14 years showed that 32% had anti-dengue antibodies, with 7% of these having a primary and 93% a secondary infection [5]. Studies in Peru and Saudi Arabia have shown an overall prevalence for dengue IgG of 29.5% and 32%, respectively [6,7]. In the present study, a strong linear association of increasing antibody prevalence with age was detected. This may result from a relatively stable transmission rate over decades. This result is in line with previous work that also showed an agedependent increase of anti-dengue antibodies in exposed populations [6,8,9]. Other results show that dengue prevalence depends on time of exposure, on the risk of getting bitten by an infected mosquito, and thus strongly on age in endemic and in epidemic countries. Higher prevalence rates are often more common in younger age groups during and after dengue epidemics and outbreaks [10]. In endemic areas, misdiagnosis of dengue, Japanese encephalitis or malaria occurs frequently. Awareness and clinical suspicion for patients

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

184 Clinical Microbiology and Infection, Volume 10 Number 2, February 2004

with similar illness (fever, headache, vomiting, etc.) are necessary. The present survey demonstrates that dengue fever may be a problem in the region of Attapeu Province. Further detailed serological studies with inclusion of other flaviviruses are recommended.

RESEARCH NOTE

Serum ferritin levels in West Nile encephalitis B. A. Cunha, B. Sachdev and D. Canario

ACKNOWLEDGMENTS We are grateful to Bounlay Phommask (Deputy Director, Ministry of Health), the Center of Malariology, Parasitology and Entomology (Vientane, Lao PDR) and the Provincial Health Service of Attapeu province for the successful cooperation. We thank the staff of the Attapeu Provincial Hospital as well the staff in the District Hospital of Xaysettha for their assistance to collect the data and the samples. The study was financially supported in part by the Friedrich Baur Institute, Munich, and by WHO.

REFERENCES 1. Gubler DJ, Clark GG. Dengue ⁄ Dengue hemorrhagic fever: the emergence of a global health problem. Emerg Infect Dis 1995; 1: 55–57. 2. Tassniyom S, Vasanawathana S, Chirawatkul A, Rojanasuphot S. Failure of high-dose methylprednisolone in established dengue shock syndrome: a placebo-controlled, double-blind study. Pediatrics 1993; 92: 111–115. 3. World Health Organization. Dengue Haemorrhagic Fever: Diagnosis, Treatment, Prevention and Control, 2nd edn. Geneva: World Health Organisation 1997; 7–9. 4. Innis B. Antibody responses to dengue virus infection. In: Gubler, DJ, Kuno, G, eds. Dengue and Dengue Haemorrhagic Fever. New York: CAB International, 1997; 221–243. 5. Vaughn DW, Green S, Kalayanarooj S et al. Dengue in the early febrile phase: viremia and antibody responses. J Infect Dis 1997; 176: 322–330. 6. Reiskind MH, Baisley KJ, Calampa C, Sparp TW, Wats DM, Wilson ML. Epidemiological and ecological characteristics of past dengue virus infection in Santa Clara, Peru. Trop Med Int Health 2001; 6: 212–218. 7. Fakeeh M, Zaki AM. Virologic and serologic surveillance for dengue fever in Jeddah, Saudi Arabia, 1994–99. Am J Trop Med Hyg 2001; 65: 764–767. 8. Chungue E, Marche` G, Plichart R, Boutin JP, Roux J. Comparison of immunoglobulin G enzyme-linked immunosorbent assay (IgG-ELISA) and hemagglutination inhibition (HI) test for detection of dengue antibodies. Prevalence of dengue IgG-ELISA antibodies in Tahiti. Trans Roy Soc Trop Med Hyg 1989; 83: 708–711. 9. Heyes CG, Phillipsa IA, Callahan JD et al. The epidemiology of dengue virus infection among urban, jungle, and rural populations in the Amazonas region of Peru. Am J Trop Med Hyg 1996; 55: 459–463. 10. Desparis X, Roche C, Murgue E, Chungue E. Possible dengue sequential infection: dengue spread in a neighbourhood during the 1996 ⁄ 97 dengue-2 epidemic in French Polynesia. Tropmed Int Health 1998; 3: 866–871.

Infectious Disease Division, Winthrop-University Hospital, Mineola and State University of New York School of Medicine, Stony Brook, New York, USA

ABSTRACT West Nile encephalitis (WNE) presents clinically as aseptic meningitis, meningoencephalitis, encephalitis, or acute flaccid paralysis. Non-specific laboratory findings, e.g., leukopenia and thrombocytopenia, accompany WNE. Lymphopenia is marked and prolonged with WNE. Three patients with WNE were found to have elevated serum ferritin levels. Severity seemed to be directly related to serum ferritin levels. Although preliminary, the results suggested that serum ferritin levels ‡ 500 ng ⁄ mL (normal range 5– 187 ng ⁄ mL) occur late with WNE, and not in a control group of patients with viral meningitis or encephalitis. Keywords Encephalitis, ferritin, viral meningitis, West Nile encephalitis Original Submission: 4 June 2003; Revised Submission: 25 July 2003; Accepted: 4 August 2003

Clin Microbiol Infect 2004; 10: 184–186 West Nile encephalitis (WNE) first appeared in the USA in the New York area, and has subsequently spread westwards across the USA, with cases in most states [1–4]. The clinical presentation may be manifested as aseptic meningitis, meningoencephalitis, encephalitis, or acute motor paralysis. The clinical course of WNE presents a spectrum of illness ranging from mild aseptic meningitis to fatal encephalitis. WNE is a diag-

Corresponding author and reprint requests: B. A. Cunha, Infectious Disease Division, Winthrop-University Hospital, Mineola, NY 11501, USA Tel: + 1 516 663 2505 Fax: + 1 516 663 2753

