Diagnosis of Streptococcus pneumoniae and Haemophilus influenzae type b meningitis by identifying DNA from cerebrospinal fluid-impregnated filter paper strips.
ABSTRACT Bacterial meningitis remains often etiologically unconfirmed, especially in resource-poor settings. We tested the potential of real-time polymerase chain reaction to identify Streptococcus pneumoniae (Pnc) and Haemophilus influenzae type b (Hib) from cerebrospinal fluid impregnated on filter paper strips.
Pnc and Hib genome equivalents were blindly quantified by polymerase chain reaction from 129 liquid cerebrospinal fluid (CSF) samples-the standard-and strips stored at room temperature for months. Genome counts were compared by simple regression.
The strips showed a sensitivity and specificity of 92% and 99% for Pnc, and of 70% and 100% for Hib, respectively. The positive and negative predictive values were 94% and 97% for Pnc, and 100% and 89% for Hib, respectively. For Pnc, the positive and negative likelihood ratio was 92 and 0.08, and the overall accuracy 98%, whereas for Hib they were 70 and 0.30, and 91%, respectively. Genome counting showed good correlation between the filter paper and liquid CSF samples, r(2) being 0.87 for Pnc and 0.68 for Hib (P < 0.0001 for both).
Although not replacing bacterial culture, filter paper strips offer an easy way to collect and store CSF samples for later bacteriology. They can also be transported in standard envelops by regular mail.
- [Show abstract] [Hide abstract]
ABSTRACT: In recent decades the treatment of childhood acute bone and joint infections has shifted towards shorter antibiotic courses and rapid transition to oral therapy. We prospectively collected 265 culture-positive cases of non-neonatal bone and joint infections in Finnish children during 1983-2005. The duration of antimicrobial treatment and the extent of surgery were defined in the study protocol, but for ethical reasons, the liaison clinician determined the time of discharge using normalization of the serum C-reactive protein (CRP) level as a yardstick. We examined changes during the study in the distribution of causative organisms, severity of disease, and length of hospital stay. Staphylococcus aureus was overwhelmingly the most common causative agent throughout the study, whereas Haemophilus influenzae type b was eliminated soon after the introduction of vaccination. The mean time from initial symptoms to presentation remained the same at 4 days, and no significant change was observed in the severity of disease, CRP, or the rate of sequelae. The mean duration of intravenous antibiotic administration was only 4 days. The average hospital stay shortened significantly from 13 days to 9 days (p = 0.0001). The shortened hospital stay was not due to a change in the anatomical site of these infections, but to simplified treatment. Considerable savings in hospital stay, and thus costs, are feasible in osteoarticular infections of childhood by using CRP in monitoring the disease and shortening intravenous treatment by a swift move to per oral administration.Scandinavian Journal of Infectious Diseases 05/2012; 44(9):683-8. · 1.71 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: We investigated whether dried cerebrospinal fluid (CSF) conserved on filter paper can be used as a substrate for accurate PCR diagnosis of important causes of bacterial meningitis in the Lao PDR. Using mock CSF, we investigated and optimized filter paper varieties, paper punch sizes, elution volumes and quantities of DNA template to achieve sensitive and reliable detection of bacterial DNA from filter paper specimens. FTA Elute Micro Card™ (Whatman, Maidstone, UK) was the most sensitive, consistent and practical variety of filter paper. Following optimization, the lower limit of detection for Streptococcus pneumoniae from dried mock CSF spots was 14 genomic equivalents (GE)/μL (interquartile range 5.5 GE/μL) or 230 (IQR 65) colony forming units/mL. A prospective clinical evaluation for S. pneumoniae, S. suis and Neisseria meningitidis was performed. Culture and PCR performed on fresh liquid CSF from patients admitted with a clinical diagnosis of meningitis (n = 73) were compared with results derived from dried CSF spots. Four of five fresh PCR-positive CSF samples also tested PCR positive from dried CSF spots, with one patient under the limit of detection. In a retrospective study of S. pneumoniae samples (n = 20), the