Journal of Medical Virology 70:228–239 (2003)

Rapid and Quantitative Detection of Human

Adenovirus DNA by Real-Time PCR
Albert Heim,* Carmen Ebnet, Gabi Harste, and Patricia Pring-Åkerblom
Institute for Virology, German National Reference Laboratory for Adenoviruses, Hannover Medical School, Germany

Rapid diagnosis of human adenovirus (HAdV) their 51 types are associated with a variety of diseases
infections was achieved by PCR in the recent affecting all organ systems [Wadell et al., 1999]. For
years. However, conventional PCR has the risk of example, acute respiratory diseases in small children
carry-over contamination due to open handling are often caused by the types HAdV-1, -2, and -5 of the
with its products, and results are only qualitative. species HAdV-C, and these viruses persist as asympto-
Therefore, a quantitative ‘‘real-time’’ PCR with matic infections of the adenoids for months and years
consensus primer and probe (dual fluorescence with occasional virus shedding. Severe pneumonia in
labelled, ‘‘TaqMan’’) sequences for a conserved young military recruits is caused by the types HAdV-4
region of the hexon gene was designed and eva- (species HAdV-E) and HAdV-7 (species HAdV-B). The
luated. Real-time PCR detected all 51 HAdV proto- latter two types are also frequent causative agents
types. Sensitivity of the assay was
15 copies/run of pharyngoconjunctival fever (PCF), as is HAdV-3
and the linear range of quantitation 1.5  101 to (species HAdV-B), which is also associated with menin-
1.5  108 copies/run. TaqMan PCR gave identical gitis. Epidemic keratoconjunctivitis (EKC), a more
results compared to an established conventional severe and highly contagious eye infection, is predomi-
one-step PCR protocol in 218 (38 positive and 180 nantly caused by HAdV-8, -19, and -37 (species HAdV-
negative) of 234 clinical samples including blood, D). HAdV-19 and -37 are also sexually transmitted and
serum, eye swabs, and feces, and had divergent associated with ulcerative genital lesions, cervicitis,
results in 16 samples (15 positive only in TaqMan and urethritis. Cystitis is caused by HAdV-11, -21, -34,
PCR, all with low copy numbers, and one positive and -35 (species HAdV-B), and these types tend to
only in conventional PCR), indicating a higher sen- persist as asymptomatic infections of the urinary tract.
sitivity of TaqMan PCR. Adenovirus viremia was HAdV-40 and HAdV-41 (species HAdV-F) are consid-
detected by TaqMan PCR in 4 of 27 (14.8%) ered to be second to rotaviruses as a cause of gastro-
paediatric and 8 of 93 (8.6%) adult stem cell enteritis in young children. HAdV-31 (species HAdV-A)
transplant recipients but only in 5 of 306 healthy has been isolated from infants with gastroenteritis in
controls (blood donors, 1.6%). Virus loads of small outbreaks.
pediatric patients (median 1.7  105) were signifi- Recently, several studies have demonstrated the
cantly higher than in adult patients (median clinical significance of life-threatening HAdV infections
2.3  103) and than in controls (all samples
1.7  in immunocompromised hosts, for example, hemato-
103 copies/ml). A few immunosuppressed chil- poetic stem cell transplant recipients [Hierholzer, 1992;
dren had very high virus loads (up to 1.1  1010 Echavarria et al., 1999; Howard et al., 1999]. In im-
copies/ml), which were associated with symp- munodeficient hosts, HAdV infections are frequently
toms of disseminated disease. In conclusion, real- disseminated as a ‘‘sepsis’’ and are associated with one or
time PCR is a sensitive and quantitative procedure several organ manifestations as pneumonia, hepatitis,
for the detection of adenovirus infections. J. Med. meningitis, diarrhea, rash, and cystitis. Considering the
Virol. 70:228–239, 2003. ß 2003 Wiley-Liss, Inc. HAdV types (HAdV-1, -2, -5, -11, -34, and -35) frequently
detected in immunosuppressed bone marrow transplant
KEY WORDS: immunosuppression; disse- recipients, it seems probable that adenovirus disease
minated infection; virus load;
Adenoviridae; polymerase
*Correspondence to: Dr. Albert Heim, Institut für Virologie,
chain reaction OE5230, Medizinische Hochschule Hannover, Carl-Neuberg-
Strasse 1, D-30625 Hannover, Germany.
E-mail: ahei@virologie.mh-hannover.de
INTRODUCTION Accepted 8 January 2003
DOI 10.1002/jmv.10382
The six species of Human Adenoviruses (HAdV, Published online in Wiley InterScience
Genus Mastadenovirus, Family Adenoviridae) with (www.interscience.wiley.com)

ß 2003 WILEY-LISS, INC.

