Standard keratic precipitates (KP)are far more heterogeneous and variable on in vivo confocal microscopy (IVCM)than with standard, low-powered slitlamp biomicroscopy (SLB). A, StandardSLB photograph of the cornea and anterior segment, revealing diffuse granulomatousKP scattered throughout the endothelium. B, The IVCM photograph reveals large,infiltrating KP that correspond to the large opacities on the SLB photograph(A). It is evident that large KP are made up of individual cellular componentsclustered together. C, The IVCM was performed on an area of the endotheliumbetween the large granulomatous KP. This area shows fibrinlike material adherentto the endothelium. ASL indicates Advanced Scanning Ltd (New Orleans, La).The depth from corneal epithelium at which the image was captured is indicatedin micrometers.
Keratic precipitates (KP) in individualpatients are consistent throughout the cornea, and in cases of bilateral disease,consistent between both eyes. A and B, Patient 1. A, The central cornea wasimaged using in vivo confocal microscopy (IVCM) and revealed multiple globularKP. B, The peripheral cornea shows consistent globular KP in the same individual.C and D, Patient 2. C, Peripheral corneal IVCM reveals a globular, stippledpattern of KP. D, A slightly more central view of the KP shows consistentmorphologic features. E and F, Patient 3. E, The central cornea shows smooth-roundedKP. F, The peripheral cornea shows similar, consistent smooth-rounded KP.ASL indicates Advanced Scanning Ltd (New Orleans, La). The depth from cornealepithelium at which the confocal image was captured is indicated in micrometers.
The morphologic features of keraticprecipitates (KP) change across time, including a change with disease progressionand treatment. A, In vivo confocal microscopy (IVCM) photograph of KP showsa stippled pattern consistent throughout the affected cornea. B, At 5 days,the stippled pattern had changed to smooth-rounded. C, Twelve days after steroidtherapy was initiated, only small pigmentations were visible on IVCM. ASLindicates Advanced Scanning Ltd (New Orleans, La). The depth from cornealepithelium at which the image was captured is indicated in micrometers.
In vivo confocal microscopy (IVCM)of keratic precipitates (KP) in infectious causes of uveitis. A and B, TheIVCM photograph of KP in a patient with toxoplasmosis. A, Infiltrating KPwith a globular central core and dendritiform pseudopodia typically seen ininfectious causes of uveitis. B, An area of cornea between the larger KP,showing multiple dendritiform bodies. C and D, The IVCM photograph of KP ina patient with cytomegalovirus retinitis, revealing an infiltrating appearancewith dendritiform pseudopodia. C, Image taken at the endothelial layer; D,image taken slightly deeper into the endothelium. E, The IVCM photograph ofKP in a patient with herpes zoster ophthalmicus. Central globular KP withinfiltrating dendritiform pseudopodia are seen. ASL indicates Advanced ScanningLtd (New Orleans, La). The depth from corneal epithelium at which the confocalimage was captured is indicated in micrometers.
In vivo confocal microscopy (IVCM)of keratic precipitates (KP) in noninfectious causes of uveitis. A, The IVCMphotograph of KP in a patient with Vogt-Koyanagi-Harada syndrome. The imagereveals globular KP without an infiltrating appearance. B, The globular appearanceof these KP is emphasized by a 3-dimensional reconstruction of the KP. C,The IVCM photograph of another patient with Vogt-Koyanagi-Harada syndrome,showing globular, noninfiltrating KP. D, The IVCM photograph of KP in a patientwith idiopathic granulomatous uveitis. The oblique image reveals smooth-roundedKP at the level of the endothelium. E, The IVCM photograph of KP in a patientwith sarcoidosis uveitis, showing multiple KP with a globular appearance.F, The IVCM photograph of another patient with sarcoidosis, showing KP witha globular appearance. ASL indicates Advanced Scanning Ltd (New Orleans, La).The depth from corneal epithelium at which the image was captured is indicatedin micrometers.
