Vitreous haemorrhage in massive hemorrhagic polypoidal choroidal vasculopathy: clinical characteristics and surgical outcomes

Background To report the outcomes of vitreous hemorrhage (VH) associated with hemorrhagic polypoidal choroidal vasculopathy (PCV). Methods A retrospective study of 28 eyes of 27 consecutive patients of hemorrhagic PCV with VH, which were managed surgically between January 2003 and December 2011, was performed. All patients underwent pars plana vitrectomy for VH associated with PCV. The main outcome measure was best-corrected visual acuity (BCVA) at baseline, at 1, 3 and 6 months post operatively and at last follow up. Results The visual acuity measured on early treatment diabetic retinopathy study (ETDRS) chart improved in 16 eyes (57.1 %) by two or more lines, remained unchanged in nine eyes (32.1 %) and decreased in three (10.7 %) after surgery when compared to baseline VA. The mean baseline VA was 2.69 ± 0.57 logMAR units (<20/2000) which improved to 1.65 ± 0.93 logMAR units (20/800) at 1 month post operative visit and was sustained at 1.72 ± 1.12 (20/800) with an improvement of 0.96 logMAR units (p < 0.001, 95 % CI 0.54–1.37). The average postoperative follow up was for 14.2 months (range 1–84). The complications noted in postoperative follow up were cataract (n = 10), macular scaring (n = 9), organised dehemoglobinised blood (n = 7), retinal tear or detachment (n = 5), recurrent VH (n = 3) and choroidal detachment (n = 1). Conclusion Majority of patients with loss of vision due to VH secondary to hemorrhagic PCV have sustained improvement in visual acuity following surgery.


Background
Polypoidal choroidal vasculopathy (PCV) is a persistent and chronic disease with a variable course [1]. Although the prognosis of PCV is commonly reported to be more favorable and the clinical course more stable as compared with exudative age-related macular degeneration (AMD) [2][3][4], the natural history of PCV, especially haemorrhagic PCV, is not clearly known and may not be as benign as previously reported [1].
The primary abnormality in PCV involves the choroidal circulation, and the characteristic lesion is an inner choroidal vascular network of vessels ending in an aneurysmal bulge or outward projection seen clinically as reddish orange polyps [5]. Leakage in the vessel wall at dilatations leads to serous pigment epithelial detachments (PED) and lipid deposition in the exudative form

Open Access
International Journal of Retina and Vitreous of PCV while recurrent serosanguinous detachments of the retinal pigment epithelium and neurosensory retina are the predominant manifestation of hemorrhagic PCV. Massive hemorrhages may develop from rupture of venules and occasionally arteries. Sequelae of this is a sudden onset severe visual loss due to break through vitreous hemorrhage (VH) [1,[6][7][8][9][10].
About half the cases of PCV have a stable course and relatively favorable visual outcome, while the others have persistent leakage, repeated haemorrhages and poor visual outcome [1,4]. Profound visual loss in hemorrhagic PCV primarily results from recurrent serosanguinous haemorrhagic detachments involving the fovea or recurrent VH.
There is a lack of understanding of hemorrhagic PCV and there are very few reports in the literature on the outcomes of conservative [1,4], or surgical management of PCV with VH [4,6]. In the current study we report the incidence, clinical characteristics, management and surgical outcome in haemorrhagic PCV presenting with vitreous haemorrhage.

