British Journal of Ophthalmology, 1987, 71, 817-825

Retinal arterial macroaneurysms: a retrospectivestudy of 40 patientsMICHAEL J LAVIN,' RONALD J MARSH,' STANLEY PEART,2AND ALIM REHMAN'

From the I Western Ophthalmic Hospital, Marylebone Road, London NWI; the 2Pickering Unit andDepartment of Medicine, St Mary's Hospital, London WI; and 'St Mary's Hospital Medical School,London WI

SUMMARY We studied 40 patients with a total of 44 retinal arterial macroaneurysms. All patientswere followed up for at least six months. Macroaneurysms (MAs) have variable clinicalpresentations and are still frequently misdiagnosed before fluorescein angiography. HaemorrhagicMAs were most frequently misdiagnosed (75%), and had a sudden onset with a relatively poorvisual outcome. Patients with these MAs had higher systolic blood pressures and significantly fewerassociated retinal vein occlusions (p<0-05) than other types of MA. Exudative MAs caused agradual onset of symptoms, were frequently associated with retinal vein occlusions, and were themost frequent indication for laser treatment. Only one of 10 quiescent MAs subsequentlydeveloped significant exudation or haemorrhage. We confirm the association of MAs with retinaland systemic vascular disease. In addition we found that MA patients had a significantly higherblood packed cell volume (haematocrit) than controls (p<0-05). Laser treatment significantlyshortened the duration ofMA patency (p=0006).

Retinal arterial macroaneurysms (MAs) are acquiredaneurysmal dilatations of the retinal arteries whichusually occur in elderly, hypertensive people. Theterm macroaneurysm was introduced in 1973 todistinguish aneurysms of the retinal arteries fromthose of capillaries and veins. '1- AlthoughRobertson's report was the first review, isolated casesof retinal arterial aneurysms have been describedsince the nineteenth century"7 and a number in thepast decade.=27 MAs appear as local arterial dilata-tions with variable degrees of artery wall hyalinisa-tion and surrounding retinal exudate or haemorr-hage. They are frequently associated with retinalvascular occlusions,' lo13 23 retinal emboli,' 3 systemichypertension,' and vascular disease."' Histologicalexamination confirms the presence ofa true aneurysmwith vessel wall thickening, hyaline change, andelastotic degeneration.1" 26Most reports have not classified MAs, although

Palestine and colleagues' study2'3 grouped MAsaccording to their position in relation to the vasculararcades and whether visual acuity was affected.Correspondence to Mr Michael Lavin, FRCS, Moorfields EyeHospital, City Road, London ECIV 2PD

Nevertheless, visual loss is influenced not only by theproximity of the MA to the macula, but also by thepresence of oedema, exudate, or haemorrhage andits severity, duration, and position. Since the MAswith visual loss and haemorrhage may have a differ-ent course and outcome than those with exudate,I"'I'' aclassification reflecting this may be useful.

Spalter24 has emphasised that MAs are oftenincorrectly diagnosed and regards them as amasquerade syndrome. The differential diagnosisincludes Coats' disease, Leber's miliary aneurysms,von Hippel-Lindau syndrome, and very rarelyvenous macroaneurysms. It is important to differen-tiate MAs from retinal capillary aneurysms whichmay at times be large but do not arise from an arteryand may therefore be treated without risk of vascularinjury.

Photocoagulation of MAs was first described byHudomel and Imre' and has since been reported byseveral authors. 621 2'25 27 Although visual acuity isoften reported as improved after laser photocoagula-tion of the MAs with macular oedema or exudate, itmay improve spontaneously. MAs are described asbeing obliterated within one to seven months of

817

on July 15, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.71.11.817 on 1 Novem

ber 1987. Dow

nloaded from

MichaelJ Lavin, Ronald J Marsh, Stanley Peart, and A lii Rehinan

treatment, though no study has compared this with alarge group of untreated cases.

Despite the fact that MAs are not rare, they arefrequently misdiagnosed, the natural history is notclear, and only two reports have included more than20 patients. Further information is required to defineprognostic features and indications for and results oftreatment. We were therefore prompted to reviewour series.

