Prognostic Factors of Fluid-gas Exchange Outcomes in Patients with Failed Primary Surgery for Idiopathic Macular Hole

This retrospective study aimed to evaluate prognostic factors associated with the success of uid-gas exchange in patients with failed primary idiopathic macular hole (IMH) surgery. Outcomes of the uid-gas exchange were categorised as unclosed, or U-type, V-type, or W-type closure. Patients were divided into the successful and unsuccessful groups according to the absence or presence of bare retinal pigment epithelium, respectively, following uid-gas exchange. Demographics, baseline characteristics, and pre-procedure characteristics were assessed. A total of 19 eyes in 19 patients that had failed IMH surgery and then underwent uid-gas exchange were included. Of those, 18 eyes had MH closure (successful, 15 eyes; unsuccessful, three eyes). One eye was unclosed after uid-gas exchange; therefore, the patient underwent additional vitrectomy for MH closure (unsuccessful). The successful group showed signicantly lower pre-procedure base diameters; higher baseline and pre-procedure macular hole index, hole form factor, and tractional hole index values; and lower baseline and pre-procedure minimum diameter and diameter hole index values. Moreover, a better visual prognosis was observed in the successful group. These results suggest that indices predicting favourable results of primary surgery for IMH are useful to predict the success of uid-gas exchange in patients with failed primary MH surgery.


Introduction
A macular hole (MH) is a full-thickness defect in the fovea. Since Kelly and Wendel 1 reported the successful closure of an MH using pars plana vitrectomy with uid-gas exchange, this surgical method combined with internal limiting membrane peeling has become the standard treatment for MH, showing an anatomical closure rate of more than 90% 2,3 . Various optical coherence tomography (OCT) parameters, such as the hole form factor (HFF), base diameter (BD), and macular hole volume (MHV), are prognostic factors that determine functional and anatomical success in primary MH surgery 4,5 .
Unfortunately, about 10% of all MH surgeries fail and, therefore, may require additional procedures.
Intravitreal gas injection or uid-gas exchange using octa uoropropane (C 3 F 8 ) or sulfur hexa uoride (SF 6 ) for MH closure has shown favourable results in such cases 6,7 . However, some of them with a failed primary MH surgery patients treated with additional procedure of injection or uid-gas exchange have needed additional surgery to close the MH 8 . To the best of our knowledge, there have been no studies regarding the effectiveness of indices in predicting the outcomes of a uid-gas exchange procedure following a failed primary MH surgery. Therefore, we aimed to evaluate the prognostic factors associated with the success of uid-gas exchange after a failed primary idiopathic macular hole (IMH) surgery.

