Journal of Neuro-Ophthalmology, September 1986, Volume 6, Issue 3

Purtscher's retinopathy secondary to pancreatitis. Aspects of the topography of retinal abnormalities.

A 22-year-old man with acute pancreatitis developed sudden bilateral blindness with funduscopic features of Purtscher's retinopathy. The topography of retinal abnormalities has been ascribed to the properties of the retinal capillary network. However, the similarity between the distribution of the retinal abnormalities and the density of the nerve fiber layer suggests that the observed topography may, in fact, result from the anatomical characteristics of the nerve fiber layer.

'(" lYHh Raven Pre"S, New York Pu rtscher' s Retinopathy Secondary to Pa ncrea ti tis Aspects of the Topography of Retinal Abnormalities Pradip K. Toshniwal, M.D., Andrew A. Berman, M.D., and Alan J. Axelrod, M. D. A 22-vt'ar-llld 111,111 wilh aClIlL' p,lncrl',ltiti~ de\'eh'ped ~lIddl;n bilatt'r,ll blindnl'ss with fllndl1sclIpic Il',lll1rl'S llf rurl~cher'~ rdinop,llhy. The Il1pll~raphv 01 rl'lin,ll ab­normalitie~ h,lS been ascribl'd III the propl'rti('~ of the rt'tinal capillary ndwork HllWt'\'l'r, tl1L' similMit\' be­twt'en the dislribution of the rdin,,1 abnorm,llilie~ ,md tht' density of the nl'rve fiber 1,1\'l'r suggests that the l)b­served to~';ography may, in filct: re~lIlt fn1l11 the an,1!om­ieal eharaelt'risties of the nerve fiber laver. From thl' Divisions lit Nl'Urologv ,mel Ophthalnllllogv, COl,k County Hospital, ChlCdgo, Illinois. Address correspondl'nn' <1nd f\'print rl'qul'sts ILl Pr,lLiip K. Toshmwal, M.D., DIvISIon ot Nl'Urolllg\', Cllok Count\' Hos-pital. IH25 W. HMrison Sln'l'l, Chic<1go, '11. hUhl2. . 160 Acute sewre visual loss in the context of pancre­atitis results from retinopathy which bears the eponym, Purtscher's retinopathy (1). Retinal ex­amination is characteristic; the abnormalities con­sist of large contluent dense exudates, edema, and superficial hemorrhages confined to the area be­tween the disc and the maculae (2). This localiza­tion has been suggested to result from the topo­graphical variation in the properties of the retinal capillar:' network, in that the capillaries in the central retina are said to be end vessels, while those in the peripheral retina are said to anasto­mose freely (3,-1). The pathogenesis of retinal abnormalities has been a matter of considerable debate. While it is accepted generally' that capillary occlusion is the underlving mechanism, the nature of the occlusive material remains controversial. Early investiga­tions suggested free fat embolization, but recent studies of Jacobs et al. offer important evidence to suggest granu!oc\'te and monocyte aggregates as an alternative (-1). In this report, we describe a patient with pan­creatitis who de\'eloped Purtscher's retinopathy and offer an alternative physiological explanation for the topograph:' of the retinal abnormalities ob­served. CASE REPORT A 22-year-old black man was hospitalized with a 3-day history of severe epigastric pain radiating to the back and associated with recurrent vomiting, No history of other medical problems was present. The patient admitted to heavy alcohol use, but de­nied any other drug habits. The diagnosis of acute pancreatitis was confirmed by high serum and uri- PW\TSClII:.'N.'S InrtNO[JI\TfIY /6/ nary amylase. Hemo~lobin electrophoresis showed evidence of sickle tr,1it, but otill' I' bloud studies indudin~ VDRL were nurm,11. The tuxi­cology screen \-."as ne~,1ti\'l' for dh,1I1UL n1l'th,1I1ul, barbiturates, and bl'nzociiazq~illl's_ On the st'cl)l1d hl)spit,11 d,n', the p,1til'nt l'Xpl'ri­enced acute blurrin~ of vision in\'olvin~ both eves, which bel.