Article Text

Ocular and cerebral trauma in non-accidental injury in infancy
  1. Department of Ophthalmology
  2. University of Naples Federico II
  3. via S Pansini, 5
  4. 580131 Napoli, Italy
    1. S J TALKS,
    2. J S ELSTON
    1. Eye Hospital, Radcliffe Infirmary
    2. Woodstock Road, Oxford OX2 6HE
      1. M A GREEN,
      2. G LIEBERMAN,
      3. C M MILROY,
      4. M A PARSONS
      1. University of Sheffield, Ophthalmic Sciences Unit
      2. Royal Hallamshire Hospital, Sheffield S10 2JAF

        Statistics from

        Editor,—We read with interest the paper by Greenet al.1 We agree with the authors about the importance of ophthalmoscopic examination in the ‘battered child syndrome’. However, we feel that some considerations on the pathogenesis should be discussed.

        Firstly, we believe, based on personal clinical and not autoptical cases, that there is a possible association between subdural and intraocular haemorrhages. Nevertheless, we want to underline that intraocular haemorrhages could be isolated manifestations in battered child syndrome; sometimes not due to a direct bulbar trauma.2

        Traumas of various types, even not ocular, may involve the retinal vascular system, as previously described by Purtscher at the beginning of this century.3

        Several unilateral or bilateral retinopathies similar to those observed by Purtscher have been reported after compressive thoracic injuries (for example, seat belt injuries), head trauma, and violent deceleration.4-6

        Various pathogenetic mechanisms of these retinal vascular alterations have been reported—sudden rise in intrathoracic venous pressure,4 arterial angiospasms, retinal vessel occlusion by gas, and lipid embolisms or aggregates of granulocytes.

        In the case of shaking, the pathogenesis of retinochoroidal haemorrhage basically can be caused by: (1) transient blood flow arrest due to rapid bending of the neck or rapid movement of the head, both resulting in direct trauma of the carotid-ophthalmic vascular system and/or retinal vasospasm. This mechanism is the same as the one that occurs in some cases of whiplash lesions; (2) acute thoracic compression probably due to a rapid muscular contraction with closed glottis, resembling a Valsalva’s manoeuvre. Such compression would give rise to a venous pressure wave transmitted to the eye, as a result of the lack of antireflux valves between the caval vein and the eye. The unilaterality or bilaterality of the symptoms may be explained by the anatomical distribution of the cervical veins and the position of the neck at the moment of shaking; (3) acute thoracoabdominal compression by catching, that leads to an event’s sequelae similar to those described in hypothesis (2); (4) a mechanism similar to that causing subdural haemorrhage according to Green et al—namely, the effect of inertial movements of the vitreous body within the eye during cycles of acceleration–deceleration during shaking. Optic nerve sheath haemorrhage is, in their opinion, the result of angular, rotational, or axial movement of the eye about a point in the most anterior part of the optic nerve, posterior to the sclera.

        Our hypothesis is based on clinical evidence of similar cases such as choroidoretinal haemorrhages in road accidents (whiplash lesion with and without seat belts).

        Therefore, we agree with Green et al about the importance of choroidoretinal haemorrhages as alarm signs for cerebral haemorrhages, but we would like to point out the medicolegal importance of the ocular lesions without other complications, for suspecting a battered child syndrome, otherwise difficult or impossible to diagnose. We suggest, therefore, that an accurate ophthalmoscopic examination should be mandatory in those cases.



        Editor,—The meticulous work of Green et al  has provided an important insight into the mechanisms responsible for the ocular signs in fatal non-accidental injury (NAI). While vitreous traction is likely to be a major factor in the pathogenesis of intraocular pathology, such as retinal detachment, there is indirect evidence that intravascular perfusion changes contribute to the characteristic intraretinal and preretinal haemorrhages.

        Firstly, haemorrhages of the same appearance and distribution as in NAI, often with vitreous haemorrhages, occur as a result of an acute rise in intracranial pressure (ICP). These signs may be seen in subarachnoid haemorrhage (Terson’s syndrome) and with an acute cerebral venous sinus thrombosis. Retinal haemorrhages are thought to occur in these cases because blood flow is occluded in the central retinal vein by the acute rise in ICP as the vein traverses the subarachnoid space in the optic nerve sheath. Flow continues in the central retinal artery, rupturing the preretinal capillary plexuses. An analogous acute rise in central retinal venous pressure may occur in shaking injuries in children. In these cases there is often evidence from ribcage bruising that the child has been gripped around the thorax, preventing venous return while cardiac output continues. This mechanism is thought to explain the occurrence of retinal haemorrhages with prolonged retching or vomiting. Moreover, intracranial pressure may rise acutely in NAI as a result of subarachnoid haemorrhage. Other causes of retinal haemorrhages associated with raised retinal venous pressure include asphyxia and epileptic fits.

        Secondly, in acute central retinal vein occlusion, which is usually due to a localised vascular event, deep retinal haemorrhages extending to the periphery are seen. A similar pattern is seen in NAI suggesting a common mechanism via perfusion/pressure changes within the central retinal vein.