median (IQR; range) CSF S. pneumoniae bacterial load was 1.1 × 10(4) GE/μL (1.2 × 10(5) ; 1 to 6.1 × 10(6) DNA GE/μL). Utilizing the optimized methodology, we estimate an extrapolated sensitivity of 90%, based on the range of CSF genome counts found in Laos. Dried CSF filter paper spots could potentially help us to better understand the epidemiology of bacterial meningitis in resource-poor settings and guide empirical treatments and vaccination policies.Clinical Microbiology and Infection 05/2013; · 4.58 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Tropical infectious diseases diagnosis and surveillance are often hampered by difficulties of sample collection and transportation. Filter paper potentially provides a useful medium to help overcome such problems. We reviewed the literature on the use of filter paper, focusing on the evaluation of nucleic acid and serological assays for diagnosis of infectious diseases using dried blood spots (DBSs) compared with recognized gold standards. We reviewed 296 eligible studies: 101 studies evaluating DBSs and 192 studies on other aspects of filter paper use. We also discuss the use of filter paper with other body fluids and for tropical veterinary medicine. In general, DBSs perform with sensitivities and specificities similar or only slightly inferior to gold standard sample types. However, important problems were revealed with the uncritical use of DBS, inappropriate statistical analysis, and lack of standardized methodology. DBSs have great potential to empower healthcare workers by making laboratory-based diagnostic tests more readily accessible, but additional and more rigorous research is needed.The American journal of tropical medicine and hygiene 12/2013; · 2.53 Impact Factor
Diagnosis of Streptococcus pneumoniae and Haemophilus
influenzae Type b Meningitis by Identifying DNA From
Cerebrospinal Fluid-Impregnated Filter Paper Strips
Heikki Peltola, MD,* Irmeli Roine, MD,† Maija Leinonen, PhD,‡ Leena Kuisma, MSc,‡
Antonio Gonza ´lez Mata, MD,§ Antonio Arbo, MD,¶ Jose ´ Goyo, MD,? and Annika Saukkoriipi, PhD‡
Background: Bacterial meningitis remains often etiologically uncon-
firmed, especially in resource-poor settings. We tested the potential of
real-time polymerase chain reaction to identify Streptococcus pneumoniae
(Pnc) and Haemophilus influenzae type b (Hib) from cerebrospinal fluid
impregnated on filter paper strips.
Methods: Pnc and Hib genome equivalents were blindly quantified by
polymerase chain reaction from 129 liquid cerebrospinal fluid (CSF)
samples–the standard–and strips stored at room temperature for months.
Genome counts were compared by simple regression.
Results: The strips showed a sensitivity and specificity of 92% and 99%
for Pnc, and of 70% and 100% for Hib, respectively. The positive and
negative predictive values were 94% and 97% for Pnc, and 100% and 89%
for Hib, respectively. For Pnc, the positive and negative likelihood ratio
was 92 and 0.08, and the overall accuracy 98%, whereas for Hib they were
70 and 0.30, and 91%, respectively. Genome counting showed good
correlation between the filter paper and liquid CSF samples, r2being 0.87
for Pnc and 0.68 for Hib (P ? 0.0001 for both).
Conclusion: Although not replacing bacterial culture, filter paper strips
offer an easy way to collect and store CSF samples for later bacteriology.
They can also be transported in standard envelops by regular mail.
Key Words: bacterial meningitis, meningitis, filter paper, diagnosis,
PCR, Hib, S. pneumoniae
(Pediatr Infect Dis J 2010;29: 111–114)
influenzae type b (Hib), Streptococcus pneumoniae, and Neisseria
meningitidis, still some 5000 patients a day worldwide1,4fall ill
with this potentially devastating disease.5–8Most of them live in
conditions where even basic bacteriology does not exist or it is
unreliable. Because the true incidence of meningitis remains un-
known, vaccines are not always used to the extent they should be.1
espite effectiveness of conjugate vaccines1–3against the 3
leading agents of childhood bacterial meningitis, Hemopilus
And without any epidemiological data, no vaccination programs
are funded. The best available treatment is often not chosen, and
unnecessarily expensive methods are used9,10assuming that all
patients require similar approach. New, simple enough methods to
identify the causative agents would be of major relevance to get a
more reliable picture of bacterial meningitis globally.
