Clinical Experience with FluorescenceRetinal Cinematography

By JOSEPH W. LINHART, M.D., HENRY D. MCINTOSH, M.D.,ALBERT HEYMAN, M.D., AND LEONARD M. HART

PHOTOGRAPHY of fluorescence in retinalvessels as described by Novotny and Alvis

affords a means of studying the dynamic char-acteristics of the retinal circulation.' Studiesin our laboratory and by other investigators 2using this method or a modification have dem-onstrated retinal vascular phenomena thatwould be difficult to recognize by ophthal-moscopic visualization alone. The single filmphotographic technic, however, has the dis-advantage of allowing exposures at relativelyslow speeds. This necessitates multiple injec-tions and grouping arrangements in order tocomplete a retinal vascular study.

In order to evaluate fully the moment-to-moment changes in the retinal circulation dur-ing various physiologic and pathologic con-ditions, a high-speed (up to 16 frames persecond) photographic method was developedin our laboratory. This report will describeour observations in 70 patients studied byfluorescence retinal cinematography. The ad-vantages and limitations of the method andthe clinical and investigative applications willalso be discussed.

Technic and ProcedureA Bausch and Lomb binocular ophthahnoscope

From the Cardiovascular Laboratory and the Neu-rology Division of the Department of Medicine, DukeUniversity Medical Center, and the Medical Illustra-tions Service of the Veterans Administration Hospital,Durham, North Carolina.

Supported in part by Grants B 669, H-6960, andHE-07563-01 from the National Heart Institute ofthe National Institutes of Health, U. S. Public HealthService, and Grants-in-Aid from the North CarolinaHeart Association.

Investigation carried out during Dr. Linhart's tenureof a Postdoctoral Fellowship from the National HeartInstitute, U. S. Public Health Service.

Dr. Linhart's present address is Cardiology Branch,National Heart Institute, Bethesda, Maryland.

Circulation, Volume XXIX, April 1964

and a modified Arriflex 16-mm. camera gearedto run between 10 and 16 frames per secondwere adapted for this study. The details of theinstrumentation have been reported elsewhere ,

and will be only briefly reviewed.The light obtained, from a 6-volt auto lamp

increased temporarily to 12 volts, was focused andpassed through a no. 47 (Kodak) blue Wrattengelatin filter (transmission, 420 to 470 milli-microns), and was then reflected into the ocularfundus. Light at this wavelength excites fluores-cence of the injected dye. Light reflected fromthe retina passed through a no. 56 (Kodak)green Wratten filter (transmission, 510 to 560millimicrons) to Kodak Cineflure film. This filterimpeded transmission of most of the reflectedblue rays from the nonvascular portions of theretina and increased the contrast between fluor-escent and nonfluorescent areas.

Full dilatation of the subject's pupils and fixa-tion of gaze are essential. A few feet of controlfilm were exposed with only the green filter inthe system. The blue filter was inserted, thecamera started, and approximately 10 ml. of5-per cent Fluorescite (Kirk and Company) wereinjected in an antecubital vein. Rapid transitthrough the venous system was encouraged byutilizing reactive hyperemia and a flushing tech-nic. The moment of injection was indicated onthe film by depressing a foot switch which mo-mentarily extinguished the illumination.The completed motion-picture study was ex-

amined in a Weinberg-Watson converted Kodakprojector. This permitted different speeds of pro-jection as well as single frame examinations. Ifmore detailed analyses were desired, film stripscould be printed on photographic paper.

Results and CommentsNormal Circulation

In the normal retinal circulation, as dem-onstrated in 15 patients, fluorescein appearedat the optic disc 8 to 10 seconds after in-jection into an antecubital vein. Arterioles inthe macular region fluoresced first, followedquickly by the remainder of the arteriolarsystem. This arteriolar phase was usually com-

577

by guest on October 4, 2017

http://circ.ahajournals.org/D

ownloaded from

LIN5LIART ET AL.