Material and methods

We retrospectively reviewed cases of angiographicallyproved retinal arterial macroaneurysms seen at theWestern Ophthalmic Hospital, London, from 1973 to1984. The cases were identified from a fluoresceinpunch card index. The patients' notes were studied,and where necessary patients were recalled for re-examination. Colour stereo photographs andfluorescein angiograms were examined and classifiedseparately by two of the authors (MJL and RJM).Particular attention was paid to MA characteristicsand positions, retinal vascular abnormalities, andto the presence of associated ocular disease. Athorough medical assessment was performed on 18patients (45%) in the medical clinic supervised by SirStanley Peart. Particular attention was paid to thecardiovascular examination, and most patients hadelectrocardiography, chest x-ray, full blood count,and tests of electrolytes and urea performed. Theblood packed cell volume and systolic blood pressureat or within six weeks of presentation were availablein a number of patients and were compared betweengroups and with age matched controls. All cases hada minimum of six months' follow-up.

Clinically and histologically MAs are characterizedby a dynamic process of formation, enlargement withor without blood-retinal barrier disruption, andgradual spontaneous resolution after a variable timeperiod. Since the extent of blood-retinal barrierdisruption determines the nature of visual impair-ment, we classified MAs in terms of their barrierfunction:

Quiescent MAs, with haemorrhage or exudateextending for less than 1 disc diameter and sparingthe macula (Fig. 1).

Exudative MAs, in which exudate is the majorcomponent measuring more than 1 disc diameter andis responsible for visual loss in cases combined withhaemorrhage (Fig. 2).Haemorrhagic MAs with haemorrhage extending

more than I disc diameter, more extensive than anyassociated exudate, and responsible for visual loss(Fig. 3).Argon laser photocoagulation was performed on

10 patients, five with exudative MAs and five with

Fig. 1 Fundus photograph of a quiescent macroaneurvsm,with minor local inirarelinnal hlaetnorrhage.

haemorrhagic MAs. In the haemorrhagic group twoMAs were treated because of the development ofmacular oedema or exudate, while three were treatedto prevent further haemorrhage. The techniqueconsisted of applying a ring of argon laser photo-coagulation (Coherent 200 source) two or threeburns in width to the retina immediately surrounding

Fig. 2 Fundus photograph ofan exudativemacroaneurysm. Hyalinised, pale macroaneurvsm is locatedat arterial bifurcation and surrounded by retinal oedema.Exudate is approaching thefovea.

818

on July 15, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.71.11.817 on 1 Novem

ber 1987. Dow

nloaded from

819Retinal arterial macroaneurysms: a retrospective study of 40 patients

Fig. 3 Fundus photograph (A) shows haemorrhagicneuroretinal detachment (arrows) with adjacent subretinalhaemorrhage. Venous phase offluorescein angiogram (B)shows macroaneurysm partially masked by overlyinghaemorrhage, with proximal arterial constriction.

the MA sufficient to obtain mild to moderate retinalblanching. Burns were then placed on the aneurysmitself using similar intensity until the MA wasblanched. Treatments employed a spot size of 100 to200 [m with a duration of 0 1 to 0-2 seconds; nineMAs required an average of 52 burns, while two

others needed over 300 burns each. The intensity andnumber of burns required depend on media clarityand local tissue swelling and uptake characteristics.

Results

A total of 44 MAs were identified in 43 eyes of 40patients, and in 37 patients the MAs were single.Three patients had more than one MA, with bilateralinvolvement in all 3 (7-5%).

Sex. A definite female excess was noted, with 28females (70%) and 12 males. The age specific incid-ence of MAs in the general population was greaterfor females (Fig. 4).

Age. The age at presentation ranged from 55 to 89years, with a mean of 66-1 years. The incidence ofMAs was found to rise with age (Fig. 4).

Clinical appearance and examination of stereocolour photographs disclosed a set of clinical signs in86% of affected eyes which included a local retinalarterial dilatation surrounded by a ring of hyalinechange and intraretinal haemorrhage with overlyingretinal oedema. In some patients these signs werepartly masked by overlying retinal haemorrhage, butin most an accurate diagnosis was possible fromexamination of the stereo colour photographs alone.Serous neuroretinal detachment was seen in fourpatients, with macular involvement in two. In mostpatients the classification as haemorragic or exuda-tive was obvious. In seven patients with haemorr-hagic MAs a minor degree of exudate was seen aboutthe peripheral margins of retinal haemorrhage, andtwo of these later required laser treatment formacular oedema or exudate. Subretinal haemorr-hage was present in 20 of the 21 eyes with haemorr-hagic MAs. This gradually resorbed over a period ofmonths, leaving a variable degree of pigment

FREQUENCYPER MILLIONPOPULATIONPERANNUM

90

80

70

6050

40

3020-10-o

FEMALES-MMALESOVERALL INCIDENCE 32.1 x 10-6 POPULATIONOVER 45yrs p.a. -

45-64 65-74 75&overFig. 4 Age specific incidence ofmacroaneurysms ascalculatedfrom catchment population.