Quantitative Adenovirus PCR 229

starts by generalisation of a persistent asymptomatic that the cloned HAdV-2 sequence was identical to the
(‘‘latent’’) infection of the adenoids or the urogenital HAdV-2 prototype Genbank sequence (number J01917).
tract. Concentration of the plasmid was determined by
Virus isolation is the gold standard for diagnosis of photometry at 260 nm and calculated as genome
HAdV infections and permits subsequent typing of the equivalents (copies)/ml as the molecular weight of the
isolate. However, it may take up to 3 weeks until a plasmid was known. A serial dilution of the HAdV
cytopathic effect (CPE) develops, and some HAdV types quantitation standard plasmid was stored at 208C in
are fastidious with slow and ineffective growth in aliquots to avoid repeated freeze/thaw cycles. For
culture or require special cell lines such as 293 Graham testing a panel of HAdV types, A549 cells (Graham 293
cells for isolation. For these reasons, several groups cells in case of HAdV-40 and HAdV-41) were infected
have developed PCR protocols for the detection of HAdV with HAdV prototype strains. At about 50% CPE, cells
in clinical samples [Allard et al., 1990; Pring-Akerblom were freeze-thawed and DNA was extracted from 200 ml
and Adrian, 1994; Echavarria et al., 1998]. Although of the lysate with the Qiagen Blood kit (Qiagen, Hilden,
PCR was a major improvement for the rapid and re- Germany). Prototype virus stocks (types HAdV-1, to -21,
liable diagnosis of localised HAdV infections, detection -23, -25, -27, -28, -30 to -41, -43) of the German national
of adenovirus DNA in blood samples of immunosup- reference laboratory and of the American Type Culture
pressed patients has only a low predictive value for Collection (Manassas, VA) (HAdV-3, -5, -7, -12, -18, -22,
disseminated adenovirus disease as occasional HAdV -24, -26, -29, -30, -35, -36, -42, -44 to -49 and the proposed
DNA-emias are also observed in healthy, persistently types -50 and -51) were included in the panel (for some
infected individuals [Horvath et al., 1986; Carrigan, HAdV types, prototype strains from both origins were
1997; Flomenberg et al., 1997]. Quantitation of virus tested). HAdV-2, -3, -4, -5, -7, -11, -21, -31, -34, -41, and
load in blood samples may solve this diagnostic problem -48 were quantified as TCID50 (50% tissue culture
[Lankester et al., 2002]. Real-time detection of PCR infective dose)/ml and tested by PCR in serial dilutions
amplicons during thermal cycling by hybridisation to evaluate the sensitivity of detection and precision of
with a double fluorescence labelled probe (‘‘TaqMan’’ quantitation for HAdV types with varying numbers
principle) which is cleaved by the 50 -30 exonuclease of mismatches in the binding sites of the primers and
activity of Taq DNA polymerase permits simple and the probe.
reliable quantitation of viral nucleic acids as the number
of amplification cycles required for increase of fluores- Clinical Samples
cence over a threshold level is correlated with the
DNA was extracted from clinical samples as blood,
template concentration [Holland et al., 1991; Kawai
serum, feces, sputum, urine, eye swabs, and nasophar-
et al., 1999; Ryncarz et al., 1999; Weinberger et al., 2000;
yngeal swabs with the Qiagen Blood kit (Qiagen, Hilden,
Stamey et al., 2001]. Although the TaqMan principle
Germany). In case of liquid samples, 200 ml sample was
was adapted successfully for the diagnostic detection of
used according to the manufacturer’s instructions.
various human pathogenic viruses by PCR [Morris et al.,
Urine samples with a volume >200 ml were concentrat-
1996; Hawrami and Breuer, 1999; Kawai et al., 1999;
ed by centrifugation through Ultrafree-15 Biomax
Laue et al., 1999; Mercier et al., 1999; Ryncarz et al.,
100 filters (Millipore, Eschborn, Germany) prior to
1999; Schweiger et al., 2000; Weinberger et al., 2000;
DNA extraction. For the detection of HAdV viremia in
Stamey et al., 2001], the ‘‘generic’’ detection of all six
immunosuppressed bone marrow transplant recipients
HAdV species has not yet been described, probably due
and a control group of blood donors the extraction
to the high sequence diversity of HAdVs, which makes it
protocol was modified as follows in order to increase the
difficult to find a well-conserved sequence for probe
concentration of the purified DNA: elution of DNA was
hybridisation flanked by appropriately well-conserved
carried out with 50 ml instead of 200 ml of the provided
sequences for primer binding. The application of the
elution buffer (prediluted 1:5 with nuclease free water).
TaqMan principle is described for the generic detection
DNA was extracted from cerebrospinal fluid samples
of all 51 HAdV types and the evaluation of this method
with the Qiagen viral RNA kit.
for diagnostic purposes, including the quantitative de-
tection of HAdV in blood samples of immunosuppressed
Conventional PCR
patients, is reported.
Two hundred thirty-four patient samples were tested
MATERIALS AND METHODS by real-time and conventional PCR protocols. DNA was
extracted from the sample as described above, and the
HAdV Quantitation Standard
eluate that contained DNA was divided for conventional
and Panel of Viruses
and TaqMan PCR. Conventional adenovirus PCR pro-
For preparing a positive control standard, a HAdV-2 tocol was performed with the generic primers hex1deg
PCR amplicon (nt. 18856–19137 of the HAdV-2 se- and hex2deg and cycling conditions as described
quence) was cloned in a pGEM-T Easy plasmid vector previously [Wadell et al., 1999; Allard et al., 2001]. The
(Promega, Madison, WI). Plasmid DNA was purified protocol was modified slightly by using a ready made
from E. coli using the Nucleobond 100 kit (Macherey and master mix with a ‘‘Hot-start’’ DNA polymerase (Qiagen
Nagel, Germany), and it was confirmed by sequencing HotStarTaq Master Mix). In addition, all samples with
230 Heim et al.

divergent results between TaqMan and conventional 0.58C/sec between the annealing and extension step.
PCR were retested with another conventional PCR After initial evaluation of TaqMan PCR, the number of
protocol using the generic adenovirus primer set Ad-1 cycles was reduced to 40 for the clinical evaluation of the
and Ad-2 as described by Xu et al. [2000]. After ampli- assay. Fluorescence data was acquired at the end of each
fication, gel electrophoresis of PCR products (10 ml) was extension step in the channel F1 of the LC instrument
performed on 2% agarose gels, stained with ethidium (acquisition mode ‘‘single’’). The crossing point (CP),
bromide, and visualised by UV illumination. the cycle number corresponding to an increase of the
fluorescence over a threshold, was calculated automa-
Molecular Typing of HAdV tically by the lightcycler software (version 3.5c, settings:
proportional baseline adjustment, threshold ¼ baseline
A multiplex PCR that amplified the fiber gene region
þ6 SD of baseline, set point calculus with two points).
was carried out with HAdV DNA positive samples to
After the final cycle, the tubes were cooled down to 308C
achieve identification of the HAdV species [Xu et al.,
and disposed without opening the capillaries.
2000]. Amplicons of conventional PCRs were sequenced
directly with rhodamine-labelled dideoxynucleotide
HAdV Viremias in Immunosuppressed
chain terminators (DNA Sequencing Kit, ABI, Foster
Bone Marrow Transplant Recipients
City, CA) on an ABI-Prism 310 automatic sequencer.
and Healthy Controls
Both conventional PCR protocols amplify regions of the
hexon gene [Xu et al., 2000; Allard et al., 2001]. The Frequency of HAdV DNA detection and virus load was
sequences obtained permit identification of the HAdV compared by using whole blood (EDTA blood) samples
type in some cases using BLAST and FASTA programs that were sent in from three groups: (I) adult bone
(criteria: sequence homology >99% and next, less hom- marrow recipients (421 samples of 93 patients), (II)
ologous databank sequence <97% homology), if suf- pediatric bone marrow recipients (117 samples of 27 pa-
ficient sequence data is available in the Genbank tients), and (III) healthy adults as a control group
database. As hexon regions of several HAdV-D types (306 samples of 306 healthy donors). This control group
were not yet sequenced, sequencing does not permit consisted of adults who donated blood during the
identification of these HAdV types, but by multiple study period (February 2001 to July 2002) at the local
alignments with databank sequences and clustering blood bank. In the two patient groups, all EDTA blood
identification is easily achieved at the species level. samples were included that were available to us during
the study period (February 2001 to July 2002) and
TaqMan PCR originated from immunosuppressed bone marrow trans-
plant recipients (allogeneic and autologous). In some of
Taqman PCR was carried out with help of the
these patients a complete follow-up for several weeks
LightCycler (LC, Roche Diagnostics, Mannheim, Ger-
after BMT was possible; in other patients only a few
many) in sealed glass capillaries with a total reaction
samples or a single sample immediately after BMT or
volume of 20 ml. The FastStart Hybridization Kit (Roche
before leaving the BMT ward were available.
Diagnostics) was used to prepare a PCR master mix.
Adenovirus specific primer sequences were as follows:
Statistical Analysis
50 -GCC-ACG-GTG-GGG-TTT-CTA-AAC-TT-30 , Adeno-
quant 1 (AQ1) and 50 -GCC-CCA-GTG-GTC-TTA-CAT- The frequencies of HAdV viremia in bone marrow
GCA-CAT-C-30 , Adenoquant 2 (AQ2). The sequence of transplant recipient groups (93 adults and 27 children)
the probe was 50 -TGC-ACC-AGA-CCC-GGG-CTC-AGG- were compared with the control group (306 blood donors)
TAC-TCC-GA-30 (Adenoprobe, AP) with FAM labelled as using Fisher’s exact test (two-sided). Differences in
a fluorescent dye on the 50 end and TAMRA as a virus loads of these groups were tested using the Mann-
fluorescence quencher dye labelled to the 30 end. All Whitney test (two-sided). This nonparametric test was
oligonucleotides were synthesized, labelled and purified used because on nonequal variances and nonequal
by Eurogentec (Seraing, Belgium). The probe, the sample sizes of groups.
primers, and MgCl2 were added to the master mix to
achieve a final concentrations of 0.4 mM, 0.5 mM (each RESULTS
primer), and 3 mM, respectively. Heat labile Uracil-
Design of Primers and Probe
DNA Glycosylase (UNG, 1 U/reaction, Roche, Man-
nheim, Germany) was added to the master mix; 8 ml of Oligonucleotide primers were designed to achieve
the master mix and 12 ml of the DNA template were DNA amplification of all 51 types of the genus HAdV. As
added in each capillary. Sealed capillaries were centri- nucleic acid sequences of human adenoviruses are
fuged in a microcentrifuge and placed into the LC. highly divergent, all five completely sequenced human
Reaction conditions were 5 min 358C for Uracil-DNA- adenoviruses, type 2 (species HAdV-C, Genbank
Glycosylase incubation followed by 958C for 10 min to #J01917), type 5 (species HAdV-C, #M73260), type 12
activate the ‘‘hot start’’ Taq-polymerase. Forty-five (species HAdV-A, #X73487), type 17 (species HAdV-D,
cycles that consisted of denaturation at 958C for 3 sec, #AF108105), and type 40 (species HAdV-F, #L19443),
annealing at 558C for 10 sec, and extension at 658C for were aligned using the clustalX software (version 1.8)
60 sec were performed with a temperature increase of [Thompson et al., 1994]. Several highly conserved
Quantitative Adenovirus PCR 231