plete wXithlinl 1 to 2 se( ondls. F1luorescein ap-peared inlitially in the C'central portionl of thearterJiole alnd xwas folloXve(l ipy completeopaci-fication of tlhe luminenj. This front" as characteTristi of laminia flow. Imminedi.atelv afterthe arteriolar phase, a fainit diffuse backgroundfluorescenice appeared, wlhichl was probablyduie to fluoresecinl in retinal capillaries and thechoroidal vesseils. The first phase of fliiores-encee of the venous circulation appearedbefore- the dye liad completely filled thearterioles in the niasal portion of the retina.Filuorescence did niot appear at the samemomeit in all of the small veins draining intoa major venouis chlannel. It mighit appear inone small veinl as mtich as three secolids beforeit appearedl in a similar sized companioniveini. As the fluorescein-laden 1)10ood draine(linto a larger vein, fluioreseence appeare(l alongthe lateral margin of the larger vein anidstreamed toward the optic (lise. As more of th:essmall veins draininig inito a larger vein coni-tained ftluorescein, the streaming effect be-camieit less apparenit (fig. 1). As prevtiouslydescribed iby Dollery anid his associates,streaming is abruptly (lisrupted at points ofairteriovenllonis nickinig xvhere turbuilence exists.Streaminig in the veinis was seen in all sttudies

Figure 1

Venous streaning. (This and .suibseqenlicTt figuires ore

enlargements of a sinrgle frame from iG-nus. film, runrait 10 frarnes per second.) The arterioles and capil-laris atie fluorescent and the first phase of venous fill-ing hias beguin. The fluoresceinl itn the reins is seen

streanming along the lateral margins and at the june-

/iontl of twvo veins (arroxi7 .

but xxas m.ost promninent in patienlts witlh theslovest retinal 1)l0(1d flo. The plheiomencionm.ay bc (duec to (liffereuces in velocity of flowthlrl(luglh segments of the retinial circulation orto (lifferences in the lenigtlh of the vaseilar eir-cuiits in varioLus areas of the retinia.The emptying of the veins followed a pat-

tern opposite to the filling phase. Fluoresceindisappeared first from the lateral margins ofthe veins as b)l0(1od withotit fluorescein en-terecl fronm the smaller xeims. The rtetinialvenotus plhase genierally lasted 15 to 20 see-ondls. The retinal background fluloresceniceslowly faded and usually disappeared alongwith the loss of xvenous fluorescence.

Abnormalities in the Retinal CirculationDiabetic Retiniop)athlyi

Hi.stologic studies liave demonstrated thatman-y microaneuirysms in diabetic retinopathyare not visil)le by conventional ophthalmo-scopic examination.4 A greater number of schlmicroaneurysms are visible by the flnoresceintechniic than by routine retinial plhotogr-aplhy(fig. 2). A significant iiinber of suielh lesionisxvas demonstrated in all but txvo of 12 diabeticpatients stuidied. Diabetic patienits also hadlvascul.ar structiures which appeared clinicallyto be microaneulrysms blut vhlich did niotfluoresce. Presuimably, these strtuctuires xvereeither small retinal hemorlrhages or oceluidedaneuirysms, as demonstrated onl postmortenmexamination by Cogan and hisi associates.' Thecapillary aneu.ryssms were freqnenitly seeni inclose proximity to flame-shaped lhenmorrlagesthat clidt not fluloresce and near soft extidatesthat did fluoresce. The microan euirysms uisuallyfilled diuring the late arteriolar or early venouisphiases. In conitrast to the rest of the eirctila-tion., they appeared to have a prolonged emp-tying time and to retain fluorescence afterthe otler vessels lhad been completely clearedof fluorescein. It is n:ot known whether suchprolonged flulorescence xvas due to sloxv cir-etilation of the dye thiroughi these struc-tuires or to staining of the walls of theinicroaneuirysms, xhich may lhave an increasedaffinity or permeability for fluorescein.

Narroxved and tortuious retinal vessels xvereseen in older diabetic subjects, consistent with

(irtliatinn. V olut XXlX,\'Y\ April 1964

578

by guest on October 4, 2017

http://circ.ahajournals.org/D

ownloaded from

RETINAL CINEMATOGRAPHY

Figure 2Diabetic nuicroauieuirys.7ns. Top. The control film (he-

fore injection of fluorescein) of the retina int a diabetic)atient. A heemorrhagic area is present (dark area

(l)bove the lighit reflex) and onte microaneu.rysiai is seent

(aIrow), Ihich7 does u-tot fluloresce. Bottom. After the

fliuoreseCeinrt injectiont Ttunuerousp1iilctuate areas of flu-orescence are seen, whiich have the hemodynamiccharacteristics of (liaIbetic uliiroau)i(,trysmts. Otne tmiicro-aTneutrysmi mtaiy Ie seen in the hemorrhkagic area

(arrow ), which. oftherwuise (Jars not fluoresce-

atherosclerotic and lhypertensive changes. Inmany patienits these changes appeared more

promninent by the flulorescein technic than withordinary liglht ophthalmoscopy, since more andsmaller vessels were visible with the formernethod. FluloreseenceIhas been] demonstratedin areas of neovasctlarization in diabetic fuindi.These new vessels apparenitly h1ave ani in-creased permeability resulting in large fluor-escing exudative lesions. Suelh exudative les-iois were not appreciated by convenitionalo:phtlh.almoiscopic examination.