In vivo confocal microscopy imagesof keratic precipitates (KP) of 5 patients with active HLA-B27–associateduveitis. All of the KP had a stippled appearance. ASL indicates Advanced ScanningLtd (New Orleans, La). The depth from corneal epithelium at which the imageswere captured is indicated in micrometers.
Wertheim MS, Mathers WD, Planck SJ, Martin TM, Suhler EB, Smith JR, Rosenbaum JT. In Vivo Confocal Microscopy of Keratic Precipitates. Arch Ophthalmol. 2004;122(12):1773-1781. doi:10.1001/archopht.122.12.1773
Copyright 2004 American Medical Association. All Rights Reserved.Applicable FARS/DFARS Restrictions Apply to Government Use.2004
To evaluate the heterogeneity of keratic precipitates (KP) in varyingsubtypes of uveitis by in vivo confocal microscopy (IVCM).
The KP were viewed with a scanning confocal microscope in patients (n = 33)who sought care at a tertiary referral uveitis service for immune-mediatedand infectious forms of uveitis, including HLA-B27–associated uveitis,sarcoidosis, Vogt-Koyanagi-Harada syndrome, juvenile chronic arthritis, Fuchsheterochromic iridocyclitis, cytomegalovirus retinitis, herpes zoster ophthalmicus,ocular toxoplasmosis, and idiopathic uveitis. Images were captured and digitalizedin real time.
Forty-two eyes of 33 patients were examined in this study. Patient ageranged from 22 to 84 years, with a mean age of 49.4 years. Seventeen (52%)of the patients were women, and 16 patients (48%) were men. The KP rangedin diameter from 10 to 350 μm. We observed the following absolute andspeculative outcomes: KP are markedly heterogeneous and variable as documentedby IVCM; KP in individual patients are consistent throughout the cornea; themorphologic features of KP change across time; infectious vs noninfectiouscauses of uveitis seem to be readily distinguishable by using IVCM; and KPmay have consistency for specific disease states and therefore may have diagnosticimportance.
To our knowledge, this is the first time that IVCM has been used todescribe the architecture and heterogeneity of KP in uveitis. Such observationsreveal a heterogeneity that could not be appreciated by conventional slitlampmicroscopy and may have diagnostic relevance.
Formation of keratic precipitates (KP) is a characteristic finding invarious forms of intraocular inflammation, including uveitis and corneal transplantrejection. Generally, KP are created by the clustering of cells with adherenceto the corneal endothelium. Typically, these clusters are composed of epithelioidcells, lymphocytes, and polymorphonuclear cells.1 Generally,KP are classified into 2 major subgroups: granulomatous and nongranulomatous.Specific nomenclature has isolated certain subtypes of KP as in the stellateKP seen in Fuchs heterochromic iridocyclitis. Limited information exists onthe histologic characteristics of KP. To our knowledge, no studies have examinedthe specific histopathologic features of KP in different disease states. Onestudy2 used specular microscopy to describethe morphologic features of endothelial deposits in patients with cytomegalovirusretinitis. The reason for this paucity in the literature is simply that samplesof KP in specific active diseases are impossible to obtain without destroyingthe corneal endothelium, performing a penetrating keratoplasty, or waitingfor postmortem examination. Therefore, KP have been seen as an interestingby-product of inflammation and have played, until now, a limited role in theactual diagnosis and management of intraocular inflammation.
As its name reveals, the confocal microscope’s success revolvesaround the fact that both the observation and illumination systems can befocused on a single (confocal) point.3 Thisallows for excellent lateral (x-axis, y-axis) and axial (z-axis) resolution.Because the confocal microscope relies on a single point of reference, thefield of view is grossly limited. This is overcome by the rapid-scanning componentof the microscope, which uses a spinning metal plate that consists of a seriesof microscopic holes called a Nipkow disk, thus enabling reconstruction ofthe image with real-time viewing on-screen.4 Confocalmicroscopy allows in vivo examination of the human cornea at all cellularlevels. Physicians have used the confocal microscope to view the differentlayers of the corneal epithelium, Bowman membrane, stromal constituents (includingkeratocytes and corneal nerves), Descemet membrane, and endothelium. No study,to our knowledge, has ever specifically examined KP on the corneal endotheliumusing this system.
