Terson syndrome (TS) was originally described by Moritz Litten in 1881 as vitreous bleeding following aneurysmal subarachnoid hemorrhage (SAH); its name “Terson syndrome” (TS) was given in 1900 by French ophthalmologist Albert Terson. The pathogenesis of TS is highly controversial. It has been reported that the higher the intracranial pressure, the higher the risk of developing TS. The source of retinal hemorrhages and their connection to the SAH in TS has been deliberated extensively and is still contentious.
Postulated theories about the pathogenesis of TS
There is voluminous literature dealing with the pathogenesis of TS. Many theories have been postulated to explain it. Following is a summary of them, and the fundamental flaws in them.
An early theory suggested that blood simply tracks from the intracranial subarachnoid space into the optic nerve sheath, then penetrates the sclera in the porous region, and finally appears in the vitreous space within the eyeball. However, no such pathway exists, and that invalidates this theory.
It has been postulated that a sudden rise in the intracranial pressure (ICP) associated with SAH results in the rupture of retinal vessels. If a sudden rise of ICP caused the rupture of retinal vessels, then TS would develop immediately, as in Valsalva retinopathy; however, TS can develop several days or even weeks after the sudden rise in ICP. An acute rise in the ICP is found in several conditions without any rupture of retinal vessels. Furthermore, no mechanism is known by which an acute rise in ICP can cause the rupture of retinal vessels.
It has been proposed that the blood may enter the vitreous cavity around the retinal vessels near the optic disc, and inside the eye, the blood may spread in the intra-retinal, sub-internal limiting membrane, or along the retinal vessels. There is no anatomical basis for blood to travel from the subarachnoid space of the optic nerve sheath to the optic disc and the eyeball along the retinal vessels.
It has also been reported that blood around the optic nerve in the subarachnoid space infiltrates the intraocular space through the perivascular space around the central retinal vessels within the optic nerve. However, the perivascular space around the central retinal vessels within the optic nerve does not extend into the eyeball. Thus, blood in the subarachnoid space cannot get to the retina.
Recently, the glymphatic system has been proposed to play a role in TS. Kumaria et al. hypothesized that the glymphatic pathway is the only extravascular anatomical conduit between the subarachnoid space and the retina. They put forward the view that raised ICP causes subarachnoid blood in the skull to be refluxed through glymphatic channels into the globe, resulting in intraocular hemorrhage. However, the existence of glymphatic channels from the subarachnoid space to the retinal vessels in the eye has never been demonstrated so far.
Other reports have described occlusion of the central retinal vein (CRV) by a variety of mechanisms in TS. These include the following:
- Raised ICP, due to SAH, results in seepage of cerebrospinal fluid (CSF) through the optic canal into the optic nerve sheath and leads to the dilation of the retrobulbar optic nerve sheath and compression of the CRV. My study of the optic nerve sheath in 20 rhesus monkeys and 80 human optic nerves showed that the dilated retrobulbar portion of the optic nerve sheath is a normal anatomical feature of the sheath in monkeys and humans, and not a pathological change. The outer layer of the optic nerve sheath is a thick fibrous tissue and not elastic tissue; therefore, it cannot distend by the increased pressure in the sheath.
- According to another theory, a sudden rise in ICP with SAH results in compression of the CRV and retinal venous stasis, and retinal hemorrhages. According to this theory, a sudden rise in ICP is thought to decrease venous return to the cavernous sinus from the veins that drain the globe.
All these theories are inadequate to explain the exact mechanism of the association between SAH and retinal hemorrhages in TS. That was my basis to conduct this experimental study.
The study was conducted in 68 rhesus monkeys by experimental occlusion of the central retinal artery (CRAO in 39 orbits), posterior ciliary arteries (PCAO in 8 orbits), and CRV (CRVO in 21 orbits) [Fig. 1]. The question arises: what was the rationale of my experimental study design?
Rationale of the experimental study design
I have been conducting basic, experimental, and clinical research on ocular vascular occlusive and optic nerve disorders for about 70 years. On rare occasions, in my past research studies, I experienced totally unexpected findings. Out of scientific curiosity, I explored those mysteries further; which in some cases led to new seminal discoveries. For example, in 1963, I experimentally investigated the pathogenesis of optic disc edema in raised ICP. The most popular theory at that time was that the raised intracranial CSF pressure, leading to raised CSF pressure in the optic nerve sheath, resulting in compression of the CRV and consequent optic disc edema production. I decided to experimentally explore this by occluding the CRV. That led to the seminal discovery that CRVO is of two types: nonischemic and ischemia, with very different clinical features, visual outcomes, complications, prognoses, and management. Currently, this is the accepted concept. My studies on the pathogenesis of optic disc edema in raised ICP finally showed that the optic disc edema was due to axoplasmic flow stasis.
