Evaluation of serological tests for the diagnosis of rickettsiosis in Denmark.
ABSTRACT Two commercially available immunofluorescence assays (IFA) were compared using historical sera evaluated for rickettsial antibodies by the Weil-Felix test. An IFA test produced by Focus Diagnostics prepared with Rickettsia rickettsii and R. typhi antigens was compared with a custom made kit from Fuller Laboratories with R. rickettsii, R. typhi, R. conorii and R. helvetica as antigens. The serum panel used for the comparison included Weil-Felix-positive and -negative samples. Cross-reactions were analyzed using serum samples from patients with clinical symptoms similar to those of rickettsiosis. When analyzing the data using the manufacturers' cut-off values, 41% of samples from presumably healthy blood donors were found positive for spotted fever group Rickettsia antibodies. This does not correlate to the general picture of rickettsiosis in Denmark. Furthermore, sera with Coxiella burnetii antibodies were found to be cross-reacting in both tests. When applying cut-off values calculated on the 95% percentile on data from blood donor serum samples, there was no significant difference between the two kits. Moreover, when using the newly established cut-off, cross-reactions were eliminated.
- SourceAvailable from: ncbi.nlm.nih.gov[show abstract] [hide abstract]
ABSTRACT: Rickettsioses are caused by species of Rickettsia, a genus comprising organisms characterized by their strictly intracellular location and their association with arthropods. Rickettsia species are difficult to cultivate in vitro and exhibit strong serological cross-reactions with each other. These technical difficulties long prohibited a detailed study of the rickettsiae, and it is only following the recent introduction of novel laboratory methods that progress in this field has been possible. In this review, we discuss the impact that these practical innovations have had on the study of rickettsiae. Prior to 1986, only eight rickettsioses were clinically recognized; however, in the last 10 years, an additional six have been discovered. We describe the different steps that resulted in the description of each new rickettsiosis and discuss the influence of factors as diverse as physicians' curiosity and the adoption of molecular biology-based identification in helping to recognize these new infections. We also assess the pathogenic potential of rickettsial strains that to date have been associated only with arthropods, and we discuss diseases of unknown etiology that may be rickettsioses.Clinical Microbiology Reviews 11/1997; 10(4):694-719. · 17.31 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Enzyme-linked immunosorbent assay (ELISA) techniques for the determination of immunoglobulin G to rickettsial lipopolysaccharides were developed. These techniques provide a simple and convenient way to serodiagnose Mediterranean spotted fever and murine typhus with a single serum dilution. The results of the ELISAs correlated with the indirect immunofluorescence assay titers of cross-reacting sera.Journal of Clinical Microbiology 05/1997; 35(4):1034-5. · 4.07 Impact Factor
Article: Rickettsia felis infection, Tunisia.[show abstract] [hide abstract]
ABSTRACT: We report, for the first time, serologic evidence of Rickettsia felis and R. aeschlimannii infections acquired in Tunisia from 1998 to 2003. We found that most patients with antibodies against both R. conorii and R. typhi had serologic evidence of R. felis infection.Emerging infectious diseases 02/2006; 12(1):138-40. · 5.99 Impact Factor
Evaluation of serological tests for the diagnosis of rickettsiosis in Denmark
Bjørn Kantsø, Claus Bo Svendsen, Charlotte Sværke Jørgensen, Karen A. Krogfelt⁎
Department of Bacteriology, Mycology and Parasitology, Statens Serum Institut, Copenhagen, Denmark
a b s t r a c ta r t i c l ei n f o
Received 26 August 2008
Received in revised form 9 December 2008
Accepted 11 December 2008
Available online 3 January 2009
Two commercially available immunofluorescence assays (IFA) were compared using historical sera evaluated
for rickettsial antibodies by the Weil-Felix test. An IFA test produced by Focus Diagnostics prepared with
Rickettsia rickettsii and R. typhi antigens was compared with a custom made kit from Fuller Laboratories
with R. rickettsii, R. typhi, R. conorii and R. helvetica as antigens.