This study evaluates the heterogeneity of KP in varying subtypes ofuveitis by in vivo confocal microscopy (IVCM). We propose that with the useof IVCM of KP, the physician can make intelligent suppositions that may helpwith the diagnosis and subsequent treatment of the patient with intraocularinflammation. This method of confocal microscopy permits the evaluation ofKP with enhanced detail so that a differential diagnosis might be possiblebased on what adheres to the corneal endothelium and what lies beyond.
This study was approved by Oregon Health & Science University’sinstitutional review board. Patients 18 years or older who attended eitherthe tertiary referral Inflammatory Eye Disease Clinic or the Adult Eye Clinicat the Casey Eye Institute and who had evidence of KP from previous or activeocular inflammation were invited to participate in a study to delineate themorphologic features and extent of their KP using IVCM. Each patient signedan informed consent form before participation. Data collection included age,sex, ethnicity, ocular diagnosis and duration thereof, disease activity, currenttopical and systemic medication, and slitlamp color photographs or detaileddocumentation and description of the KP.
The ocular diagnoses were made using standard departmental protocols.This included a thorough history, including extensive review of systems, ophthalmicexamination, and pertinent laboratory or radiologic investigations. The generalapproach of this institution to differential diagnosis has been previouslydescribed.5- 8 Someexamples include the following. Ocular toxoplasmosis was diagnosed with thecombination of a typical clinical picture of a chorioretinal scar and positivetiters of immunoglobulins. Cytomegalovirus retinitis was diagnosed based onclinical appearance in an immunocompromised host and response to appropriatetherapy. We diagnosed HLA-B27–associated uveitis in patients with acute-onset,unilateral, anterior uveitis who had a positive test result for the HLA-B27antigen with or without associated back symptoms. Vogt-Koyanagi-Harada syndromewas diagnosed according to international consensus group recommendations.8
Patients’ corneas and KP were viewed with a scanning confocalmicroscope (ASL-1000; Advanced Scanning Ltd, New Orleans, La). This is a tandemscanning light microscope that incorporates a spinning Nipkow disk. The whitelight source was derived from a xenon lamp. An applanating objective restedcomfortably on the cornea once topical anesthetic was applied. The area observedwas approximately 500 × 500 μm at approximately ×400 magnification.Depth of field was 10 to 12 μm, and angular resolution was approximately1 to 2 μm.
Patients were seated comfortably at the IVCM with their chin restingon a standard slitlamp frame. Topical anesthetic in the form of 0.5% proparacainehydrochloride (Alcaine; Alcon Laboratories Inc, Fort Worth, Tex) was instilledin the affected eye. A drop of hydroxypropyl methylcellulose (GenTeal Gel;Novartis Ophthalmics, Duluth, Ga), used as an immersion substance, was appliedto the applanation head of the IVCM. The applanator was then gently placedon the patient’s cornea. The patient was examined in the primary positionand/or in elevation, depending on the location of the KP. The entire examination,on average, took less than 10 minutes.
Images were captured through a digitally enhanced black-and-white camera(KAPPA opto-electronics GmbH, Gleichen, Germany). Video capture ranged from1 to 5 minutes at 30 frames per second. These videos were then viewed andedited using Adobe Premiere version 6.5 (Adobe Systems Inc, San Jose, Calif),and individual frames were captured that showed KP at the level of the cornealendothelium or further into the anterior chamber. These individual frameswere then saved and viewed on their own. To delineate further the architectureof certain KP, 3-dimensional views were rendered from z-stacks of images obtainedfrom suitable video clips.
Forty-two eyes of 33 patients were examined in this study. The agesof patients ranged from 22 to 84 years, with a mean age of 49.4 years. Seventeen(52%) of the patients were women, and 16 (48%) patients were men. All but5 patients were white and had varying diagnoses (Table). The KP varied in diameter from 10 to 350 μm. Becausethis is a unique study that describes high-powered images of KP, we proposeour own illustrative and morphologic terms and use them descriptively in thetable and figures that follow. After reviewing all of the images, we decidedon 6 umbrella terms to describe the KP: globular, infiltrating, smooth-rounded, stippled, dendritiform, and cruciform.