Several years ago, when I experimentally occluded the central retinal artery occlusion (CRAO) in rhesus monkeys, I found occasional eyes developed a dome-shaped macular hemorrhage, typically seen in TS. That was a completely unexpected, finding, because it is well established that CRAO always produces retinal infarction, but never retinal hemorrhages. This unexpected finding prompted me to explore further in the present study.
The central retinal artery, posterior ciliary arteries, and CRV were exposed by lateral orbitotomy in rhesus monkeys. In most monkeys, after orbitotomy slow leakage of blood from the orbital venous plexuses kept obscuring my views of these vessels. To have a clear view of the vessels, first I needed a bloodless clear field; for that, all animals required a tamponade of the orbital apex by cotton swabs plugs – that also exerted pressure on the various tissues at the apex of the orbit and the CRV lying outside the optic nerve posteriorly [Fig. 1]. The exerted compression needed by the tamponade to achieve a bloodless clear field varied markedly among different orbits; which caused a variable degree of compression of the CRV. There was no clinical method to calibrate and evaluate the extent of CRV compression during surgery. Post-op fundus examination showed markedly variable retinal venous dilation in all these eyes– definite proof of CRV compression.
The central retinal arteries and CRV were separated from one another by microdissection near their site of penetration into the optic nerve sheath (PPS in Fig. 1). The central retinal artery or CRV were occluded at their points of entry/exit at the optic nerve sheath, and posterior ciliary arteries were occluded at their point of entry into the eyeball [Fig. 1].
Color fundus photography and fluorescein fundus angiography were performed before, soon after occluding the vessels, and serially thereafter to document the fundus and vascular changes, including retinal findings and ocular circulation.
Retinal hemorrhages following CRAO
Fig. 2 shows some examples of various types and locations of hemorrhages in this group. There were 39 eyes in this group. Retinal hemorrhages were seen soon after the procedure in 7 eyes, and on follow-up in a total of 15 eyes (38%). The hemorrhages were in the macular region in seven [Fig. 2a, b, c, f], perimacular region in seven [Figs. 2c, d, f], the peripheral retina in four [Figs. 2c, e, f], and on the optic disc in six [Figs. 2g and h]. The eyes also had engorged retinal veins.
The follow-up length of the eyes after the procedure was 3–350 (median 61.5 ± 57.2) days.
Retinal hemorrhages following posterior ciliary arteries occlusion
There were eight eyes in this group. Retinal hemorrhages were seen soon after the procedure in one eye, and on follow-up in a total of three eyes (37.5%). Figs. a and b in Fig. 3 show examples of fundus changes, such as some hemorrhages and engorged retinal veins, in this group. Fig. (b) also shows choroidal infarcts. The hemorrhages were in the macular region in two, and on the optic disc in two.
The follow-up length of the eyes after the procedure was 15–150 (median 98 ± 49.1) days.
Retinal hemorrhages following central retinal vein occlusion (CRVO)
Fig. c and d in Fig. 3 show examples of fundus change due to CRV—including some examples of various types and locations of retinal hemorrhages in this group. Fig. (c) also shows cottonwool spots, as are seen in ischemic CRVO. Fig. (d) shows a large macular subhyaloid hemorrhage, in addition to other retinal hemorrhages. Many of these eyes also developed macular and optic disc edema.
There were 21 eyes in this group. All eyes developed engorged retinal veins and scattered retinal hemorrhages, such as those seen in TS. It could be argued that in my previous report of six animals with identical CRVO procedures, there were engorged retinal veins, but developed little or no retinal hemorrhages. How to reconcile this discrepancy?
My first CRVO study was in young healthy monkeys, and CRVO was of two types: (a) with CRVO alone, and (b) with both CRV and central retinal artery occluded. CRVO alone produced engorged retinal veins and little or no retinal hemorrhages, that is, “nonischemic CRVO.” However, CRVO combined with central retinal arterial occlusion (ischemia) produced engorged retinal veins and extensive retinal hemorrhages, that is, “ischemic CRVO.” As I mentioned above, that led to the seminal discovery that CRVO is of two types: nonischemic and ischemia.