Gram-staining is the traditional method to identify bacteria
in cerebrospinal fluid (CSF), but it requires an experienced mi-
croscopist, and detects bacteria only if they are in large quantities
(?105cfu/mL).11Latex agglutination12–14is other simple method,
but the volume needed is not minuscule, reagents are costly,
specificity of antisera varies, and interpretation requires experi-
ence. Furthermore, the reagents for agglutination have a short
shelf-life, which adds to the cost.
Being an extremely sensitive assay, polymerase chain reac-
tion (PCR) has opened new doors in the diagnostics of bacterial
meningitis.15–17The transport of CSF samples is always cumber-
some, but nowadays, international shipping of biologic specimens
has become very costly, let aside difficulty with proper storage
through the journey from the “bush” to metropolis.
We examined whether CSF samples impregnated onto filter
paper strips might be a useful and reliable method in the diagnos-
tics of bacterial meningitis.
As a part of our large treatment study on pediatric bacterial
meningitis in Latin America,18129 CSF samples were collected
from 82 (70 from Venezuela, 12 from Paraguay) children aged
from 2 months to 15 years. Diagnostic lumbar punctures were
derived from 82 samples, and 47 were obtained from the taps
performed 24 to 36 hours after start of treatment. All samples were
examined for S. pneumoniae and Hib. The study was approved by
the institutions’ ethical committees, and the child was included
only with consent of a legal guardian. The protocol stated that a
part of the CSF samples was to be frozen for further analysis.
Conventional bacteriology was performed in the local lab-
oratories. Of the 82 cases, 73 (89%) fulfilled the preset criteria for
bacterial meningitis,18and in those, an agent was identified locally
in 62 cases (S. pneumoniae, 30; Hib, 28; N. meningitidis, 3; and
Salmonella enteritidis, 1). Pneumococcal meningitis was diag-
nosed by CSF culture in 19 cases, by latex agglutination in 5,
gram-staining in 4, and by blood culture with concordant CSF
staining in 2. For Hib meningitis, the respective numbers were 21,
1, 5, and 1.
Preparation for PCR Testing
The liquid CSF samples were kept frozen at ?20°C, also
when transported in a triple safety package to Finland, where the
liquid CSF specimens were thawed, and 200 ?L of each specimen
was taken for DNA extraction using QIAamp DNA Mini Kit
(QIAGEN, Hilden, Germany). The manufacturer’s instructions
Accepted for publication June 26, 2009.
From the *Hospital for Children and Adolescents, Helsinki University Central
Hospital, Helsinki, Finland; †Facultad de Ciencias de la Salud, Universidad
Diego Portales, Santiago, Chile; ‡Child and Adolescent Health and Well-
being Unit, National Institute for Health and Welfare, Oulu, Finland;
§Hospital Pediatrico Dr. Agustin Zubillaga, Barquisimeto, Venezuela; ¶In-
stituto de Medicina Tropical, Universidad Nacional de Asuncio ´n, Asuncio ´n,
Paraguay; and ?Hospital Universitario de los Andes, Me ´rida, Venezuela.
Supported by Alfred Kordelin, Pa ¨ivikki and Sakari Sohlberg, and Sigfrid
Juse ´lius Funds, and the Foundation for Pediatric Research, Finland.
Address for correspondence: Heikki Peltola, MD, HUCH Hospital, Hospital for
Children and Adolescents, PO Box 281 (11 Stenba ¨ck St), 00029 HUS
Helsinki, Finland. E-mail: firstname.lastname@example.org.
Supplemental digital content is available for this article. Direct URL citations
appear in the printed text and are provided in the HTML and PDF versions
of this article on the journal’s Web site (www.pidj.com).