Keratic precipitates are far more heterogeneous and variable on IVCMcompared with standard, low-powered slitlamp biomicroscopy (SLB) (Figure 1). Figure1 shows images of an 84-year-old man who was diagnosed as havingpostoperative endophthalmitis. The detail of the architecture and heterogeneityof KP are far more discernible using IVCM compared with standard SLB. Thestandard SLB photograph of the cornea and anterior segment (Figure 1A) reveals diffuse granulomatous KP scattered throughoutthe endothelium. In vivo confocal microscopy reveals large, infiltrating KP(Figure 1B) that correspond to the largeopacities seen on the SLB photograph (Figure 1A). These large KP are evidently made up of individual cellularcomponents clustered together. Note the soft, infiltrating border to the KP;this was a frequent finding in patients with infectious causes of uveitis.Between the large KP, IVCM reveals patches of fibrinlike material and cellularcomponents adherent to the endothelium (Figure1C). This type of detail is not discernible with SLB. The patternseen in Figure 1B and C was consistentthroughout the affected areas of the cornea.
Keratic precipitates in individual patients are consistent throughoutthe cornea, and in cases of bilateral disease, consistent between both eyes(Figure 2). An interesting observationwas that KP were consistent, in affected areas of the cornea, in individualpatients. The IVCM was performed in a 35-year-old African American woman (Figure 2A and B). The patient was diagnosed ashaving bilateral granulomatous uveitis secondary to sarcoidosis. She had 2+cellular activity in her anterior chamber at the time of IVCM and was notreceiving topical or systemic therapy. The central cornea was imaged usingIVCM and revealed multiple globular KP (Figure2A). The peripheral cornea showed consistent globular KP in thesame individual (Figure 2B). We performedIVCM in a 59-year-old white woman who had the diagnosis of idiopathic granulomatousuveitis (Figure 2C and D). Disease activityat the time of IVCM was 2+ anterior chamber cells, and she was taking 1% prednisoneacetate hourly for the affected eye. The IVCM of her peripheral cornea revealeda globular, stippled pattern of KP (Figure 2C).A slightly more central view of the KP showed consistent morphologic features(Figure 2D). The IVCM was performedin a 31-year-old Asian man who also had the diagnosis of idiopathic granulomatousuveitis (Figure 2E and F). He had milddisease activity at the time of IVCM, with only trace cells in the anteriorchamber, and was taking 1% prednisone acetate 3 times per week for the affectedeye. The central cornea shows smooth-rounded KP (Figure 2E), and the peripheral cornea shows a similar, consistentsmooth-rounded KP (Figure 2F).
The morphologic features of KP change across time, including a changewith disease progression and treatment (Figure3). Figure 3 shows IVCM imagestaken during a 3-week time frame of a 53-year-old white man with the diagnosisof idiopathic unilateral panuveitis. He received follow-up in the clinic fora 3-week period. At initial examination and before topical treatment, theKP showed a stippled pattern and were consistent throughout the affected cornea(Figure 3A). He began using topical1% prednisolone acetate hourly and was evaluated 5 days later. At 5 days afterthe initiation of therapy, his stippled pattern had changed to a smooth-roundedpattern (Figure 3B), and this too wasconsistent throughout the affected cornea. Twelve days after he initiatedsteroid therapy, his uveitis showed only mild activity, and his KP had nearlydisappeared. Only small, pigmented KP were observed on SLB. Pigment tendsto shine brightly on IVCM (Figure 3C).The shape and morphologic features of KP tend to change across time. Not onlyhave these changes been seen in treated patients (Figure 3), we have also noted them in disease progression with IVCM.