By contrast, the current CRVO study was done in old, atherosclerotic, and hypertensive monkeys. It is well-established that these general health problems collectively make humans and monkeys predisposed to ischemic vascular disorders. Because of those associated ischemic factors, CRVO in this group of monkeys resulted in the development of “ischemic CRVO,” with extensive retinal hemorrhages [Fig. 3c, d] and even cotton wool spots [Fig. 3c].
Why did I have old, atherosclerotic, and hypertensive monkeys? My primary research interest deals with ocular vascular occlusive diseases. It is well-established that ocular vascular occlusive disorders are most seen in elderly, atherosclerotic, and hypertensive persons. To have experimental research findings valid to humans from this group of monkeys, I decided to produce a large colony of rhesus monkeys with old age, atherosclerosis, and hypertension.
Many theories have been postulated to explain the development of retinal hemorrhages in TS, but all of them have a variety of problems. To understand the pathogenesis of TS, it is essential to consider the following issues.
TS in 1900 was characterized by the presence of intravitreal hemorrhage. This statement by Terson in 1900 is no longer valid; since then, several studies have reported isolated intraretinal hemorrhages in TS as well. Scientific knowledge advances constantly, and, with that, definitions of diseases change.
The experiments conducted in this study were aimed at investigating the pathogenesis of TS, not previously published, with the purpose of determining whether TS’s pattern of retinal hemorrhages can be caused by experimental vaso-occlusive lesions in the orbit.
Role of the optic canal in TS
The optic canal plays a critical role in the development of TS. In cases of raised ICP or intracranial SAHs, the region of the optic canal is crucial to the dynamics of transfer of the CSF and SAH from the cranial cavity into the sheath of the optic nerve. I investigated the optic sheath anatomy in a comprehensive study of 80 human specimens. The dura of the sheath in the optic canal is firmly bound to the adjoining bone by numerous thick, fibrous bands [Fig. 4], which reduces the space of the sheath to a fine capillary-sized subarachnoid space [Fig. 4b], and the space appears as a trabecular meshwork.
To reach the orbital part of the sheath, the CSF and hemorrhages in the cranial cavity must percolate through the capillary subarachnoid space meshed trabecular network in the region of the optic canal [Fig. 4]. The facility of communication from the cranial cavity to the sheath of the optic nerve shows marked interindividual variations in the canal from free communication to almost none. This has the following implication. Unilateral or bilateral absence of CSF and the number of hemorrhages in the optic nerve sheath may be due to a difference in the facility of their transmission through the canal.
The sheath is a little loose behind the eyeball compared to elsewhere. Due to this looseness of the sheath near the eyeball and the space available here, any blood which enters the sheath from the cranial cavity tends to accumulate in a larger amount in this region behind the eyeball than in the posterior orbital part [Fig. 5].
An anatomical study of 100 human specimens showed that no communication existed between the perivascular space around the central retinal vessels in the optic nerve and the subarachnoid space of the sheath.
Incidence of TS
All patients with SAH do not develop TS. As discussed, the characteristics of the optic canal must play a major role in the incidence of TS.
In my experimental study, the incidence depended upon the amount of CRV compression by the tamponade at the apex of the orbit. That varied markedly.
Laterality of TS in Relation to Subarachnoid Hemorrhages
TS may be unilateral or bilateral in patients, with a rise in ICP and the amount of SAH. Although there is a bilateral rise in ICP and SAH, intraocular hemorrhages are usually present in only one eye. Regarding this issue, it is relevant to note the finding of my experimental study in rhesus monkeys, where I produced high CSF pressure and optic disc edema by gradually inflating an intracranial balloon. In that study, the amount of optic disc edema was usually not equal on the two sides and the optic disc edema generally appeared first on the side of the intracranial balloon and was also more marked on that side. To investigate whether the location of a brain tumor determines the severity of optic disc edema, I injected two monkeys with Prussian blue solution into the cerebellomedullary cistern just before their death. There was a marked difference in the severity of edema of the optic disc between the two sides. More filling of the sheath of the optic nerve with the dye was seen on the side with more marked edema of the disc [Fig. 6], which was also the side of the balloon. A similar mechanism may also be playing a role in the development of unilateral or bilateral retinal hemorrhages in TS despite raised ICP and subarachnoid hemorrhages in TS.
