Chorioretinal Toxicity of Perfluorooctane (Ala Octa)


To compare outcomes of patient who underwent surgery using perfluorooctane (PFO; C 8 F 18 ; Ala Octa) with those who underwent surgery with perfluorodecalin (F-Decalin).


Retrospective, consecutive, comparative, interventional case series.


A total of 48 eyes that underwent vitrectomy with PFO were compared to 29 eyes that underwent vitrectomy with perfluorodecalin. Two experienced surgeons performed vitrectomies at the Geneva University Eye Clinic. Visual acuity before, at 8 and 24 weeks after surgery, was documented, and spectral domain optical coherence tomography (SD-OCT) images were analyzed for abnormalities.


Two patients experienced severe retinal toxicity, including 1 with severe vision loss. However, no statistical differences in VA were observed between the PFO and perfluorodecalin patients. Analysis of SD-OCT images showed differences in occurrence of several abnormalities, for example, inner segment-outer segment alterations were found in 60.4% of eyes treated with PFO and in 10.3% of perfluorodecalin-treated eyes; retinal atrophic areas were found in 41.7% of PFO and in none of the perfluorodecalin eyes; inner limiting membrane contraction was found in 58.4% of PFO and in none of perfluorodecalin eyes; inner retina cystic alterations were found in 58.3% of PFO eyes and 17.2% of perfluorodecalin eyes; outer retina cystic alterations were found in 39.6% of PFO eyes and 13.8% of perfluorodecalin eyes; retinal holes were found in 14.6% of PFO eyes and in none of the perfluorodecalin eyes; and outer retinal inclusions were found in 20.8% of PFO eyes and in 3.45% of perfluorodecalin eyes.


Perfluorooctane caused significantly more toxic damage than perfluorodecalin. Special consideration should be given to develop a central European Union (EU) control agency for medical devices and to reevaluate safety procedures currently accepted by the EU and International Organization for Standardization for intraocular surgery.

Since their introduction, perfluorocarbon liquids (PFCLs) have been extensively used to relocate and fixate detached retinas during surgery as well as to displace subretinal or subchoroidal fluid, to facilitate visualization of the retinal periphery (Schlieren effect), for the detection of discrete proliferative vitreoretinopathy and protection of the macula from retained lens fragments or foreign bodies during surgery.

PFCLs are not recommended for long-term tamponade and are removed at the end of surgery, because toxic effects to retinal pigment epithelial cells and retinal ganglion cells in culture have been reported after several days of exposure. Perfluorooctane (PFO; C 8 F 18 ; Ala Octa; Alamedics, GmbH Co. KG, Domstadt, Germany), a frequently used PFCL, is an alkane containing carbon and fluorine. It is transparent and has the capacity to transport oxygen. The physical properties of PFO are useful in intraocular surgery because its low viscosity and low surface tension allow maintenance of a single confluent bubble and the ability to pass above dehiscences while its high density confers a tamponade effect. Recently, PFO has been reported to be toxic to the retina of some patients after short-term intraocular use during surgery and has been withdrawn from the market. Reported toxic effects in patients included anterior chamber inflammation, iris rubeosis, and retinal re-detachments.

Market approval was given by the European Medicines Agency (EMA) and national regulatory agencies even though PFO was shown to induce severe degenerative changes and cause retinal pigment epithelium and ganglion cell death in culture in human and swine cell lines of retinal pigment epithelium.

In the authors’ clinic, in addition to PFCLs from other manufacturers, PFO was used from April 2014 to October 2015. The authors analyzed charts and spectral domain optical coherence tomography (SD-OCT) (Spectralis; Heidelberg Engineering, Heidelberg, Germany) images of patients who were given PFCLs during intraocular surgery to quantify potential anatomical abnormalities on SD-OCT and postoperative outcomes (i.e., best-corrected visual acuity [BCVA], retinal re-detachment, and additional surgical interventions, such as phacoemulsification and epiretinal membrane [ERM] peeling). Comparison of data from patients whose eyes were treated with PFO with patients operated with perfluorodecalin (F-Decalin; Fluoron, GmbH, Ulm, Germany) during surgery has been used to uncover potential toxic complications of PFO and to identify patients at risk for such toxic effects that should be carefully monitored.