Infectious vs noninfectious causes of uveitis seem to be readily distinguishableby using IVCM (Figure 4 and Figure 5). Infectious causes of KP seem to havean infiltrating and dendritic appearance (Figure4). Walter et al2 performed a postmortemexamination in a patient with cytomegalovirus retinitis and observed chainsof dendritiform macrophages and fibrin adherent to the apical surface of thecorneal endothelium. We observed similar appearances in our patients withinfectious disease etiologies. Noninfectious causes appear to result in KPthat tend to be more smooth-rounded and globular (Figure 5). The IVCM was performed in a 62-year-old white man witha diagnosis of unilateral ocular toxoplasmosis. He had positive IgM and IgGantibodies with a typical chorioretinal scar (Figure 4A and B). Figure 4Ashows infiltrating KP with globular central cores and dendritiform pseudopodia,which we typically see in infectious causes of uveitis. Figure 4B reveals an area of the cornea between the larger KP, showingmultiple dendritiform bodies. The IVCM was performed in a 71-year-old whiteman with stage III multiple myeloma undergoing chemotherapy. He had a classicdiagnosis and appearance of cytomegalovirus retinitis with a positive serumtest result for antibodies (Figure 4Cand D). Disease activity at the time of IVCM showed 1+ anterior chamber cellswith flare. The IVCM revealed a similar infiltrating appearance to that seenin the toxoplasmosis case with dendritiform pseudopodia (Figure 4C). The image shown in Figure4C was taken at the endothelial layer; the image shown in Figure 4D was taken at a slightly deeper levelto the endothelium. We performed IVCM on a 59-year-old white woman with thediagnosis of herpes zoster ophthalmicus (Figure4E). The typical infectious signs are seen yet again in this example:central globular KP with infiltrating dendritiform pseudopodia. There is astriking similarity among all the images in Figure4, and there is an obvious difference when comparing them with theKP in the noninfectious causes of uveitis (Figure5).
Noninfectious granulomatous conditions are illustrated in Figure 5. At initial glance, the differences between the imagesin Figure 4 (infectious causes) and Figure 5 are striking and obvious. The infiltratingor dendritiform appearance gives way to globular and smooth-rounded KP. TheIVCM was performed in a 32-year-old white woman with the diagnosis of Vogt-Koyanagi-Haradasyndrome with serous retinal detachments (Figure5A). Her inflammation was not active at the time of IVCM. The imagereveals globular KP without an infiltrating appearance. The globular appearanceof the KP is emphasized in a 3-dimensional reconstruction of the KP (Figure 5B). This 3-dimensional image was renderedfrom z-stacks taken with the confocal microscope. The IVCM was performed ina 60-year-old white woman who also had a diagnosis of Vogt-Koyanagi-Haradasyndrome with serous retinal detachments (Figure5C). The image was taken while there was significant inflammation,with 3+ anterior chamber and 2+ vitreous activity. The IVCM shows globular,noninfiltrating KP. We performed IVCM in a 31-year-old Asian man with thediagnosis of idiopathic granulomatous uveitis (Figure 5D). This oblique image reveals smooth-rounded KP at thelevel of the endothelium. He had mild disease activity at the time with onlytrace cells in the anterior chamber. Figure 5E represents an image of KP in a 35-year-old African American womanwith the diagnosis of bilateral granulomatous uveitis secondary to sarcoidosis.A separate image from this patient is seen in Figure 2A and B. The KP imaged are multiple with a globular appearance.This patient had 2+ cellular activity at the time of IVCM. The final image, Figure 5F, is an IVCM image of a 45-year-oldwhite man with the diagnosis of biopsy-proven sarcoidosis. He had mild diseaseactivity at the time of imaging. The imaged KP have a globular appearance.
The KP may have consistency for specific disease states and thereforemay have diagnostic importance (Figure 6).We used IVCM to image the KP of 6 patients with active HLA-B27–associateduveitis (Figure 6A-E). The prominentfeature noted was that all of the KP had a stippled appearance and that theyhad no obvious similarity to the other KP imaged. These KP seem to be madeup of individual cells, as seen most clearly in Figure 6A and B. Lymphocytes measure between 8 and 12 μm, andpolymorphs measure approximately 10 to 12 μm in diameter. These sizescorrespond to the stippled areas seen in Figure6A-E. We therefore believe that the KP in HLA-B27–associateduveitis are made up of mainly single but at times small clusters of individualleukocytes. For 1 of the 6 patients with HLA-B27–associated uveitis,the images were of poor quality and are therefore not reproduced.