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

Research Note 185

nostic possibility in patients with otherwise aseptic meningitis or encephalitis, particularly if accompanied by weakness and ⁄ or muscle tremors. The possibility of WNE is further suggested by the non-specific laboratory findings that frequently accompany this arboviral infection, e.g., leukopenia, lymphopenia and thrombocytopenia. Unlike other causes of viral meningitis or encephalitis, the lymphopenia in WNE is marked and prolonged [5]. Non-specific laboratory abnormalities have been described in cases of WNE. This report describes three patients admitted to Winthrop-University Hospital with a confirmed diagnosis of WNE in whom serum ferritin levels were elevated late in the course of infection. The three patients were confirmed as suffering from WNE on the basis of a positive WNE PCR result and elevated IgM ELISA levels obtained by the New York State Department of Health Laboratories. The patients presented with the typical clinical findings of WNE, including encephalitis, weakness and muscle tremors. The three patients also had leukopenia, thrombocytopenia and prolonged lymphopenia. Serum transaminases were mildly increased in two of the three patients. Serum ferritin levels were highly elevated in all three patients late (i.e., weeks 3–4) in the course of WNE. Peak serum ferritin levels (normal range 5–187 ng ⁄ mL) ranged between 594 ng ⁄ mL and 3077 ng ⁄ mL in the three patients: patient 1, 373–594 ng ⁄ mL (mild WNE); patient 2, 1170– 1660 ng ⁄ mL (moderate WNE); patient 3, 886– 3077 ng ⁄ mL (fatal WNE). These elevated serum ferritin levels peaked later and higher than might be expected if ferritin was an acute-phase reactant. None of the patients had disorders of iron overload associated with high serum ferritin levels, e.g., megaloblastic sideroblastic anaemias, thalassaemia, hereditary spherocytosis, porphyria, primary or metastatic carcinomas, lymphomas, leukaemias, haemosiderosis, chronic liver or renal disease, hyperthyroidism, or Gaucher’s disease. Serum ferritin levels in the WNE patients remained elevated for days to weeks and gradually decreased with time (Fig. 1). Ferritin is an iron storage protein complex of iso-ferritins, produced by the reticuloendothelial (RE) system. The RE system plays a critical role in iron metabolism by processing the haemoglobin from senescent red blood cells. In the absence of disorders that affect iron metabolism, iron equilibrium is regulated by macrophages of the RE

Fig. 1. Late elevations of serum ferritin levels with mild (patient 1), moderate (patient 2) and fatal (patient 3) West Nile encephalitis. Days on the x-axis refer to hospital days.

system which modulate iron release and iron storage. There is an inverse relationship between early release of iron from senescent red blood cells and serum ferritin levels. Acutely, inflammation and infection induce RE blockade of iron release. This has the effect of decreasing serum iron levels in inflammation, infection and neoplasia. The decrease in serum iron levels is part of the febrile response to deny serum iron, a virulence factor for many microorganisms, early in infection. The mild and early increases in serum ferritin during acute infection reflect the transfer of serum iron into RE system storage to deprive microorganisms of serum iron [6–11]. The late and sustained elevations in serum ferritin levels observed in the three WNE patients cannot be ascribed to the acute-phase response. As an acute-phase reactant, ferritin may be transiently or mildly elevated early in the course of infection [6–8]. However, the elevated ferritin levels seen in the WNE patients occurred late (during the third to fourth weeks), and were highly elevated beyond the levels attributable to an acute-phase reactant of early infection. Six control patients (age range 22–69 years) with viral meningitis or encephalitis, who presented in the same season, i.e., the summer, had normal or near-normal serum ferritin levels, i.e., range 87–389 ng ⁄ mL. The reason why serum ferritin levels are elevated in WNE is not clear. Serum ferritin levels ‡ 500 ng ⁄ mL appear to be a late non-specific laboratory feature of WNE. The magnitude of

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186

186 Clinical Microbiology and Infection, Volume 10 Number 2, February 2004

serum ferritin levels observed in the three patients seemed to be related to the severity of WNE. The results from these preliminary findings in three patients need to be confirmed in a larger cohort of WNE patients.

5.

6. 7.

REFERENCES 1. Petersen LR, Marfin AA, Gubler DJ. West Nile virus. JAMA 2003; 290: 524–528. 2. Centers for Disease Control and Prevention. West Nile virus activity United States, 2001. MMWR 2002; 23: 497– 501. 3. Cunha BA. West Nile encephalitis. Infect Dis Pract 1999; 23: 85–90. 4. Asnis DS, Conetta R, Teixeira AA, Waldman G, Sampson BA. The West Nile virus outbreak of 1999 in New York.

8. 9. 10. 11.

The Flushing Hospital experience. Clin Infect Dis 2000; 30: 413–418. Cunha BA, Minnaganti V, Johnson DH, Klein NC. Profound and prolonged lymphocytopenia with West Nile encephalitis. Clin Infect Dis 2000; 31: 1116–1117. Weinberg ED. Iron and infection. Microbiol Rev 1978; 42: 45–66. Deiss A. Iron metabolism in reticuloendothelial cells. Semin Hematol 1983; 20: 81–89. Wooldridge KG, Williams PH. Iron uptake mechanisms of pathogenic bacteria. Microbiol Rev 1993; 12: 325–348. Griffiths E. Iron and bacterial virulence: a brief overview. Biol Metals 1991; 4: 7–13. Worwood M. Ferritin. Blood Rev 1990; 4: 259–269. Krol V, Cunha BA. Diagnostic significance of serum ferritin levels in infectious and non-infectious diseases. Infect Dis Pract 2003; 27: 199–200.

 2004 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 10, 171–186