The serum panel used for the comparison included Weil-Felix-positive and -negative samples. Cross-
reactions were analyzed using serum samples from patients with clinical symptoms similar to those of
When analyzing the data using the manufacturers' cut-off values, 41% of samples from presumably healthy
blood donors were found positive for spotted fever group Rickettsia antibodies. This does not correlate to the
general picture of rickettsiosis in Denmark. Furthermore, sera with Coxiella burnetii antibodies were found to
be cross-reacting in both tests.
When applying cut-off values calculated on the 95% percentile on data from blood donor serum samples,
there was no significant difference between the two kits. Moreover, when using the newly established cut-
off, cross-reactions were eliminated.
© 2008 Elsevier B.V. All rights reserved.
Bacteria of the genus Rickettsia (R) are Gram-negative, non-motile,
intracellular bacteria (Raoult and Roux,1997) with endothelial cells as
their primary target (Walker, 1984). Rickettsiae are phylogenetically
classified into four groups: spotted fever group (SFG), typhus group
(TG), transitional group, and ancestral group (Walker and Ismail,
2008). Serologically, the SFG rickettsiae are similar to the transitional
and ancestral groups but all are distinguishable from the TG
rickettsiae. Within the SFG, R. conorii is the etiological agent of
Mediterranean spotted fever also termed Boutonneuse fever (Anton
et al., 2003), R. rickettsii, the etiological agent of Rocky Mountain
spotted fever (Dantas-Torres, 2007); and R. helvetica (a rickettsial
species found in Denmark (Nielsen et al., 2004;Skarphedinsson et al.,
2007;Svendsen et al., 2009)) has been suspected as an agent involved
in sarcoidosis (Nilsson et al., 2002; Svendsen et al., 2008) Among the
TG, R. typhi is the etiological agent of murine typhus (endemic typhus)
Serologic diagnosis of rickettsial infections was first discovered by
Weil and Felix in 1916 (Weil and Felix,1916). The Weil-Felix (WF) test
is based on agglutination between Proteus vulgaris OX2, P. vulgaris
OX19, P. mirabilisOXK(WeilandFelix,1916), andrickettsialantibodies.
These bacteria primarily agglutinate with antibodies from SFG (with
the exception of R. rickettsii), TG and Orientia tsutsugamushi (Amano
et al., 1992; Amano and Williams, 1998).
Rickettsial antibodies can be detected by complement fixation
(Philip et al.,1977), latex agglutination (Kaplan,1986), enzyme-linked
immunosorbent assay (Keysary,1997) and immunofluorescence assay
(IFA) (La Scola and Raoult, 1999). IFA is frequently used and regarded
as the gold standard for serological diagnosis of rickettsial infections
(La Scola,1997). Studies on the sensitivity and specificity indicate that
IFA has the highest sensitivity and specificity among the evaluated
methods (Kaplan 1986; Newhouse et al., 1979; Philip et al., 1978).
The aim of this study was to compare two commercially available
IFA tests used for rickettsial serology (IgG-antibodies) with regard to
their diagnostic accuracy as compared to the traditional Weil-Felix
test. Furthermore, a cut-off value was established for the diagnosis of
rickettsiosis in a presumed low prevalence country such as Denmark.
2. Materials and methods
2.1. Patient material
Historical patient sera, previously evaluated using the Weil-Felix
test, were divided into four groups: A) 19 sera with positive titers
towards OX2 or OX19 at or above 200; B) 15 sera with positive titers
towards OX2 or OX19 equal to 100 (the current laboratory cut-off); C)
52 sera with titers towards OX2 or OX19 between 25 and 50 (clinically
regarded as negative); and D) 19 sera with titers below 25. Plasma
samples from 106 presumably healthy blood donors from the Blood
Bank at Rigshospitalet (Copenhagen University Hospital) in
Journal of Microbiological Methods 76 (2009) 285–288
⁎ Corresponding author. Statens Serum Institut, Department of Bacteriology,
Mycology and Parasitology, 5 Artillerivej, DK-2300 Copenhagen S, Denmark. Tel.: +45
32 68 37 45; fax: +45 32 68 31 47.