Copyright © 2010 by Lippincott Williams & Wilkins
The Pediatric Infectious Disease Journal • Volume 29, Number 2, February 2010
www.pidj.com | 111
were followed with the following modifications: lysis of the
sample was undertaken overnight, and after incubation at room
temperature for 5 minutes, extracted DNA was eluted into 100 ?L
(instead of 200 ?L) of Buffer AE. This sample concentration was
taken into account when calculating genome counts. General
precautions were taken to avoid contamination including the use of
a separate room only for DNA extraction. When extracting DNA,
every 12th sample served as a negative control (sterile distilled
water). The extracted DNA was stored at ?20°C.
When thawed for DNA extraction and PCR analysis, a
25-?L specimen was spotted on a filter paper strip (Whatman,
Middlesex, UK). The strips were not autoclaved, but handled with
gloves on. They were labeled with the numbers of the conventional
samples and allowed to dry in room air. Finally, the strips were put
in small plastic bags (in which they could have been sent by
regular mail), and stored at room temperature protected from light.
After storage for at least 8 months, the PCR analysis was
done in 2005, blinded from the earlier results. One-fourth of the
slip was cut with a sterile scalpel, and from this small piece of filter
paper DNA was extracted by putting the piece in a microcentrifuge
tube into which 200 ?L of Buffer ATL and 20 ?L of Proteinase K
(both from QIAGEN) were added. The specimens were incubated
overnight at ?56°C.
The following day, the supernatant was collected and DNA
was extracted from it as described for the liquid CSF samples.
Purified DNA was eluted in 100 ?L of Buffer AE and stored at
?20°C. When calculating genome counts, the dilution during the
DNA extraction (one-fourth of the strip with 25 ?L of CSF to 100
?L of purified DNA) was taken into account.
DNA was searched for with 2 real-time PCR assays. For S.
pneumoniae, we used our own method19for detection of the
pneumolysin (ply) encoding gene. The sample volume was 8 ?L
and sequence-specific hybridization probes were used.19
For Hib, a conventional PCR method20was modified for
real-time PCR by using a new reverse primer (sequence 5?
GCTAAGATGAAGTTATGGCGAA 3?). A pair of fluorescent-
labeled, sequence specific hybridization probes (designed by O.
Landt, Tib-Molbiol, Germany; sequences 5? CATTTACAGAC-
GACCAAAGGTATCTTG- fluorescein 3?, and 5? LCRed640-
GTATAGCCTCGCCCCCAGAATTC-phosphate 3?) detected the
amplification products. The method is based on the published
sequence of the capB locus (GenBank accession no X78559).
The 20-?L Hib PCR reaction volume contained 1 ? Fast-
Start DNA Master Hybridization Probes solution (Roche Diagnos-
tics, Mannheim, Germany), 4 mM MgCl2, 0.5 ?M of each primer,
0.2 ?M of each hybridization probe and 8 ?L of sample DNA. The
Hib PCR program comprised initial denaturation at 95°C for 10
minutes followed by 50 cycles of amplification each consisting of
denaturation at 95°C for 10 seconds, annealing at 59°C for 10
seconds, and elongation at 72°C for 14 seconds. In both PCR
assays, at least every seventh specimen was a no-template control
(sterile distilled water).
External standard curves were used for quantification, cre-
ated by preparing 6 serial dilutions of purified pneumococcal or
Hib DNA and amplifying 4 replicates of each dilution. In each run,
a standard was included as a reference and it also served as the
The results read from the strips were compared with the
gold standard, the PCR results of the liquid CSF-samples
deriving from the same patients and specimens. Sensitivity and
specificity were defined as the percentage of positive (sensitiv-
ity) or negative (specificity) findings of the filter papers from all
the positives or negatives, respectively, detected by the gold
standard. The positive and negative predictive values were
defined as the percentage of the by filter paper correctly
identified positives or negatives, respectively, from all positives
or negatives, respectively. Finally, the positive likelihood ratio
was obtained from the formula of sensitivity/1-specificity, the
negative likelihood ratio deriving from the formula of 1-sensi-
tivity/specificity. A positive likelihood ratio of at least 10
indicates such a high degree of test reliability that further
testing can be suspended. Similarly, a negative likelihood value
below 0.1 justifies the same conclusion for a negative result.21
Because of the non-normal genome count distribution, log
transformation was done before simple regression, used in the
comparison of the results from the conventional versus filter paper
samples. A PCR result was taken as positive when at least 5
genomes/uL of S. pneumoniae or Hib were detected.22
For S. pneumoniae, the strips proved very reliable, as the
sensitivity was 92%, specificity 99%, and the positive and negative
predictive values of 94% and 97%, respectively (Table 1). The
positive and negative likelihood ratio was 92 and 0.08, respec-
tively, while the overall accuracy was 96%.