I a I

on July 15, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.71.11.817 on 1 Novem

ber 1987. Dow

nloaded from

MichaelJ Lavin, RonaldJ Marsh, Stanley Peart, and Alim Rehman

42% 23266%

42% 44%> C6 4

31%

Fig. 5 Distribution ofretinal arterial macroaneurysms.

epithelial disturbance and subretinal scarring in itswake. Haemorrhagic MAs resolved spontaneously inall but two patients, both ofwhom suffered recurrentbleeds. One of these had received laser photocoagu-lation to the MA some weeks before recurrenthaemorrhage.

Site. The right eye was affected in 17 patients andthe left in 20. The remaining three patients hadbilateral involvement. MAs were as frequent on thesuperotemporal vessels as on the inferotemporalvessels (Fig. 5). Most MAs (84%) were seen on thetemporal vascular arcades. Haemorrhagic MAs werelocated significantly closer to the optic disc (40%within 1-5 disc diameters) as compared with otherMAs (16%, p<005). MAs were commonly found inassociation with arterial bifurcations or arterio-venous crossings.

Presentation. The commonest presenting featurewas acute loss of vision, either central or generaliseddue to haemorrhage, and occurred in 21 eyes of 20patients (52-5%). Visual loss was central as a result ofmacular haemorrhage in 15 eyes, and was generalised

Fig. 6 Initial andfinal visualacuities in eyes with quiescent(dashed lines) and exudative(continuous lines)macroaneurysms: 22 eyes.

owing to vitreous haemorrhage in six. The meanduration of symptoms before presentation inhaemorrhagic MAs was 1-4 weeks.A total of 13 eyes of 13 patients (32.5%) were

found to have exudative MAs. These patients had asignificantly longer duration of symptoms beforepresentation (mean 16-6 weeks, p<0O05), withgradual and often progressive central visual impair-ment. No patient in this group experienced latehaemorrhage or recurrent exudation after apparentresolution of the MA. Four patients had a minordegree of associated intraretinal haemorrhage. Afurther patient had a significant degree of subretinalhaemorrhage, though macular exudate was themajor feature.A total of 11 quiescent MAs were identified in

10 eyes of eight patients. One eye included twoquiescent MAs. One patient had a quiescent MAwhich subsequently ruptured, and this patient isincluded in the haemorrhagic group for analysis.No other quiescent MAs developed exudation orhaemorrhage. Quiescent MAs were discoveredincidentally in three patients who presented withretinal vein occlusions, two with macular degenera-tion, one with diabetic retinopathy, a fellow eye to ahaemorrhagic MA, and one patient with lensopacities.

Visual outcome. Initial and final visual acuities aresummarised in Figs. 6 and 7. Initial visual acuities ofcounting fingers or worse were common in haemorr-hagic MAs (52%) but unusual in exudative MAs(15%). In general the outcome was relatively benign,with 54% of exudative MAs and 48% of haemorr-hagic MAs retaining an acuity of 6/12 or better. Poorvisual outcome (counting fingers or worse) was seenin 24% of haemorrhagic MAs and 15% of exudativeMAs and was related to the severity and duration ofmacular involvement. Two eyes with exudative MAs

2 Senile Maculardegeneration

VISUALACUITY

6/66/96/126/186/24

Quiescent MA 6/36ExudativeMA - 6/60

1 Diabetic retinopathy CFtreated

820

on July 15, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.71.11.817 on 1 Novem

ber 1987. Dow

nloaded from

Retinal arterial macroaneurysms: a retrospective study of40 patients

Haemorrhagtc6/6 MAs- 6/66/9 7T1 6/96/12 6/12

6/18 6/18

6/24 6/24

6/36 6/36

6/60 6/60

CF _CF

Initial Fincdacuity acuity

Fig. 7 Initial andfinal visual acuities in eyes withhaemorrhagic macroaneurysms: 21 eyes. L- laser treatedeye.

suffered a reduction in visual acuity of two or morelines owing to macular exudation; in both cases lasertreatment had been recommended but was refused.Only two eyes with quiescent MAs suffered reducedacuity, due to senile macular degeneration in both(Fig. 6). The causes of poor final visual acuity are

summarised in Table 1. When laser treated groupswere compared with untreated groups for differenttypes of MA, no difference in visual outcome was

found.Duration. Follow-up visits for all MAs ranged from

a minimum of six months to seven and a half years,with a mean of two years and five months. Theduration ofMA patency from diagnosis to resolutioncould be established in seven treated and 14untreated cases. The total duration of MA patencywas significantly shorter in laser treated cases (mean3-5 months) as compared with untreated cases (mean13 6 months, p=0.006, Fisher's exact test), MAs in

six out of seven treated eyes resolved within one

month of laser treatment, as proved on post-treatment angiography. The duration of untreatedMAs ranged from three months to seven years.

