Air vs Gas Tamponade During Macular Hole Repair Surgery

Abstract

Purpose:

This work investigates the visual and anatomical outcomes of full-thickness macular hole (FTMH) repair surgery using air in comparison to gas tamponade.

Methods:

A retrospective consecutive review of medical records was undertaken of all patients undergoing pars plana vitrectomy for idiopathic FTMH at an academic practice from January 2010 to May 2017. Each operative report was reviewed to investigate the agent used for tamponade at the end of the surgery. Preoperative hole duration and size as measured using optical coherence tomography as well as successful postoperative hole closure were recorded. Use of gas or air was not randomized and was instilled at surgeon discretion.

Results:

The final analysis included 211 eyes. Gas was used as the tamponade agent in 171 of the 211 eyes; most of these eyes (144 of 171) received sulfur hexafluoride (SF₆) and the remainder received perfluoropropane (C₃F₈). Forty eyes underwent only a complete fluid-air exchange without any gas placement following vitrectomy. There was no statistically significant difference between the 2 groups in mean preoperative macular hole size (P = .43). Nine of the 171 macular holes receiving gas tamponade failed to close (5.3%). One of the 40 macular holes receiving only air failed to close (2.5%). There was no statistically significant difference in hole closure rates between the 2 groups (P = .45).

Conclusions:

Air served as an equally efficacious internal tamponade agent in comparison to nonexpansile gas following idiopathic FTMH repair surgery.

Introduction

Since the landmark pilot study by Kelly and Wendel in 1991 ¹ proposing a technique for macular hole (MH) repair surgery, there has been a paradigm shift in the approach to treating a previously untreatable condition. ² Their technique, incorporating pars plana vitrectomy (PPV) with removal of posterior cortical vitreous, 20% sulfur hexafluoride (SF₆) gas tamponade, and postoperative face-down positioning, resulted in a 58% hole closure rate. With subsequent refinements in technique and advancements in instrumentation, this procedure has become the mainstay of MH treatment with reported closure rates surpassing 90%. ^3,4

Debate exists regarding certain nuances of MH surgery and postoperative care, including extent and technique of internal limiting membrane (ILM) peeling ⁵ and need for face-down positioning. ⁶ The high success rate of MH closure across the board makes it difficult to prove that certain variations in technique represent a clinically meaningful upgrade over others.

There is consensus among vitreoretinal surgeons that an integral step in MH repair is insufflation with a tamponing agent to facilitate hole closure. The choice of endotamponade is typically based on surgeon preference and experience. We know that both nonexpansile SF₆ and perfluoropropane (C₃F₈) are highly effective in promoting MH closure following vitrectomy. ^{7 -9} The data on MH closure rates using only air as an internal tamponade are less convincing. A few series have shown air to be specifically effective in closing MHs smaller than 400 μm in diameter. ^{10 -13} The data supporting air use in this setting are not as rigorous, thus hindering its routine application in MH repair. The goal of the present study is to further investigate MH closure rates with air fill vs gas fill following small-gauge vitrectomy and ILM peeling.

Methods

The medical records of patients undergoing elective PPV surgery for idiopathic full-thickness macular hole (FTMH) at Vanderbilt University Medical Center from January 2010 to May 2017 were retrospectively reviewed. Only patients undergoing elective small-gauge PPV for idiopathic FTMH (Common Procedural Technology code 67042) confirmed on preoperative optical coherence tomography (OCT) were included. Main exclusion criteria were patients with previous incisional vitreoretinal surgery in the study eye, traumatic or pediatric MH, and associated retinal detachment or tractional macular membranes from secondary retinopathy (eg, proliferative diabetic retinopathy).

Surgery was performed at 1 site by 5 vitreoretinal surgeons in the Retina Division at the Vanderbilt Eye Institute. All operations were performed in a similar fashion using the Alcon vitrectomy system with 23-, 25-, or 27-gauge vitrectomy probes. Surgical approach and technique were per the individual operating surgeon’s protocol, including use of tissue-staining dyes, choice of endotamponade, and postoperative face-down positioning regimen. All surgical procedures were performed with the assistance of a vitreoretinal fellow. A peripheral vitreous shave was performed in all cases. Operative reports were reviewed to confirm successful peeling of ILM around the MH. All patients were patched with follow-up on postoperative day 1. Postoperative medications, positioning regimens, and follow-up were implemented per individual surgeon protocol.

OCT images were reviewed to measure preoperative hole size and postoperative anatomic result. The OCT caliper function was used to measure aperture size—the minimum linear diameter of the MH in the midretina by creating a line parallel to the retinal pigment epithelium. ¹⁴ Patients without macular OCT at least 3 months postoperatively were excluded from analysis.

Based on previously published success rates for gas and air tamponade for MH surgery, we determined that a sample size of 204 patients would be required to detect a 10% difference in MH closure rates with a certainty of at least 80%. Data collection and statistical analysis were performed in Microsoft Excel (Microsoft Corp) and GraphPad (GraphPad Software, Inc). A P value of less than .05 was regarded as statistically significant.

