SURVEY OF OPHTHALMOLOGY VOLUME 46 • NUMBER 3 • NOVEMBER–DECEMBER 2001

DIAGNOSTIC AND SURGICAL TECHNIQUES

MARCO ZARBIN AND DAVID CHU, EDITORS

Applications of the Polymerase Chain Reaction to

Diagnosis of Ophthalmic Disease
Russell N. Van Gelder, MD, PhD

Department of Ophthalmology and Visual Sciences, Department of Molecular Biology and Pharmacology, Washington
University School of Medicine, St. Louis, Missouri, USA

Abstract. The polymerase chain reaction (PCR) is a powerful molecular biologic technique for the
analysis of very small amounts of DNA. This technique has found increasing use in the past 10 years for
the detection of pathogenic organisms associated with many forms of ocular inflammatory and infec-
tious disease. PCR has shown utility in the diagnosis of viral uveitis, infectious endophthalmitis, and par-
asitic eye disease. The strengths and weaknesses of this diagnostic technique are discussed.
Additionally, uses of PCR in linking known pathogens to disease, and to discovering novel pathogens,
are addressed. (Surv Ophthalmol 46:248–258, 2001. © 2001 by Elsevier Science Inc. All rights
reserved.)

Key words. diagnostic technique • endophthalmitis • pathogen discovery • polymerase

chain reaction • uveitis

The polymerase chain reaction (PCR) is an enor- mologist has a working knowledge of the uses and
mously powerful molecular biologic technique that misuses of PCR. This review discusses the basic bio-
allows the rapid production of analytic quantities of chemistry of PCR, how its use has impacted oph-
DNA from infinitesimal amounts of starting mate- thalmic practice (particularly in the analysis of uvei-
rial. Since the introduction of its modern form in tis and infectious disease), and ways in which PCR is
1988,77 PCR has revolutionized much of molecular improving our understanding of the mechanisms of
biology and has greatly accelerated the development ophthalmic disease.
of molecular diagnostics. Kary B. Mullis received a
Nobel prize in 1993 for inventing this technique. Basic Biochemistry
Many clinical uses of PCR are transparent to the PCR is DNA replication stripped to its bare essen-
physician. When a physician requests genetic testing tials. Arthur Kornberg showed over 40 years ago that
for a disease, for example, the patient’s DNA is usu- DNA can be replicated in a test tube using only a
ally amplified by PCR prior to sequencing or restric- DNA template to be copied, short oligonucleotide
tion mapping. For a growing subset of diseases, the primers, the enzyme DNA polymerase, nucleotide
physician must specifically request diagnostic PCR triphosphates, and the appropriate salts.43 PCR is
testing. It is important that the practicing ophthal- performed using two specific primers that flank the

248
© 2001 by Elsevier Science Inc. 0039-6257/01/$–see front matter
All rights reserved. PII S0039-6257(01)00274-0
PCR AND OPHTHALMIC DISEASE DIAGNOSIS 249

DNA region of interest. After enzymatic synthesis of amplified. This sequence may be a gene in the pa-
the replicated strand is complete, the DNA is dena- tient’s DNA (for example, the sequence surrounding
tured into single strands. This allows the newly syn- a suspected mutation in the mitochondrial DNA for a
thesized strand to serve as template for subsequent patient with Lebers optic neuropathy), or a gene
synthesis of new strands (Fig. 1). By using a thermo- from a suspected pathogen (for instance, the UL97
stable DNA polymerase (isolated from a bacterial gene of cytomegalovirus). One can also use “degener-
species—Thermus aquaticus—which lives in the geo- ate” primers or primers to highly conserved se-
thermal vents in Yellowstone National Park), one quences to amplify DNA from any bacterium in the
can perform the denaturation step using heat, with- sample. Some protocols require partial purification of
out also denaturing the polymerase. The annealing, the DNA to be amplified, whereas other protocols al-
synthesis, and denaturation steps can be directed by low PCR to be performed directly on the patient sam-
simply changing the temperature. Using an auto- ple. To perform PCR, one begins with the initial sam-
mated thermal heat block, 30 to 40 rounds of repli- ple containing the target DNA and mixes in the
cation can be performed in just a few hours. appropriate primers, DNA polymerase, nucleotide
The degree of amplification that can be achieved, triphosphates, and buffered salts. Following perfor-
and thus the sensitivity of PCR, is remarkable. Theo- mance of PCR in the thermal cycler, the products
retically, the molar amount of PCR product doubles may be detected in one of several ways. Generally, gel
with each round of replication. Thirty-five cycles are electrophoresis, with use of acrylamide or agarose, is
typically used for diagnostic PCR; this corresponds employed to determine if a DNA fragment of expected
to 1010 amplification of the starting material. If a size has been produced. Confirmation of the identity
single molecule of pathogen DNA is present in a of the PCR product can be achieved by digesting the
sample, after such a PCR there will theoretically be product with restriction endonuclease and observing
over 34 billion copies produced in just a few hours. the restriction digest pattern, a technique called fin-
For a 500 base-pair fragment, this would correspond gerprinting. Alternatively, the PCR-produced DNA
to about 20 ng of DNA, a sufficient quantity to di- fragment can be blotted to nylon or nitrocellulose fil-
rectly visualize by agarose gel electrophoresis and ter paper and specifically hybridized with a labeled
ethidium bromide staining. DNA that will anneal only to the correct product. Ulti-
In order to perform PCR, one must have a source mate identification of a DNA fragment can be
of DNA (for example, DNA extracted from an aque- achieved by sequencing the PCR product DNA.
ous or vitreous tap specimen, or from the peripheral Although the original PCR technique was de-
blood), and some knowledge of the sequence to be signed for the amplification of DNA, PCR can be

Fig. 1. Schematic of the polymerase chain reaction. Starting with a patient sample (i.e., aqueous or vitreous biopsy, con-
junctival swab, or blood), DNA is partially purified. Oligonucleotide primers (red  upstream, blue  downstream) are syn-
thesized and allowed to anneal to the target DNA. As the temperature is increased, the DNA polymerase in the reaction
becomes active, allowing template-directed DNA replication of the target sequence. On further elevation of the reaction
temperature, the original and newly synthesized DNA strands denature and separate. When the temperature is again low-
ered, a set of primers can anneal to both the original target (black) and newly synthesized (red) DNA strands, and the
process is repeated. PCR will thus give a theoretical yield of 2n molecules for each starting molecule, amplified through n
rounds of annealing, elongation, and denaturation.
250 Surv Ophthalmol 46 (3) November–December 2001 VAN GELDER

