Examination: In routine clinical practice, visual acuity is most commonly measured using a Snellen chart comprising a series of letters for the patient to read which gradually decrease in size. Good vision is documented 6/6 in the UK, referring to the 6 m distance at which one of the lower lines of letters should be legible. In the USA, the same acuity is documented 20/20, referring to the equivalent distance measured in feet. If a patient’s vision is so poor that they are only able to read the largest letter on the top line of the chart at 6 m, acuity is documented as 6/60 or 20/200. If even reading the top letter at 6 m proves impossible, the chart is moved closer. Reading the top letter at 1 m is documented 1/60 or 3/200. Acuities worse than this are consecutively described as counting fingers (CF), hand movements (HM), perception of light (PL) or no perception of light (NPL). Alternative scoring methods exist, such as logMAR (logarithm of the minimum angle of resolution), which are often used in research practice. Letter characteristics, for example, spacing, are more standardised in this system, but the principles are otherwise identical. LogMAR scores of 0, 1 and 1.7 correspond to 6/6, 6/60 and NPL, respectively.

Regardless of the scoring system used, if visual impairment is identified, it is important to use a pinhole to correct any refractive errors and also to take a careful history for preexisting amblyopia.

Testing colour and low-contrast acuity using Ishihara plates or Sloan charts can increase sensitivity to detect mild optic nerve pathology, but visual loss in pituitary apoplexy is often marked and may be dramatic, so these techniques tend to be more useful in a less acute setting.

Relative afferent pupillary defects are not frequently reported in the apoplexy literature. This could be because the visual apparatus is commonly involved bilaterally at the chiasm. Oculomotor nerve damage may result in an efferent pupillary defect.

Reduction in visual acuity was noted in some of the earliest descriptions of pituitary apoplexy (Kux 1931) and was evident in two of the five cases from the original series describing the clinical syndrome (Brougham et al. 1950). In the accompanying postmortem pathological descriptions, compression of the right optic nerve against the overlying anterior cerebral artery was noted.

One of the early studies with a sizeable cohort came from a group in Tokyo and investigated the incidence of apoplexy in patients with pituitary tumours (Wakai et al. 1981). They reported a retrospective case series of 560 patients with pituitary tumours operated over a 30-year period. Ninety-three of these cases (16.6 %) had either clinical or surgical evidence of haemorrhage, but only 51 (9.1 %) presented with acute symptoms. They further subdivided this group into “major” and “minor” attacks. A major attack was defined as disturbance of consciousness, hemiparesis, visual loss or ocular palsy; there were 38 patients in this group (6.8 %). The other 2.3 % of patients with a minor attack had headache, nausea, vomiting or vertigo. Of the group with “major” symptoms typical of the clinical pituitary apoplexy syndrome, 79 % had visual loss, the nature of which was not specified.

Another study from the same year came from a German group (Muller-Jensen and Ludecke 1981). They investigated 586 surgical cases and found 72 with evidence of cystic necrosis with or without haemorrhage. Fifty-eight of these cases had clinical symptoms, and the authors report rates of 70 and 58 % visual disturbance depending on whether the tumour had ruptured or not. The nature of the visual disturbance was not defined further.

An important study from the USA focussed on clinical presentation and visual outcome (Bills et al. 1993). They studied 37 patients who presented with an acute syndrome characterised by abrupt onset headache and/or visual disturbance in the presence of a pituitary adenoma between 1975 and 1991. They reported reduced visual acuity in 52 %. All of the patients except one underwent transsphenoidal surgery. Visual acuity improved following surgery in 88 %, and surgery within a week of symptom onset was associated with better visual recovery. Two of three patients who presented with blindness (defined as CF or worse) improved to 20/20 following early surgery within a week; the third had delayed surgery and recovered to 20/80. The authors advocated early neurosurgical intervention to optimise visual outcome.

A Polish study published the same year analysed 799 patients with pituitary adenoma, 14.4 % of whom had histological evidence of apoplexy and 5 % of whom presented with acute clinical apoplexy (Bonicki et al. 1993). Of the 39 patients with acute clinical apoplexy, 38 % reported blurred vision and 5 % sudden loss of vision. Visual loss and disturbance of consciousness were considered the main indications for surgery. Twenty-three patients were operated, 19 urgently. The authors reported full neurological recovery including visual deficits in 14/19. Two further patients made a partial recovery, one developed consecutive hypopituitarism and two patients died from cerebral oedema and hypothalamic damage. Sixteen patients without severe visual deficits were treated conservatively. The authors reported that some clinical improvement occurred in all members of this group but that a poor outcome resulted in several cases at long-term follow-up; this was not defined further.