E-mail address: firstname.lastname@example.org (K.A. Krogfelt).
0167-7012/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Journal of Microbiological Methods
journal homepage: www.elsevier.com/locate/jmicmeth
Copenhagen were used as negative controls. We had access to the
donors' age and sex, but no clinical information on them. Cross-
reactions were analyzed on 80 serum samples; 10 samples from each
of the following groups: patients found positive for antibodies against
either Salmonella spp., Helicobacter pylori, Coxiella burnetii, Campylo-
bacter jejuni, Mycoplasma pneumoniae, Chlamydophila pneumoniae,
Bartonella spp., or Borrelia burgdorferi. The cross-reaction samples
where chosen from infections with clinical symptoms resembling
those of rickettsiosis, with the exception of Salmonella spp., C. jejuni
andH.pylori,whichwere chosenbecauseof their–in ourexperience–
tendency to cross-react broadly. Sera for the cross-reaction test were
obtained from the routine diagnostic laboratory at Statens Serum
Institut. The study was planned as a quality control study, why ethics
committee approval was not required by Danish law.
We used kits specifically designed for detection of IgG antibodies
to Rickettsia species. Kit A was purchased from Focus Diagnostics
(IF0100G, Rickettsia IFA IgG; Focus Diagnostics, Inc., Cypress, CA, USA),
with antigens from R. typhi (TG) and R. rickettsii (SFG). Kit B was
custom made by Fuller Laboratories (Fullerton, CA, USA) containing
Fig.1. Shows the geometric mean titers of the different groups. A: WF 200 consists of serum samples with Weil-Felix titer equal to or above 200; WF 100 consists of serum samples
with titers equal to 100 (the cut-off); WF 25–50 consists of serum samples with titers of 25 or 50; WF neg sera had been tested negative/non-reactive with the Weil-Felix test. Donor
samples have not been tested with the Weil-Felix test. B: Geometric mean titers from the cross-reaction panel. The eight bacterial species,10 samples of each with antibodies, tested
for cross-reactions in the two commercial kits.
B. Kantsø et al. / Journal of Microbiological Methods 76 (2009) 285–288
antigens from R. typhi (TG), R. rickettsii (SFG), R. conorii (SFG), and R.
helvetica (SFG). Both kits were used according to the manufacturer's
specifications. All samples were screened at dilutions of 1:64 and
1:128 and titrated to end-point fluorescence in two-fold dilutions. Kit
A used a positive cut-off of ≥1:256, and kit B≥1:128. All sera were
2.3. Statistical analysis
For comparisons between the kits and WF, we dichotomized the
end-point titers according to the cut-off. Spearman's rank correlation
titers were calculated for comparison between groups because of the
logarithmic nature of the data. All statistical analyses were performed
using Microsoft Excel 2003 for Windows and SPSS statistical software
(SPSS 18.104.22.168 for Windows, SPSS Inc. [Chicago, IL, USA], 1989–2006).
Significance level was set at p=0.05. For the statistical analysis, IFA
titers b1:64 and WF titers b1:25 were set as 2 before logarithmic
transformation to prevent left-censoring of the data set.
3.1. WF vs. IFA
A positive correlation between the WF titer and the IFA titers were
seen (Fig. 1A, WF vs. kit A: Spearman's rho=0.400, pb0.001, WF
sensitivity 74%, WF specificity 79%; WF vs. kit B: Spearman's
rho=0.299, pb0.001, WF sensitivity 60%, WF specificity 73%). The
group with titers below 25 stood out from the general picture.