modified to detect specific RNAs, by converting the all RNA and some DNA. Different laboratories
RNA to DNA with use of the enzyme reverse tran- employ different methods for purifying the nucleic
scriptase. This technique is called reverse transcrip- acids from the sample and inactivating any intrinsic
tion polymerase chain reaction (RT-PCR), and it al- inhibitors of PCR. Several authors have found that
lows for detection of viruses with RNA genomes such boiling the frozen sample for 15 minutes is sufficient
as HIV. PCR can also be performed using internal to release nucleic acids and remove the inhibitors.16,42
controls to generate a quantitative estimate of the
amount of pathogen DNA in the starting material; Power and Limits of PCR
this technique is used, for example, for measure- Among the most powerful uses of PCR in medical
ment of peripheral viral load in HIV. Newer genera- practice is in the detection of foreign organisms. By
tion thermal cyclers with fluorescence detectors al- designing primers for the DNA of a particular organ-
low for real-time monitoring of PCR product ism, one can perform PCR on very small amounts of
accumulation and markedly improved quantitation tissue. A large number of organisms that cause oph-
of products. thalmic disease have been detected by PCR; these
are summarized in Table 1. An example of a positive
Obtaining Specimens for PCR PCR reaction for herpes simplex virus from a patient
PCR can be performed on nearly any ocular speci- with acute retinal necrosis is shown in Fig. 2.
men or biopsy. For diagnosis of ocular infectious dis- The sensitivity for detection of foreign DNA is very
ease, the obtained sample is usually a conjunctival high. Most PCR assays for viral pathogens have a sensi-
swab, anterior chamber paracentesis, or vitreous tap. tivity of 10–100 genomes (i.e., the number of copies
For the former, the swab should be immediately of the pathogenic organism’s DNA), which typically
placed in 0.1 ml of balanced salt solution and mate- corresponds to less than 1 plaque-forming unit in a vi-
rial aseptically “milked” out of the swab. Anterior ral culture (i.e., the number of pathogens required
chamber paracentesis of 50 l is usually sufficient for detection in culture).2,3,16,42,54,56,89 Thus PCR poten-
for diagnostic purposes. For vitrectomy specimens, tially is more sensitive than culture for detection of
the initial preinfusion aspirate (100–500 l) is pre- many organisms. By utilizing a secondary detection
ferred, although we have been able to detect patho- system in concert with the initial PCR reaction (for in-
gen DNA from the vitrectomy infusion cassette in stance, Southern blotting of the PCR specimen with a
some cases. Specimens should be aseptically trans- pathogen-specific probe, performing nested PCR with
ferred to a sterile, capped tube (i.e., a 1.5-ml mi- internal primers, or restriction digestion of the prod-
crofuge tube) and quick-frozen on dry ice or in liq- uct), perfect specificity can be assured. Although PCR
uid nitrogen. The sample should remain frozen would seem to have nearly ideal characteristics for a
until processed by the accepting laboratory; freeze- diagnostic test, the high sensitivity and specificity can
thaw cycles will release nucleases that will degrade cause significant pitfalls.

TABLE 1
Organisms Responsible for Ocular Disease Detected by Polymerase Chain Reaction
Organism Representative References
Viruses
Cytomegalovirus 18, 26, 28, 47, 54, 55, 56, 71
Herpes Simplex Virus 3, 15, 18, 19, 24, 29, 42, 56, 81, 89
Varicella Zoster Virus
Epstein Barr Virus 7
HTLV-1 58, 59, 67, 68
Human Herpes Virus-8 (periocular) 40
Adenovirus 13, 29, 36, 39, 45, 61, 78, 84
Bacteria and fungii
Bacterial endophthalmitis organisms (Staphylococcus, streptococcus, 11, 41, 48, 64, 66, 87
Pseudomonas spp., etc.)
Propionibacterium acnes 33, 49
Mycobacterium spp. 8, 34, 44, 50
Borrelia burgdorferi 31, 79
Bartonella spp. 20, 27, 38
Tropheryma whippeli 74
Candida albicans and other fungii 30, 49, 57, 65
Parasitic disease
Toxoplasma gondii 5, 25, 37, 60
Onchocerca volvulus 22, 91
PCR AND OPHTHALMIC DISEASE DIAGNOSIS 251

rus (EBV), for example. Because most American pa-

tients have been exposed to EBV, and because the
virus remains latent in a subset of white blood cells,
almost any PCR reaction on a patient sample con-
taining white blood cells will yield a positive result.
The results of PCR must be kept in clinical context.
We recently analyzed a sample from a 66-year-old im-
munocompetent patient with persistent vitritis. The
PCR results showed a positive reaction for cytomega-
lovirus, but this likely represents latent viral DNA in
the white blood cells comprising the vitritis.

SPECIFICITY
The match between primer sequence and host
DNA needs to be nearly perfect in order to produce
Fig. 2. Example of PCR diagnosis in a patient with severe a product. Sequence polymorphisms between strains
media opacity. A 35-year-old immunocompromised man of the organism can lead to poor priming and false-
presented with acute onset redness and blurring of vision negative PCR. For example, a primer set in our labo-
in his right eye. Clinical examination revealed hand mo- ratory that recognized the standard AD169 strain of
tions vision, mild anterior chamber inflammation, and cytomegalovirus (CMV) failed to recognize a strain
dense vitritis precluding view of the retina. Ultrasound ex-
amination revealed a retinal detachment. The differential of CMV isolated from the vitreous of a patient,
diagnosis included acute retinal necrosis syndrome, cy- whereas a second primer set recognized either
tomegalovirus retinitis, and ocular toxoplasmosis. Vitreous strain. All primer sets that are used must be validated
sample was obtained at the time of retinal detachment re- on patient samples, preferably from a number of
pair. PCR for VZV, Toxoplasma gondii, and CMV were nega- sources. Use of multiple primer sets for each organ-
tive. PCR is shown for HSV virus. The first lane is a refer-
ence DNA ladder. Dilutions refer to positive control ism under consideration can also decrease the
purified HSV DNA concentrations; the assay has a sensitiv- chance of a false-negative result from a polymor-
ity of 10 molecules. Negative control contains no patient phism in the primer sequence.
sample. Arrow indicates positive result for PCR, indicating
that the cause of his inflammation was acute retinal necro-
sis syndrome due to HSV. Based on the results of the PCR, AVOIDING THE PITFALLS
the patient was treated with a course of intravenous acyclo- Several techniques can be used to minimize the
vir. The fellow eye did not become involved, and the clini- potential pitfalls of PCR diagnostics in uveitis. Strict
cal course was consistent with ARN syndrome once the vit-
ritis had cleared. adherence to rigorous laboratory technique can
minimize the risks of false positives arising from car-
ryover contamination. The use of deoxyuracil base,
Pitfalls of PCR along with treatment of all PCR reactions with uracyl
DNA glycosylase (which degrades previous PCR
SENSITIVITY products so they cannot serve as templates for subse-
The extremely high sensitivity of PCR can produce quent reactions) also decreases the likelihood of car-
false-positive results in several ways. When a tech- ryover contamination.88 Negative and positive con-
nique is capable of detecting as little as one molecule trol samples must be tested with each diagnostic
of foreign DNA, laboratory contamination can be a PCR, to establish sensitivity and specificity for each
substantial hazard. Carryover of DNA via pipettes, or reaction. Ideally, the use of a real-time PCR, which
even viral shedding by the laboratory technician allows real-time quantitative analysis of reaction
(who may be a carrier of herpes simplex, for exam- products by monitoring of accumulation of a fluo-
ple) can lead to false-positive results. It is essential rescent DNA intercalating agent,14,21 can allow for
that all pathogen samples be stored separately from discrimination of commensal and low-level contami-
laboratory reagents. Rigorous negative controls must nants. Although nested PCR (where first-round am-
be performed with each PCR reaction. A second plification products are subjected to a second round
source of false-positive results is internal contamina- of PCR with use of primers internal to the original
tion from latent host DNA. Many herpes viruses, in set) has greater sensitivity for detection of pathogen,
71,72
particular, can become incorporated into the host routine use of this technique may yield increased
genome; detection of a few copies of this host DNA false-positive results, and should be avoided if possi-
can yield a false-positive result. This makes PCR es- ble. Finally, all PCR diagnostic results must be con-
sentially useless for the detection of Epstein–Barr vi- sidered in a clinical context.
252 Surv Ophthalmol 46 (3) November–December 2001 VAN GELDER