regions, which may allow annealing not only of the (four of ten runs) and higher dilutions were negative, as
primers but also of the labelled probe, were found in the were negative controls containing only distilled water
hexon gene. As additional hexon sequence data of (Table I). The serial dilution of plasmid DNA was also
several HAdV types are available, a multiple alignment used for determining the dynamic range of quantitation.
of the hexon gene including sequences of all 6 HAdV Regression analysis of the crossing points vs. log HAdV
species was generated and used for primer design DNA concentration resulted in a very high correlation
(Fig. 1). Amplification primers were selected according coefficient (0.99 to 1.0) for the concentration range
to the guidelines set up for Taqman PCR [N.N.], between 1.5  101 to 1.5  108 copies HAdV DNA. The
spanning a third highly conserved region that served resulting regression lines (n ¼ 10) had a slope of 3.50
as a putative binding site for the probe. In spite of the (SD ¼ 0.075), indicating that a 10-fold increase of
selection of conserved regions, there are minor sequence the template (quantitation standard) concentration is
diversities in the binding sites of the primers. Consensus directly related to an increase of 3.5 cycles in the CP.
primer sequences were designed to balance the max- Amplification efficiency of the PCR was high with a
imum number of mismatches by calculating the melting value of 1.93 (calculated with the formula: effi-
temperatures for the interaction of both primers with ciency ¼ 10-1/slope, theoretical maximum 2.0 that stands
each adenovirus sequence with help of the Metcalc for a doubling of the DNA with each PCR cycle).
software (Fig. 1) [Schütz and von Ahsen, 1999]. Intraassay variability (standard deviation of CP values)
Successful amplification of all 51 HAdV prototypes was low, for example 2.7% for 1.5  107 copies/run and
(including the proposed types 50 and 51) using the 1.1% for 1.5  104 copies/run. Interassay variability was
primers AQ1 and AQ2 in a conventional PCR indicated slightly higher; Table I gives the average crossing point
that the primer set was capable to amplify all human (CP) and standard deviation (SD) of the CP determined
adenoviruses (data not shown). Compared to conven- in ten different runs on different days. In each experi-
tional PCR, real-time detection of PCR amplicons with a ment, HAdV DNA concentrations of the plasmid serial
Taqman probe requires an almost complete hybridisa- dilution were set as a standard and HAdV DNA
tion of a double fluorescence-labelled probe to achieve concentrations of each point were calculated automati-
digestion of the probe by the nuclease activity of Taq cally by the LC software (version 3.5c) under the assum-
polymerase. The design of the probe followed guidelines ption of a semi-logarithmic relation between crossing
set up for Taqman probes [N.N.], and the number of points and HAdV DNA concentration. Calculated HAdV
mismatches for binding to each adenovirus sequence of DNA concentrations and SD of calculated concentra-
the multiple alignment was minimised similar to the tions indicate a dynamic range of HAdV DNA quantita-
strategy used for the primers (Fig. 1). The melting tion of at least 6 logs (1.5  108 to 1.5  102 copies HAdV
temperatures of the probe and the primers for hybridi- DNA), whereas with lower template concentrations (for
sation to each adenovirus sequence were calculated and example 1.5  101 copies/run) the standard deviation
reaction conditions in real-time PCR were adjusted to increased and with even lower template concentrations
permit amplification and detection of probably all results were only qualitative (positive or negative) and
human pathogenic adenoviruses. Real-time PCR gave Poisson distributed (Table I). As quantitation of HAdV
positive results for all prototype strains of the genus DNA in clinical samples was the main goal of our study
HAdV including the isolated, proposed new types HAdV- and this goal was achieved with less than 40 amplifica-
50 and HAdV-51. CP values (defined as the cycle number tion cycles (Table I), it was decided to carry out only 40
with fluorescence increasing over background) were cycles for the evaluation of the assay with clinical
low (<20) for all prototype strains suggesting effective samples, restricting the sensitivity to about 1.5  101
amplification and sensitive detection. Moreover, 17 copies/run. Spiking of the HAdV-2 DNA serial dilution
clinical isolates [HAdV-1 (two isolates), -2 (three with human genomic DNA (500 ng/run) did not interfere
isolates), -3 (two isolates), -4 (three isolates), -5 (two with the sensitivity and quantitation of HAdV DNA
isolates), -7 (two isolates), -31, -34, and -41] were tested detection as the increase of fluorescence over threshold
by TaqMan PCR, and all isolates were positive with low (CP values) was unchanged by hDNA. In contrast to CP
CP values (<20), indicating efficient amplification. Real- values, end point fluorescence data indicated some
time PCR was always negative with 100 ng DNA negative interference of 500 ng hDNA with amplifica-
(hDNA) extracted from cultured human cells (MRC5) tion of HAdV DNA. Serial dilutions 11 HAdV serotypes
or blood of an healthy adult as a template (n ¼ 38). representing the six HAdV species were used to deter-
mine whether HAdV types with more mismatches to the
consensus primer and probe sequences than HAdV-2
Sensitivity and Dynamic Range of Quantitation
were amplified as efficiently as HAdV-2 and quantified
A plasmid DNA containing a partial HAdV 2 hexon precisely using the HAdV-2 plasmid as a quantitation
sequence was serially diluted and used as a template for standard (Fig. 2, Table II). For example, the HAdV-3 and
real-time PCR (1.5  108 to 1.5  101 HAdV 2 genome HAdV-7 have mismatches comparatively close to the 30
equivalents (copies)/run) in ten repeated runs to estab- end of primer AQ1 and HAdV-41 has three mismatches
lish the sensitivity of TaqMan PCR. As few as 1.5  101 compared to the sequence of primer AQ1 which may
copies were reliably detected by real-time PCR (n ¼ 10), result in inefficient amplification (Fig. 1). Nevertheless,
whereas 1.5  100 copies were only occasionally detected the sensitivity of detection of these viruses was in the
232 Heim et al.