(ircelation, Vol/um XXIXl, APril 1964

The retinial circulalctioni timlPe noraially rangedfrom 15 to 20 second(Is wlhen iineasure(l frcomiithe first appearance of fluioresceini in an arteryto its dlisappearac.ieie ini a corresponding xveili.In fotour of six dialetic patienlts in wliilch thiscould le measuired, the time xvas prolonged(from 25 to 95 seconds) in the absence ofany other factor that wouild tend to distorta dye curve. In one patient with mild diabetesof 9 years' duiiationi, it was prolonged 3.5 to40 seconds, dlespite ain otherwise o-inormal ret-inial cireclation.

Hypertensive RetitopathylTen patients with hyper-tenisive vascuilar dis-

ease were studied. Arteriolar tortuosity andspasm were clearly seen by the fluoresceintechnic. Flame-shaped lhemorrlhages did notfluoresce. Exudates in more advanced stagesof hypertension could be divided inito twogroups as regards flulorescence. Those thathlad the appearance of hard exudates did notdevelop fluorescence. The new, soft, or so-called cotton-wool lesions were noted to fluor-esce brightly and to retain this fluorescenceafter the rest of the retinal structtlres lhad beencleared of dye. In many cases this would lastfor hours. As noted by Dollery, Il-odge, andEnigel," these lesions may be resorbed clinically duiring the course of treatment forhypertenision, but fluorescenice may still ap-pear at the previous site. Also a flutoreseniitarea cani precede a visible soft exludate, dem-o.strating in each instance an increase incapillary perrmeability, which is otlhervise notvisible. It is not uniicommon to see the area offiluorescence extend oveir a larger area thanthat covered by a xisil)le exuidate.

Maligniaint lhypertension with papilledemapresented a striking fluloreseence. There wasa marked increase in vascularization of theedematotus nierve head anid, suil)sequently, adiffuise fluloresceciee (fig. 3) which imiight lastfor hours. It appeared that there vas a markeddegree of capillary permeability in the areaabouit the optic disc. The extent of the vas-cularization and diffuse fluiorescence seemedto parallel the clinical severity of the papill-edema. These vessels disappear as the papill-edema regresses."

579

by guest on October 4, 2017

http://circ.ahajournals.org/D

ownloaded from

LINIHART ET A.

The vasctular strtictiuires in a patient witlhLinden-von Ilippel (lisease xxer]e (lifthetilt tosee l)x ordinary ophthlalimioscol,xpv hut w(er(eclearly demionstrated iby thfluoesei techiimec. The structure seems to l)e a vast net-work witlh a (liffiusion of (lye into the sur-r_otundiing areat. The fiuorescence remnained atleast 5 minutes after the rest of the vascuilar.struictuires laid returned to normal.

It lhas l)een possible in otler patients todemonistrate vascuilar disruption in ar-eas ofchorioretinitis and abl)n ormal fillinig patternis inpatien.ts xvitl vascuilar occlusions. Thie teclhnicmiay he uisefutl in distinigtuislinig large retinallhemorrhagres fromii tuimnors suchl as mnelanomas.

Figure 3

Ppi.)llederoa secondlarl to itncr.ea,eseld intr.acraniail pies-su.e. Top. Early venous phase. The neovascudarizationof the disc is present aind initial exudation of fltiores-ceini is seent. Bottom. The mo/or vessels arc notecleared of flinorescein ( dark lin-es) anid cliff t.sc fluior.es-c&ncc of the dlisc renintins.