For Hib, the filter paper PCR sensitivity was no more than
70%, but the specificity was 100%. The positive and negative
predictive values were 100% and 89%, respectively, and the
positive and negative likelihood ratio 70 and 0.30 (Table 1). The
overall accuracy was 91%.
PCR for both types of samples was positive in 33 cases of
Pnc. The bacterial genome range was enormous (Fig., Supplemen-
tal Digital Content 1, http://links.lww.com/INF/A183): in strips,
the median count was 5.960/?L with a range from 19 to 4020.000
genomes per ?L, while in the liquid samples it was 16 438/?L
with a range from 5 to 9187.500 per ?L.
Also for Hib (N ? 26), the range of genome counts
was very wide (Fig., Supplemental Digital Content 2,
http://links.lww.com/INF/A184). In filter papers, the median
count was 476/?L and the range from 7 to 120.800 genomes per
?L, while in the conventional samples the median was
2.381/?L and the range from 18 to 9250.000 genomes per ?L.
High correlation (P ? 0.0001) between the strip and liquid
sample counts prevailed for both agents.
Versus Conventional Liquid CSF Samples to Detect
Streptococcus pneumoniae and Haemophilus influenzae
Type b by Real-time PCR in Childhood Meningitis
Accuracy of the Filter Paper Samples
(N ? 128)
(N ? 129)
(N ? 257)
Predictive value %
*Pneumolysin encoding gene.
†Type b capsulation locus.
Peltola et al
The Pediatric Infectious Disease Journal • Volume 29, Number 2, February 2010
© 2010 Lippincott Williams & Wilkins
112 | www.pidj.com
In liquid CSF, genome count for S. pneumoniae, log 2.83 ?
2.36/?L, did not differ from that for Hib, 2.27 ? 2.35/?L (P ?
0.24). However, the counts of both agents were significantly (P ?
0.0001) higher in the liquid samples if filter paper was positive
than if it was not. For Pnc, the filter paper-positive cases showed
a liquid sample count of log 4.04 ? 1.88/?L versus log 0.34 ?
0.68/?L in the filter paper-negative cases. For Hib, the counts were
log 3.85 ? 1.77/?L versus log 0.40 ? 1.36/?L, respectively.
When searching for the agent causing bacterial meningi-
tis, merely the “yes or no” result regarding the pathogen usually
suffices while awaiting antibiotic susceptibility testing results.
Filter paper PCR does not necessarily disclose the agent on site,
but offers a chance to get this information months, perhaps
years, later if needed.
PCR has been used in bacterial meningitis in developing
countries,23,17but we are aware of only one study from Nigeria
in the 1970s in which CSF samples dried on filter paper were
used for the etiologic diagnosis.24Counterimmunoelectro-
phoresis detected capsular polysaccharide antigens in 55%
(78/142) of purulent CSF samples; relation to the culture-
positivity was not reported in that study.
As one might expect, PCR was more sensitive than
counterimmunoelectrophoresis more than 30 years ago; PCR
found 85% to 90% of the culture-positive cases. The association
was good, but equally clear was that PCR is not a substitute for
bacterial culture. The only major advantage of the filter paper
methodology is that it keeps the agents identifiable long after-
ward. Some loss of sensitivity due to high ambient temperature
is likely to occur if the strips are stored for a very long time.