Fig. 1. Multiple alignment of partial hexon sequences of various Accession numbers are HAdV-2 (NC 001405), HAdV-3 (X76549),
HAdV-types and the consensus sequences of the TaqMan PCR primers HAdV-4 (AF06062), HAdV-5 (NC 00146), HAdV-7 (Z48571), HAdV-12
(AQ1 and AQ2) and the probe AP. Melting temperatures were (AF065065), HAdV-16 (X74662), HAdV-17 (AF108105), HAdV-21
calculated for the binding of AQ1, AQ2, and AP to each sequence con- (AY008279), HAdV-34 (AB052911), HAdV-40 (L19443), and HAdV-
sidering the mismatches between the consensus sequence and each 41 (M21163).
HAdV sequence. Base numbering according to the HAdV-2 sequence.
Quantitative Adenovirus PCR 233

TABLE I. Interassay Variability of HAdV TaqMan PCR Determined in Different Runs on

Different Days

Genome Proportion Crossing point SD Calculated concentration SD

equivalents positive (mean) % (mean) %
1.5  108 10/10 15.47 4.30 1.91  108 15.66
1.5  107 10/10 19.27 4.09 1.56  107 13.88
1.5  106 10/10 22.96 3.07 1.39  106 15.58
1.5  105 10/10 26.59 2.27 1.26  105 11.25
1.5  104 10/10 29.93 2.03 1.47  104 32.73
1.5  103 10/10 33.37 1.77 1.49  103 28.13
1.5  102 10/10 37.23 2.03 1.22  102 35.13
1.5  101 10/10 39.63 2.97 2.43  101 49.95
1.5  100 4/10 >41 nd* nd* nd*
1.5  101 0/10 nd* nd* nd* nd*
*nd, not determined.

same range as for HAdV-2, and quantitation of all and bands of conventional PCR amplicons were faint;
viruses using the HAdV-2 plasmid standard varied thus direct sequencing was not successful. In one of
less than 1 log (Table II). Slopes (CP vs. log TCID50 these four samples, multiplex PCR was positive, in-
concentration) were also similar, indicating almost dicating a species HAdV-D. In 3 EDTA blood samples,
equally efficient amplification of different HAdV types only the TaqMan PCR gave a positive result indicating
(Table II). low copy viremia (1  103, 1.1  103, and 2.2  103 copies/
ml blood) and conventional PCR and multiplex PCR
were negative. Sixty serum and plasma samples were
Evaluation of HAdV TaqMan PCR
tested, 57 had a concordant result (2 positive and 55
in Comparison to Conventional PCR
negative), and 3 were positive only with the TaqMan
Two hundred thirty-four clinical specimens [58 assay. One of the concordantly positive samples had a
EDTA-blood, 60 serum and plasma, 21 throat washes, virus load of 4.5  103 copies/run (about 9.0  104 copies/
5 combined nasopharynx swabs, 17 eye swabs, 26 ml blood); this sample originated from an immunosup-
cerebrospinal fluid, 22 fecal samples, 12 bronchoalveo- pressed bone marrow transplanted child (‘‘patient A’’)
lar lavage and tracheal aspirates, and 13 other materi- described in detail later. All other positive serum and
als (as pericardial, pleural, and peritoneal fluids, urine), plasma samples had low virus concentrations (3.6  102,
and biopsies from lymph nodes and bowel] were tested 3.8  102, and 4.8  102 copies/ml serum), including the
for HAdV DNA both by TaqMan PCR and an establish- other serum sample concordantly positive in both assays
ed conventional diagnostic PCR protocol. Specimens had (originating from an immunocompetent male infant
been collected either from patients who were admitted to suffering from conjunctivitis and myocarditis). Twenty-
the Hannover Medical School clinical centre with four of 26 CSF samples had concordant results in both
suspected adenovirus infection or were sent in from all PCR protocols (1 positive with 1.6  104 copies/ml CSF,
over Germany to the national adenovirus laboratory for 23 negative). Two CSF samples were only positive by
diagnostic purposes. After extraction of DNA, samples TaqMan PCR; however virus loads were low (8.1  102
were divided into aliquots that were used for TaqMan and 2.5  103 copies/ml CSF). These patients suffer-
PCR, conventional PCR, and in case of positive samples ed from aseptic meningitis and meningo-encephalitis,
also for identification of HAdV species by multiplex PCR. and no other infectious agent was identified in these
In 218 samples, both PCRs had a concordant result (38 samples (HSV, VZV, CMV, and enterovirus tested by
positive samples and 180 negative samples). Sixteen PCR; bacterial cultures were all negative). All 21 throat
samples had divergent results between both assays. In wash samples had concordant results (8 positive and
15 of these samples real-time PCR was positive with 13 negative). Virus loads in concordant throat wash
high CP values (CP <37, which equals approximately samples were about 3.3  102 to 3.2  105 copies/run.
<150 copies HAdV DNA/run), whereas conventional Because the quality of throat wash samples varies
PCR was negative, indicating a higher sensitivity of widely, quantitation is not feasible and results were not
TaqMan PCR compared to the conventional PCR pro- calculated as copies/ml. In five of these samples typing
tocol. Only one sample (EDTA-blood) was positive in the was achieved by nucleic acid sequencing (HAdV-3,
conventional PCR protocol (faint band on the agarose HAdV-4, and HAdV-5 in three samples). All naso-
gel) but negative by real-time PCR. This result was pharynx swabs had concordant results (1 positive and
confirmed by multiplex PCR and the virus turned out 4 negative). The positive sample had a high virus DNA
to be species HAdV-D. Of the 57 other EDTA blood load (>1.5  108 copies/run), and was typed as HAdV-3
samples, 54 had concordant results (4 positive and 50 by nucleic acid sequencing. Absolute quantitation of
negative). Concentrations detected by TaqMan PCR HAdV copies in swabs is not a sensible strategy as the
were low in these 4 cases (3.3  102 to 8.4  102 copies/ material is not homogenous and varies widely. Fourteen
run equivalent to 6.6  103 to 1.6  104 copies/ml blood) of 17 eye swabs had a concordant result (9 positive and
234 Heim et al.