Miscellaneous StudiesRetinal cinematography was carried out in

a patient with leuLkemia (fig. 4). Oni clinicalgrounds, the vaseuilar lesions present wereconsidered to be retinal hemorrhages. Injec-tion of fihioresceiin, hiowever, revealed some ofthese structures to he in continuity with thevascular system filling during the late arteriolarphase, presenting the appearance of diabeticmicroaneeurysms. In contrast to the diabetic Figure 4lesion, hovever, the fluiorescence was of sshort \nIicronrqstns in lenket>inia. ToT1/c cdark pnnctatedulration, disappearinig with the venouis phlase. aireas itn thuL cortiol filni itece clrko recl in vito. Bot-Fluorescence did not appear in otlher similair tom. After flinoresceia inijectior tian appearanice siunilat

to chah.7etic waisnhiteroanerisnis i.s .sc cn hc s'1/pnnc()tatc(tdlesions in this patienit wlhichl mav havxe been cire i c is. s i Tt/ f/ictcnnhemorrlhages o(r occlded mi-ncraenry.smies. pi/iys

(iual,iion., Vo/iim \XIX, /prii 1964

.51380

by guest on October 4, 2017

http://circ.ahajournals.org/D

ownloaded from

RETINAL CINEMATOGRAPHY

Such a differentiation should be possible butexperience is limited at the present time.A fluorescein injection was carried out in a

patient who had a "Hollenhorst plaque" in aretinal arteriole. The fluorescein injection dem-onstrated adequate, although delayed, flowpast the plaque. These plaques often causevascular occlusion.7One of the first uses for fluorescein retinal

cinematography was in the study of the effectsof hypothermia and high- and low-flow car-diopulmonary bypass in the dog. The differ-ences in the rate of flow under these variousconditions were clearly evident.8

DiscussionRetinal cinematographic studies in our

laboratory have demonstrated the value ofthis technic with use of filming rates up to16 frames per second. The entire sequence ofretinal flow is dramatically and continuouslyportrayed without resort to multiple injectionsand grouping arrangements necessary withthe still photographic technic. The details ofthe finer vascular structures at present arebetter with the still technic, for the electronicflash permits the films to be exposed at 1/1,000of a second. The detail of the cinematogramis usually satisfactory, however, so that rarelyis it necessary to resort to the still technic. Astill color photograph of the retina is essentialfor interpreting a fluorescent cinematogram.The still color photograph aids in orientationof the cinematogram. Even more importantis recognition of structures that have failed tofluoresce (i.e., hemorrhages) or fluorescencethat was not visible on the color photo (i.e.,microaneurysm).We are investigating methods of increasing

the motion picture speed without loss of de-tail. This may be possible by improving thelight source. Increasing the concentration ofdye at the retinal site should also improve thedetail. With use of an indwelling needle inconjunction with reactive hyperemia and aquick saline flush, a good bolus of dye reachesthe retinal vessels in most patients. Othermethods for insuring a high concentration ofdye in the retinal vessels may be used in pa-tients with larger blood volumes, low cardiacCirculation, Volume XXIX, April 1964

output, or polycythemia. Dollery and asso-ciates have used a polyethylene catheter in thesuperior vena cava or innominate artery2 5

with good results, and we have made directinjections into the root of the aorta.

Fluorescein cinematography has proved val-uable in several ways. First of all, it has al-lowed us to evaluate the normal retinal cir-culation. That the flow in arterioles is laminaris clearly demonstrated. The streaming offluorescein seen in large veins graphically dem-onstrates that the retinal circulation is com-posed of a number of segments, each ofwhich may have different flow characteristics,depending on a difference in velocity of flowin certain segments or a difference in segmentlength.

This technic should lead to a more thoroughunderstanding of the sequence of events inthe pathogenesis of retinal hemorrhages, exu-dates, microaneurysms, vascular occlusions,and other disease states in which the conse-quences of vascular change may be manifest.The changes in the blood vessels in arteri-osclerotic, diabetic, and hypertensive patientshave been clearly demonstrated and flow char-acteristics were elucidated.

It also seems probable that the differentia-tion of retinal or choroidal hemorrhages fromtumors such as melanomas may be possible.In fact, any vascular lesion of the retina orchoroid, such as those with Lindau-von Hippeldisease or vascular occlusions, may be photo-graphed with this method. Follow-up studiesafter surgery can then be accomplished andthe results evaluated.The cinematographic technic offers an ad-

vantage in elucidating changes in caliber andespecially flow in retinal vessels and indirectlyin cerebral vessels under a variety of physio-logic stimuli. Cinematographic technics permitan accurate timing of the various phases ofthe retinal circulation. The retinal circulationtime has been measured as that time from thefirst appearance of dye in a retinal arterioleuntil its disappearance from a correspondingvein.