However, Leishmania was identified from Giemsa-stained le-
sion imprint slides stored for up to 36 years,25and “suicide
PCR” from the tooth pulp of human bodies documented Yer-
sinia pestis as the agent causing black death in the 14th
century.26PCR has also been used for identifying H1N1influ-
enza A virus in Spanish Flu.27No doubt, PCR has brought new
sights in the diagnostics of infectious diseases.
A laboratory test is considered reliable if it has high sensi-
tivity combined with high negative predictive value. For S. pneu-
moniae, filter paper showed a 92% sensitivity and a negative
predictive value of 97%, the overall accuracy being 98%. Thus, if
a strip PCR detects pneumococcus, no confirmation is needed.
Filter paper sensitivity was only 70% for Hib, but the
specificity was 100%, the negative predictive value 89%, and
the overall accuracy 91%. Thus, conditions permitting, a neg-
ative filter paper result should preferably be checked (because of
somewhat lower sensitivity), but this must be weighed against the
true needs. Lower genome counts in the conventional CSF sample
for Hib than Pnc may partly explain the lower sensitivity. It might
also be possible (though unlikely) that the filter papers we used
contained some PCR inhibitors bound to DNA which caused
reduced amplification in Hib PCR.
The PCR technology is rather costly, but it also can be
brought on site.17However, if transport of the specimens is
needed, we see no biologic risk in the regular mail, especially if the
strips are sealed in plastic bags; the relevant bacteria do not survive
on filter paper. If storage cannot be avoided, the strips should
preferably be kept in a desiccator with silica gel, because humidity
might cause decomposition.28Contamination is hardly an issue
since pneumococcus and Hib (and meningococcus) are rarely
found in ambience or on human skin or hands. For, say, staphy-
lococci, this is not the case. Simple and inexpensive filter paper
strips offer a handy tool for epidemiologic studies and to clinicians
who treat severe infections in difficult conditions.
The authors thank Elsi A¨ija ¨la ¨ for her expertise in the PCR
analyses and also Dr. Ralf Clemens, then with GlaxoSmithKline,
who organized the initial grant for this nonprofit making study.
1. Peltola H. Worldwide Haemophilus influenzae type b disease at the begin-
ning of the 21st century: global analysis of the disease burden 25 years after
the use of polysaccharide vaccine and a decade after the advent of conju-
gates. Clin Microbiol Rev. 2000;13:302–317.
2. Tsai CJ, Griffin MR, Nuorti P, et al. Changing epidemiology of pneumo-
coccal meningitis after the introduction of pneumococcal conjugate vaccine
in the United States. Clin Infect Dis. 2008;46:1664–1672.
3. Center or Disease Control and Prevention. Prevention and control of
meningococcal disease: recommendations of the Advisory Committee on
Immunization Practices (ACIP). Morb Mortal Wkly Rep. 2005;54(RR-7):
4. Peltola H. Burden of meningitis and other severe bacterial infections of
children in Africa: implications for prevention. Clin Infect Dis. 2001;31:
5. Daoud AS, AL-Sheyyab M, Batchoun RG, et al. Bacterial meningitis: still
a cause of high mortality and severe neurological morbidity in childhood.
J Trop Pediatr. 1995;41:308–310.
6. Goetghebuer T, West TE, Wermenbol V, et al. Outcome of meningitis
caused by Streptococcus pneumoniae and Haemophilus influenzae type b in
children in the Gambia. Trop Med Int Health. 2000;5:207–213.
7. Koedel U, Scheld WM, Pfister HW. Pathogenesis and pathophysiology of
pneumococcal meningitis. Lancet Infect Dis. 2002;2:721–736.
8. Pelkonen T, Roine I, Monteiro L, et al. Acute childhood bacterial menin-
gitis in Luanda, Angola. Scand J Infect Dis. 2008;40:859–866.
9. Macfarlane JT, Anjorin FI, Cleleand PG, et al. Single injection treatment of
meningococcal meningitis 1. Long-acting penicillin. Trans R Soc Trop Med
10. Roine I, Ledermann W, Foncea LM, et al. Randomized trial of four vs.
seven days of ceftriaxone treatment for bacterial meningitis in children with
rapid initial recovery. Pediatr Infect Dis. 2000;19:219–222.