BLAST and FASTA alignments of the nucleic acid

sequence (Genbank #AJ 344624) as a homologous
sequence was not in the database. However, by multiple
alignments and clustering, the sequence was identified
as a species HAdV-D sequence. This result was con-
firmed by multiplex PCR. In two conjunctival swab
samples with lower virus DNA concentrations (about
3.7  104 and 1.8  102 copies/run), HAdV-7 was found,
and in two other samples with concordant positive
results typing failed. Three of 17 eye swab samples were
primarily positive by TaqMan PCR only (3.1  102,
1.5  102, and 3.4  101 copies/run); however retesting
with another generic primer pair in a conventional PCR
protocol [Xu et al., 2000] gave positive results and
nucleic acid sequencing of the amplicons demonstrated
HAdV-7. Twenty-two fecal samples were tested in
parallel by TaqMan PCR and conventional PCR giving
identical results in 21 samples (10 positive and 11
negative). Quantitative values were not calculated per g
fecal sample as feces is not a homogenous material.
Three of the positive samples had very high virus loads
with crossing points <15, indicating virus loads
>1.5  108 copies/run and HAdV-40, HAdV-41 and
HAdV-1 (sample of immunosuppressed patient ‘‘A’’)
were identified by nucleic acid sequencing. HAdV-41
was also found in another sample with a high virus load
of 7.6  105 copies/run, whereas virus loads in two faecal
samples with HAdV-7 sequences were lower (1.2  104
and 2.6  104 copies/run). In four fecal samples with
concordantly positive results but lower virus loads
(4.2  102 to 2.6  104 copies/run), we were unable to
type the virus. One fecal sample had low HAdV DNA
loads (8.6  101 copies/run) by TaqMan PCR and was
negative by conventional PCR. Ten of 12 bronchoalveo-
lar lavage samples had concordant results (2 positive
and 8 negative) but virus loads were low in positive
samples (<1  103 copies/run); one of these samples
contained HAdV-D. Two other bronchoalveolar lavage
samples were only positive by TaqMan PCR, and HAdV
DNA concentrations were very low. Eleven of 13 sam-
ples from various other sites were concordantly negative
by both PCR assays. A lymph node biopsy was positive
by both assays with a DNA concentration of 1.7  103
copies/run and identified as HAdV-D by multiplex PCR.
Moreover, a peritoneal effusion sample was only positive
by TaqMan PCR with a low HAdV DNA concentration.
Fig. 2. Comparison of amplification plots of 10-fold serial dilutions of
(A) the quantitation standard plasmid DNA (HAdV-2 sequence,
1.5  108 copies/run, approximately equivalent to 3  109 copies/ml of Frequency and DNA Load of HAdV Viremias
a virus stock considering the DNA extraction procedure used for virus
stocks) and (B) DNA extracted from a HAdV-2 virus stock with HAdV-2 in Immunosuppressed Bone Marrow
(4.3  108 TCID50/ml, approximately equivalent to 4.3  109 copies/ml Transplant Recipients
considering the usual particle/TCID50 ratio) and (C) DNA extracted
from a HAdV-3 virus stock (5  108 TCID50/ml, approximately HAdV viremia was investigated comparatively with
equivalent to 5  109 copies/ml). Horizontal line indicates threshold of
fluorescence.
EDTA blood samples of in three groups of patients: adult
bone marrow recipients (421 samples of 93 patients),
pediatric bone marrow recipients (117 samples of 27 pa-
5 negative). In three samples with a high virus load tients), and healthy blood donors as a control group
(about 6.1  104, 1.1  106, and 5.5  106 copies/run), (306 samples of 306 donors). For a direct comparison of
HAdV-37 was typed by nucleic acid sequencing. In the frequency of HAdV viremias in immunosuppres-
another sample with a high virus DNA concentration sed adult bone marrow transplant recipients with the
(1.2  105 copies/run), identification of the HAdV type by control group (5 of 306 blood donor samples positive,
Quantitative Adenovirus PCR 235

TABLE II. Quantitation of Various HAdV Stock Virus Solutions Representing the Six
Species of Human Adenoviruses*

Sensitivity Quantitation Amplification efficiency

Type Species (TCID50/ml) (copies/TCID50) (slope)
2 C 4.3  101 38.8 1.93 (3.50)
3 B (B1) 5.0  101 30.8 1.95 (3.46)
4 E 6.3  101 10.8 1.87 (3.68)
5 C 1.5  100 20.8 1.84 (3.78)
7 B (B1) 6.3  100 17.2 2.0 (3.23)
11 B (B2) 7.5  100 13.0 1.98 (3.36)
21 B (B1) 6.3  101 12.0 1.98 (3.43)
31 A 6.3  100 138.0 1.98 (3.46)
34 B (B2) 7.5  100 90.0 1.98 (3.65)
41 F 3.2  101 11.4 1.98 (3.82)
48 D 2.7  100 20.4 1.93 (3.49)
*Approximate sensitivity (TCID50/ml) was determined by serial dilution of the HAdV stock virus.
Quantitation of each sample was performed using the HAdV-2 plasmid standard, and a ratio of the copy
number determined by PCR per TCID50 was calculated. Amplification efficiency (x-fold increase of DNA/
PCR cycle) was calculated using the slope value as described in Material and Methods.

1.6%), only the first sample obtained of each patient was in the control group (>2  104 copies/ml blood). Patient A
evaluated, with 8 of 93 samples being positive (8.6%, was a 5-year-old male child who had been liver
P ¼ 0.0030, Fisher’s exact test, two-tailed). Twenty- transplanted and bone marrow transplanted. On day
eight of 93 adult patients were HAdV positive in at least 41 after bone marrow transplantation, an EDTA blood
one sample (with 1 to 12 samples/patient available). sample was HAdV positive (5.8  103 copies/run) and a
If all samples of adult bone marrow recipients (40 of serum sample obtained 1 day later was also positive
421 samples positive, 9.5%) were compared with the (Fig. 4). At that time, leukocyte counts were normal but
blood donor samples, an even more clear difference was the patient suffered from diarrhea and a fecal sample
found (P < 0.0001, Fisher’s exact test). However, the was also positive for HAdV DNA. In the following days,
latter analysis of the data may be biased as there may HAdV DNA loads increased as depicted in Figure 4,
have been more samples available of bone marrow leukocyte counts dropped, and fever, cholestasis, and an
transplant recipients who had clinical complications
than from those who did well. HAdV viremia was also
observed more frequently in pediatric bone marrow
recipients compared to the adult blood donor control
group; 4 of 27 first samples were positive (14.8%,
P ¼ 0.0033, Fisher’s exact test). Eight of 27 pediatric
patients were HAdV positive in at least one sample (with
1 to 14 samples/patient available adding up to 117
samples). Pediatric samples were significantly more
often positive (21 of a total of 117, 18.8%) compared to
the adult blood donor control group (P < 0.0001), and
compared to all samples of adult bone marrow trans-
plant recipients (P ¼ 0.0137, Fisher’s exact test).
The highest viremia observed in a healthy, nonimmu-
nosuppressed blood donor of the control group was
1.7  103 copies/ml blood (Fig. 3). In adult bone marrow
transplant recipients the median virus load was 2.3 
103 copies/ml blood (range 3  102 to 5.8  105, mean
5,0  104, SD ¼ 1.4  105). Considering only the first
sample of each patient, the median virus load was
2.0  103 copies/ml blood (range 6.5  102 to 4.6  104,
mean 7.9  103, SD ¼ 1.5  104). Virus loads of pediatric
bone marrow transplant recipients (median 1.7  105
copies/ml blood, range 4  102 to 1.1  1010, mean
1.2  109, SD ¼ 3.9  109) were significantly higher com-
pared to blood donors (P ¼ 0.009, Mann-Whitney test) Fig. 3. HAdV DNA loads of immunosuppressed bone marrow
and adult bone marrow recipients (P < 0.001). Two transplant recipients compared to a control group of healthy blood
pediatric (A and B) and four adult bone marrow donors: (A) all samples of adult patients, (B) all samples of paediatric
patients, (C) control group of healthy blood donors, (D) only first
transplant recipients (C, D, E, F) presented with high samples of adult patients, and(E) only first samples of paediatric
virus loads at least one magnitude higher than observed patients.
236 Heim et al.