Generally, the injection of fluorescein pro-duced no symptoms. In patients with faircomplexions, a pale yellow hue was observed

581

by guest on October 4, 2017

http://circ.ahajournals.org/D

ownloaded from

LINHART ET AL.

for 8 to 12 hours with subsequent discolora-tion of the urine for approximately 24 hours.In 10 per cent of our patients nausea andvomiting occurred within 1 minute of the in-jection and wvas usually associated with thelarger doses (10 ml.). The nausea was usuallytransitory and subsided spontaneously within5 minutes. Four milliliters of 5-per cent fluor-escein injected into the root of the aorta pro-duced no nioticeable effects in one patient,and injections of the ininominate artery havebeen well tolerated.)Poor retin-al cinematographs are occasional-

ly produced for reasons other than thosedirectly related to camera technic. Poor pa-tient fixation or inability to cooperate, slowgeneral circulation, and various cardiovasculardisturbances may result in poor contrast films.

SummaryClinical experience with fluorescence retina]

cinematography in 70 patients is presented.The sequence of normal retinal vascular

filling is described. Of particular interest isthe demonstration that the retinal circulationconsists of a number of vascular segments,each of which may have different flow charac-teristics.The pathologic alterations in patients with

diabetic and hypertensive retinopathy, papill-edema, and other disease states are discussed.The dynamic portrayal of the retinal circula-

tion, enabling moment-to-moment evaluation,and the efficiency of the cinematographictechnic are distinct advantages over the single-film method.

References1. NOVOTNY, H. R., AND ALVIS, D. L.: A method of

photographing fluorescence in circulating bloodin the human retina. Circulation 24: 82, 1961.

2. DOLLERY, C. T., HODGE, J. V., AND ENGEL, M.:Studies of the retinal circulation with fluores-cein. Brit. M. J. 2: 1210, 1962.

3. HART, L. M., HEY,MAN, A., LINHART, J. WV.,MCINTOSH, H. D., AND DAVID, N. J.: Fluores-cence motion picture photography of theretinal circulation: A description of techniqueand normal retinal circulation. J. Lab. & Clin.Med. 62: 703, 1963.

4. COGAN, D. G., ToUSSAINT, D., AND KUWABARA, T.:Retinal vascular patterns IV: Diabetic retin-opathy. Arch. Ophth. 66: 367, 1961.

5. DOLLERY, C. T., HODGE, J. V., AND ENGEL, M.:Retinal photography using fluorescein. Med.Biol. Illus. 13: 4, 1963.

6. DAVID, N. J., HEYMAN, A., AND HART, L. M.:Fluorescence photography of retinal vesselsand optic discs in papiledema. Tr. Am.Neurol. A. In press.

7. HOLLENHORST, R. WV.: Significance of brightplaques in the retinal arterioles. J.A.M.A. 178:123, 1961.

8. WVHALEN, R. E., BROWN, I. W., JR., SMITH, XV. XV.,MCINTOSH, H. D., AND MARGOLIS, G: Perfu-sion hypothermia studies in dogs; cineflu-orographic and hemodynamic stuidies. Arch.Surg. 86: 118, 1963.

Circidalion, Volmnie XXIX. April 1964

582

by guest on October 4, 2017

http://circ.ahajournals.org/D

ownloaded from

LEONARD M. HARTJOSEPH W. LINHART, HENRY D. MCINTOSH, ALBERT HEYMAN and

Clinical Experience with Fluorescence Retinal Cinematography

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1964 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/01.CIR.29.4.577

1964;29:577-582Circulation. 

http://circ.ahajournals.org/content/29/4/577located on the World Wide Web at:

The online version of this article, along with updated information and services, is

  http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. and Rights Question and Answer

Permissionsthe Web page under Services. Further information about this process is available in thewhich permission is being requested is located, click Request Permissions in the middle column ofClearance Center, not the Editorial Office. Once the online version of the published article for

can be obtained via RightsLink, a service of the CopyrightCirculationoriginally published in Requests for permissions to reproduce figures, tables, or portions of articlesPermissions:

by guest on October 4, 2017

http://circ.ahajournals.org/D

ownloaded from