PCR Diagnosis of Posterior Uveitis PCR for Endophthalmitis

The initial application of PCR diagnostics to oph- Determination of a causative organism in postop-
thalmic disease was in the detection of viral uveitis.2,3, erative bacterial endophthalmitis is frequently diffi-
15,42,56,89
Knox et al42 performed PCR on aqueous or cult. The Endophthalmitis Vitrectomy Study re-
vitreous samples of 37 eyes of 38 patients referred to ported culture yields of only 70%.1 Culture results
a specialty uveitis practice with “diagnostic dilem- are also slow to return, thus requiring patients be
mas” in posterior uveitis. These were cases in which treated with broad spectrum antibiotics for several
media opacity either precluded clinical diagnosis, or days, even for relatively indolent bacteria. All bacte-
where the natural history of the disease was inconsis- ria share common, highly repetitive DNA sequences
tent with clinical appearance. Of these cases, a defin- for their 16S ribosomal RNA. By designing primers
itive diagnosis of a viral infection could be made by to these conserved 16S sequences, PCR can be per-
PCR in 25 eyes. Of the PCR-negative cases, a number formed on biopsy material from eyes with suspected
were ultimately diagnosed to be toxoplasmosis endophthalmitis, with the results available within
(which was not included in the PCR panel), and the hours. Therese et al demonstrated the utility of this
remainder had natural histories inconsistent with vi- approach for culture-negative endophthalmitis.87
ral retinitis. Thus, both positive and negative PCR re- This group was able to determine a bacterial cause
sults likely had diagnostic significance in this study. for endophthalmitis in 100% of culture-positive and
Probably the most common indication for perform- 44% of culture-negative cases. Of the remaining cul-
ing diagnostic PCR for posterior uveitis is the pres- ture-negative cases, one-third were found to be fun-
ence of media opacity. Significant media opacity from gal. Recently, PCR primers for intraocular fungal
cataract or dense vitritis can make otherwise straight- disease have also been introduced,30 suggesting that
forward diagnoses difficult. Mitchell et al developed a cause of upwards of 90% of endophthalmitis cases
PCR primers with a sensitivity of 93% and specificity may be detectable by PCR. These results suggest that
of 98% for the detection of CMV, and they used prim- PCR can contribute significantly to the diagnosis of
ers with similar sensitivity and specificity for Varicella bacterial endophthalmitis.
Zoster virus (VZV) to analyze vitreous biopsies from The primary drawback to performance of PCR for
patients with posterior uveitis and media opacity pre- bacterial endophthalmitis has been the labor and time
venting definitive diagnosis.56 Of the nine patients involved in determining the bacterium responsible for
tested, four tested positive for CMV, and three for a positive 16S PCR product. Initially, the PCR products
VZV. The remaining two were subsequently judged to were subjected to DNA sequencing;41,48 although this
have toxoplasmosis. In all cases, the clinical course technique yields a definitive identification of the caus-
was consistent with the PCR-based diagnosis. ative organism, sequencing of PCR products can be
The clinical diagnosis of atypical toxoplasmosis technically challenging and requires specialized equip-
can also be problematic. Classical reactivation toxo- ment. However, several new techniques will allow
plasmosis can be diagnosed by clinical examination, much more rapid determination of causative bacteria.
but primary toxoplasmosis can resemble a number Carroll et al used nested PCR primers to distinguish
of other infectious acute retinitides.32,75,76 Initial Gram-positive from Gram-negative 16S bacterial DNA
studies of PCR diagnosis of Toxoplasma gondii were sequences.11 The test had nearly perfect sensitivity, and
disappointing, showing sensitivities less than 50%.5,25 could be completed in as little as 3.5 hours. Okhravi et
The use of intraocular antibody titers appeared to al66 used a different approach, performing restriction
give increased sensitivity, although in at least one endonuclease digest fingerprinting of the 16S PCR
study there was a complementary relationship be- product. Using this method, they were able to distin-
tween intraocular antibody production and PCR, guish all major bacterial causes of postoperative en-
with the two techniques together having a sensitivity dophthalmitis, except for being able to distinguish be-
of 72%.5 However, recent advances in primer design, tween E. coli and S. marcescens. The restriction-digestion
utilizing highly repetitive pathogen DNA sequences, fingerprinting method adds only 1 hour to perfor-
have greatly improved yields for PCR of T. gondii. mance of PCR, allowing results to be obtained within
Montoya et al60 were able to detect Toxoplasma about 4 hours.
DNA in nearly 80% of patients with suspected ocular Delayed-onset postoperative endophthalmitis is a
toxoplasmosis and positive serum IgG titers. Using a vision-threatening complication of cataract surgery
similar PCR assay, Bou et al9 were able to detect Toxo- and presents even further diagnostic challenges.
plasma gondii DNA in the peripheral blood of most Causative organisms include Propionibacterium acnes,
patients with active ocular toxoplasmosis, raising the Staphylococcus epidermidis, Actinomyces israelli, and
possibility that in the future, reactivation disease fungi. These organisms are frequently present in low
could be diagnosed via a blood test. numbers, and they can be difficult to culture. Yields
PCR AND OPHTHALMIC DISEASE DIAGNOSIS 253