Allard et al., 1992; Hierholzer et al., 1993; Pring-

Akerblom and Adrian, 1994]. However, all these proto-
cols include labour intensive and time consuming steps
for the detection of PCR amplicons, as for example
ethidium bromide-stained gel electrophoresis or hybri-
disation procedures requiring open handling of PCR
products that bears a high risk of template carry over
contamination. In contrast, real-time detection of the
amplicon using the ‘‘TaqMan’’ principle during thermal
cycling avoids these drawbacks [Holland et al., 1991].
However, the ‘‘generic’’ detection of HAdV DNA had not
yet been achieved by TaqMan PCR, probably because of
the high genetic diversity of the 51 HAdV types [Wadell
et al., 1996, 1999]. So far, a real-time protocol for the
detection of only a single HAdV type had been published
in detail [Ma et al., 2001]. In the present study, real-time
Fig. 4. Time course of HAdV DNA loads of patient ‘‘A.’’
detection of HAdV DNA of all 51 HAdV prototypes was
achieved by a strategy of constructing consensus seq-
atypical pneumonia developed. HAdV DNA was identi- uences that balanced mismatches with various so far
fied as HAdV-1 by conventional PCR and nucleic acid sequenced HAdV types achieving similar melting temp-
sequencing from a blood sample and a fecal sample. eratures (Fig. 1) and setting reaction conditions accord-
Patient A died on day 75 post bone marrow transplanta- ingly. Using this approach, HAdV DNA of various types
tion of respiratory insufficiency and circulatory failure. (including HAdV-1, -3, -4, -5, -7, -34, -37, -40, -41, and
Patient B was a 15-months-old female child who under- various viruses of the species HAdV-D) was detected
went bone marrow transplantation for hemophagocytic directly in clinical samples (without virus isolation).
lymphohistiocytosis. Six consecutive positive blood Simple and rapid detection of HAdV DNA including
samples were obtained with a peak virus load of 3.6  types 4 and 7 as achieved by TaqMan PCR should also be
105 copies/ml when the patient suffered from fever and helpful in potentially life threatening cases of acute
diarrhoea. HAdV was also detected in stool samples and respiratory distress in young adults, especially military
typed as HAdV-2. Virus loads decreased subsequently recruits [CDC, 2001]. A panel of 234 clinical samples was
and the patient recovered. The four other patients with used to determine the sensitivity and specificity of real-
comparatively high HAdV loads did not present with time PCR in comparison to an established conventional
any symptoms of localised or disseminated HAdV PCR protocol. Real-time PCR detected HAdV DNA in
disease but in adult patient C high virus loads were 53 of 54 positive clinical samples, whereas the conven-
associated with increasing the immunosuppressive tional PCR protocol detected only 38 of 54 positive
therapy. The first three blood sample obtained over a samples. All samples positive in real-time PCR but
period of 17 days were HAdV negative or had low virus negative in conventional PCR had HAdV DNA concen-
loads (2.3  103 copies/ml blood). Meanwhile, immuno- trations <103 copies/run, indicating a slightly lower
suppressive therapy was increased by adding predniso- sensitivity of the conventional one step PCR proto-
lone because of a graft vs. host disease of the skin. Six col with the same number of cycles (40) as used in real
days later, the virus load had increased to 5.0  105 time PCR [Wadell et al., 1999; Allard et al., 2001]. The
copies/ml and two other samples obtained during the sensitivity of the conventional PCR can be enhanced
following 2 weeks had also high virus loads (5.8  105 by performing an additional nested PCR amplifica-
and 5.5  105 copies/ml). From adult patient D, eight tion [Allard et al., 2001]. However, we were afraid of
samples were received over a period of 28 days, only the cross-contamination problems, which are inherent to
fourth sample was HAdV positive with a virus load of nested PCR protocols, and therefore did not perform
3.9  104 copies/ml. A similar pattern was observed in nested PCR.
adult patient E; six samples were obtained over a period In general, cross-contamination is a serious risk for all
of 3 weeks with only the last sample positive (4.5  104 sensitive assays that may detect as few as a single copy
copies/ml). Adult patient F (nine samples available over of target DNA. Contamination by amplicon carry-over
a period of 28 days) was occasionally positive (samples 3 seems to be almost impossible by the TaqMan principle
and 6 with low virus loads of 9.5  102 and 8.3  102 itself, which avoids open handling of PCR products.
copies/ml) and sample 9 with a high virus load, 5.5  104 Furthermore, UNG treatment was undertaken and
copies/ml. guidelines for avoiding contamination were followed
[Kwok and Higuchi, 1989]. However, cross-contamina-
tion during DNA extraction and PCR set-up cannot be
DISCUSSION
excluded completely, considering the highly positive
Various PCR protocols for the ‘‘generic’’ detection of feces and eye swab samples containing >108 copies
HAdV DNA are sensitive and rapid procedures for the HAdV DNA template/run. Therefore, a negative control
diagnosis of HAdV infections [Echavarria et al., 1998; (water) was extracted with each batch of clinical
Quantitative Adenovirus PCR 237