-anw worse over the next il.'w hours. Ex­aminatil) n l)l1 the fl)urth Iwspit,11 d,l\' reve,1ll'd vi­sual acuity tl) be counting fingl'rs ,1t 2 il.'l'l bil,1!l'r­ally. Nl) impwven1l'nt W,1S ,Khil'\'I.'d with pinhull.' or rt'iractil)n. Pupils Wt'rl' ::; mm ruund, equal ,1nd reactive tl) light withl)ut ,1I1V rl'I,1tivl' ,1iierent pu­F'illar~' ddect. Ex.tr,wcul.n muvements were normal. The extt'rnall'x,lminatiun and lenses were l1l)rmal. Intraocular pressures Wt're 12 mm Hg in both eves, and the media were dear. The fundus examination was remarkable for the presence of large. dense, coniluent "exudates" most promi­nently invoh'ing the area around and between the discs and the maculae (Fi~. 1). The peripheral retina was relativelv nl)rmal. The discs showed some blurring of the margins, and the retinal vessels were normal except for being obscured by overlying exudate in places. A single nerve fiber layer hemorrhage was seen on the left side. Both maculae had a cherry red spot appearann'. Fluorescein angiography was done on the same day and revealed absence of capillary iluorescence in the same areas as the exudates described above (Fig. 2). This could have been secondary to capil­lary dropout or obscuration by overlying exudates or both. The choroidal tluorescence also was af­fected similarly. Early lye leakage and occlusion of small terminal peripapillary arterioles were noted. On follow-up, the patient's vision did not im­prove. At 4 weeks, the exudates had resolved par­tially (Fig. 3). At 16 weeks, fundoscupy demon­strated pale discs, vascular sheathing, and no exu­dates. The retina in the peripapillary region had a mottled appearance (Fig. 4). A iluorescein angio­gram at this time revealed dt'layed retinal arteri­olar filling with pigment mottling and are,lS of hyper- and hypofluoresence. The previuusly oc­cluded arterioles now filled with fluorescein, and there was evident widening uf the perifuVt'ill c'lpil­lary net (Fig. 5). DISCUSSION In the patient described, acute bilatl'rill Sl'vere visual loss, confluent dense exudiltes locillized to the posterior pole, peripapillary retinill hemor­rhage, and fluorescein angiographic demonstril-tiun of capillary dosure ilnd retinal non perfusion ilre chilractl'ristic of Purtscher's retinopathy (2). Besides pilncreiltitis, similar changes have been described with major injuries with multiple frac­tures, as well ilS with minor trauma, head injury, surgl'ry of fatty tissues, chest compression, and ophthillmologicill hydrostiltic pressure syndrome. The rl'tinal abnormalities are thought to result from capillMY occlusion and infarction. Fluores­cein ilngiogrilphic findings are consistent with this postula tion. Fu rther su pport is provided by the experimental studies of Ashton and Henkind, in which infusion of glass microspheres with diam­eters of 15-75 f..lm into the carotid artery of an an­imal resulted in a similar fundus picture (5). Sim­ilar observations have been made