from diagnostic vitreous biopsies in this condition the adenoviral and herpes PCR reactions). With the
are less than 50%. Usually, the diagnosis is made on development of more sophisticated multiplex PCR
clinical appearance, and patients are treated with reactions, it may be possible to test an entire differ-
pars plana vitrectomy, capsulectomy, and IOL re- ential diagnosis in a single reaction.17
placement regardless of causative organism. How- Acanthamoeba is an uncommon cause of corneal
ever, medical therapy alone may possibly be effective infection, but can have devastating clinical conse-
in a subset of patients with S. epidermidis infection. quences, particularly in contact lens wearers. The di-
Lohmann et al49 used 16S ribosomal primers as well agnosis of this disease can be challenging, as patients
as fungal PCR primers, in concert with culture and often present in the early stages of infection with
stain for 25 eyes with delayed-onset endophthalmitis. pain out of proportion to findings. The organism is
Aqueous culture and microscopy each had a 0% difficult to culture, and indirect means of detection
yield, whereas vitreous culture had a 24% yield in (such as staining with calcufluor white) offer low
these patients. PCR of the aqueous yielded a diagno- sensitivity. Lehmann et al46 developed a PCR diag-
sis in 84% of the cases, and PCR of a vitreous tap nostic test for Acanthamoeba, and demonstrated 84%
yielded a diagnosis in 92%. PCR thus has clear supe- sensitivity for detection of Acanthamoeba from epithe-
riority to any other available diagnostic technique lial scrapings (compared with 53% for culture); even
for diagnosis of delayed-onset endophthalmitis. tear samples had a 66% sensitivity in this assay. Spec-
ificity was 95% in this study. Further confirmation of
PCR for Anterior Segment and the utility of PCR for diagnosis of Acanthameoba was
External Disease demonstrated by Mathers et al,53 who were able to
Most causes of bacterial conjunctivitis or keratitis detect Acanthamoeba DNA in 24 of 33 epithelial
are readily cultured or detected by Gram-staining; scrapings, with use of a combination of primers. In-
thus, there have been relatively few efforts to apply terestingly, on sequencing these products, the group
PCR-based diagnostics to external ocular disease. discovered heterogeneity of sequence, suggesting
Adenoviral conjunctivitis, however, is not readily cul- that not all cases were caused by the species Acanth-
tured. Although ELISA assays for adenovirus may be amoeba castellanii, but that other previously unidenti-
used to confirm infection, these tests generally do fied species of the Acanthamoeba genus may be re-
not allow serotyping. Several groups have developed sponsible for some cases.
PCR primer sets for adenovirus,13,39,78 demonstrating
sensitivity of a few viral genomes. Using techniques
similar to those described above for bacterial en- Unique Uses of PCR in
dophthalmitis, Takeuchi et al84 were able to perform Ophthalmic Diagnostics
restriction fragment length polymorphism (RFLP) The uses of PCR described in the preceding sec-
analysis of the PCR products and determine a sero- tions are all applications to diagnostic problems that
type for adenoviral swabs from most patients tested. can be addressed (to greater or lesser extents) by
Interestingly, following DNA sequence analysis, this other diagnostic tests, such as culture or serologic
group discovered a novel subtype associated with an testing. PCR is advantageous in these situations
epidemic keratoconjunctivitis outbreak in Japan;85 largely because of its high sensitivity and rapidity.
this level of subtype analysis would not be possible However, there are also a number of emerging uses
with standard reagents. Other infectious causes of of PCR to make diagnoses that cannot be readily ac-
keratitis and conjunctivitis, such as ocular herpes complished by any other technique. These include
simplex29 and adult inclusion conjunctivitis caused detection of organisms that cannot be cultured, mo-
by Chlamydia species,45 have also been detected with lecular phenotyping of infectious disease, and study-
use of PCR. However, as the differential diagnosis of ing host gene alterations in masquerade syndromes.
organisms grows, so does the cost and time of per- Several organisms that are difficult or impossible
forming individual PCR reactions. To address this to culture and for which serology is not available can
potential obstacle to the routine use of PCR for diag- be detected by PCR. Whipples disease was long sus-
nosis of anterior segment disease, Jackson et al36 de- pected to have a bacterial cause on the basis of bac-
signed a multiplex PCR reaction for adenovirus and teria visible in stains of jejunal biopsies. However,
herpes simplex virus that used material from con- the organism could not be cultured, and no sero-
junctival swabs. In multiplex PCR, primer sets repre- logic evidence of infection could be detected. Jeju-
senting multiple pathogens are used in a single reac- nal biopsy was thus required for diagnosis of this un-
tion. This group showed comparable sensitivity for common but devastating cause of uveitis. Relman et
testing between monoplex and multiplex primer sets al73 performed PCR on jejunal biopsies, using prim-
(although addition of a third primer set for host ers to the conserved bacterial 16S ribosomal subunit.
DNA recovery substantially lowered the sensitivity of This group found a sequence for a novel bacterium
254 Surv Ophthalmol 46 (3) November–December 2001 VAN GELDER

