362 Journal of the Royal Society of Medicine Volume 72 May 1979

Pathogenesis of hypertensive retinopathy: a review'

Professor A Garner MD MRCPathProfessor N Ashton CBE FRS DSCDepartment of Pathology, Institute of OphthalmologyJudd Street, London WCIH 9QS

Although retinal vascular changes occur in benign hypertension, the term hypertensiveretinopathy usually refers to the more florid retinal changes found in the malignant oraccelerated phase. Indeed, examination of the fundus is a useful way of detecting at a very earlystage the potentially disastrous transition of benign hypertension to a malignant phase.

Clinical pictureThe earliest clinical features of accelerated hypertension are small linear haemorrhagesaccompanied by cotton-wool spots close to the optic disc and posterior pole. Exudates areinconspicuous at first but common a few weeks later: they are often found in the macularregion where they tend to be arranged in a star-figure but may also occur in other parts of theposterior retina.

Increasing haziness of the posterior retina commonly presages oedema of the disc, whichbecomes congested and swollen.

Direct changes in the retinal vessels include diffuse arteriolar narrowing, widened andenhanced light reflexes due to increased thickness of the vessel wall, and arteriovenous crossingchanges. Nipping of the veins is commonly an illusion related to the effect of arteriolar muralsclerosis in concealing blood flow through the vein as it passes behind the arteriole. However,none of these vascular changes is peculiar to malignant hypertension, identical appearancesdeveloping in the benign phase and in a proportion of ageing normotensive people.Conversely, some hypertensive individuals may have insignificant arteriolar changes as gaugedby ophthalmoscopy.

Pathogenesis of the vascular disorderWhile systemic arteries respond to increases in blood pressure by hypertrophy and evenhyperplasia of their smooth muscle coat, the smaller arterioles are liable to develop destructiveas well as adaptive changes, since it is here that the main step down from arterial to capillarypressure takes place.

Arteriolar smooth. muscle responds to increases in intraluminal tension by contracting andconsequently, although such vasoconstriction tends to be more pronounced in the malignantphase where pressures are higher, it is not peculiar to malignant hypertension. The distinctivemanifestation of the malignant phase occurs when the constricted vessel succumbs todisruptive forces. The resultant necrosis is usually focal and, in the retina, occurs principallywithin the terminal precapillary arterioles. Experimental studies using cynomolgus monkeyswith blood pressure increase secondary to induced renal ischaemia have shown that theessential feature of these arterioles in hypertensive retinopathy is leakage of plasma into thevessel wall, the leakage being related to breaks or holes in the endothelial lining (Ashton 1972,Garner et al. 1975). This accumulation ofplasma in the vessel wall has been termed an insudateand is the basis of fibrinoid necrosis, long recognized in histological sections as the hallmark ofthe malignant phase. Incidentally, since the material includes fibrinogen polymerized to formbanded fibrin it is probably more appropriate to speak of fibrinous necrosis.

1 Paper read to Section of Ophthalmology, 13 October 1977. Accepted 11 November 1977

IF" 1979 The Royal Society of Medicine0 1 41-0768/79/050362-04/$O 1.00/0

Journal of the Royal Society of Medicine Volume 72 May 1979 363

The initial break in the endothelial lining of the vessel appears to be caused by stretching andattenuation related in turn to degeneration of the surrounding smooth muscle coat (Figure 1).Loss of support from the muscle coat allows focal rupture of the lining endothelium. In studiesof tissues other than those of the eye separation of the junctions linking individual endothelialcells has been described (Ooneda et al. 1965, Giacomelli et al. 1970), but we saw no evidence ofthis in our studies of the retinal vasculature: nor did we see any convincing sign of increasedpinocytotic transport of plasma across intact endothelium. Although it is conceivable thattransient separation of the intercellular junctions occurred, we were not surprised thatpermanent separation was not found since, in contrast to vessels outside the central nervoussystem, the endothelial cells of the retinal vessels are joined by tight encircling junctions whichare unusually resistant to disruption.Once gaps within the cytoplasm of attenuated endothelial cells have been created plasma

seeps into the wall, displacing the degenerate muscle cells as it does so, and the muralthickening which results narrows the lumen and further impedes blood flow.The steps we have outlined are essentially beyond dispute. The only matter for debate is the

cause of the muscle degeneration preceding and disposing to the endothelial damage. On thebasis that the arteriolar constriction is an adaptive response to the rising blood pressure, it isconceivable that beyond a certain level the capacity to adapt will be exceeded and that themuscle coat will yield to further stress. This concept is supported by some clinical studies on thecerebral circulation (Lassen & Agnoli 1972, Strandgaard et al. 1973) and has been referred toas 'breakthrough of autoregulation'. Alternatively, one explanation of this failure might bethat the muscle degenerates as a result of ischaemic damage from prolonged and excessive focalvasoconstriction with its attendant metabolic aberrations. Support for this second possibility