samples, this extraction control and two additional immunosuppressed patients. Six patients had virus
negative controls (water and hDNA) were amplified loads at least 10-fold higher than the highest HAdV
with each run. DNA extraction from feces and eye swab DNA concentration observed in healthy controls.
samples was performed in separate batches than from Although not a primary goal of the study, it was
blood, plasma, and serum samples. In a single batch of investigated whether high HAdV loads were associated
DNA extraction from fecal samples, the extraction with disseminated adenovirus disease, as in other virus
control was false positive, as were four of seven fecal infections of immunosuppressed patients, for example,
samples. Three of these samples and the extraction CMV, high DNA loads are associated with disease [Zaia
control had low HAdV concentrations of up to 5  102 et al., 1997; Ferreira-Gonzalez et al., 1999; Emery et al.,
copies/run, whereas the fourth fecal sample had a 2000]. Symptoms suggestive for a disseminated or loca-
very high HAdV concentration (>1.5  108 copies/run). lized HAdV infection were observed only in two of these
Nucleic acid sequencing demonstrated identical HAdV- six patients with high virus loads (patient A and B, both
41 sequences in these samples and the extraction children) but not in any other patients. Patient A had
control. DNA was extracted from the original fecal also a fatal outcome after presenting with extremely
samples a second time (only positive the high virus load high virus loads (>1  1010 copies/ml blood). Similar
sample) and run together with the original DNA ex- cases of fatal disseminated HAdV infection associated
traction samples (results as before), demonstrating a with very high virus loads were reported recently but
cross-contamination event during the DNA extraction the method used for HAdV quantitation was not
procedure. In future, automated DNA extraction may described in detail [Lankester et al., 2002]. Virus loads
minimise the risk of sample cross contamination. of patient A were so high that direct complement
Quantitative generic detection of HAdV DNA is not activation by HAdV-1 may be suspected as cause of
yet possible, and thus the virus load dynamics of human multi-organ failure and septic shock [Cichon et al.,
adenovirus infections are not known precisely. In this 2001]. However, it is not completely clear whether
study, it was demonstrated that very high HAdV HAdV was the only cause of the fatal outcome because
DNA concentrations were detected in highly contagious patient A survived high and increasing HAdV loads for
adenovirus diseases, for example, epidemic keratocon- at least 3 weeks and suffered also from a generalised
junctivitis caused by HAdV-37 and diarrhea caused by candida albicans infection. Therefore, further studies
HAdV-40 and -41, whereas less severe and contagious are required on HAdV loads in disseminated HAdV
eye infections as PCF caused by HAdV-7 were associated disease and the prognostic significance of high HAdV
with lower HAdV DNA concentrations in eye swabs. loads. Previously, the detection of HAdV from different
True quantitation of HAdV loads by quantitative PCR is body sites (e.g., in feces, throat washing, urine, and
feasible for clinical samples as blood, plasma, serum, blood) was proposed for the diagnosis of disseminated
and CSF, which are homogenous and can be sampled in a HAdV disease in immunodeficient patients but this
highly reproducible manner. Quantitation of HAdV approach seems to be quite awkward [Carrigan 1997;
DNA in blood samples is an interesting point consider- Howard et al., 1999]. More recently, qualitative detec-
ing the fact that HAdV DNA was detected in blood tion of HAdV DNA in serum was proposed as an indi-
samples of healthy persons with a widely varying cator of fatal disseminated HAdV infections [Echavarria
frequency. Using a nested PCR method very sensitive et al., 2001]. Although patient A was HAdV DNA
for HAdV-C (but less sensitive for HAdV-B2), HAdV positive in serum (9  104 copies/ml) and suffered from
DNA was detected in 1 of 73 healthy persons (about a disseminated disease with a fatal outcome, our data
1.4%) [Flomenberg et al., 1997], which is about the same does not support the high predictive value of qualitative
frequency as that observed by TaqMan PCR in this HAdV DNA detection in serum in general. For example,
study (1.6%), whereas in a previous study 13 of 17 two patients were HAdV DNA positive in serum with
lymphocyte samples were positive [Horvath et al., 1986]. low copy numbers and did not suffer from any HAdV
This study used a Southern blot hybridisation proto- disease.
col for HAdV-C DNA but the high incidence of HAdV There is no antiviral therapy for disseminated HAdV
DNA may have been caused by insufficient stringency infections, although in vitro activity against HAdV was
[Horvath et al., 1986]. In general, HAdV DNA detection demonstrated with various antiviral agents. Cidofovir
in blood samples of healthy persons is related to a latent and ribavirin are used occasionally in life-threatening
or persistent infection of lymphoid cells, which can also HAdV infections in immunosuppressed patients
be persistently infected by HAdV in vitro [Andiman and [Cassano, 1991; Sabroe et al., 1995; Ribaud et al., 1999;
Miller, 1982; Chu et al., 1992; Flomenberg et al., 1996]. Arav-Boger et al., 2000; Miyamura et al., 2000; Bordi-
Therefore, the clinical significance of qualitative HAdV goni et al., 2001, Legrand et al., 2001;] but controlled
DNA detection in blood is difficult to interpret in clinical trials have nit been carried out. Moreover,
immunosuppressed individuals who are at risk of a antiviral effects have not been demonstrated in a clinical
life-threatening disseminated HAdV infection. HAdV setting as it should be possible in future by following
DNA quantitation demonstrated now that HAdV DNA virus loads with quantitative PCR. Therefore, the
loads in healthy immunocompetent individuals are quantitative TaqMan PCR for HAdV holds promise of
low (
1.7  103 copies/ml blood), whereas HAdV DNA value in clinical studies on the antiviral therapy of life-
was observed with significantly higher virus loads in threatening HAdV infections.
238 Heim et al.