(related to Actinomyces), which they called Tro- In addition to detection of infectious disease, PCR
pheryma whippelii. Once molecular identification of can also be useful for the diagnosis of masquerade
the bacterium was in hand, the species was success- syndromes. B cell lymphoma (formerly called reticu-
fully cultured in vitro. T. whippelii has been identi- lum cell sarcoma) can mimic posterior uveitis, and is
fied by PCR in ocular specimens of patients with a significant presentation of ocular inflammation in
Whipples disease–related uveitis.74 Similarly, the older patients. Lymphomas frequently have gene re-
causative agent for cat-scratch disease, bacteria of ge- arrangements, placing oncogenes under the control
nus Bartonella, escaped detection by culture until of immunoglobulin gene-expression elements. Shen
about 10 years ago; at about the same time, the or- et al80 devised a PCR-based assay to detect IgH gene
ganism was identified by PCR.4 More recently, DNA rearrangements in intraocular lymphoma, and they
from these organisms has been PCR amplified from were able to identify rearragements in four of four
ocular tissues in patients with orbital granulomas,20 pathologic specimens.
Parinauds oculoglandular syndrome,27 and vitreous
samples of neuroretinitis.38 Linking an Organism to a Disease
Classic microbiologic differentiation of strains re- Koch’s postulates remain the gold standard for
quires either an identifiable change in the metabolic causally linking an organism to a disease.10,23,51,86 Fu-
or physical properties of a microorganism (i.e., the filling Koch’s postulates requires identifying the or-
ability to ferment a particular sugar), or the avail- ganism of interest from all cases of the disease, prop-
ability of antisera that differentially recognize sub- agating the organism in vitro, causing the disease by
types of an organism. However, not all organisms reintroduction into the naïve host, and recovering
showing identical culture properties and serologic the organism from the infected host. Although prac-
profiles are identical. Toxoplasma gondii, the caus- tical for some diseases, Koch’s postulates cannot be
ative organism of ocular toxoplasmosis, has tradi- used in cases where the causative organism cannot
tionally been treated as a single strain. The organism be cultured (as in Whipples disease, above), or
is very difficult to culture; therefore, metabolic dif- where no animal model exists for disease. In these
ferences among isolates are not easily identified. cases, we are left with indirect evidence, such as tem-
There is a single dominant epitope for serology as poral correlation of seroconversion, linking disease
well, making all T. gondii appear identical to the im- to organism.
mune system. Using PCR, Sibley and Boothroyd82 PCR can provide valuable evidence in linking an
have demonstrated that there are, in fact, at least organism to disease. By isolating the observed organ-
three molecularly unique strains of T. gondii. The ism in most cases of a disease, a causal relationship
differences can be identified only by PCR of particu- can be hypothesized. Granulomatocyclitic crisis is an
lar genes. When grown in vitro, these three strains uncommon, unilateral episodic anterior uveitis asso-
have markedly different antibiotic sensitivities to ciated with marked increases in intraocular pressure.
sulfa medications, which may explain the clinically The presenting signs and symptoms are very reminis-
observed heterogeneity in response to treatment. In cent of herpetic disease. Yamamoto et al90 developed
the future, strain typing of T. gondii by PCR may be- a very sensitive PCR reaction for HSV, then tested
come a routine part of the workup for this disease. aqueous humor from patients with active disease,
Similarly, antibiotic resistance of microorganisms compared with non-affected controls. Three of three
can be problematic. Although testing for bacterial patients with active disease had positive PCR for her-
sensitivity to antibiotics is quite straightforward (al- pes simplex virus, whereas none of 10 patients with
beit time consuming), testing viruses is difficult. other forms of uveitis had this finding. Although not
Clinically, however, identification of resistant strains fulfilling Koch’s postulates, this finding is certainly
is extremely important, as some viruses (such as suggestive of a causal link, and it is potentially
CMV) have high rates of antiviral resistance to com- grounds for attempting antiviral therapy in the treat-
mon agents. One might not place a ganciclovir ment of this disorder.
implant, for example, into an eye that had known Perhaps the most impressive use of PCR to detect
ganciclovir-resistant CMV. Ganciclovir resistance is a causative organism was the demonstration by
frequently due to mutations in the cytomegalovirus Chang et al12 of an infectious basis to Kaposi sar-
UL97 gene. Liu et al47 devised a PCR assay that assays coma. This vascular neoplasm is frequently seen in
for four well-described point mutations conferring immunocompromised patients, particularly those
ganciclovir resistance. They tested this assay on 11 with AIDS. However, it was seen more commonly in ho-
vitreous samples from eyes with clinically resistant mosexual men with AIDS, suggesting a second infec-
CMV retinitis, and six were shown to have one of the tious agent was the cause. Chang et al used a power-
four known mutations. Such rapid strain typing is ful variant of PCR, called representational difference
not possible with any other technique. analysis (RDA), to compare Kaposi sarcoma tissue
PCR AND OPHTHALMIC DISEASE DIAGNOSIS 255

from uninvolved tissue in the same patient. RDA is a been used to suggest M. tuberculosis as a cause for
form of subtractive hybridization, and can be used to Eales disease.8,50 Sarcoidosis and pars planitis simi-
isolate DNA sequences found in one tissue but not larly have have epidemiologic or physiologic charac-
in another. After performing RDA on the Kaposi sar- teristics suggesting the possibility of an infectious
coma tissue, they discovered a novel sequence be- cause, and they are excellent candidates for patho-
longing to a new member of the herpes virus family, gen panning using PCR.35,69
which was dubbed human herpes virus 8 (HHV-8).
HHV-8 was subsequently found to also underlie Cas-
tleman disease, a lymphoproliferative disorder.40 Al- Extending the Use of Diagnostic PCR in
though Koch’s postulates have not been fufilled for Ophthalmologic Diagnosis
HHV-8 and Kaposi sarcoma, additional supporting PCR is an extremely powerful molecular biologic
evidence strongly supports the link. In the ganciclo- technique, whose potential for improving manage-
vir implant trial, Martin et al52 found a profound de- ment of ophthalmic disease is only beginning to be
crease in incidence of Kaposi sarcoma in patients re- realized. At present, PCR has demonstrated utility for
ceiving systemic ganciclovir, compared with those the diagnosis of viral retinitis, conjunctivitis, and de-
receiving local therapy alone. layed-onset endophthalmitis. The spectrum of uveitic
Similarly, PCR can be used to identify previously disease detectable by PCR is broadening steadily.
unknown bacteria that may cause disease, which can Several obstacles are currently limiting the in-
be useful as 30% of infectious postoperative endoph- creased usage of PCR for clinical diagnostics. Most
thalmitis cases are culture negative.1 Okhravi et al63 current studies of PCR in clinical diagnostics have
used PCR for the conserved 16S RNA subunit to test been performed under Institutional Review Board
for the presence of bacteria in 17 culture-negative (IRB) approval. Widespread clinical usage in the
endophthalmitis cases. All samples yielded positive United States will require adoption of Food and Drug
PCR results for the presence of bacteria. By sequenc- Administration–approved diagnostic kits, with proce-
ing the DNA amplification products, they deter- dures performed in College of American Patholo-
mined that eight of these cases harbored previously gists–approved laboratories. Certification of assays
unidentified bacterial species. Although this finding and facilities for routine molecular diagnostic use is
does not conclusively demonstrate that these bacte- expensive, and will likely be prohibitively so for organ-
ria are causative of the observed infection, it does isms whose pathology is primarily ocular and whose
generate a testable hypothesis that may lead to im- prevalence is low (i.e., Acanthamoeba, P. acnes, Toxo-
proved diagnosis and treatment of culture-negative plasma gondii). Whereas the incentive for commercial-
bacterial endophthalmitis. ization of PCR assays for HIV or hepatitis C virus are
What other diseases may have previously unsus- clear, it is unclear if the same companies will invest re-
pected infectious causes? Rheumatoid arthritis (RA) sources into commercialization of diagnostic kits for
is a multi-organ disease with significant ocular mor- the far less common ophthalmic pathogens.
bidity, particularly in its juvenile form. RA is consid- Standardization across laboratories is also nonex-
ered a canonical autoimmune disease, with different istent. Many groups have developed independent as-
subtypes demonstrating different classes of auto- says for individual pathogens, but the relative sensi-
antibodies. However, there has been some historical tivities and specificities of these tests on the same
suggestion that infection with a common parvovirus, reference material has not been established. Indeed,
B19, is associated with disease. Using PCR, Taka- the precise criteria for confirmation of a positive re-
hashi et al83 were able to identify parvovirus B19 sult has not yet been established, which has led to
DNA in the inflamed joints of 30 of 39 patients with cases of conflicting laboratory results on identical
rheumatoid arthritis, but in only 4 of 26 with os- ocular specimens. Despite these organizational and
teoarthritis, and 5 of 31 with traumatic joint disease. political hurdles, however, the immense power of
Although not definitive proof of causation, this cor- PCR will make it an increasingly common tool in the
relation supports the hypothesis that the virus is in- armament of ophthalmic diagnostic tests.
volved in the pathogenesis of the disease, and it will
lead to intensified work exploring this relationship.
Similar studies have strongly suggested a role for Method of Literature Search
Chlamydia pneumonia in the pathogenesis of athero- The Medline database, from 1966 to 2000, was
sclerosis.6,62,70 Although distinction of primarily im- searched with use of the Ovid and PubMed search
mune-mediated and primarily infectious causes of engines for the terms polymerase chain reaction or
uveitis is blurred, several ophthalmic diseases have PCR, and uveitis, endophthalmitis, retinitis, keratitis, and
epidemiology or clinical presentations that are very ocular infection. Additional references were derived
suggestive of infectious disease. Such analysis has from the cited references in each chosen paper. En-
256 Surv Ophthalmol 46 (3) November–December 2001 VAN GELDER