3.2. Kit A vs. Kit B
In the groups with WF titers ≥200 and =100, the agreement
between the two kits was 100% when dichotomizing the results
according to the manufacturers' cut-offs.
Within the positive sera from the groups WF≥200 and WF=100,
10% (2/19) and 33% (5/15) respectively, showed no reactivity in any of
the IFAs. The group with WF titers of 25–50 (clinically regarded as
negative) exhibited a discrepancy between the two kits: 38% (20/52)
were positive with kit A; 56% (29/52) with kit B. Of the 19 samples
from the WFb25 group,1 (5%) was positive in kit A, and 13 (68%) were
positive in kit B, for either SFG or TG rickettsiae when using the
3.3. Cross-reaction of other antibodies
Sera from patients with antibodies to Coxiella burnetii showed the
highest cross-reactivity (see Fig. 1B). One of 10 samples with
antibodies against Bartonella henselae was found to cross-react. In all
samples from the cross-reactions panel, the geometric mean titer of
antibodies to the SFG rickettsiae was higher than the titer towards TG
3.4. Blood donor sera
Among the 106 samples from the blood donors, we found 14%
positive for R. rickettsii and 6% positive for R. typhi in kit A when using
the manufacturer's cut-off. In kit B, 4 samples (4%) were positive for R.
rickettsii at 1:64 dilution (below cut-off), and 1 (1%) was positive at
1:128 (cut-off). For R. conorii, 41% were positive; for R. typhi, 39%.
The appropriate cut-off for the Danish population was determined
from the titers of the presumably healthy blood donors. A cut-off of
1:512 was chosen, corresponding to the 95% percentile of logarith-
mically transformed titers.
When raising the cut-offs of both kits to 1:512, 5% were positive to
R. rickettsii and none to R. typhi in kit A. In kit B, R. conorii and R. typhi
were positive in 6%. Using 1:512 as cut-off, there was no significant
difference between the two kits when comparing dichotomized
results (below vs. at or above cut-off).
To our knowledge, this is the first study to compare commercially
available immunofluorescence kits used for rickettsial serology.
The results from our study showed a clear positive correlation
between the Weil-Felix titer and the IFA titer.
As seen in Fig. 1A, the geometric mean titer in the WFb25 group
and the group with WF titers above or equal to 200 stand out from the
rest by having the highest geometrical means in kit A (R. rickettsii) and
kit B (R. typhi). This is indicative of the lowspecificityand sensitivityof
the Weil-Felix test, also illustrated in Fig. 2.
Cross-reactions between SFG and TG were also observed, the most
frequently observed was SFG seemingly cross-reacting with TG as
recently described elsewhere (Znazen et al., 2006).
Fig. 2. Scatter plot of reactive Weil-Felix samples, plotted as either positive or negative
(at or above 100 vs. below 100), and their equivalent IFA titer. A: from kit A (Focus). B:
from kit B (Fuller). Black line shows the newly established cut-off (1:512) fitted to the
general population in Denmark.
B. Kantsø et al. / Journal of Microbiological Methods 76 (2009) 285–288
When using the manufacturers' cut-off values (kit A: 1:256; kit B:
1:128), a relatively high number of positive sera among the 106 blood
donors was found, none was found positive for R. rickettsii with kit B.
We have no certain explanation for this, but we suspect it to be due to
Based on the 95% percentile, the cut-off was raised to 1:512
thereby achieving a more acceptable number of positive serum
samples compared to our general presumption of rickettsial disease in
Denmark, where only imported cases of rickettsiosis have been
We expected to find similar titer values between the antigens from
R. conorii and R. rickettsii in kit B due to cross-reactions between SFG
rickettsiae. However, this was not the case. The R. rickettsii antigen
from kit B did not perform as expected from the kit B R. conorii titer
and the kit A R. rickettsii titer. Kit B R. rickettsii provided no reactivity
in the group of blood donors and the cross-reactions group. The same
pattern was observed with R. helvetica from kit B. We expect it relates
to the quality of the present antigen formulation. However, it could
also be due to low prevalence of R. helvetica infections or altered
kinetics of the serological response to R. helvetica, such as that
previously described with R. africae (Fournier et al., 2002). R. helvetica
is the only rickettsial species found by PCR in Danish ticks, with a
prevalence of 1.1–13% as evaluated in three recent studies (Nielsen
et al., 2004; Skarphedinsson et al., 2007; Svendsen et al., 2009).