11. Kaplan SL. Rapid diagnostic techniques. In: Feigin RD, Cherry JD, eds.
Textbook of Pediatric Infectious Diseases. 2nd ed. Philadelphia, PA: WB
12. Severin WP. Latex agglutination in the diagnosis of meningococcal men-
ingitis. J Clin Pathol. 1972;25:1079–1082.
13. Leinonen M, Ka ¨yhty H. Comparison of counterimmunoelectrophoresis,
latex agglutination and radioimmunoassay in detection of soluble capsular
polysaccharide antigens of Haemophilus influenzae type b and Neisseria
meningitidis of groups A and C. J Clin Pathol. 1978;31:1172–1176.
14. Sanborn WR, Toure ´ IM. A simple kit system for rapid diagnosis of
cerebrospinal meningitis in rural areas of developing countries. Bull World
Health Organ. 1984;62:293–297.
15. Ni H, Knight AI, Cartwright K, et al. Polymerase chain reaction for
diagnosis of meningococcal meningitis. Lancet. 1992;340:1432–1434.
16. Corless CE, Guiver M, Borrow R, et al. Simultaneous detection of Neisseria
meningitidis, Streptococcus pneumoniae, and Haemophilus influenzae in
suspected cases of meningitis and septicemia using real-time PCR. J Clin
17. Parent du Cha ˆtelet I, Traore Y, Gessner BD, et al. Bacterial meningitis in
Burkina Faso: surveillance using field-based polymerase chain reaction
testing. Clin Infect Dis. 2005;40:17–25.
18. Peltola H, Roine I, Ferna ´ndez J, et al. Adjuvant glycerol and/or dexameth-
asone to improve the outcomes of childhood bacterial meningitis: a pro-
spective, randomized, double-blind, placebo-controlled trial. Clin Infect
19. Saukkoriipi A, Palmu A, Kilpi T, et al. Real-time quantitative PCR for the
detection of Streptococcus pneumoniae in the middle ear fluid of children
with acute otitis media. Mol Cell Probes. 2002;16:385–390.
20. Falla TJ, Crook DW, Brophy LN, et al. PCR for capsular typing of
Haemophilus influenzae. J Clin Microbiol. 1994;32:2382–2386.
21. Deeks JJ, Altman DG. Education and debate. Statistics Notes. Diagnostic
tests 4: likelihood ratios. BMJ. 2004;329:168–169. Available at: http://
The Pediatric Infectious Disease Journal • Volume 29, Number 2, February 2010 Diagnosis of Meningitis
© 2010 Lippincott Williams & Wilkins
www.pidj.com | 113
www.bmj.com/cgi/content/full/329/7458/168. Accessed November 10,
22. Roine I, Saukkoriipi A, Leinonen M, et al. Microbial genome count in
cerebrospinal fluid compared with clinical characteristics in pneumococcal
and Haemophilus influenzae type b meningitis in children. Diagn Micro
Infect Dis. 2009;63:16–23.
23. Sidikou F, Djibo S, Taha MK, et al. Enhancement of the surveillance of
bacterial meningitis in remote areas in Niger: relevance of PCR assay.
Emerg Infect Dis. 2003;9:486–488.
24. Whittle HC, Greenwood BM, Gabrielse L, et al. The bacteriological
diagnosis of pyogenic meningitis from cerebrospinal fluid dried on filter
paper. Trans R Soc Trop Med Hyg. 1976;70:217–218.
25. Volpini AˆC, Marques MJ, Lopes dos Santos S, et al. Leishmania identifi-
cation by PCR of Giemsa-stained lesion imprint slides stored for up to 36
years. Clin Microbiol Infect. 2006;12:815–818.
26. Raoult D, Aboudharam G, Crube ´zy E, et al. Molecular identification by
“suicide PCR” of Yersinia pestis as the agent of medieval black death. Proc
Natl Acad Sci USA. 2000;97:12800–12803.