In conclusion, TaqMan PCR protocol permits rapid Flomenberg P, Gutierrez E, Piaskowski V, Casper JT. 1997. Detection
and quantitative detection of HAdV DNA in clinical of adenovirus DNA in peripheral blood mononuclear cells by
polymerase chain reaction assay. J Med Virol 51:182–188.
samples. Detection of high virus loads in blood holds Hawrami K, Breuer J. 1999. Development of a fluorogenic polymerase
promise for a simplified and earlier diagnosis of dis- chain reaction assay (TaqMan) for the detection and quantitation of
seminated adenovirus disease in immunosuppressed varicella zoster virus. J Virol Methods 79:33–40.
patients. Hierholzer JC. 1992. Adenoviruses in the immunocompromised host.
Clin Microbiol Rev 5:262–274.
Hierholzer JC, Halonen PE, Dahlen PO, Bingham PG,
ACKNOWLEDGMENTS McDonough MM. 1993. Detection of adenovirus in clinical speci-
mens by polymerase chain reaction and liquid-phase hybridization
We thank Dr. L. Grigull and Dr. B. Hertenstein for quantitated by time-resolved fluorometry. J Clin Microbiol 31:
1886–1891.
helpful discussion, Dr. C. Henke-Gendo and Dr. J. Holland PM, Abramson RD, Watson R, Gelfand DH. 1991. Detection of
Sheldon for critical reading of the manuscript, and G. specific polymerase chain reaction product by utilizing the 50 —-30
Scheller for technical assistance. exonuclease activity of Thermus aquaticus DNA polymerase. Proc
Natl Acad Sci USA 88:7276–7280.
Horvath J, Palkonyay L, Weber J. 1986. Group C adenovirus DNA
REFERENCES sequences in human lymphoid cells. J Virol 59:189–192.
Howard DS, Phillips IG, Reece DE, Munn RK, Henslee-Downey J,
Allard A, Girones R, Juto P, Wadell G. 1990. Polymerase chain reaction Pittard M, Barker M, Pomeroy C. 1999. Adenovirus infections in
for detection of adenoviruses in stool samples. J Clin Microbiol hematopoietic stem cell transplant recipients. Clin Infect Dis
28:2659–2667. 29:1494–1501.
Allard A, Albinsson B, Wadell G. 1992. Detection of adenoviruses in Kawai S, Yokosuka O, Kanda T, Imazeki F, Maru Y, Saisho H. 1999.
stools from healthy persons and patients with diarrhea by two-step Quantification of hepatitis C virus by TaqMan PCR: comparison
polymerase chain reaction. J Med Virol 37:149–157. with HCV Amplicor Monitor assay. J Med Virol 58:121–126.
Allard A, Albinsson B, Wadell G. 2001. Rapid typing of human Kwok S, Higuchi R. 1989. Avoiding false positives with PCR. Nature
adenoviruses by a general PCR combined with restriction endonu- 339:237–238.
clease analysis. J Clin Microbiol 39:498–505.
Lankester AC, van Tol MJ, Claas EC, Vossen JM, Kroes AC. 2002.
Andiman WA, Miller G. 1982. Persistent infection with adenovirus Quantification of adenovirus DNA in plasma for management of
types 5 and 6 in lymphoid cells from humans and woolly monkeys. infection in stem cell graft recipients. Clin Infect Dis 34:864–867.
J Infect Dis 145:83–88.
Laue T, Emmerich P, Schmitz H. 1999. Detection of dengue virus RNA
Arav-Boger R, Echavarria M, Forman M, Charache P, Persaud D. 2000. in patients after primary or secondary dengue infection by using
Clearance of adenoviral hepatitis with ribavirin therapy in a the TaqMan automated amplification system. J Clin Microbiol 37:
pediatric liver transplant recipient. Pediatr Infect Dis J 19:1097– 2543–2547.
1100.
Legrand F, Berrebi D, Houhou N, Freymuth F, Faye A, Duval M,
Bordigoni P, Carret AS, Venard V, Witz F, Le Faou A. 2001. Treatment Mougenot JF, Peuchmaur M, Vilmer E. 2001. Early diagnosis of
of adenovirus infections in patients undergoing allogeneic hemato- adenovirus infection and treatment with cidofovir after bone
poietic stem cell transplantation. Clin Infect Dis 32:1290–1297. marrow transplantation in children. Bone Marrow Transplant 27:
Carrigan DR. 1997. Adenovirus infections in immunocompromised 621–626.
patients. Am J Med 102:71–74. Ma L, Bluyssen HA, De Raeymaeker M, Laurysens V, van der Beek N,
Cassano WF. 1991. Intravenous ribavirin therapy for adenovirus Pavliska H, van Zonneveld AJ, Tomme P, van Es HH. 2001. Rapid
cystitis after allogeneic bone marrow transplantation. Bone Marrow determination of adenoviral vector titers by quantitative real-time
Transplant 7:247–248. PCR. J Virol Methods 93:181–188.
CDC. 2001. Two fatal cases of adenovirus-related illness in previously Mercier B, Burlot L, Ferec C. 1999. Simultaneous screening for HBV
healthy young adults: Illinois, 2000. MMWR Morb Mortal Wkly Rep DNA and HCV RNA genomes in blood donations using a novel
50:553–555. TaqMan PCR assay. J Virol Methods 77:1–9.
Chu Y, Sperber K, Mayer L, Hsu MT. 1992. Persistent infection of Miyamura K, Hamaguchi M, Taji H, Kanie T, Kohno A, Tanimoto M,
human adenovirus type 5 in human monocyte cell lines. Virology Saito H, Kojima S, Matsuyama T, Kitaori K, Nagafuji K, Sato T,
188:793–800. Kodera Y. 2000. Successful ribavirin therapy for severe adenovirus
Cichon G, Boeckh-Herwig S, Schmidt HH, Wehnes E, Muller T, Pring- hemorrhagic cystitis after allogeneic marrow transplant from close
Akerblom P, Burger R. 2001. Complement activation by recombi- HLA donors rather than distant donors. Bone Marrow Transplant
nant adenoviruses. Gene Ther 8:1794–1800. 25:545–548.
Echavarria M, Forman M, Ticehurst J, Dumler JS, Charache P. 1998. Morris T, Robertson B, Gallagher M. 1996. Rapid reverse transcription-
PCR method for detection of adenovirus in urine of healthy and PCR detection of hepatitis C virus RNA in serum by using the
human immunodeficiency virus-infected individuals. J Clin Micro- TaqMan fluorogenic detection system. J Clin Microbiol 34:2933–
biol 36:3323–3326. 2936.
Echavarria MS, Ray SC, Ambinder R, Dumler JS, Charache P. 1999. N.N. Sequence detection systems quantitative assay design and
PCR detection of adenovirus in a bone marrow transplant recipient: optimization. (www.appliedbiosystems.com, application note num-
hemorrhagic cystitis as a presenting manifestation of disseminated ber 77101-006).
disease. J Clin Microbiol 37:686–689. Pring-Akerblom P, Adrian T. 1994. Type- and group-specific polymer-
Echavarria M, Forman M, van Tol MJ, Vossen JM, Charache P, ase chain reaction for adenovirus detection. Res Virol 145:25–35.
Kroes AC. 2001. Prediction of severe disseminated adenovirus Ribaud P, Scieux C, Freymuth F, Morinet F, Gluckman E. 1999.
infection by serum PCR. Lancet 358:384–385. Successful treatment of adenovirus disease with intravenous
Emery VC, Sabin CA, Cope AV, Gor D, Hassan-Walker AF, cidofovir in an unrelated stem-cell transplant recipient. Clin Infect
Griffiths PD. 2000. Application of viral-load kinetics to identify Dis 28:690–691.
patients who develop cytomegalovirus disease after transplanta- Ryncarz AJ, Goddard J, Wald A, Huang ML, Roizman B, Corey L. 1999.
tion. Lancet 355:2032–2036. Development of a high-throughput quantitative assay for detecting
Ferreira-Gonzalez A, Fisher RA, Weymouth LA, Langley MR, Wolfe L, herpes simplex virus DNA in clinical samples. J Clin Microbiol
Wilkinson DS, Garrett CT. 1999. Clinical utility of a quantitative 37:1941–1947.
polymerase chain reaction for diagnosis of cytomegalovirus dis- Sabroe I, McHale J, Tait DR, Lynn WA, Ward KN, Shaunak S. 1995.
ease in solid organ transplant patients. Transplantation 68:991– Treatment of adenoviral pneumonitis with intravenous ribavirin
996. and immunoglobulin. Thorax 50:1219–1220.
Flomenberg P, Piaskowski V, Harb J, Segura A, Casper JT. 1996. Schütz E, von Ahsen N. 1999. Spreadsheet software for thermodynamic
Spontaneous, persistent infection of a B-cell lymphoma with melting point prediction of oligonucleotide hybridization with and
adenovirus. J Med Virol 48:267–272. without mismatches. Biotechniques 27:1218–1224.
Quantitative Adenovirus PCR 239

Schweiger B, Zadow I, Heckler R, Timm H, Pauli G. 2000. Application Wadell G, Allard A, Hierholzer H. 1999. Adenoviruses. In: Murray PR,
of a fluorogenic PCR assay for typing and subtyping of in- Baron EJ, Pfaller MA, Tenover CA, Yolken RA, editors. Manual of
fluenza viruses in respiratory samples. J Clin Microbiol 38:1552– clinical microbiology. Washington, D.C.: ASM Press. p 970–981.
1558. Weinberger KM, Wiedenmann E, Bohm S, Jilg W. 2000. Sensitive and
Stamey FR, Patel MM, Holloway BP, Pellett PE. 2001. Quantitative, accurate quantitation of hepatitis B virus DNA using a kinetic
fluorogenic probe pcr assay for detection of human herpesvirus 8 fluorescence detection system (TaqMan PCR). J Virol Methods 85:
dna in clinical specimens. J Clin Microbiol 39:3537–3540. 75–82.
Thompson JD, Higgins DG, Gibson TJ. 1994. CLUSTAL W: improving Xu W, McDonough MC, Erdman DD. 2000. Species-specific identifica-
the sensitivity of progressive multiple sequence alignment through tion of human adenoviruses by a multiplex PCR assay. J Clin
sequence weighting, position-specific gap penalties and weight Microbiol 38:4114–4120.
matrix choice. Nucleic Acids Res 22:4673–4680. Zaia JA, Gallez-Hawkins GM, Tegtmeier BR, ter Veer A, Li X, Niland
Wadell G, Allard A, Hierholzer JC. 1996. Adenoviruses. In: Fields BN, JC, Forman SJ. 1997. Late cytomegalovirus disease in marrow
Knipe DM, Howley PM, editors. Virology. Philadelphia, New York: transplantation is predicted by virus load in plasma. J Infect Dis
Lippincort-Raven. vol 2, p 2111–2148. 176:782–785.