Cross-reactions of the IFA kits towards other bacteria using
antibody-positive sera were tested. Antibodies to Coxiella burnetii
were found to be the most cross-reacting with the rickettsial antigens:
4 samples out of 10 were found positive, using the newly described
cut-off value (1:512). This was to be expected due to the relative
phylogenetic proximity of the two species; yet to our knowledge, only
one case of serological cross-reaction between Coxiella burnetii and
Rickettsia spp. has been described to date: a 25-year-old female with
Coxiella burnetii endocarditis who showed serological cross-reactions
towards R. rickettsii (Graham et al., 2000). One sample antibody
positive (IgA, IgM and IgG) for Campylobacter jejuni was found SFG
positive in both kit A and kit B, according to our new cut-off adjusted
to the Danish population. One sample highly antibody-positive for
Bartonella henselae was found highly positive against both SFG and
TG rickettsiae in both kits. This possible cross-reaction between Bar-
tonella and Rickettsia has only been described once in the literature,
confirmed by cross-absorption of antibodies (Takeda et al., 2007).
The present study had the following strengths. We performed an
evaluation of both kits with regard to both positive and negative
samples, donor samples and potentially cross-reacting samples. Our
study is the first of its kind in rickettsial serology and we used a
relatively large sample size. Previous studies on rickettsial serology
have either used in-house assays or not taken into account the
necessity of determining a local cut-off. This has probably resulted in
some over-estimation of the true prevalence of rickettsial disease in
certain studies. Our study also suffers from certain weaknesses. We
did not take issues such as pricing and readability of the slides into
account in the present study. This must of course be done when
introducing a new assay as a routine procedure. Our positive samples
were chosen from a bank of sera previously tested with the Weil-Felix
assay. This gives a possible bias in the selection, as the Weil-Felix test
is not a gold standard in rickettsial serology. We would have preferred
to have had information on culture and PCR on the patients as well,
but these have only recently been introduced as diagnostic modalities
Conclusively, the cut-off values recommended by the manufac-
turers did not apply for the general population in Denmark and
resulted in many false positive results. When we determined our own
cut-off value at 1:512, based on healthy blood donors, there was no
significant difference between the two kits. Furthermore, at this cut-
off level cross-reactions to antibodies from other infections were
The study was partly sponsored by unrestricted grants to Bjørn
Kantsø and Claus Bo Svendsen from Civil Engineer Johannes E.
Ormstrup and wife Grete Ormstrup's Fund, Aase & Einar Danielsen's
Fund, Erik Birger Christensen's Fund, and The Danish Microbiological
Society. The sponsors of the study had no influence on the design of
Amano, K., Hatakeyama, H., Okuta, M., Suto, T., 1992. Serological studies of antigenic
J.Clin.Microbiol. 30, 2441–2446.
typhi and Rickettsia prowazekii and their chemical similarity to the lipopolysaccharide
from Proteus vulgaris OX19 used in the Weil-Felix test. Infect. Immun. 66, 923–926.
Anton, E., Font, B., Munoz, T., Sanfeliu, I., 2003. Clinical and laboratory characteristics of 144
Baxter, J.D., 1996. The typhus group. Clin. Dermatol. 14, 271–278.
Dantas-Torres, F., 2007. Rocky Mountain spotted fever. Lancet Infect. Dis. 7, 724–732.