        We suggest that while vitreous traction forces, as a result of vitreous inertia, may well be the cause of the more severe ocular injuries such as retinal detachment, changes in retinal venous pressure are a cause of the retrohyaloid, preretinal, and intraretinal haemorrhages.


        Authors’ reply

        Editor,—We are grateful to Liguori et al and to Talks and Elston for their helpful comments on our paper on non-accidental injury (NAI) in infancy, in particular those related to the possible effect of increased vascular pressure in the pathogenesis of retinal haemorrhages in this condition.

        All of our cases died as a result of their injuries, and we are interested to hear that Liguori et al believe that they have a similar ‘possible association between subdural haemorrhage and intraocular haemorrhages’ in non-fatal cases of NAI, although it is unclear whether their cases are the result of direct head or eye trauma. It is important to emphasise that in all our cases brain injuries were as a result of indirect trauma, with no evidence of direct trauma either to the head or the eyes.

        We agree that it is well established that retinal haemorrhages can be associated with a range of conditions, including subarachnoid haemorrhage (Terson’s syndrome) and other causes of raised intracranial pressure; and with raised intraocular or intrathoracic venous pressure, such as in central retinal vein thrombosis or acute thoracic compression injuries (such as may occur if the chest of an infant is compressed during shaking). While we agree that all of these conditions can be associated with choroidal, retinal, and preretinal haemorrhages, the haemorrhages associated with such conditions tend to be most severe in peripapillary areas, and decrease in intensity towards the retinal periphery (with no equatorial sparing); haemorrhages in these conditions are not associated with focal areas of retinal detachment.

        In our study we have shown that in NAI due to violent shaking of the child the equatorial zone of the fundus is relatively spared, and that haemorrhages are most frequent and severe at the retinal periphery, followed by peripapillary areas (this distribution is usually easily seen on macroscopic examination of the retina). In addition, however, there are often focal areas of retinal detachment related to haemorrhages, and in the same zonal distribution. In particular, we frequently see a ‘compound retinal lesion’, consisting of focally coincident subhyaloid (preretinal) haemorrhage, intraretinal haemorrhage, and haemorrhagic retinal microdetachment in NAI. We believe that such compound retinal lesions, in a distribution which spares equatorial areas, are a highly specific feature of NAI in infants. This particular distribution of haemorrhages and associated retinal detachment indicates, we believe with little room for doubt, that the primary pathogenesis of the injuries is via vitreoretinal traction. The association of such retinal injuries with subdural haemorrhages strengthens this belief, as these are caused by a similar relative motion of the brain with respect to fixed points of the skull and meninges.

        It remains theoretically possible that raised intracranial or intravascular pressure transmitted to intraocular vessels might be a contributory component in the pathogenesis of intraocular haemorrhages in NAI, although this cannot explain either the equatorial retinal zone sparing or focal retinal detachment seen in our series. Talks and Elston argue in their letter that ‘changes in retinal venous pressure are a cause of retrohyaloid, preretinal, and intraretinal haemorrhages’ but that ‘vitreous traction forces, as a result of vitreous inertia, may well be the cause of more severe ocular injuries, such as retinal detachment’. The problem with this argument is that the most severe cases of trauma would also have the haemorrhages associated with less severe trauma and, if the less severe injury is due to raised venous pressure, the equatorial zone would not be spared (see above)—to be so a mechanism of removal of equatorial haemorrhage (seen as part of the distribution of retinal haemorrhages in raised venous pressure) would have to be invoked. A hypothetical combination of raised intravascular pressure and vitreoretinal rotational traction might conceivably lead to very severe haemorrhages affecting all zones of the retina and the vitreous (in extremely severe shaking of an infant), producing haemorrhages similar, perhaps, to those seen in other non-traumatic causes. In such a case, however, areas of haemorrhagic retinal microdetachment in peripheral and peripapillary retina would provide evidence of rotational trauma, and the patient’s history would determine whether this was accidental or NAI.

        In summary, we believe that overwhelming evidence points towards vitreoretinal traction as the major cause of intraocular injuries in NAI. We agree with Liguori et al (as stated in our paper) that accidental causes of whiplash (such as could occur in severe seat belt injuries, head trauma, and violent deceleration) could cause similar ocular injuries, related to similar rotational vitreoretinal traction forces. Raised intracerebral or vascular pressure may be a relatively minor factor contributing to intraocular haemorrhage, but in the absence of vitreoretinal traction does not explain the relative sparing of the equatorial fundus.

        Finally, we have recently found (unpublished observations) that some infants dying of head injury due to NAI have, in addition to recent intraocular haemorrhages, areas of Perls’ Prussian blue staining of haemosiderin deposits in ocular tissues, indicating earlier episodes of haemorrhage, from which the child recovered. This emphasises the point made by Ligouri et al in their letter that ‘intraocular haemorrhages could be isolated manifestations of “battered child syndrome”, presumably due to direct or indirect trauma’.

        View Abstract

        Request permissions

        If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.