Subjects and Methods

This retrospective, consecutive, comparative, interventional case series was approved by the Swiss Commission Cantonale d’Ethique de la Recherche (CCER) Genève (reference number 2020-00735).

The study was based on all available cases with complete medical dossiers during the period of use of each of the two PFCLs in patients who met the inclusion and exclusion criteria.

Patients were selected who underwent pars plana vitrectomy (PPV) between April 2014 and April 2017 at the Eye Clinic of the University Hospitals of Geneva (HUG), Geneva, Switzerland, using PFO or perfluorodecalin as an endotamponade.

Patients’ medical records consisted of funduscopic images, optical coherence tomography (OCT) images, surgical videos, and patient demographics.

Inclusion Criteria

  • PPV surgery with intraocular PFO between April 2014 and October 2015 or perfluorodecalin between November 2015 and April 2017 (perfluorodecalin has been used in the authors’ hospital since PFO was withdrawn from the market).

  • Indications for PPV were rhegmatogenous retinal detachment (RD), recurrent RD, intraocular lens luxation, traumatic RD, and intraocular foreign body extraction.

Exclusion Criteria

  • Incomplete medical record.

  • Follow-up inferior to 6 months.

  • Written or oral documented refusal of using data for research purposes.

Between April 2014 and October 2015, a total of 67 eyes of 55 patients underwent PPV for complex pathologies using intraocular PFO, and 42 eyes of 42 patients underwent PPV for complex pathologies between November 2015 and April 2017 using perfluorodecalin, at the HUG Eye Clinic. A total of 48 eyes of 41 patients treated with PFO and 29 eyes of 29 patients treated with perfluorodecalin satisfied the inclusion and exclusion criteria.

According to the manufacturer’s certificate of analysis, PFO contains 99.0% purified PFO (C 8 F 18 ). Its physical and chemical properties consist of packaging in 5-mL vials; minimum purity of 99%; density of 1.76 g/cm 3 ; and vapor pressure of 18.5 mbar. Perfluorodecalin treatment was chosen as the control for comparison because the only other perfluorooctan used in the authors’ clinic after PFO withdrawal and was known as a well-tolerated and safe intraoperative tool for retinal surgery.

All surgeries were performed by 2 experienced retinal surgeons. Following vitrectomy, PFCL was injected slowly into the vitreous cavity, avoiding the optic disc and macula to prevent retinal stress and dispersion of the PFCL. At completion of the surgery, PFO or perfluorodecalin were removed and exchanged by balanced salt solution, air, or silicone oil. All retinal detachments underwent surgery within 72 hours of diagnosis and within 24 hours if the macula was attached.

Preoperative and postoperative VA at 8 and 24 weeks, SD-OCT readings with the description of all abnormalities, and intraoperative surgical records were documented for all patients. All patients were imaged using the Spectralis machine (Heidelberg Engineering) in SD-OCT mode. VA was measured using the Monoyer chart and converted to logarithm of minimal angle of resolution (LogMAR) units for statistical analysis and Snellen. The following LogMAR denotations were used for non-numeric VAs: finger counting = 1.7 LogMAR; hand movement = 2.0 LogMAR; light perception = 2.3 LogMAR; and no light perception = 3.0 LogMAR.

Conventionally, LogMAR is used for statistical analysis as it enables quantification of VA, whereas the Snellen chart is commonly used to measure VA in clinic. Both units were used in this study. VA is shown in LogMAR followed by Snellen in brackets.

Preoperative acuity was defined as the VA measured before surgery. Differences between pre- and postoperative VA were calculated for each patient by subtracting the postoperative BCVA in LogMAR at the last follow-up from the BCVA at presentation.


Assuming a gaussian distribution, mean VA values (in logMAR) of different groups were analyzed for statistically significant differences ( P < .05) using parametric tests: analysis of variance (ANOVA) with the Dunnett multiple comparison test for macula-off (refers to retinal detachment involving the macula) primary rhegmatogenous RD and the t -test for macula-on (refers to retinal detachment not involving the macula) primary rhegmatogenous RD (Prism version 7.02 software; GraphPad, LaJolla, California). Nonparametric OCT abnormality values of different groups were analyzed for statistically significant differences ( P < .05) using a Fisher exact test (Prism version 7.02 software; GraphPad).


Patient Demographics

Patient demographics and surgical data are shown in Table 1 .