Cause of retinal hemorrhages in TS
The circumstances and amount under which and the number of retinal hemorrhages developing in human CRVO and TS are very different. Evidence suggests that in TS sudden compression of the CRV during its course in the subarachnoid space [Fig. 7], closes the lumen of the vein ◊ raised venous pressure ◊ venous stasis and retinal hemorrhages. In human CRVO, in contrast, there has been progressively increasing narrowing of the CRV lumen by arteriosclerosis/atherosclerosis over months or years, with a gradually and progressively increasing impediment to the retinal venous circulation, which gives time for development collaterals to compensate for the block. However, in the case of TS, in contrast, there is sudden acute compression of the CRV, resulting in the sudden onset of the severe rise of venous pressure and venous stasis, with no time to develop collaterals. In this experimental model, the compression of the CRV was by the tamponade effect of the cotton wool plugs at the apex of the orbit. Therefore, the monkey experimental model used in this study simulates TS and not CRVO. Also, in the human TS, the site of compression of the CRV is by the large, accumulated blood in the optic nerve sheath [Figs. 5 and 7] and raised CSF pressure. The effect of compression of the CRV and the resulting marked retinal venous stasis is identical, irrespective of the site of compression of the vein in the sheath of the optic nerve or at the apex of the orbit.
Relevance of my experimental study to the pathogenesis of TS
To put my novel experimental study in proper perspective for the pathogenesis of TS, one must consider the following facts.
- Clinical and experimental studies in CRAO have shown that eyes with CRAO immediately develop retinal ischemia and infarction, but NO retinal hemorrhages.
- PCAO is commonly seen in giant cell arteritis; those eyes show choroidal infarction and arteritic anterior ischemic optic neuropathy, but no retinal hemorrhages. Similarly, experimental PCAO developed choroidal infarction, but no retinal hemorrhages.
Therefore, the question arises, why did experimental CRAO and PCAO in this study show development of retinal hemorrhages such as those reported in TS [Figs. 2, 3a, b] For that, one must look at the study design and its relevance to the pathogenesis of TS. In the current experimental study, to stop the constant orbital venous leak during surgery, the apex of the orbit was plugged with cotton wool swabs. Because of the narrow apex of the orbit, which compressed the CRV in the posterior part of the orbit [Fig. 1]; compression of the CRV was confirmed by the development of the retinal venous stasis, marked engorgement of the retinal veins, rupture of retinal venous capillaries and retinal hemorrhages. The almost invariable absence of development of ischemic CRVO confirmed that the CRV was not completely closed.
Pathogenesis of TS
The following concept of the pathogenesis of TS is based on the findings of this experimental study, and on my basic, experimental, and comprehensive clinical studies on CRVO. All this information suggests the following pathogenesis of TS.
Compression of the CRV causing raised central retinal venous pressure and venous stasis plays a crucial role in the development of TS. In the current study, the site of compression of the CRV was outside the optic nerve in the narrow posterior part of the orbit [Fig. 1], by the pressure from cottonwool swab plugs in that narrow orbital region to control oozing of the blood. However, in TS in humans, the CRV, is compressed as it lies in the subarachnoid space of the optic nerve sheath [Figs. 5 and 7]. SAH is associated with intracranial hypertension. It has been reported that the higher the intracranial pressure, the higher the risk of developing TS. For transfer of the CSF and SAH from the cranial cavity into the subarachnoid space of the optic nerve sheath, the extent of patency of the optic canal [Fig. 4b], plays a critical dynamic role. Rarely, the optic canal may be completely closed, so that CSF and hemorrhages cannot infiltrate into the optic nerve sheath◊ no development of TS. These findings explain the frequency and the amount of blood reaching the optic nerve sheath. The raised CSF pressure and/or accumulated blood in the optic nerve sheath, separately or collectively, compress the CRV in the subarachnoid space in the sheath [Figs. 7 and 8].
It could be argued that my experimental model does not resemble TS, with no blood in the subarachnoid space of the optic nerve sheath; hence it cannot provide the required information about the pathogenesis of TS. However, my experimental study has shown that the basis for retinal hemorrhages in TS is retinal venous stasis and raised venous pressure produced by compression of the CRV. In TS patients, a large amount of accumulated blood in the sheath of the optic nerve[Fig. 6], and/or raised CSF pressure in the sheath of the optic nerve, compress the CRV to produce raised pressure in the retinal veins and retinal venous stasis and retinal hemorrhages. In my experimental study, the tamponade at the apex of the orbit compressed the CRV and produced retinal venous stasis, raised venous pressure, and retinal hemorrhages. The crucial issue of CRV compression ◊ retinal venous stasis ◊ raised venous pressure ◊ retinal hemorrhages exactly mimic TS in the experimental study.
