Method

A literature review was undertaken using the MEDLINE database (1950–2010), supplemented by further review of any additional relevant publications quoted in the papers. The initial search was based on the keywords ‘retrobulbar haemorrhage/retrobulbar haemorrhage’.

Inclusion criteria were: human studies; English language publications; full articles. Exclusion criteria were: non-English language publications; animal studies; letters; trauma in children. Only those publications relating to blunt facial/craniofacial trauma were analysed. Penetrating trauma was excluded as bleeding would be much more likely (and obvious).

For each case reported the diagnostic criteria of RBH was placed into one of four groups. Group 1, clinical signs only. Group 2, clinical signs followed by clearly documented decompression of haematoma. Group 3, clinical signs confirmed by CT or other imaging (pictures unavailable). Group 4, clinical signs confirmed by CT or other imaging (pictures available). In groups 1 and 2 no CTs were carried out on the cases reported. In group 1, diagnosis was based solely on clinical signs and could, in theory, be questioned. Cases in which the successful evacuation of blood (during surgical decompression) was clearly stated, were regarded as good evidence for the diagnosis of RBH. Groups 3 and 4 included cases in which imaging (usually CT) had ‘confirmed’ the presence of RBH, although in group 3 the images were not shown and could therefore have been incorrectly interpreted.

Results

From the 186 papers initially retrieved, 32 met the inclusion/exclusion criteria. Some publications were of isolated case reports, others were of cases series. They yielded 82 cases of RBH secondary to blunt facial trauma that were suitable for analysis, and they were grouped as outlined above. The results are displayed in Table 2 .

Results

From the 186 papers initially retrieved, 32 met the inclusion/exclusion criteria. Some publications were of isolated case reports, others were of cases series. They yielded 82 cases of RBH secondary to blunt facial trauma that were suitable for analysis, and they were grouped as outlined above. The results are displayed in Table 2 .

Discussion

Traditionally the tense, proptosed, non-seeing eye with a non-reacting, dilated pupil, following facial trauma (or its repair) is considered to be due to acute bleeding within the orbit – retrobulbar haemorrhage. This requires prompt recognition and treatment , which should not be delayed to undertake investigations. Lateral canthotomy and cantholysis are performed as soon as possible and medical measures commenced while preparing the patient for theatre . This management strategy is based on the reported successes in the management of RBH in the literature. The purpose of this study was to review this literature to see how accurately cases had been diagnosed.

From the results of this study it can be seen that in only just over half of cases reported was the diagnosis of RBH seemingly established beyond doubt (by either CT evidence or documented evacuation of haematoma). It can be said that in just under half of the reported cases (group 1), the diagnosis may have been presumed, or questionable, due to a lack of evidence. Some other cause, notably retrobulbar oedema, may have been the underlying problem, not haemorrhage. In two cases ( Fig. 1 ) in group 4 the CT image shown did not show any blood although the diagnosis of RBH was made. This questions the reliability of CT interpretation in those cases reported in group 3.

(i) and (ii) Two examples of erroneous interpretation of CT in reported retrobulbar haemorrhage. Although there is proptosis, there is no blood behind the globe (the arrow is not demonstrating blood). Compare with (iii) where blood is obviously recognised.

RBH is generally thought to be a common cause of acute proptosis in trauma. Yet it could be argued that it is commonly diagnosed because it is thought to be a common diagnosis. The authors’ results and experiences are in keeping with the possibility that RBH may not be as common as currently thought and that other causes must be considered and excluded (ideally by CT), although time constraints may make this difficult or impossible in the early stages of patient management. CT has revolutionised the assessment of the multiply injured patient and is now essential for most craniofacial injuries . CT technology has evolved rapidly, significantly reducing radiation dosages and acquisition times, such that imaging of the face is now safely possible during the assessment of life-threatening injuries. This avoids additional transfers later and delays in management, while facilitating comprehensive treatment planning. On this basis, any patient requiring an urgent brain CT who has suspected midface injuries, or a proptosis, should also undergo imaging of (at least) the orbits (although ideally the whole face should be scanned). Skull base, orbital apex and ocular injuries need rapid identification.