with experi­mental fibrin clot and air embolism. In the only report on the retinal pathology of Purtscher's reti­nopathy with pancreatitis, occlusion of both ret­inal and choroidal arterioles was demonstrated (6). It has been proposed previously that the topog­raphy of the retinal changes in Purtscher's reti­nopathy results from the peculiarities of the retinal capillary network, where the central area is sup­plied by end vessels and is thought to be more susceptible to ischemia, while the peripheral retina is supplied by an anastomosing plexus of capillaries (3,4). However, in a generalized process of embolic ocdusion (as opposed to a localized process), all similarly sized capillaries are likel~' to be affected equally and, therefore, the presence or absence of anastomoses may be inconsequential. Since the experimental studies of McLeod et aI., cotton wool spots are thought to result from the blockade ofaxoplasmic tlow and accumulation of organelles in the axons (7,8). In the present case, as well as in the ones previouslv reported, obscur­ation of the retinal vessels bv the exudates sug­gests their superficial localization in the retinal nerve fiber layer (NFL). Ii ll(1t:' cl.H1siders the topog­raphy of the NFL as elucidated bv Radius and An­derson, the greatest density and thickness of fibers is found in the arcuate bundles and the pa­pilll) macular bundle in the peripapillarv region (l)- I I). It is h)r this reaSl)n that the NFL defects beCllme appilrent in the same regilm (12,13). Simi­lilrly, in the case of bll)cked axoplasmic tlow and ilxon,ll swelling, Cottlll1 wl)ol spots are likely to be found in the areas with the greatest axon density, where they become obvious because of the con­iluence of swollen axons. The composite fundus photograph illustrates the remilrkable similarity bl'tween the topography of the exudates and the retinal areas with ~reater nerve fiber density (Fig. I). It is importilnt to note that this distribution is (b) I'. 1\. TOSffNIWAL £T AL. FIG. 1. Composite fundus photographs of left (a) and right (b) eyes show bilateral. large. confluent, dense retinal opacities, predominantly involving the areas be­tween and around the disc and the maculae. PURTSCH£WS RLTlNOPATHY /63 FIG.2. Fluorescein angiogram shows lack of capillary fluorescence in the same areas as the retinal opac­ities. Early dye leakage and occlusion of small ter­minal peripapillary arterioles are noted. different from that of the radial peripapillary capil­laries described by Wise et al. and is said to have fewer anastomoses (3). These latter vessels follow the temporal vascular arcades and are not present in the perifoveal region, an area prominently in­volved in the present and other cases of Purtscher's retinopathy. Some prominence of the NFL, probably secondary to swelling, is seen in FIG. 4. Fundus photograph at 16 weeks shows pale disc, vascular sheathing, and mottled appearance of peripapillary retina. Retinal opacities observed earlier have resolved. the peripheral retina also, but no cotton wool spots are observed. Sparing of the macula and its cherry red spot appearance is seen in this patient and also has been commented on previously (2). Since no nerve fibers are present over the macula, lack of an over­lying exudate and the transmission of choroidal color easily explain the cherry red spot. FIG. 3. Fundus photograph at 4 weeks shows resolving exudates and vascu lar sheathing. I elill NCllrtl-OI,litlwllllol. Vol. 6, No.3, 1986 1M I). 