Methods
This study was a retrospective, interventional consecutive case series. The study was approved by the local ethics committee of L.V. Prasad Eye Institute (LEC-11-160) and all participants gave consent for using patient data for research purpose. Medical records of all cases having a diagnosis ICD coding of PCV and vitreous haemorrhage who underwent pars plana vitrectomy (PPV) were identified and studied. Twenty-eight eyes of 27 consecutive patients treated surgically for PCV associated with VH were included.
Due to the presence of VH, the media clarity did not allow the visualization of the fundus or angiography in most eyes at presentation. Although the EVEREST trial laid down guidelines for the diagnosis of PCV [11], massive hemorrhagic PCV may not show the same features at presentation in the affected eye. The vitreous haemorrhage was attributed to PCV on the basis of preoperative B-scan features [6], and confirmed with demonstration of polypoidal lesions and/or branching vascular network in the study or fellow eye on indocyanine angiography (ICGA) in the past or post-operatively, as laid down in the EVEREST trial [11]. Patients were excluded if the same or fellow eye had any other vascular or degenerative retinal or macular condition, or any other ocular pathology which could limit their vision.
After a preoperative written informed consent all patients underwent standard 20 or 23 gauge PPV. Removal of subretinal blood was not attempted in this series.
The data pertaining to age, gender, duration of diabetes and hypertension, duration of symptoms, best-corrected visual acuity (BCVA) at presentation, 1, 3, 6 months postoperative and at the last follow up, intraocular pressure (IOP, Goldman applanation tonometry), surgical procedures performed, surgical complications and duration of follow up were collected and are summarized in Table 1.
The primary outcome measure of this series was the number of patients who improved by two or more lines on early treatment diabetic retinopathy study (ETDRS) chart after surgery. Secondary outcome measures were mean change in logMAR visual acuity, and complications of surgery.
The visual improvement post operatively and the differences between BCVA at 1 month post operative visit and final follow up from the baseline pre operative visual acuity were compared using paired t test. Statistical analysis was done using Statistical software R, version 2.14.1.

Results
Twenty-eight eyes of 27 patients (21 men, 6 women) with a mean age of 58.89 ± 13.13 years (range 36-82) underwent PPV for vitreous haemorrhage associated with PCV between January 2003 and December 2011. Their baseline demographic characteristics are shown in Table 1. Ten (37 %) patients had a history of diabetes mellitus for an average of 13 years (range 3-25 years) and 16 (59 %) patients had a history of hypertension for an average 10.5 years (range 1-36).  The mean duration of symptoms at presentation was 6.2 ± 3.3 months with sudden onset profound painless diminution of vision being the most common presenting complaint. The mean preoperative baseline logMAR visual acuity was 2.69 ± 0.58 (<20/2000, range 20/400 to perception of light, PL). At 1 month postoperative follow up, the mean BCVA was 1.69 ± 0.93 (20/979, range 20/20 to PL) with a mean improvement of 1.00 logMAR units from the baseline (p < 0.001, 95 % CI 0.64-1.34). At 3 and 6 months follow up the BCVA was 1.46 ± 0.93 (20/576, range 20/20 to PL) and 1.57 ± 1.26 (20/743, range 20/20 to no PL) respectively. The mean BCVA at the last post operative follow up visit was 1.73 ± 1.14 (20/1074, range 20/20 to no PL) with an improvement of 0.96 logMAR units from the baseline (p < 0.001, 95 % CI 0.54-1.37). The BCVA improved by two or more lines on ETDRS chart in 16 of 28 eyes (57.1 %), stabilized or improved by less than two lines in nine (32.1 %) and worsened in three (10.7 %) at a mean follow up of 14.28 months. The three eyes which worsened post-operatively had a preoperative vision of hand movement which decreased to PL in one and loss of PL in the other two eyes. The results of visual outcome have been summarized in Table 2. The trend of BCVA change from baseline preoperative to post operative last visit is shown in Fig. 1.
Prior to surgery, five study eyes had been treated with intravitreal bevacizumab, one received intravitreal triamcinolone, three had undergone focal laser treatment, two eyes had undergone sclera buckling in the same eye in the past and three eyes had a history of photo dynamic therapy (PDT). Three patients in whom B-scan showed exudative retinal detachment were treated with oral steroids preoperatively. Nine (33.3 %) patients had bilateral PCV of which four (14.8 %) also had associated VH with PCV in the fellow eye.
At presentation, fundus evaluation was not possible due to dense vitreous haemorrhage in 21 of the 28 eyes (75 %). B-scan showed vitreous haemorrhage in all 28 eyes, extensive exudative retinal detachment in six and suspected mass lesion in one eye. Diagnosis of underlying cause being PCV was based on demonstration of polyps or branching vascular on previous or post operative ICG-Angiography (ICG-A) in seven study eyes, four fellow eyes (Fig. 2) and on fluorescein angiography in two study eyes. The rest of the cases were diagnosed on the basis of presence of multiple sero-sanguinous RPE detachments suggestive of PCV. Two of the cases where polyps had been demonstrated on ICG-A before VH had regression of the previously noted polyps, but developed polyps in new locations subsequently (Fig. 3).