glish abstracts were utilized for papers in non- 20. Dondey JC, Sullivan TJ, Robson JM, Gatto J: Application of
polymerase chain reaction assay in the diagnosis of orbital
English languages. granuloma complicating atypical oculoglandular cat scratch
disease. Ophthalmology 104:1174–8, 1997
References 21. Dworkin L, Schmitz T, Van Gelder R: A real-time PCR assay
to identify and quantify pathogens responsible for infectious
1. ___: Microbiologic factors and visual outcome in the en- posterior uveitis. Invest Ophthalmol Vis Sci 42:S572, 2001
dophthalmitis vitrectomy study. Am J Ophthalmol 122:830– 22. Erttmann KD, Unnasch TR, Greene BM: A DNA sequence
46, 1996 specific for forest form Onchocerca volvulus. Nature 327:
2. Abe T, Sato M, Tamai M: Correlation of varicella-zoster virus 415–7, 1987
copies and final visual acuities of acute retinal necrosis syn- 23. Fredericks DN, Relman DA: Sequence-based identification
drome. Graefes Arch Clin Exp Ophthalmol 236:747–52, 1998 of microbial pathogens: a reconsideration of Kochs postu-
3. Abe T, Tsuchida K, Tamai M: A comparative study of the lates. Clin Microbiol Rev 9:18–33, 1996
polymerase chain reaction and local antibody production in 24. Ganatra JB, Chandler D, Santos C: Viral causes of the acute ret-
acute retinal necrosis syndrome and cytomegalovirus retini- inal necrosis syndrome. Am J Ophthalmol 129:166–72, 2000
tis. Graefes Arch Clin Exp Ophthalmol 234:419–24, 1996 25. Garweg J, Boehnke M, Koerner F: Restricted applicability of
4. Anderson B, Sims K, Regnery R: Detection of Rochalimaea the polymerase chain reaction for the diagnosis of ocular
henselae DNA in specimens from cat scratch disease patients toxoplasmosis. Ger J Ophthalmol 5:104–8, 1996
by PCR. J Clin Microbiol 32:942v8, 1994 26. Garweg J, Fenner T, Bohnke M, Schmitz H: An improved
5. Aouizerate F, Cazenave J, Poirier L: Detection of Toxo- technique for the diagnosis of viral retinitis from samples of
plasma gondii in aqueous humour by the polymerase chain aqueous humor and vitreous. Graefes Arch Clin Exp Oph-
reaction. Br J Ophthalmol 77:107–9, 1993 thalmol 231:508–13, 1993
6. Arbustini E, Morbini P, Bello BD: From plaque biology to 27. Grando D, Sullivan LJ, Flexman JP: Bartonella henselae asso-
clinical setting. Am Heart J 138:S55–60, 1999 ciated with Parinauds oculoglandular syndrome. Clin Infect
7. Bassili SS, Peyman GA, Gebhardt BM: Detection of Epstein- Dis 28:1156–8, 1999
Barr virus DNA by polymerase chain reaction in the vitreous 28. Hansen KK, Ricksten A, Hofmann B: Detection of cytomega-
from a patient with Vogt-Koyanagi-Harada syndrome. Retina lovirus DNA in serum correlates with clinical cytomegalovi-
16:160–1, 1996 rus retinitis in AIDS. J Infect Dis 170:1271–4, 1994
8. Biswas J, Therese L, Madhavan HN: Use of polymerase chain 29. Hidalgo F, Melon S, de Ona M: Diagnosis of herpetic kerato-
reaction in detection of Mycobacterium tuberculosis com- conjunctivitis by nested polymerase chain reaction in human
plex DNA from vitreous sample of Eales disease. Br J Oph- tear film. Eur J Clin Microbiol Infect Dis 17:120–3, 1998
thalmol 83:994, 1999 30. Hidalgo JA, Alangaden GJ, Eliott D: Fungal endophthalmitis
9. Bou G, Figueroa MS, Marti-Belda P: Value of PCR for detec- diagnosis by detection of Candida albicans DNA in intraocu-
tion of Toxoplasma gondii in aqueous humor and blood lar fluid by use of a species-specific polymerase chain reac-
samples from immunocompetent patients with ocular toxo- tion assay. J Infect Dis 181:1198–201, 2000
plasmosis. J Clin Microbiol 37:3465–8, 1999 31. Hilton E, Smith C, Sood S: Ocular Lyme borreliosis diag-
10. Burmester GR, Solbach W: Hit and run, hit and hide or per- nosed by polymerase chain reaction on vitreous fluid. Ann
manent hit: why it is premature to dump Kochs postulates in Intern Med 125:424–5, 1996
rheumatic diseases. J Rheumatol 19:1173–4, 1992 32. Holland GN, Muccioli C, Silveira C: Intraocular inflamma-
11. Carroll NM, Jaeger EE, Choudhury S: Detection of and dis- tory reactions without focal necrotizing retinochoroiditis in
crimination between gram-positive and gram-negative bacte- patients with acquired systemic toxoplasmosis. Am J Oph-
ria in intraocular samples by using nested PCR. J Clin Micro- thalmol 128:413–20, 1999
biol 38:1753–7, 2000 33. Hykin PG, Tobal K, McIntyre G: The diagnosis of delayed
12. Chang Y, Cesarman E, Pessin MS: Identification of herpesvi- post-operative endophthalmitis by polymerase chain reac-
rus-like DNA sequences in AIDS-associated Kaposis sarcoma. tion of bacterial DNA in vitreous samples. J Med Microbiol
Science 266:1865–9, 1994 40:408–15, 1994
13. Cooper RJ, Yeo AC, Bailey AS, Tullo AB: Adenovirus poly- 34. Ikonomopoulos JA, Gorgoulis VG, Zacharatos PV: Multiplex
merase chain reaction assay for rapid diagnosis of conjunc- polymerase chain reaction for the detection of mycobacte-
tivitis. Invest Ophthalmol Vis Sci 40:90–5, 1999 rial DNA in cases of tuberculosis and sarcoidosis. Mod
14. Costa JM, Pautas C, Ernault P: Real-time PCR for diagnosis Pathol 12:854–62, 1999
and follow-up of Toxoplasma reactivation after allogeneic 35. Ishige I, Usui Y, Takemura T, Eishi Y: Quantitative PCR of
stem cell transplantation using fluorescence resonance en- mycobacterial and propionibacterial DNA in lymph nodes of
ergy transfer hybridization probes. J Clin Microbiol 38:2929– Japanese patients with sarcoidosis. Lancet 354:120–3, 1999
32, 2000 36. Jackson R, Morris DJ, Cooper RJ: Multiplex polymerase
15. Cunningham ET Jr, Short GA, Irvine AR: Acquired immuno- chain reaction for adenovirus and herpes simplex virus in
deficiency syndrome–associated herpes simplex virus retini- eye swabs. J Virol Methods 56:41–8, 1996
tis. Clinical description and use of a polymerase chain reac- 37. Jones CD, Okhravi N, Adamson P: Comparison of PCR de-
tion–based assay as a diagnostic tool. Arch Ophthalmol 114: tection methods for B1, P30, and 18S rDNA genes of T. gon-
834–40, 1996 dii in aqueous humor. Invest Ophthalmol Vis Sci 41:634–44,
16. Cunningham ET Jr, Short GA, Irvine AR: Acquired immuno- 2000
deficiency syndrome–associated herpes simplex virus retini- 38. Kerkhoff FT, Bergmans AM, van Der Zee A, Rothova A: Dem-
tis. Clinical description and use of a polymerase chain reac- onstration of Bartonella grahamii DNA in ocular fluids of a
tion–based assay as a diagnostic tool. Arch Ophthalmol 114: patient with neuroretinitis. J Clin Microbiol 37:4034–8, 1999
834–40, 1996 39. Kinchington PR, Turse SE, Kowalski RP, Gordon YJ: Use of
17. Dabil H, Boley ML, Schmitz TM, Van Gelder RN: Validation of polymerase chain amplification reaction for the detection of
a diagnostic multiplex polymerase chain reaction assay for infec- adenoviruses in ocular swab specimens. Invest Ophthalmol
tious posterior uveitis. Arch Ophthalmol 119:1315–22, 2001 Vis Sci 35:4126–34, 1994
18. de Boer JH, Verhagen C, Bruinenberg M: Serologic and 40. Knowles DM, Cesarman E: The Kaposis sarcoma-associated
polymerase chain reaction analysis of intraocular fluids in herpesvirus (human herpesvirus-8) in Kaposis sarcoma, ma-
the diagnosis of infectious uveitis. Am J Ophthalmol 121: lignant lymphoma, and other diseases. Ann Oncol 8(Suppl
650–8, 1996 2):123–9, 1997
19. Dlugosch D, Eis-Hubinger AM, Kleim JP: Diagnosis of acute 41. Knox CM, Cevallos V, Margolis TP, Dean D: Identification of
and latent varicella-zoster virus infections using the polymer- bacterial pathogens in patients with endophthalmitis by 16S
ase chain reaction. J Med Virol 35:136–41, 1991 ribosomal DNA typing. Am J Ophthalmol 128:511–2, 1999
PCR AND OPHTHALMIC DISEASE DIAGNOSIS 257