A group from Oxford, UK studied 35 patients (Randeva et al. 1999). They reported a 66 % prevalence of reduced acuity. Four people were managed conservatively and none of these patients had ocular sequelae. Eighty-six percent of the surgically treated group improved, and outcome was found to be better in patients treated within 8 days, all of whom had complete resolution of acuity (defined as 6/6 or a return to premorbid acuity).

Another UK study is especially notable for pursuing a conservative management strategy in 40 % of the 45 patients studied (Sibal et al. 2004). In this series, decreased visual acuity was found in 46 % including four patients managed conservatively. Ninety-three percent of surgically treated patients had either complete or near-complete resolution of visual acuity but so did all four of the patients with visual loss in the conservative group. The authors advocated a conservative strategy in patients with mild neuro-ophthalmological signs.

Another British study supports this approach (Gruber et al. 2006). They reported complete blindness in 4 of their 30 patients, monocular blindness in 2 and reduced visual acuity in 12. Two-thirds of the patients were managed conservatively, including two of the six blind patients. Three of the blind patients regained partial vision, and this included the two in the conservative group.

The largest number of patients from a primary investigative study comes from a Chinese group (Mou et al. 2009). They reported 83 patients, of whom 81 % had visual loss. However, these patients were defined on the basis of histology at surgery and/or radiological features, rather than a clinical presentation with acute apoplexy, so is likely to have included patients with subclinical onset. For this reason, these patients have not been included in our prevalence estimate meta-analysis.

A fairly large joint South African-American clinical study of 62 patients found reduced visual acuity in 56 % of their patients overall (Semple et al. 2004). This figure rose to 61 % when patients were excluded if their vision could not be accurately assessed, for example, if comatose at presentation. The severity of visual dysfunction was graded by the authors. It was recorded as reduced but functional in 44 %, reduced and nonfunctional in 7 % and blind in 10 %. Ninety-five percent of these patients were treated surgically. Follow-up data was available for 55/62 patients at an average of 56 months. At this time, acuity was normal in 69 %, improved but not normal in 16 % and unchanged in 15 %. The six blind patients remained blind.

A Belgian group studied 24 patients (Dubuisson et al. 2007). Fifty percent of these patients had a visual deficit, which could have been either a reduction in acuity or a field defect. Twenty-one of the patients were operated, and 92 % of these patients’ visual deficits resolved postoperatively.

A group from Taiwan studied the differences in patients presenting with clinical versus subclinical apoplexy (Liu et al. 2010). Sixty-five patients were identified with histopathological confirmation of pituitary haemorrhage at the time of surgery. Only 25 had symptoms consistent with acute pituitary apoplexy. In the other 40 there was radiological evidence of pituitary haemorrhage. Visual impairment was defined as either a reduction in Snellen acuity or a field defect on automated perimetry performed by an ophthalmologist. This was present in 88 % of the clinical group and 70 % of the subclinical group. All patients were treated surgically. Vision improved in 64 % of the clinical group and 93 % of the subclinical group. Only the clinical group has been included in our meta-analysis.

Two studies have focused specifically on the neuro-ophthalmic manifestations of pituitary apoplexy, an early study by a Scottish group (McFadzean et al. 1991) and a recent study by an Australian group (Simon et al. 2011). McFadzean et al. (1991) quantified loss of vision in their group of 15 patients, 12 of whom had reduction in acuity. In four patients, it was mild (6/9–6/12), in one moderate (6/18–6/36), in one severe (6/60 to CF) and in four very severe (HM or less) emphasising the variability that is seen.

Simon et al. (2011) studied 23 patients. All except one presented as acute apoplexy. Reduced visual acuity was present in 55 % (11/20) of patients for whom data was available. Bilateral involvement was present in more than half (55 %). One patient was NPL. Eighteen patients were operated. After a median follow-up of 11 months, all of the patients with reduced acuity had made some visual improvement but this was complete in only 25 %.