1\.. T05ffNIWAL ET AL. FIG. 5. Fluorescein angiogram at 16 weeks shows pigment mottling and areas of hyper- and hypofluo­rescence. Arteriolar filling is now normal and perlfo­veal capillary net shows apparent widening. There has been considerable interest in the na­ture of occlusive material. Demonstration of fat emboli in the systemic circulation led to the impli­cation of fat as the embolic material (l.t, 15). Re­cently, Jacobs et al. and others have accumulated evidence to show complement activation, prob­ably secondary to release of proteolytic enzymes (e.g., trypsin) from the inflamed pancreas (-1,16­18). Complement factor C5a in vitro has been shown to cause transient leucoaggregation (lll). In vivo, infusion of a small bolus of C5a into the aorta leads to leucocyte margination and aggregation, which can be demonstrated by laser intravital mi­croscopy of fluoresceinated leucocytes in mt'sen­teric microvasculature (20). The size of the aggre­gates may be up to 80 fJ.m, which is well within the size range found by Ashton and Henkind tl) be functionally important in capillary occlusion (5). Activated leucocyte aggregaks have been shown to damage cultured endothelial cells through the production of oxygen frel' radicals (21,22). In vivo, leakage of plasma and red cells into the inkrstitial tissues has been demonstrated after intraaortic C5a infusion (20). While in the illdex case of Jacob et a!., studies to show complement activation and leucoaggrega­tion were not carried out, in a second case, the same group has demonstrated these abnormalities (4,23). Leucoaggregation and consequent endo­thelial damage also havl' been shown to be the probable cause of pulmonary dYsfunction and acute respiratory distress syndrom.e. associated with pancreatitis. Besides pancreatitIs, com~le­ment activation also is known to occur followmg trauma, the other common precipitating cause of Purtscher's retinopathy (24). It is of interest that both leucoaggregation and free radical mediated endothelial injury can be inhibited by large doses of steroids (25,26). There have been, however, no clinical studies yet to show their therapeutic effi­cacy in a case of retinal injury. Acknowledgment: We would like to thank Dr. James A. Goodwin for reviewing the manuscript. REFERENCES Purhcher 0 :-';"ch unbekanntt.' Befunde nach Schadel­trauma \'l''''1II111l DI"iI Llpiltilallll,'1 Ge, 1910;36:294-301. ~ Inkell" D~1. Wabh jB Retinal fat emboli as a sequela to acute pancrealItl' .'1111 I Llpiltilallliol IYi5;l'\U:Y35-S, 3 Whe G:-';. D"lIe[\' CT. HenkInd P Till' Retlll'll ClrculatlOll, :-';l'\\' Y"rk Harp~r and R"". IYi12U-31 4, jacllb HS. G"ld,teln 1~1. Shaplr" I. t.'t al. Sudden blindness In dcute pdncredllll', P''''lbl,' wle ut cumplement induced rt.'tIndlleuk",'mbohLallun :\r,'iI IlIt,'m .