42. Knox CM, Chandler D, Short GA, Margolis TP: Polymerase 66. Okhravi N, Adamson P, Matheson MM: PCR-RFLP-mediated
chain reaction-based assays of vitreous samples for the diag- detection and speciation of bacterial species causing en-
nosis of viral retinitis. Use in diagnostic dilemmas. Ophthal- dophthalmitis. Invest Ophthalmol Vis Sci 41:1438–47, 2000
mology 105:37–44, 67. Ono A, Ikeda E, Mochizuki M: Provirus load in patients with
43. Kornberg A: DNA replication. J Biol Chem 263:1–4, 1988 human T-cell leukemia virus type 1 uveitis correlates with
44. Kotake S, Kimura K, Yoshikawa K: Polymerase chain reaction precedent Graves disease and disease activities. Jpn J Cancer
for the detection of Mycobacterium tuberculosis in ocular Res 89:608–14, 1998
tuberculosis. Am J Ophthalmol 117:805–6, 1994 68. Ono A, Mochizuki M, Yamaguchi K: Increased number of
45. Kowalski RP, Uhrin M, Karenchak LM: Evaluation of the circulating HTLV-1 infected cells in peripheral blood mono-
polymerase chain reaction test for detecting chlamydial nuclear cells of HTLV-1 uveitis patients: a quantitative poly-
DNA in adult chlamydial conjunctivitis. Ophthalmology 102: merase chain reaction study. Br J Ophthalmol 79:270–6,
1016–9, 1995 1995
46. Lehmann OJ, Green SM, Morlet N: Polymerase chain reac- 69. Ormerod LD, Puklin JE, Giles CL: Chronic Propionibacte-
tion analysis of corneal epithelial and tear samples in the di- rium acnes endophthalmitis as a cause of intermediate uvei-
agnosis of Acanthamoeba keratitis. Invest Ophthalmol Vis tis. Ocul Immunol Inflamm 5:67–8, 1997
Sci 39:1261–5, 1998 70. Pai J, Knoop FC, Hunter WJ 3rd, Agrawal DK: Chlamydia
47. Liu W, Kuppermann BD, Martin DF: Mutations in the cy- pneumoniae and occlusive vascular disease: identification
tomegalovirus UL97 gene associated with ganciclovir-resis- and characterization. J Pharmacol Toxicol Methods 39:51–
tant retinitis. J Infect Dis 177:1176–81, 1998 61, 1998
48. Lohmann CP, Heeb M, Linde HJ: Diagnosis of infectious en- 71. Porter-Jordan K, Rosenberg EI, Keiser JF: Nested polymerase
dophthalmitis after cataract surgery by polymerase chain re- chain reaction assay for the detection of cytomegalovirus
action. J Cataract Refract Surg 24:821–6, 1998 overcomes false positives caused by contamination with frag-
49. Lohmann CP, Linde HJ, Reischl U: Improved detection of mented DNA. J Med Virol 30:85–91, 1990
microorganisms by polymerase chain reaction in delayed en- 72. Pujol-Rique M, Derouin F, Garcia-Quintanilla A: Design of a
dophthalmitis after cataract surgery. Ophthalmology 107: one-tube hemi-nested PCR for detection of Toxoplasma
1047–51; discussion 1051–2, 2000 gondii and comparison of three DNA purification methods.
50. Madhavan HN, Therese KL, Gunisha P: Polymerase chain J Med Microbiol 48:857–62, 1999
reaction for detection of Mycobacterium tuberculosis in 73. Relman DA, Schmidt TM, MacDermott RP, Falkow S: Identi-
epiretinal membrane in Eales disease. Invest Ophthalmol fication of the uncultured bacillus of Whipples disease. N
Vis Sci 41:822–5, 2000 Engl J Med 327:293–301, 1992
51. Marshall BJ: Helicobacter pylori in peptic ulcer: have Kochs 74. Rickman LS, Freeman WR, Green WR: Brief report: uveitis
postulates been fulfilled? Ann Med 27:565–8, 1995 caused by Tropheryma whippelii (Whipples bacillus). N
52. Martin DF, Kuppermann BD, Wolitz RA: Oral ganciclovir for Engl J Med 332:363–6, 1995
patients with cytomegalovirus retinitis treated with a ganci- 75. Ronday MJ, Ongkosuwito JV, Rothova A, Kijlstra A: Intraocu-
clovir implant. Roche Ganciclovir Study Group. N Engl J lar anti-Toxoplasma gondii IgA antibody production in pa-
Med 340:1063–70, 1999 tients with ocular toxoplasmosis. Am J Ophthalmol 127:294–
53. Mathers WD, Nelson SE, Lane JL: Confirmation of confocal mi- 300, 1999
croscopy diagnosis of Acanthamoeba keratitis using polymerase 76. Ronday MJ, Ongkosuwito JV, Rothova A, Kijlstra A: Intraocu-
chain reaction analysis. Arch Ophthalmol 118:178–83, 2000 lar anti-Toxoplasma gondii IgA antibody production in pa-
54. McCann JD, Margolis TP, Wong MG: A sensitive and specific tients with ocular toxoplasmosis. Am J Ophthalmol 127:294–
polymerase chain reaction-based assay for the diagnosis of cy- 300, 1999
tomegalovirus retinitis. Am J Ophthalmol 120:219–26, 1995 77. Saiki RK, Gelfand DH, Stoffel S: Primer-directed enzymatic