Haemorrhagic MAs resolved spontaneously in all buttwo patients, both of whom suffered recurrenthaemorrhage (5%). One of this pair had receivedlaser photocoagulation to the MA.

Fluorescein findings. Angiography showed thatMA sizes were variable, with larger MAs on the

Table I Causes ofpoorfinal visual acuity (countingfingersor less)

Fig. 8 Fluorescein angiogram in a patient with vitreoushaemorrhagefrom a macroaneurysm discloses multipleareas ofcapillary telangiectasia and coarsening ofthecapillary network (arrows).

larger arteries. Aneurysms appeared to be fusiformor saccular. Fusiform MAs showed rapid filling in theearly arterial phase and involved most of the arterialcircumference. Saccular MAs showed minimal earlyfilling, with good filling in the middle or late phases ofthe angiogram. The rate of filling seemed to bedetermined by the size of the aneurysm neck.Fluorescence of MAs was not necessarily homo-geneous and many showed irregular inhomogeneousfilling, thought to represent intraluminal clot forma-tion. Occasionally the MA was masked by overlyingblood, but subsequent angiograms demonstrated theMA. Focal areas of capillary dilatation, telangiec-tasia, or closure were seen in 48% of eyes. Althoughfine capillary changes were frequently seen in theretina adjacent to the MA, in many patients thesecapillary abnormalities were also located well awayfrom the MA (Fig. 8).

Previous fundus photographs were available from10 patients, six of whom had had previous angio-

graphy for a variety of conditions, most frequentlyretinal vein occlusion. Careful examination of theseangiograms failed to disclose any reliable indicator offuture aneurysm formation.

Associated ocular disease. Retinal vein occlusionswere seen in the same or fellow eye of 13 patients(32 5%), one of whom had a central retinal vein

occlusion. A further patient had a branch arteryocclusion. Other associated ocular diseases includeddrusen (19 patients), lens opacities (nine), andepiretinal membranes (two). Retinal arterial plaques

Macular hemorrhage with pigment epithelial scars 4 eyesMacular exudate 2 eyesPersistent vitreous haemorrhage 1 eyeSenile macular degeneration 2 eyes

821

on July 15, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.71.11.817 on 1 Novem

ber 1987. Dow

nloaded from

MichaelJ Lavin, RonaldJ Marsh, Stanley Peart, and Alim Rehman

were seen in the same or fellow eye of eight patients(20%Yo), while irregular arterial calibres were noted in16 patients (40%/). Patients with haemorrhagic MAshad significantly fewer associated retinal vein occlu-sion (15%) as compared to non-haemorrhagic MAs(50%, p<0 05). Half the vein occlusions were in thesame eye as affected by the MA. In three branch veinocclusions the MA was situated at the point of veinocclusion, suggesting a causal role in the genesis ofthe occlusion. Two other MAs were found in the areaof retina involved by vein occlusions.

Systemic disease (Fig. 4). Of the 18 patientsexamined in the medical clinic 17 were found to haveevidence of systemic vascular or relevant metabolicdisease. A survey of the patients general practition-ers found that, of the remaining 22 patients, only fourhad no known vascular disease. Systolic bloodpressures of 200 mmHg or more were more frequentin patients with haemorrhagic MAs (40%) than withother MAs (20%), but the difference was not signifi-cant. Blood packed cell volume (haematocrit) levelswere available in 20 of our patients and were com-pared with those in a group of 40 cataract patientsmatched for age and sex. Haematocrits over 45%were significantly more frequent in MA patients(55%) than controls (20%, p<0 (5). A history ofsmoking was available in 15; four were non-smokersand two others had not smoked within the past fiveyears. Two patients died during the study period,both of complications of vascular disease, yielding amortality of 4-8%. One of these patients died beforesix months of follow-up and is therefore not includedfor further analysis.