Demographic characteristics
A total of 19 eyes in 19 patients were included in this study. The uid-gas exchange procedure following the failed primary IMH surgery resulted in MH closure in 18 eyes (successful in 15 eyes and unsuccessful Page 3/13 in three eyes, Fig. 1). The MH was unclosed in one eye after the uid-gas exchange, and the patient underwent additional vitrectomy to close the MH (unsuccessful). All patients in the successful group showed a U-type closure, whereas all in the unsuccessful group presented a W-type closure 9 . The demographic data of all patients are shown in Table 1. The mean age was 67 ± 6 years in the successful group and 73 ± 7 years in the unsuccessful group; the difference was not statistically signi cant.
Similarly, the affected eye and sex did not differ signi cantly between the two groups. BCVA, best-corrected visual acuity, IOP, intraocular pressure, SD, standard deviation.
Baseline clinical characteristics of the successful and unsuccessful groups A comparison of the baseline clinical characteristics between the two groups is shown in Table 1. A longer duration of decreased visual acuity was observed in the unsuccessful group (87 ± 56 days); however, this difference was not statistically signi cant, compared with that in the successful group (57 ± 59 days). The total follow-up period, baseline best-corrected visual acuity (BCVA), mean intraocular pressure (IOP), mean spherical equivalent, number of phakic patients at baseline, gas type of primary tamponade, percentage of primary gas tamponade, and mean central sub eld retinal thickness (CSRT) at baseline were not signi cantly different between the two groups. The mean baseline minimum diameter (MD) was signi cantly lower in the successful group (363.87 ± 124.95 µm) than in the unsuccessful group (612.00 ± 11.43 µm); however, no difference was observed in the mean baseline BD between the two groups. The mean MH height also showed nonsigni cant differences between the two groups. A signi cantly higher mean baseline macular hole index (MHI) was observed in the successful group (0.558 ± 0.159) than in the unsuccessful group (0.386 ± 0.041). The mean baseline HFF was signi cantly higher in the successful group than in the unsuccessful group (0.876 ± 0.154 versus 0.463 ± 0.204). The mean baseline diameter hole index (DHI) was signi cantly lower in the successful group (0.470 ± 0.182) than in the unsuccessful group (0.669 ± 0.067). The mean baseline tractional hole index (THI) was signi cantly higher in the successful group than in the unsuccessful group (1.404 ± 0.843 versus 0.587 ± 0.129). However, the mean baseline MHV was not different between the two groups.
Comparison of pre-procedure clinical characteristics between the successful and unsuccessful groups A comparison of the pre-procedure clinical characteristics between the two groups is shown in Table 2. A longer interval between primary surgery and uid-gas exchange was observed in the successful group (17 ± 8 days) than in the unsuccessful group (12 ± 3 days); however, this difference was not statistically signi cant. The mean pre-procedure CSRT also showed no difference between the successful group and unsuccessful group. Both mean pre-procedure MD and mean pre-procedure BD values were signi cantly lower in the successful group (276.13 ± 89.59 µm versus 633.00 ± 182.42 µm, and 638.20 ± 274.34 µm versus 831.00 ± 113.18 µm, respectively). In contrast, the mean pre-procedure MH height was not signi cantly different between the two groups. A signi cantly higher mean pre-procedure MHI was observed in the successful group (0.624 ± 0.193) than in the unsuccessful group (0.452 ± 0.054). The mean pre-procedure HFF was signi cantly higher in the successful group than in the unsuccessful group (0.881 ± 0.204 versus 0.511 ± 0.140). The mean baseline DHI was signi cantly lower in the successful group (0.474 ± 0.161) than in the unsuccessful group (0.751 ± 159), whereas the mean baseline THI was signi cantly higher in the successful group than in the unsuccessful group (1.419 ± 0.561 versus 0.640 ± 0.245). Moreover, the mean baseline MHV was signi cantly lower in the successful group than in the unsuccessful group (0.411 ± 0.334 versus 1.032 ± 0.426). The type of procedure gas (SF 6 ) did not differ between the two groups. Visual prognosis of the successful and unsuccessful groups The successful group presented signi cantly better nal BCVA values (0.33 ± 0.18 logMAR) than the unsuccessful group (0.78 ± 0.15 logMAR, Table 3). Furthermore, the nal BCVA in the successful group was signi cantly improved compared to baseline BCVA (0.85 ± 0.22 logMAR), whereas the unsuccessful group demonstrated a poorer nal BCVA compared to baseline (0.70 ± 0.14 logMAR); however, this difference was not statistically signi cant (Table 3).

Discussion
This study retrospectively analysed the baseline and pre-procedure characteristics of patients with primary unclosed MH treated with uid-gas exchange and showed that several baseline and preprocedure OCT parameters were associated with the success of the uid-gas exchange procedure after a failed primary IMH surgery. Various studies have reported relationships between preoperative OCT parameters and the prognosis of MH patients. Shpak et al. 10 demonstrated that a higher preoperative CSRT is related to the anatomical success of surgical IMH treatment. Similarly, our study showed increased baseline and pre-procedure CSRT values in the successful group compared to those in the unsuccessful group, although this difference was not statistically signi cant. Wakely et al. 11 reported that the parameters BD, MH inner opening, and MD were related to the anatomical and functional success in MH surgery. In patients with primary unclosed IMH treated with uid-gas exchange, a smaller baseline MD, pre-procedure MD, and pre-procedure BD were also related to success. Kusuhara et al. 12 showed that IMH patients with MHI values ≥ 0.5 had a better postoperative BCVA than those with MHI values < 0.5. In this study, baseline and pre-procedure MHI values ≥ 0.5 were observed in the successful group with a better nal BCVA than those in the unsuccessful group (baseline and pre-procedure MHI < 0.5). Several studies have reported that a higher HFF increases the closure rate and improves the postoperative visual outcome in MH patients 4 . In agreement with this publication, signi cantly higher baseline and preprocedure HFF values with better nal BCVA were observed in the successful group of our study. Ruiz-Moreno et al. 13 and Dai et al. 14 demonstrated that a higher THI is correlated with good visual prognosis, whereas DHI is not. However, our study revealed that both higher THI and lower DHI were related to the success of the uid-gas exchange in patients with primary unclosed MH. Unsal et al. 5 reported that a higher MHV is associated with a poorer postoperative BCVA, whereas Ozturk et al. 15 detected no signi cant correlation. In this study, no difference in baseline MHV was noted between the two groups, but a signi cantly higher pre-procedure MHV was observed in the unsuccessful group. These ndings suggest that MH indicators that are related to the success of primary surgery tend to be also correlated with the success of uid-gas exchange in patients with failed primary MH treatment. Zou et al. 16 reported that a shorter symptom duration is associated with an earlier reconstruction of the external limiting membrane, resulting in better visual outcomes. Shorter symptom duration was also observed in the successful group of our study; however, compared with the unsuccessful group, the difference was not statistically signi cant.
The current consensus recommends internal limiting membrane peeling during a MH surgery; however, the face-down position after primary vitrectomy remains controversial 17 . Nonetheless, a previous metaanalysis concluded that the face-down position is highly recommended when the MH size exceeds 400 µm 18 . In this study, the MH size in all patients at baseline was > 400 µm, and all patients were positioned face-down after the primary vitrectomy. Among them, 94.74% (18 of 19 patients) had MH closure after the uid-gas exchange procedure, with ≥ 7 days of additional period in face-down position. These ndings highlight that an inadequate period of gas tamponade with face-down position might be associated with the failure of primary surgery in IMH patients, and the uid-gas exchange procedure with the additional period in face-down position is useful in these failed primary IMH patients. One previous study mentioned that SF 6 and C 3 F 8 gas tamponades during surgery show similar MH closure rates 19 . In agreement with these ndings, we observed that the type of gas used for the uid-gas exchange was not different between the successful and unsuccessful groups.
This study has some limitations. First, the retrospective design may have resulted in a patient selection bias. Second, this study included only a small number of patients. Therefore, further studies with larger numbers of patients with failed primary IMH treatment will be needed. Despite these limitations, to the best of our knowledge, this is the rst study to evaluate the factors associated with the success in patients with failed primary IMH surgery treated with uid-gas exchange.
In conclusion, lower pre-procedure BD, lower baseline and pre-procedure DHI and MD, and higher baseline and pre-procedure MHI, HF, and THI values are associated with the success of uid-gas exchange in patients with failed primary IMH treatment. These results suggest that indices predicting the success of primary IMH surgery may also be useful to predict the success of uid-gas exchange in patients with failed primary surgery.