Figure 1. Electron micrograph showing smooth muscle(SM) degeneration in the wall of a retinal arteriole which isassociated with attenuation and focal necrosis of the liningendothelium (arrow). x 7500

364 Journal of the Royal Society ofMedicine Volume 72 May 1979

comes from our own electron microscopical studies, in which we observed pronouncedarteriolar constriction with incipient degenerative changes in the cytoplasm of some individualmuscle cells (Garner & Ashton 1970, Garner et al. 1975).To summarize (Figure 2), we consider that the sequence of events in the production of

fibrinous necrosis in retinal arterioles is: (1) A phase of intense vasoconstriction affectingchiefly the precapillary arterioles. (2) Smooth muscle degeneration leading to a loss of supportfor the endothelium. (3) Focal breaks in the endothelium allowing plasma to seep into thevessel wall. (4) Mural necrosis and plasmatic insudation leading to obliteration of the lumen.

Figure 2. Possible sequence in the pathogenesis of hyper-tensive fibrinous necrosis. After an initial phase of extremevasoconstriction smooth muscle necrosis supervenes in theface of continued rise in blood pressure. This leads todilatation and leakage of plasma into the necrotic wallthrough focal breaks in the unsupported endothelial lining

Pathogenesis of the extravascular changesHaemorrhagesBleeding occurs through the walls of necrotic arterioles and, since vessels of this type arelocated within the nerve-fibre layer, the haemorrhages assume a linear or flame-shapeddistribution.

Cotton-wool spotsOcclusion of precapillary arterioles as a result of fibrinous necrosis gives rise to focal ischaemiain the dependent capillary bed. Increasing metabolic embarrassment occurs over the next fewhours within the inner layers of the retina, culminating in a zone of nerve fibre swelling.Eventually many swollen axons disrupt because of ischaemic necrosis, and the residual stumpsundergo further swelling and form bulbous structures recognized in histological sections ascytoid bodies (Ashton & Harry 1963). The swelling, which in part is likely to be due directly toanoxia and intracellular oedema, is compounded by interference with axoplasmic transport(McLeod 1976). Tracer studies indicate that there is interruption of flow in both orthogradeand retrograde directions, axoplasm entering the disrupted regions and accumulating becausethe local source of energy required to propel fluid away is lacking. The swollen regioneventually undergoes extensive necrosis, with the lipoprotein membranes of degenerateorganelles congregating in the centre of the terminal bulb to constitute the pseudonucleus ofthe cytoid body. Subsequently the necrotic axons are removed by retinal macrophages andreplaced by glial tissue.

Journal of the Royal Society of Medicine Volume 72 May 1979 365

ExudatesAlthough exudates almost certainly originate along with the haemorrhages they tend topresent somewhat later in the course of the retinopathy. This is because initially theirconsistency is little different from that of transudative oedema and only later, as they pool inthe outer plexiform layer or in Henle's layer at the macula, do they become progressivelyinspissated and recognizable as hard creamy-yellow exudates.

PapilloedemaSwelling of the optic nerve head is associated with plasma leakage, as fluorescein studiesconfirm, but the major factor appears to be swelling of the nerve fibres as they pass through thedisc region. Like the cotton-wool spot, this represents ischaemic swelling and consequentinterference with axoplasmic flow in the nerve head. Both the exudation and the ischaemia arerelated to fibrinous necrosis of the smaller arterioles supplying the optic disc, although otherfactors, such as impeded venous outflow, complicate the pathogenetic mechanism of thepapilloedema.

ReferencesAshton N (1972) Transactions of the American Academy of Ophthalmology and Otolaryngology 76, 17-40Ashton N & Harry J (1963) Transactions of the Ophthalmological Societies of the United Kingdom 83, 91Garner A & Ashton N (1970) XXI Concilium Ophthalmologicum, Mexico, Excerpta Medica, Amsterdam; pp 583-600Garner A, Ashton N, Tripathi R, Kohner E M, Bulpitt C J & Dollery C T (1975) British Journal ofOphthalmology 59,

3-44Giacomelli F, Wiener J & Spiro D (1970) American Journal of Pathology 59, 133-159Lassen N A & Agnoli A (1972) Scandinavian Journal of Clinical and Laboratory Investigation 30, 113-116McLeod D (I1976) Transactions of the Ophthalmological Societies of the United Kingdom 96, 313-316Ooneda G, Ooyama Y, Matsuyama K, Takatama M, Yoshida Y, Sekiguchi M and Arai I (1965) Angiology 16, 8-17Strandgaard S, Olesen J, Skinhoj E & Lassen N A (1973) British Medical Journal i, 507-510