55. Mitchell SM, Fox JD: Aqueous and vitreous humor samples amplification of DNA with a thermostable DNA polymerase.
for the diagnosis of cytomegalovirus retinitis. Am J Ophthal- Science 239:487–91, 1988
mol 120:252–3, 1995 78. Saitoh-Inagawa W, Oshima A, Aoki K: Rapid diagnosis of
56. Mitchell SM, Fox JD, Tedder RS: Vitreous fluid sampling adenoviral conjunctivitis by PCR and restriction fragment
and viral genome detection for the diagnosis of viral retinitis length polymorphism analysis. J Clin Microbiol 34:2113–6,
in patients with AIDS. J Med Virol 43:336–40, 1994 1996
57. Mochizuki K, Matsushita H, Hiramatsu Y: Evaluation of fun- 79. Schmidt BL: PCR in laboratory diagnosis of human Borrelia
gal DNA sequence using polymerase chain reaction in infec- burgdorferi infections. Clin Microbiol Rev 10:185–201, 1997
tion-free vitreous humor. Jpn J Ophthalmol 42:213–7, 1998 80. Shen DF, Zhuang Z, LeHoang P: Utility of microdissection
58. Mochizuki M, Watanabe T, Yamaguchi K: HTLV-I uveitis: a and polymerase chain reaction for the detection of immuno-
distinct clinical entity caused by HTLV-I. Jpn J Cancer Res globulin gene rearrangement and translocation in primary
83:236–9, 1992 intraocular lymphoma. Ophthalmology 105:1664–9, 1998
59. Mochizuki M, Yamaguchi K, Takatsuki K: HTLV-1 and uvei- 81. Short GA, Margolis TP, Kuppermann BD: A polymerase
tis. Lancet 339:1110, 1992 chain reaction-based assay for diagnosing varicella-zoster vi-
60. Montoya JG, Parmley S, Liesenfeld O: Use of the polymerase rus retinitis in patients with acquired immunodeficiency syn-
chain reaction for diagnosis of ocular toxoplasmosis. Oph- drome. Am J Ophthalmol 123:157–64, 1997
thalmology 106:1554–63, 1999 82. Sibley LD, Boothroyd JC: Construction of a molecular karyo-
61. Morris DJ, Bailey AS, Cooper RJ: Polymerase chain reaction type for Toxoplasma gondii. Mol Biochem Parasitol 51:291–
for rapid detection of ocular adenovirus infection. J Med Vi- 300, 1992
rol 46:126–32, 1995 83. Takahashi Y, Murai C, Shibata S: Human parvovirus B19 as a
62. Movahed MR: Infection with Chlamydia pneumoniae and causative agent for rheumatoid arthritis. Proc Natl Acad Sci
atherosclerosis: a review. JSC Med Assoc 95:303–8, 1999 USA 95:8227–32, 1998
63. Okhravi N, Adamson P, Carroll N: PCR-based evidence of 84. Takeuchi S, Itoh N, Uchio E: Serotyping of adenoviruses on
bacterial involvement in eyes with suspected intraocular in- conjunctival scrapings by PCR and sequence analysis. J Clin
fection. Invest Ophthalmol Vis Sci 41:3474–9, 2000 Microbiol 37:1839–45, 1999
64. Okhravi N, Adamson P, Lightman S: Polymerase chain reac- 85. Takeuchi S, Itoh N, Uchio E: Adenovirus strains of subgenus
tion in the diagnosis of bacterial endophthalmitis. Br J Oph- D associated with nosocomial infection as new etiological
thalmol 83:378–80, 1999 agents of epidemic keratoconjunctivitis in Japan. J Clin Mi-
65. Okhravi N, Adamson P, Mant R: Polymerase chain reaction crobiol 37:3392–4, 1999
and restriction fragment length polymorphism mediated de- 86. Tan PL, Skinner MA: The microbial cause of rheumatoid ar-
tection and speciation of Candida spp causing intraocular thritis: time to dump Kochs postulates. J Rheumatol 19:
infection. Invest Ophthalmol Vis Sci 39:859–66, 1998 1170–2, 1992
258 Surv Ophthalmol 46 (3) November–December 2001 VAN GELDER

87. Therese KL, Anand AR, Madhavan HN: Polymerase chain 91. Zimmerman PA, Dadzie KY, De Sole G: Onchocerca volvulus
reaction in the diagnosis of bacterial endophthalmitis. Br J DNA probe classification correlates with epidemiologic pat-
Ophthalmol 82:1078–82, 1998 terns of blindness. J Infect Dis 165:964–8, 1992
88. Wang X, Chen T, Kim D, Piomelli S: Prevention of carryover
contamination in the detection of beta S and beta C genes
by polymerase chain reaction. Am J Hematol 40:146–8, 1992
89. Yamamoto S, Pavan-Langston D, Kinoshita S: Detecting her- Russell N. Van Gelder, MD, PhD, is a Research to Prevent
pes virus DNA in uveitis using the polymerase chain reaction. Blindness Career Development Awardee and is also supported by
Br J Ophthalmol 80:465–8, 1996 the Bernard Becker Clinician-Scientist Award and NEI K08-00343
90. Yamamoto S, Pavan-Langston D, Tada R: Possible role of from the National Eye Institute.
herpes simplex virus in the origin of Posner-Schlossman syn- The author has no proprietary or commercial interest in any
drome. Am J Ophthalmol 119:796–8, 1995 concept or product discussed in this article.