In vivo confocal microscopy has been effectively applied to diagnosingand describing corneal disease. Numerous case reports and studies9- 26 haveused IVCM to concentrate on keratitis, keratopathies, and other inflammatorydiseases of the cornea and endothelium. However, to our knowledge, there hasnever been a case report or study that concentrates solely on the appearanceof KP by IVCM in different disease states.
Specular microscopy has been used to image the corneal endothelium invarious diseases.2,27 One studyexamined KP by using specular microscopy and focused largely on the appearanceand consequence of the surrounding endothelium in 13 patients with differingdiagnoses.27
Although this is an initial study of KP, several novel conclusions areapparent. First, not surprisingly, KP show far greater heterogeneity usingIVCM compared with SLB. Despite their heterogeneity, morphologic featuresin individual patients and their fellow eyes are consistent. Certainly, thissuggests that the adhesion molecules and inflammatory mediators are consistentfor different locations in the cornea in a single patient. Second, our longitudinalexperience with patients is that the morphologic features of KP change withtime. Accordingly, in diseases such as Fuchs heterochromic iridocyclitis,we have observed diversity in morphologic features, which we believe reflectsdisease duration and treatment rather than heterogeneity in origin. Furtherexperience will adequately test this speculation.
Preliminary experience in observing patients with infectious causesof uveitis has shown that in all 9 patients with active infection of diverseetiology, the morphologic features of KP are distinctive with an infiltratingor dendritiform appearance. In our practice, we commonly encounter patientsfor whom an infectious cause of uveitis is suspected but not proved. Clearly,treatment is influenced by this distinction. Only much greater experiencewith IVCM will allow us to determine if the appearance of KP secondary toinfection is sufficiently consistent such that therapeutic decisions can bebased on the image.
The HLA-B27–associated uveitis was the disease that we had themost opportunity to evaluate. We have observed great consistency in theseKP. This observation leads us to speculate that IVCM may become an importanttool in differential diagnosis. However, we have not had the opportunity toimage acute-onset, nongranulomatous iritis in a patient who has a negativetest result for the HLA-B27 antigen. Additional study will be required todetermine the specificity of the stippled, unicellular appearance of KP thatwe see in patients who have a positive test result for this antigen.
To date, KP have received scant scrutiny. This study makes clear thata wealth of information could be gained by categorizing and defining theirdetailed structure. Morphometric analysis and correlative histologic and immunohistochemicalanalysis of individual KP in certain disease states may give further cluesto the specific composition and pathogenesis of KP. With this basic informationand a database of KP images in different diseases, noninvasive IVCM may becomean adjunctive investigational and diagnostic tool. The relatively small numberof patients in each clinical category in this study indicates that furtherstudies are required to more clearly define the KP in each condition.
Correspondence: Michael S. Wertheim, MBChB,MRCOphth, Casey Eye Institute, Oregon Health & Science University, 3375SW Terwilliger Blvd, Portland, OR 97201 (email@example.com).
Submitted for Publication: December 17, 2003;final revision received May 20, 2004; accepted May 27, 2004.
Financial Disclosure: None.
Reprints: James T. Rosenbaum, MD, Casey EyeInstitute, Oregon Health & Science University, 3375 SW Terwilliger Blvd,Portland, OR 97201 (firstname.lastname@example.org).
Funding/Support: This study was supported bya grant from the National Eye Institute, Bethesda, Md (RO3-EY014013), awardedto Dr Mathers. This study was also supported by unrestricted funds from Researchto Prevent Blindness, New York, NY, and the Stan and Madelle Rosenfeld FamilyTrust, Portland, Ore. Dr Rosenbaum is a senior scholar supported by Researchto Prevent Blindness. Drs Martin and Smith are supported by Research to PreventBlindness Career Development Awards.