These findings have significant implications in cases in which surgical decompression is undertaken on clinical grounds alone. If CT scans have not been possible, a ‘risk/benefit analysis’ is necessary to decide whether decompression is appropriate, safe and should be attempted, based on the clinical findings only. Retrobulbar haemorrhage can be intraconal or extraconal, the former is thought to have a worse prognosis. Extraconal blood can collect anywhere within the extraconal space and may be difficult to access through an inappropriately sited incision. Lateral surgical approaches are considered to be safest. If proptosis is due to bleeding medially or is secondary to oedema, a lateral approach will not evacuate any blood and may result in diagnostic confusion. Failure to clearly evacuate haematoma should prompt consideration of orbital oedema or other less common causes. These include air (emphysema), bone (blow-in fractures) and brain herniation (orbital roof fractures) all of which are well recognised causes of proptosis and can complicate orbital fractures. Clinically these may not be distinguishable. Additional approaches may need consideration. Ideally patients should not leave the operating table until the proptosis has shown obvious signs of improvement, or the clinician feels that the sight is no longer salvageable, or the general status of the patient precludes further treatment.

In patients in whom definitive diagnostic investigations are not immediately available a number of questions are raised, each of which complicates management. These may also be useful topics for future research.

First, at what point does a proptosis (from any cause) become severe enough to warrant the need for urgent decompression based on clinical examination alone? Imaging may not be available and even if it is, does not quantify the effects of any cause of proptosis. The decision to decompress is often made following clinical examination, which is a highly subjective assessment and at risk of over- or under-estimating, the severity of proptosis. Tonometry may be helpful , in that some of the increased retrobulbar pressure is transferred to the proptosed globe and can be measured. Tonometry is designed to measure intraocular pressures and can be misinterpreted in the presence of other causes of raised intraocular pressure (e.g. glaucoma). Nevertheless it is a useful adjunct, the value of which lies in comparative measurements, identifying trends and anticipating the need for surgical intervention. Owing to the progressive nature of swelling in the early stages of trauma, repeated examination is necessary in patients in whom decompression is not undertaken.

Second, when does retrobulbar swelling reach its maximum limit and present its greatest risk to vision? This is not clear in the literature, although some consider it is 36 h. During this time very close observation and repeated examination is necessary to ensure that the proptosis does not become severe enough to indicate decompression. Patients must be kept painfree and as quite as possible because coughing, vomiting, intubation, intravenous fluids and vasoactive drugs can all precipitate further intraocular bleeding and possibly retrobulbar bleeding.

Third, in a proptosed globe in which decompression is not undertaken, can chronic ischaemia of the retrobulbar tissues occur with time, resulting in extraocular muscle fibrosis and contractures, in the same way that it can occur in limbs (Volkmann’s ischaemic phenomenon)? If so, should the threshold for decompression (or medical measures) be lower than it is, in order to prevent this ‘orbital Volkmann’s’?

Fourth, how much of an effect does the patient’s systemic blood pressure have on orbital perfusion pressure and the decision to decompress? By analogy, considering compartment syndrome in the limbs, a reduction in systemic blood pressure can result in critical ischaemia at higher systemic pressures than in the non-swollen limb. Similarly, irreversible ischaemia in the swollen orbit may be analogous to that seen in the head-injured patient, with raised intracranial pressure. In the head-injured patient the mean systolic blood pressure must be maintained above 100 mm Hg to ensure cerebral perfusion. Any fall below this, increases mortality. This raises the question that in the presence of a proptosis, from any cause, does a low mean systolic blood pressure (as may occur in the hypovolaemic shocked, multiply injured patients) have the same effect on visual disability as it does with neurological disability in head injuries? Is there a minimal ‘orbital perfusion pressure’ necessary to prevent visual deterioration? This has not been reported in the literature, but has significant implications on how (or whether) to treat any proptosis. Sudden and profound episodes of hypotension from a variety of causes have also been reported to result in loss of sight all in the absence of craniofacial trauma. Are these risks greater in the presence of a significant proptosis? Are we wasting our time (and putting patients at risk) by attempting decompression in such cases? In many patients outcomes are still disappointing.

In conclusion, some conditions are commonly diagnosed because they are thought to occur commonly. It is important that alternative diagnoses are kept in mind, particularly in patients who do not respond as well as anticipated to treatment. Failure to clearly evacuate a haematoma from the orbit and achieve decompression is an example of this following the surgical management of a presumed RBH. The evidence base for the existence of RBH following trauma may not be as strong as is currently portrayed in standard texts and can result in diagnostic confusion. RBH is a real entity, but other conditions also exist that are clinically indistinguishable, but nevertheless affect management and its success.