Discussion
Several reports and case series describing the diverse clinical spectrum, natural course, and role of various modalities of treatments for exudative PCV have been published in the last three decades [1, 2, 6-9, 11, 12], but there are few reports in the literature on the management of vitreous haemorrhage associated with PCV [6,8,9,13]. Vitreous haemorrhage associated with PCV differs from other manifestations of PCV as these eyes are not amenable to ICG angiography. Unless the affected eye has been previously diagnosed as PCV or the fellow eye has features of PCV, it is difficult to diagnose them. B-scan features that can differentiate vitreous haemorrhage secondary to PCV from other causes has been reported and are very useful in this situation. B-scan can demonstrate defining features like blood lining an elevated membrane, hemorrhagic pigment epithelial detachment and absence of any acoustic shadowing. The natural course or role of any treatment modality is unknown for vitreous haemorrhage in PCV as no large series of cases has been studied.
Vitreous haemorrhage associated with PCV has also been reported to occur following PDT [7,17]. In our series only one patient had history of recent PDT and developed sudden onset loss of vision and VH 1 week after the treatment.
The earliest descriptions of vitreous haemorrhage in PCV was in 1985 [9,10]. In previous reports by Kleiner et al. and Perkovich et al. [8,9], visual recovery after vitrectomy was found to be equivocal. Two of the three eyes gained useful vision in one series [8], whereas the other studies reported worsening or no change in the visual acuity [4,9]. These studies had small series ranging from one to three eyes that underwent vitrectomy. Recent reported outcomes of vitrectomy in PCV did not find a higher incidence of retinal breaks [13,18].
In our study, significant visual improvement occurred in 16 of 29 eyes (57.1 %). Three patients recovered a visual acuity of better than 20/40, one eye had a visual We did not attempt to remove the subretinal blood in our study. All cases in our study would have required subretinal tissue plasminogen activator with incubation of approximately half an hour. In our experience, the blood decolorizes rapidly in these cases and does not liquefy with conventional doses of tPA. In such cases, the dose of tPA is much more than the recommended upper limit of 50 μg. Further, washing of blood in such massive hemorrhage under the macula tends to cause photoreceptor damage. The peripheral visual field would also have improved after the clearance of vitreous haemorrhage but would be difficult to document in most cases due to poor central fixation.
The commonest intra operative complication was iatrogenic retinal breaks, which occurred during induction of PVD which was also been reported by us in our previous study [6]. Hence it is possibly better to wait for spontaneous PVD to occur before attempting surgery. The cases which had no visual recovery or experienced worsening of vision had macular scar, recurrent VH, post operative retinal detachment and cataract.
In a few cases in our study, polypoidal lesions which were seen on ICG-A before vitreous haemorrhage were not seen postoperatively but new lesions polypoidal lesions developed in a new location on subsequent follow up (Fig. 3). This disappearance of classic polypoidal lesions and replacement by RPE atrophy and scarring has been described in previously reported cases [1,9]. The development of the haemorrhagic PEDs from the preexisting polypoidal lesions, followed by breakthrough VH and then subsequent scarring may be the natural course of the condition.

Conclusion
Our study shows that vitrectomy can produce significant visual improvement in patients with vitreous haemorrhage due to PCV. The incidence of retinal breaks during vitrectomy is high in PCV, and caution should be exercised while inducing PVD in such cases.