The findings of this experimental study, and my basic, experimental, and comprehensive clinical studies on CRVO, suggest the following concept of the pathogenesis of TS: Compression of the CRV plays a crucial role in the development of TS. The CRV is compressed, as it lies in the subarachnoid space of the optic nerve sheath, by raised cerebrospinal fluid pressure and/or accumulated blood leading to retinal venous stasis and raised venous pressure in the retinal veins, retinal venous engorgement, rupture of the retinal capillaries and retinal hemorrhages.
The clinical importance of compression of the CRV and NOT occlusion of CRV in TS is that the optic nerve sheath decompression by opening it and releasing the blood and raised CSF pressure, would result in immediate decompressing of the CRV in the subarachnoid space and restoration of normal circulation and prevent visual loss, a phenomenon like the optic nerve sheath decompression in raised CSF pressure relieving the optic disc edema in intracranial hypertension.
Financial support and sponsorship
Supported by grant EY-1576 from the National Institutes of Health, USA.
Conflicts of interest
There are no conflicts of interest.
1. Litten M. Über Einigevom Allgemein-Klinischen Standpunktaus Interessante Augenveränderungen. Berl Klin Wochenschr 1881; 18:23–7.
2. Terson A. De l'hémorrhagie dans le corps vitréau cours de l'hémorrhagiecérébrale. Clin Ophthalmol 1900; 6:309–12.
3. Czorlich P, Skevas C, Knospe V, Vettorazzi E, Westphal M, Regelsberger J. Terson's syndrome-Pathophysiologic considerations of an underestimated concomitant disease in aneurysmal subarachnoid hemorrhage. J Clin Neurosci 2016; 33:182–6.
4. Joswig H, Epprecht L, Valmaggia C, Leschka S, Hildebrandt G, Fournier JY, et al. Terson syndrome
in aneurysmal subarachnoid hemorrhage-its relation to intracranial pressure, admission factors, and clinical outcome. Acta Neurochir 2016; 158:1027–36.
5. Hayreh SS, Vrabec F. The structure of the head of the optic nerve in rhesus monkey. Am J Ophthalmol 1966; 61:136–50.
6. Muller P, Deck JHN. Intraocular and optic nerve sheath hemorrhage in cases of sudden intracranial hypertension. J Neurosurg 1974; 41:160–6.
7. Kumaria A, Gruener AM, Dow GR, Smith SJ, Macarthur DC, Ingale HA. An explanation for Terson syndrome
at last:The glymphatic reflux theory. J Neurol 2022; 269:1264–71.
8. Jessen NA, Munk ASF, Lundgaard I, Nedergaard M. The glymphatic system:A beginner's guide. Neurochem Res 2015; 40:2583–99.
9. Hayreh SS. The sheath of the optic nerve. Ophthalmologica 1984; 189:54–63.
10. Hayreh SS. Pathogenesis of optic disc edema in raised intracranial pressure. Prog Retin Eye Res 2016; 50:108–44.
11. Hayreh SS. Occlusion of the central retinal vessels. Br J Ophthalmol 1965; 49:626–45.
12. Hayreh SS. Central retinal vein occlusion. In:Ocular Vascular Occlusive Disorders. Springer International Publishing 2015 621–743.
13. Singh S, Dass R. The central artery of the retina I. Origin and course. Br J Ophthalmol 1960; 44:193–212.
14. Hayreh SS, Zimmerman MB. Fundus changes in central retinal vein occlusion. Retina 2015; 35:29–42.
15. Hayreh SS, Zimmerman MB. Fundus changes in central retinal artery occlusion. Retina 2007; 27:276–89.
16. Hayreh SS, Weingeist TA. Experimental occlusion of the central artery of the retina l. Ophthalmoscopic and fluorescein fundus angiographic studies. Br J Ophthalmol 1980; 64:896–912.
17. Hayreh SS. Giant cell arteritis:Its ophthalmic manifestations. Indian J Ophthalmol 2021; 69:227–35.
18. Hayreh SS, Baines JAB. Occlusion of the posterior ciliary artery II. Chorio-retinal lesions. Br J Ophthalmol 1972; 56:736–53.
19. Hayreh SS. Central retinal vein occlusion:Differential diagnosis and management. Trans Sect Ophthalmol Am Acad Ophthalmol Otolaryngol 1977; 83:OP379–91.
20. Medele RJ, Walter Stummer W, Mueller AJ, Steiger H-J, Reulen H-J. Terson's syndrome in subarachnoid hemorrhage and severe brain injury accompanied by acutely raised intracranial pressure. J Neurosurg 1998; 88:851–4.