Fournier, P.E., Jensenius, M., Laferl, H., Vene, S., Raoult, D., 2002. Kinetics of antibody
Graham, J.V., Baden, L., Karchmer, A.W., 2000. Q fever endocarditis associated with
extensive serological cross-reactivity. Clin. Infect. Dis. 30, 609–610.
Kaplan, J.E., 1986. The sensitivity of various serologic tests in the diagnosis of Rocky
Mountain spotted fever. Am J Trop Med Hyg. 35, 840–844.
Keysary, A., 1997. Use of enzyme-linked immunosorbent assay techniques with cross-
La Scola, B.,1997. Laboratory diagnosis of rickettsioses: current approaches to diagnosis
of old and new rickettsial diseases. J. Clin. Microbiol. 35, 2715–2727.
La Scola, B., Raoult, D., 1999. Serologic diagnosis of rickettsiosis. Rickettsiae and
Newhouse, V.F., Shepard, C.C., Redus, M.D., Tzianabos, T., 1979. A comparison of the
complement fixation, indirect fluorescent antibody, and microagglutination tests for
the serological diagnosis of rickettsial diseases. Am. J. Trop. Med. Hyg. 28, 387–395.
and molecular evidence of Rickettsia helvetica in Denmark. Scand. J. Infect. Dis. 36,
Nilsson,K., Pahlson, C., Lukinius, A., Eriksson,L., Nilsson, L., Lindquist, O., 2002. Presenceof
Philip, R.N., Casper, E.A., MacCormack, J.N., Sexton, D.J., Thomas, L.A., Anacker, R.L.,
Burgdorfer, W., Wick, S., 1977. A comparison of serologic methods for diagnosis of
Rocky Mountain spotted fever. Am. J. Epidemiol. 105, 56–67.
of rickettsiae of the spotted fever group by microimmunofluorescence. J. Immunol.121,
Raoult, D., Roux, V., 1997. Rickettsioses as paradigms of new or emerging infectious
diseases. Clin. Microbiol. Rev. 10, 694–719.
Skarphedinsson, S., Lyholm, B.F., Ljungberg, M., Sogaard, P., Kolmos, H.J., 2007. Detection
and identification of Anaplasma phagocytophilum, Borrelia burgdorferi, and Rickettsia
helvetica in Danish Ixodes ricinus ticks. APMIS 115, 225–230.
Svendsen, C.B., Krogfelt, K.A., Jensen, P.M., 2009. Detection of Rickettsia spp. in Danish
ticks (Acari: Ixodes ricinus) using Real Time PCR. Scand J Inf Dis. 41, 70–72.
Svendsen, C.B., Milman, N., Høier-Madsen, M., Dziegiel, M.H., 2008. Determination of
rickettsial and antinuclear antibodies in Danish patients with sarcoidosis. The
Clinical Respiratory Journal 2, 202–207.
Takeda, N., Ishiwada, N., Fukasawa, C., Furuya, Y., Tsuneoka, H., Tsukahara, M., 2007.
Pediatric pneumonia, pleural effusion, and pericarditis following cat scratch disease
and serological cross-reactions among Bartonella henselae and Rickettsia japonica
determined by indirect fluorescence antibodies. Kansenshogaku Zasshi 81, 206–209.
Walker, T.S., 1984. Rickettsial interactions with human endothelial cells in vitro:
adherence and entry. Infect. Immun. 44, 205–210.
Walker, D.H., Ismail, N., 2008. Emerging and re-emerging rickettsioses: endothelial cell
infection and early disease events. Nat. Rev. Microbiol. 6, 375–386.
Znazen, A., Rolain, J.M., Hammami, A., Jemaa, M.B., Raoult, D., 2006. Rickettsia felis
infection, Tunisia. Emerg. Infect. Dis. 12, 138–140.
B. Kantsø et al. / Journal of Microbiological Methods 76 (2009) 285–288