27. Taubenberger JK, Reid AH, Krafft AE, et al. Initial genetic character-
ization of the 1918 “Spanish” influenza virus. Science. 1997;275:1793–
28. Fa ¨rnert A, Arez AP, Correia AT, et al. Sampling and storage of blood and
the detection of malaria parasites by polymerase chain reaction. Trans R
Soc Trop Med Hyg. 1999;93:50–53.
Edited by: Robert J. Leggiadro, MD
Cronobacter Species Isolation in Two Infants—New Mexico,
Centers for Disease Control and Prevention. MMWR. 2009;58:1179–1183.
Cronobacter spp. (formerly Enterobacter sakazakii) are rare causes of
infant septicemia and meningitis, resulting in death in approximately 40%
of cases. Powdered infant formula (PIF), which is not sterile, has been
implicated repeatedly as a vehicle of Cronobacter infection. This report
describes isolation of Cronobacter spp. in 2 nonhospitalized, unrelated
infants (one male and one female) in New Mexico in 2008. The female
infant developed severe brain injury and hydrocephalus as a result of
Cronobacter meningitis and brain abscesses, diagnosed at age 6.5 weeks.
The male infant died of apparent sudden infant death syndrome (SIDS) at
7 months of age; postmortem blood culture grew Cronobacter. The 2
infants lived approximately 200 miles apart in different counties of mostly
rural southeastern New Mexico.
Ingestion of PIF was the only identified risk factor for Cronobacter
exposure for the 2 infants. The 2 infants had consumed the same brand of
PIF, but had no other common exposures. PIF preparers reported washing
their hands before preparing formula and washing bottle components by
hand between feedings for the female infant and in the dishwasher for the
male infant. Additionally, they reported following the manufacturer’s
instructions for preparing and handling PIF.
In the female infant’s household, water used for PIF reconstitution
had been boiled, cooled, and then stored in the kitchen in a plastic
container. In the male infant’s household, PIF had been reconstituted by
using refrigerated bottled water or tap water. The water was not boiled or
heated before use.
Although the 2 infants had consumed the same brand of formula,
their clinical Cronobacter isolates had different pulsed-field gel electro-
phoresis (PFGE) patterns. None of the samples obtained from the home of
the female infant yielded Cronobacter spp. However, PIF from an opened
canister and the vacuum cleaner filter from the home of the male infant
yielded Cronobacter spp. The clinical Cronobacter isolate from the male
infant and the Cronobacter isolate from the PIF canister sample from his
home had indistinguishable PFGE patterns. However, the vacuum cleaner
Cronobacter isolate from the same home had a different PFGE pattern.
Cronobacter spp. were not isolated from any of the unopened canisters of
PIF in either home.
Comment: Although a Cronobacter organism was isolated from the
male infant at autopsy, the role of this organism in the infant’s apparent
death from SIDS is unknown. Isolation of Cronobacter spp. in association
with SIDS has not been previously reported.
Previous investigations have found Cronobacter spp. cultured from
prepared formula, unopened PIF containers, and the environment where
PIF was reconstituted, clearly implicating PIF as the source of outbreaks.
Although Cronobacter spp. have been isolated from household vacuum
cleaners previously, the source of the Cronobacter organisms in such
settings and the clinical implications are unknown. Enhanced surveillance
might produce a more accurate estimate of the Cronobacter disease burden
and insight into risk factors. Accordingly, the Foodborne Diseases Active
Surveillance Network (Food Net) of CDC’s Emerging Infections Program
is expected to soon begin piloting surveillance for Cronobacter infections
in the United States.
The World Health Organization has developed guidelines for the
preparation of PIF, including reconstitution with water hot enough to
inactivate Cronobacter organisms. In the United States and elsewhere,
recommendations are to breast-feed infants when possible, use sterile
liquid infant formula in high-risk settings (eg, neonatal intensive care units
and hospital nurseries), and adhere to the safest available PIF preparation
Peltola et al
The Pediatric Infectious Disease Journal • Volume 29, Number 2, February 2010
© 2010 Lippincott Williams & Wilkins
114 | www.pidj.com