Discussion

We found that MAs are still often misdiagnoseddespite a number of reports describing their clinicalfeatures.'2' 24 7 We emphasise that in most MAsthe dominant feature is haemorrhage into the retinalor vitreous spaces and, less commonly, retinaloedema or exudate. These changes may be wide-spread and often partly obscure the underlying MA,which is therefore easily overlooked. Thus haemorr-hagic MAs often masquerade as disciform macularlesions, and MAs with exudation may be misdiag-nosed as branch vein occlusions or diabetic retino-pathy. Further diagnostic difficulties may arisebecause branch vein occlusion, senile maculardegeneration, or diabetic retinopathy may be seen inthese patients and the presentation ascribed to theseconditions. It is most important to differentiate theseentities because of the very different treatmentsavailable. The diagnosis of MA should be consideredin all elderly patients with intraocular haemorrhage,retinal oedema, or exudate, particularly if atypical

features are present. Subretinal haemorrhage inMAs is typically extensive, dense, overlies an artery,is not centred on the macula, and has vitreousextension. Careful fundus examination supple-mented by fluorescein angiography, which may haveto be repeated several times, usually establishes thediagnosis.The age range and sex ratios seen in the study are

almost identical to those found on pooling thoserecorded in the world literature.'17 The predomin-ance of females in this study is not due only to theexcess of females in the older population (see Fig. 1).Cerebral aneurysms are most frequent in womenafter the age of 40, and it is likely that similarpathogenetic factors operate in these two types ofaneurysm.

Palestine and colleagues' classification-' of MAsdid not differentiate between different types ofmacular involvement. Our classification was simpleand identified groups of patients who differed interms of visual outcome and associated diseases.Although several MAs had mixed haemorrhagic andexudative features, it was usually not difficult todetermine the major feature. Subretinal macularhaemorrhage is not remediable by laser treatmentand usually results in severe visual loss. Macularoedema or exudate on the other hand is eminentlytreatable, particularly if seen early. Clearly quiescentMAs had the best outcome.

Isolated MAs are almost always seen in the elderly.Aging is associated with replacement of the smoothmuscle of the arterial media by collagen and anincrease in intimal collagen. The arteries lose theirelastic recoil and become rigid and relatively dilated.These changes are similar to those seen in thesclerotic phase of hypertensive retinopathy, andGarner and Klintworth" noted that many age relatedchanges may be the result of haemodynamic stress.

Haemorrhagic MAs were located significantlycloser to the optic disc than other MAs. Arteriesclose to the disc have larger diameters and increasedflow rates than peripheral vessels. These factors willincrease transmural stress in these arteries, and maycontribute to haemorrhage.MAs frequently occur at arteriovenous crossings.

At the point where arterial and venous walls are incontact the adventitial layer is absent, and the twovessels share a common coat." The arterial wall hasless structural support at this point and may be proneto aneurysm formation. Any associated venousdisease may cause local arterial wall damage, con-tributing to MA formation.

Retinal vein occlusions are frequently seen inassociation with MAs, and are also described in otherarterial diseases such as Coats' disease"' and Leber'smilitary aneurysms."l They were uncommon in

822

on July 15, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.71.11.817 on 1 Novem

ber 1987. Dow

nloaded from

Retinal arterial macroaneurysms: a retrospective study of40patients

association with haemorrhagic MAs but commonlyseen in association with non-haemorrhagic MAs. Thissuggests that different factors are important in MArupture, and mechanical factors may be significant.Perhaps these associations reflect the differing effectsof widespread retinal vascular disease on highpressure vessels (arterial aneurysm formation) ascompared with low pressure vessels (resulting in veinocclusion).

Hypertension is a prominent feature in MApatients. "' Hypertensive retinopathy is associatedwith poor retinal vascular autoregulation, plasmatransudation, and subsequent vessel hyalinisation. Arigid, dilated arteriole with abnormal wall results."233The higher systolic blood pressures seen in patientswith haemorrhagic MAs may have contributed tohaemorrhage both by the increased transmuraltension and by a greater amplitude of pulsations,causing mechanical vascular damage. We could notdetermine the patients' activities at the time ofrupture in a large enough sample. However, studiesof cerebral aneurysms have shown that up to 50%ruptured during events associated with transientarterial hypertension, emphasising the importance ofmechanical events,'3 " and studies of experimentalaneurysms have emphasised the roles of both hyper-tension and increased haemodynamic stress.6 In thisrespect macroaneurysms have been compared toCharcot's cerebral aneurysms, which may be impli-cated in the genesis of cerebrovascular accidents.37A compromised vessel wall of any aetiology may