Study design and subjects
This study was designed as a retrospective, comparative case series conducted in a single hospital. The study adhered to the tenets of the Declaration of Helsinki and was approved by the institutional review board of Keimyung University Dongsan Hospital (IRB no. 2020-12-069). The patients signed informed consent for the use of their data. We retrospectively reviewed the electronic medical records of patients with a history of a failed primary IMH surgery who were treated with uid-gas exchange between January 2013 and December 2019. The patients with at least 6 months of follow-up period after the procedure were enrolled in the study. The exclusion criteria were as follows: myopia > 6 dioptres, presence of rhegmatogenous retinal detachment, diabetic retinopathy, age-related macular degeneration, or a history of previous vitrectomy due to causes other than IMH.
All patients had a full-thickness IMH and underwent primary vitrectomy with internal limiting membrane peeling using the inverted ap technique. Cataract surgery was performed simultaneously when the cataract impaired posterior visualisation. The patients were instructed to remain in a face-down position for ≥ 7 days. The uid-gas exchange procedure was performed on 19 eyes in which an unclosed MH was noted in the spectral-domain OCT (Spectralis OCT/SLO; Ophthalmic Technologies, Toronto, Ontario, Canada) or swept-source OCT (DRI OCT-1; Topcon, Tokyo, Japan) during the follow-up period. The procedure was based on a previous report by Jang et al. 20 , and the patients were instructed to stay in the face-down position for 7 days. The type and percentage of gas used in the primary surgery were determined by the surgeon's preference. During the uid-gas exchange, 16% of SF 6 gas or 14% of C 3 F 8 gas was used. All aforementioned surgical and other procedures were performed by a single clinician (Y.C.K.).
Based on the classi cation by Imai et al. 6 , the outcomes of the uid-gas exchange procedure were classi ed as U-type closure, V-type closure, W-type closure, or unclosed. The patients were divided into the successful and unsuccessful groups according to the absence or presence of bare retinal pigment epithelium after the uid-gas exchange procedure, respectively (Fig. 1). Demographic and clinical characteristics were compared between the two groups to assess the prognostic factors.

Clinical data collection
The demographic characteristics including age, sex, and laterality were assessed in all patients. Furthermore, the duration of decreased visual acuity, baseline IOP measured by a non-contact tonometer (Canon TX-20, Canon Inc, Kanagawa, Japan), spherical equivalent, lens status (phakic or pseudophakic), axial length, type of primary surgery (vitrectomy alone or combined cataract surgery) with the type of gas and the percentage used for tamponade during primary surgery, interval period between the primary surgery and uid-gas exchange, and the type and percentage of gas used in the uid-gas exchange were recorded. The baseline and nal logMAR BCVA values were also assessed. OCT images of all patients were recorded at baseline, pre-procedure, and post-procedure until the last follow-up. CSRT, BD, MD, MH height, MHV height, left arm length, and right arm length were measured using the OCT software measuring tool at baseline and pre-procedure. In addition, the MHI based on the publication by Kusuhara et al. 12 , the THI and DHI by Ruiz-Moreno et al. 13 , the HFF by Ullrich et al. 4 , and the MHV by Ozturk et al. 15 were calculated.