Table 1

Patient Demographics

All Lots Lot 110213 Lot 170713 Lot 180214 Lot 261112
Number of treated eyes 48 10 25 10 3 29
Number of patients 41 10 23 10 3 29
Males 29 5 16 7 2 18
Females 12 5 7 3 1 11
Age at first examination 56 57 55 59 51 60
Preoperative VA in RD in LogMar (Snellen)
Mean 1.86 (≈20/1600) 1.63 (≈20/800)
Median 2 (≈20/2000) 1.7 (≈20/1000)
Mean 0.19 (≈20/32) 0.21 (≈20/32)
Median 0.2 (≈20/32) 0.2 (≈20/32)
Preoperative lens status
Phakic 28 4 14 8 2 19
Pseudophakic 15 5 7 2 1 10
Aphakic 5 1 4 0 0 0
Surgery indication
Rhegmatogenous RD 27 5 12 8 2 23
Macula-on 12 2 5 4 1 14
Macula-off 15 3 7 4 1 9
Recurrent RD 15 5 7 2 1 5
Macula-on 9 4 3 1 1 4
Macula-off 6 1 4 1 0 1
IOL luxation 4 0 4 0 0 0
Traumatic RD 1 0 1 0 0 1
Macula-on 0 0 0 0 0 0
Macula-off 1 0 1 0 0 1
IOFB extraction 1 0 1 0 0 0
Endotamponade, RD
SF6 23 14
C2F6 0 1
C3F8 14 11
Densiron 68 1 1
Silicone 1000 5 2
RD surgery time (in minutes)
Macula-on 65 59 70 68 56 85
Macula-off 95 95 100 79 53 92
Macula on/off 80 73 88 73 72 88
Mean follow-up, days 457 595 406 375 696 182

Ala = Alamedics; IOFB = intraocular foreign body; IOL = intraocular lens; OCTA = optical coherence tomography angiography; RD = retinal detachment; VA = visual acuity.

Table 1 shows patients’ characteristics, preoperative status, and surgery information.

SD-OCT Abnormalities

A summary of all observed SD-OCT abnormalities is shown in Table 2 . Differences in ERM, which were present in 20 of 48 patients who underwent surgery with PFO (41.7%) and 10 of 29 patients (34.5%) in whom perfluorodecalin was used were not statistically significant. Preretinal abnormalities ( Figure 1 ) in the form of inner limiting membrane (ILM) contraction or under-ILM bubbles of PFO were found in 28 of 48 eyes (58.4%), whereas none were found in eyes treated with perfluorodecalin ( P < .0001). Inclusions ( Figure 2 ) were observed in the inner retina in 8 of 48 eyes (16.7%) and in the outer retina in 10 of 48 eyes (20.8%) treated with PFO but in only 1 of 29 inner retinas (3.5%) and 3.5% of the outer retina of eyes treated with perfluorodecalin ( P = .0447). Macular cystic alterations ( Figure 3 ) were found in 28 of 48 inner retinas (58.3%) ( P = .0007) and in 19 of 48 outer retinas (39.6%) ( P = .021) of eyes treated with PFO but only in 5 of 29 retinas (17.2%) of Decalin-treated eyes for inner macular cystic alterations and only in 4 of 29 (13.8%) for outer macular cystic alterations. Inner segment-outer segment (IS-OS) alterations were observed in 29 of 48 eyes (60.4%) in which PFO was used and in 3 of 29 eyes (10.3%) treated with perfluorodecalin ( P < .0001). Retinal atrophy ( Figure 4 ) was found in 20 of 48 eyes (41.7%) that underwent surgery with PFO and none in eyes in which perfluorodecalin was used ( P < .0001). A retinal hole ( Figure 5 ) was found in 7 of 48 eyes (14.6%) in which PFO was used and in none of the eyes treated with perfluorodecalin ( P = .0407). Choroidal alterations, in the form of choroidal atrophy or folds, were observed in 6 of 48 eyes (12.5%) in which PFO was used and in none of the eyes in which perfluorodecalin was used.

Aug 17, 2020 | Posted by in OPHTHALMOLOGY | Comments Off on Chorioretinal Toxicity of Perfluorooctane (Ala Octa)

Full access? Get Clinical Tree

Get Clinical Tree app for offline access