be less able to withstand further insults such asembolic damage, particularly in the presence of ahigh intraluminal pressure. Although MA formationis often ascribed to the mechanical effects of hyper-tension,"'4 several authors have drawn attention tothe role of local vascular damage." 2"I Retinalarterial emboli are frequently seen in eyes with MAs,which have been shown to develop at sites of previousembolic occlusion." Indeed, a detailed study ofretinal emboli reported one patient with an arterialexudative lesion 'of unknown significance', and thepublished photograph was very suggestive of anMA.' We note that the distribution ofMAs parallelsthat described for retinal emboli,3" showing a markedpreference for the temporal retinal vessels andposterior pole. Histologically, MAs are found inareas of arterial disease,"' and one eye had evidenceof focal small vascular occlusions."' The frequentfinding in our study of focal capillary abnormalities inareas adjacent to and away from MAs may be asequel of small platelet-fibrin emboli. These embolican travel far down the arterial tree, breaking intosmall fragments which occlude vessels and subse-quently resorb, leaving focal zones of capillaryclosure. Vascular remodelling can occur after

embolic occlusions,"9 and this may explain the fre-quent finding of focal capillary dilatation. Thesecapillaries may leak and cause visual loss if located inthe perifoveal area,"' though this was not seen in ourseries.

Systemic vascular disease other than hypertensionwas a prominent feature of our MA patients as it wasin other studies. In some patients the MA was the firstmanifestation of underlying systemic vasculardisease. In a few patients systemic vascular diseasewas not found at their initial visit, but appeared atlater follow-up. We are not aware of haematocritmeasurements in previous studies of MAs. Althoughraised haematocrits have been described in patientswith retinal vein occlusions,"' this did not account forthe raised haematocrits seen in our MA patients.Although our sample was small, the raised haema-tocrits did not appear to be due to smoking. Raisedhaematocrits are commonly seen in systemic vasculardisease.4' 42 The raised haematocrits in our MApatients may reflect their underlying systemicvascular disease.Although some authors have suggested that MA

patients have a high death rate after diagnosis,' "1 thiswas not our experience. The 4-8% mortalityobserved over the study period is not alarming giventhe ages of our patients.The management of MA patients requires a full

investigation for systemic disease, with particularattention to hypertension and systemic vascular andembolic disease. MAs were seen almost exclusivelyin the sclerotic phase of hypertension, with evidenceof long-standing retinal vascular changes. In the faceof long-term retinal vascular damage control ofhypertension is not likely to cause resolution of theMA, though it might reduce the risk of haemorrhage.

Ocular management depends on MA behaviour.Haemorrhagic MAs with intact maculae should beobserved closely in the initial period, as occasionallypatients develop macular exudate requiring treat-ment. In general these MAs resolve spontaneouslyand do not require treatment. It has been suggestedthat haemorrhagic MAs should be treated to preventrecurrent haemorrhage.27 Both this and a previousstudy' have found that recurrent haemorrhage isextremely uncommon, and that there are no reliableindicators of impending haemorrhage. In particular,MA pulsation is not a reliable warning of impendinghaemorrhage.'3 On the basis of this evidence webelieve that the major indication for treatment ofhaemorrhagic MAs is the development of macularodema or exudate.Our study shows that treatment results in a signifi-

cantly shorter duration of MA patency as comparedto untreated cases. Since visual loss from macularoedema or exudate depends on severity and dura-

823

on July 15, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.71.11.817 on 1 Novem

ber 1987. Dow

nloaded from

MichaelJ Lavin, Ronald J Marsh, Stanley Peart, and Alim Rehman

tion, early photocoagulation of selected cases shouldbe beneficial. When treated cases were comparedwith untreated cases, Palestine et al. found noimprovement in visual outcome after photocoagula-tion. Case selection is particularly important, aspatients with long-standing exudate or oedema areunlikely to improve. This was certainly a factor in ourpatients and explains our finding that photocoagula-tion did not improve visual outcome. Complicationsof photocoagulation included occlusion of the distalarteriole and an increase in exudate accumulationsoon after photocoagulation, necessitating great carein the treatment of lesions close to the macula.