\!t'd IYSI,141:134-6. ; A,hlun:-';. HenkInd P E\'perimental ocelu'lOn of retinal artt.'rlllle, USing gradl'd gldSS baliotIni Br I LlI'iltilalllJLlI 14"5;44225-34, ", Kincaid \1C. Green WR. Kno\' DL et al. A e1intcopatho­iL) glCdl Cdse rep,)rt ,)1 relln,)path\' ,)t pancredlIlIs, Br , 01"1­til, I/IIl"11482;bb:214-2,, i \lcLe"d D. \lar-hdll I. Kuhner E\1. et .11. The r,)le ofaxo­pl.) smlc transpurt In the path,)genesh ,)t retinal cotton ,,0,,1 'pelts Br' l)I,I,til,II"I,': 14;-;-:bl 1;-;--41 5 \lcLe"d D Ophthdlm,"c,)pIC 'Igns ot ubstructed a\'o­pI. lsmlC trdn'p,)rt alter ,)cular \'ascular ,)celuslon, Br' Opil­til. lI"It,114;-b;bU:551-b 4 RadiUS RL Anderst'O DR: The course ,)t a\','Os through the retInd and ')PIIC m'[\e head Ar,iI l)piltllallll,'1 19i9;9i; 1154-5 III Radlu, RL Anderst'n DR. The hlst,)I,)~\' ,)t reltnal nerve lIbl'r I.l\·er bundles .1Od bundle detects', .-'Ird, l)phtlllllm,)1 14;-4;4:441'-5(1 II RadiUS RL: ThlCkne,s "I the' retinal ne[\'e fiber laver in pri­mate l'\'eS, .-'Irdll)l,iTth,llII'll! 14S0;4S:lb25-9, 12 H,)vt WI'. Fn,en L: :-,;"" man :-';\1: Funduscopv of nerve lIbl'r ddec!,; 10 gl.lUC,)m.l 1'1,',',1 L)phtll'llm"1 19i3;12;814­24. D :-';l'II'01.1I1 :-';~1. T,)rnamb,' rE, C,)rbdt II: Ophthalmoscopv "t the' rdlO.)1 n,'[\'l' lIbl'r Id\','r u,e 10 detecltun l)f neuro­I,' gl( di,ea,,' ..-\rd, '\'<'11",1 14l'\2;39:22b-33 14, Ednwnds,)n HA, FIl'1d, lA: Relatwn of calcium and lipids ttl .\Cult' p.lncreatlc necr,)si" Ar,iT Illt<,m .\lcd 1942;69;177­4( 1 I=. Lvnch till :""phr')'I' .md tat t'mboli,m in acute hemor­rhagIC p.mcrealttls, .-'Ird, Illtem .\lcd 1954;94:i09-li. Ib, Coldstl'lO 1~1. Cal.) D. R.ldin A. d al.: Evidence of comple­nwnl c.)t.lb,)h,01 111 .)cult' pancreatitis. AIIl , ,\fed Sci IQ7H;27:;:2:i7-h-.l. Ii Fouli, AK, ~lurr.1\' \\'R. Galloll'av D, et al.: Endotoxaemia ,1I1d complement 'activation in acute pancreatitis in man. ClIt 141'2;23:944-50, Il'\ Wich,'r)T, Barnt" Mr, Cooper M}, et al.: Complement ac­lI\' dllon and complemt'nt control proteins in acute pancre­atitis, C'/I 1982;23:944-50. 14, Craddock PR, Hammt'rschmidt D, Whitt' IG, et a\.; Com­pl,' ml'nl (C5a) induced granulocyte aggrt'gation in vitro; a PUT?TSCHU\'S /nJINOPATf-fY /65 possible mech,1I\ism ,.i c"mpl"nwnl nll'di.lll·d k·u""sl.lsis and leukopeni,l. , Clill III,'c,1 1l)77;hll::!hll-.t. 20. Hammerschmidt DE, H.uris I'D, W,wl,lnd H, d ,II.: Com­plement indul"l'd gr.1I\ulocvt,' ,Iggrl'g'ltion in "i"". Alii / Pellll,,1 \'~!\I;Ill:!:I.tI'>-Sll. 21. Sacks T, Mold,.\\, CI', Cradd"ck PR, d.ll.: lhvg,'n r,ldi'\lls mediate end,.tlwli,11 c,'11 d'lmage bv ,','mpl,;nll'nt slimu­1, lted ~ranul,.c"tes: .11\ in "itf<' m"d"1 "I immulll' ",lSnll,lr dama~t'. / Cli,i l'm',1 Ilj7~;I'>I:111'>1-7. " Yamad,l Y, M,.ld,.\\' CI', ~,Kks I. d .11.: D,'ldL'rious L'tleds "i endotL"in ,In '-llltun'd L'ndotheli,11 <""lis: ,In in vilrLl m"del of \·,lscul.u injury. 1"/111'"111,'1/"11 Iq~I;S:IIS-:!h. 23, Jacob liS, Cr,lddock PR, Hammerschmidt DE, et al.: Com­piL'l1lL'nl induced gr,lI1ulocyle aggregation. N E'IX/ , Mcd I':1HO;]02:7H'i ·'i4. 24. Ruddy S, Cigli I, Austen KI': The complement system of 01.111. N LIIXI I Mc'! 1':I72;2H7:4IN-95, 545-9, 592-6, 642-6. 2S. Hammerschmidt DE, White JG, Craddock PR, et al.: Curti­costeroids inhibit compiL'ment induced granulocyte aggre­gation: .I possible mechanism for their efficacy in shock states. , Clill IIIl'csl 1':17'i;h3:7':1H-H03. 211. Goldstein 1M, Roos 0, Weissman G, et al.: Influence of corticosteroids on human polymorphonuclear leukocyte function in vitro: reduction of lysosomal enzyme release dnd sllperoxide production. ill/lelllllllelli,," 1':176;1:305-15. , (f'1/ N,'IIro-"l'lrtllllllll"l, V"I. 6, No.3, 1986