Direct photocoagulation of the MA may con-tribute to occlusion of the distal arteriole and wasseen in 27% of our 11 treated cases. However thisevent may occur spontaneously in the natural historyof an MA, and indeed was found in 25% of Cleary'suntreated cases." Only a small number of our un-treated cases underwent fluorescein angiographyafter resolution of their MA, and these cannotusefully be compared to our treated group. We founddistal arteriolar narrowing in 19% of our untreatedMAs on fluorescein angiography. Most MAs resolvespontaneously, and the process of vascular healing isinevitably accompanied by local vessel wall constric-tion, seen clearly on angiography. Distal occlusionmay occur as a result, and presumably depends on theseverity and extent of the repair response.There are no histological studies of the mechanism

of laser induced resolution of a MA. Clinical experi-ence suggests that the argon blue-green laser resultsin thermal uptake both in the retinal pigmentepithelium surrounding the MA (causing injury tothe outer vessel wall) and in the haemoglobin andoxyhaemoglobin43 in the arterial lumen (causingendothelial injury). The aim of perilesional retinalphotocoagulation is to aid resolution of oedema,though there is evidence that photocoagulated retinalpigment epithelium may trigger proliferation ofnearby retinal vascular endothelial cells.44 Thismechanism could also contribute to resolution of theMA. The immediate vasospasm and reduced flow aresuperseded by a vascular repair response whichresults in local vessel wall constriction and MAresolution.We are grateful to Ms Sue Ford for photographic assistance, to MsRuth Tobias for statistical assistance, and to Ms Karen Johnstone forthe figures. We thank the consultant ophthalmologists at theWestern Ophthalmic Hospital for allowing us to review their cases.

References

1 Robertson DM. Macroaneurysms of the retinal arteries.Ophthalmology 1973; 77: 55-67.

2 Schulman J. Jampol LM, Goldberg MF. Large capillaryaneurysms secondary to retinal venous obstruction. Br JOphthalnol 1981; 65: 36-41.

3 Henkind P. Walsh JB. Retinal vascular anomalies: pathogenesis,appearance and history. Trants Ophthalmnol Soc UK 1980; 100:425-33.

4 Doyne RW. Case of peculiar condition of the retina due possiblyto the formation of small aneurysms and large extravasation ofblood which has become decolourised. Traits Ophthalmnol SocUK 1896; 16: 94.

5 Loring FB. Peculiar anatomical development of one of thecentral arteries of the retina. Traits Ain Ophthalmnol Soc 1881; 3:40-2.

6 Story JB, Benson AH. Aneurysms on retinal vessels in a peculiarcase of retinitis. Traits Ophthalmol Soc UK 1883; 3: 1(8- 1(.

7 Jennings JE. Aneurysms of the retinal arteries. Ain JOphthalmnol 1918; 1: 12-3.

8 Hudomel J. Imre G. Photocoagulation treatment of solitaryaneurysm near the macula lutea. Acta Ophthalmol (Kbh) 1973;51: 633-8.

9 Schultz WT, Swan KT. Pulsatile aneurysms of the retinal arterialtree. A]nJ Ophthalmol 1974; 77: 3(14-9.

10 Cleary PE, Kohner EM, Hamilton AM, Bird AC. Retinalmacroaneurysms. Br J Ophthalfnol 1975; 59: 335-61.

11 Cleary PE. Retino-vascular malformations. Trans OphthalmolSoc UK 1976; 96: 213-5.

12 Gold DH, Walsh JB. Fluorescein angiographic patternsof retinal arterial aneurysms. Proc Int Symp Fluores-cein Angiography, Ghent. Doc Ophthalmol Proc Ser 1976;9: 541-7.

13 Lewis RA. Norton EWD, Gass JDM. Acquired arterial macro-aneurysms of the retina. Br J Ophthalmol 1976; 60: 21-30.

14 Nadel AJ, Gupta KK. Macroancurysms of the retinal arteries.Arch Ophthalmol 1976; 94: 1092-6.

15 Godel V, Blumenthal M, Regenbogen I. Arterial macro-aneurysm of the retina. Ophthalmologica 1977; 175: 125-9.

16 Tobari 1, Asso S. Yokoro M. Clinical findings and photocoagula-tion in eight cases of retinal arterial aneurysms. Jpn J ClinOphthalmol 1977; 31: 803-9.

17 Yoshiodo H, Sugita T, Yamaguchi Y. Seven cases of retinalmacroaneurysms. Jpn J Clin Ophthalmol 1977; 31: 175-86.

18 Asdourian GK, Goldberg MF, Jampol LM, Rabb M. Retinalmacroaneurysms. Arch Ophthalmol 1977; 95: 624-8.

19 Fichte C, Streeten BW, Freedman AH. A histopathologic studyof retinal arterial aneurysms. Am J Ophthalmol 1978; 85: 509-18.

20 Khalil M, Lorenzetti DWC. Acquired retinal macroaneurysms.Can J Ophthalmol 1979; 14: 163-8.

21 Franqois J. Acquired macroaneurysms of the retinal arteries. IntOphthalmol 1979; 1: 153-61.

22 Kayazawa F. Bilateral retinal arterial macroaneurysms. AnnOphthalmol 1980; 12: 18-22.

23 Palestine AG, Robertson DM, Goldstein BG. Macroaneurysmsof the retinal arteries. Am J Ophthalmol 1982; 93: 164-71.

24 Spalter HF. Retinal macroaneurysms: a new masquerade syn-drome. Trans Am Ophthalmol Soc 1982; 80: 113-30.

25 Van Nouhuys E, Deutman AF. Argon laser treatment of retinalmacroaneurysms. Int Ophthalmol 1980; 2: 45-53.

26 Gold DH, LaPiana FG, Zimmerman LE. Isolated retinal arterialaneurysms. AmJ Ophthalmol 1976; 82: 848-57.

27 Abdel-Khalek MN, Richardson J. Retinal macroaneurysm:natural history and guidelines for treatment. Br J Ophthalmol1986;70:2-11.

28 Garner A, Klintworth G. Pathobiology of ocular disease. NewYork: Dekker, 1982: 1480-525.

29 Wise GN, Dollery CT, Henkind P. The retinal circulation.Hagerstown: Harper and Row, 1971: 28-32.

3(1 Egerer 1, Tasman W, Tomer TL. Coats' disease. ArchOphthalmol 1974; 92: 109-12.

31 Wegener JK. Leber's retinal degeneration with miliaryaneurysms. Acta Ophthalmol (Kbh) 1969; 47: 108-14.

32 Jampol LM, White S. Cunha Vaz J. Vitreous fluorophotometryin patients with hypertension. Arch Ophthalmol 1983; 101:888-90.

824

on July 15, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.71.11.817 on 1 Novem

ber 1987. Dow

nloaded from

Retinal arterial macroaneurysms: a retrospective study of40patients

33 Tso MOM, Jampol LM. Pathophysiology of hypertensive retino-pathy. Ophthalmology 1982; 89: 1132-45.

A4 Locksley HB. Natural history of subarachnoid haemorrhage,intracranial aneurysm and arteriovenous malformations: basedon 6368 cases in the cooperatives study. In: Sahs AL, Perrett GE,Locksley HB, Nishioka H, eds. Intracranial aneurysms andsubarachnoid haemorrhage: a cooperative study. Philadelphia:Lippincott, 1969: 37-108.

35 Komatsu S, Seki H, Uneoka K, Takaku A, Suzuki J. Rupturingfactors of intracranial aneurysm: season, weather and psycho-somatic strain. In: Suzuki J, ed. Cerebral aneurysms: experiencein 1000 directly operated cases. Tokyo: Neuron, 1979: 25-31.

36 Sekhar LN, Heros RC. Origin, growth and rupture of saccularaneurysms: a review. Neurosurgery 1981; 8: 248-60.

37 Russell RWR. Observations on intracranial aneurysms. Brain1963; 86: 425-42.

38 Arruga J, Saunders MD. Ophthalmic findings in 70 patients withevidence of retinal embolism. Ophthalmology 1982; 89: 1336-47.

39 Klein R, Klein B, Henkind P. Bellhorn R. Retinal collateralvessel formation. Invest Ophthalmol Vis Sci 1971; 10: 471.

40 Trope GE, Lowe GDO, McArdle BM, et al. Abnormal bloodviscosity and haemostasis in long-standing retinal vein occlusion.Br J Ophthalhnol 1983; 67: 137-42.

41 Lowe GDO, Forbes CD. Blood rheology and thrombosis. ClinHaematol 1981; 10: 343-67.

42 Lowe GDO. Laboratory evaluation of hypercoagulability. ClinHaematol 1981; 10: 40)7-42.

43 L'Esperance FA Jr. Ophthalmic lasers: photocoagulation, photo-radiation and surgery. St Louis: Mosby, 1983: 93.

44 Clover MMG. Laser treatment of diabetic maculopathy and theimplications for retinal vascular barriers. PhD thesis: Universityof London, 1984.

Acceptedfor publication 26 September 1986.

825

on July 15, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.71.11.817 on 1 Novem

ber 1987. Dow

nloaded from