Results

The medical records of 464 patients undergoing PPV with ILM peeling were reviewed. After refining based on exclusion criteria, the final analysis included 211 eyes of 205 patients. Successful MH closure was based on postoperative OCT confirmation of MH resolution. Any MH that remained open after primary vitrectomy surgery, regardless of any further attempts to close the hole, was deemed a failure.

The average preoperative size of the MH in the air group was 282 ± 97 μm (range, 132-493 μm). Three of these holes were larger than 400 μm at the time of surgery. The average preoperative size of the MH in the gas group was 310 ± 168 μm (range, 91-795 μm). There was no statistically significant difference between the 2 groups in mean preoperative MH size (P = .43). However, the gas group had a significantly higher number of preoperative MHs larger than 400 μm at the time of surgery (n = 24, P = .003). See Table 1.

Table 1.

Patient Demographics and Macular Hole (MH) Characteristics.

Characteristics	Air (n = 40)	Gas (n = 171)	P
Sex
Male, n (%)	15 (37.5)	56 (32.7)	.56
Female, n (%)	25 (62.5)	115 (67.3)	.56
Age at surgery, mean, y (SD)	66.8 (6.8)	67.2 (8.6)	.79
Preoperative lens status
Phakic, n (%)	21 (52.5)	85 (49.7)	.75
Pseudophakic, n (%)	19 (47.5)	86 (50.3)	.75
MH characteristics
Duration of MH, wks (SD)	12.4 (15.6)	12.4 (13.4)	.99
Preoperative MH size, μm (SD)	282 (97)	310 (168)	.43
No. of holes > 400 μm at time of surgery	3	24	.003
MH failures
Total No. of failures, n (%)	1 (2.5)	9 (5.3)	.45
Duration of MH symptoms, wks	6	11.6
Preoperative MH size, microns (SD)	420	489 (149)
Vision, Snellen
Preoperative	20/130	20/170	.10
At last follow-up	20/50	20/80	.08

Of the 211 eyes, 171 eyes were infused with gas at the end of their MH surgery (gas group; 144 eyes with SF₆, 27 eyes with C₃F₈). SF₆ was diluted to 20% to 26%. C₃F₈ was diluted to 12% to 16%. Forty eyes underwent complete fluid-air exchange at the end of their MH surgery without additional gas infusion (air group). The overall single surgery MH closure rate for the entire series was 95.3% (201/211). In the gas group, there were 9 failures for a success rate of 94.7% (162/171). All 9 failures occurred in the SF₆ group. In the air group, there was 1 failure for a success rate of 97.5% (39/40). See Table 2. There was no statistically significant difference in success rates between the gas and air group (P = .45). Final MH closure rate after subsequent surgical procedures was 100% in each group.

Table 2.

Macular Hole (MH) Closure Rates.

	Successful MH Closure	Failures	Totals	Success Rate, %
Gas	162	9	171	94.7
SF6	135	9	144
C3F8	27	0	27
Air	39	1	40	97.5
Combined	201	10	211	95.3

Abbreviations: C₃F₈, perfluoropropane; SF₆, sulfur hexafluoride.

Looking at the 9 MHs that failed to close in the SF₆ group, there were no significant demographic differences compared with other patients in the gas cohort. The average age was 66.2 ± 9.3 years (P = .76), and the average duration of MH was 11.6 weeks (P = .77). Two of these patients were men. Average preoperative vision was similar at 20/189 (P = .67). All eyes receiving C₃F₈ had successful primary MH closure.

Conclusions

Our review of 211 eyes undergoing FTMH surgery demonstrates a comparable success rate between air and gas fill. Air and gas closure rates, along with total MH closure rates, were both near 95%. There were no statistically significant demographic differences between these 2 cohorts. Additionally, mean preoperative MH sizes were statistically similar between the 2 cohorts, further validating the findings.

The results of our study yield a few important observations. Air endotamponade can yield high success rates with MHs of all sizes, including those larger than 400 μm. Earlier studies noted improved success associated with prolonged gas tamponade. ¹⁵ More recently, using postoperative OCT, Eckardt et al demonstrated 54.5% of MHs were closed within 24 hours following PPV with air tamponade. ¹⁶ Thus there is rationale for an air fill over a longer-acting gas fill. In our case series, only 1 MH failed to close using air. Notably, this eye demonstrated only a 60% air fill on postoperative day 1. The partial air fill likely contributed to failure of the MH to close. Although holes typically close by postoperative day 1, this study did not control for surgeon preference regarding face-down positioning. The benefit of face-down positioning is debated and not necessarily crucial for holes smaller than 400 μm. Presently there is a lack of high-quality evidence to suggest its benefit in the setting of air fill vs gas fill.

Overall hole closure rate in our series of small-gauge PPV was 95.3%. This reinforces the existing data on the high anatomic success rate achieved by retina surgeons with primary MH surgery. However, persistent or reopened MHs represent a challenge, and many techniques have been proposed to address them following primary PPV. Regardless, anatomic and visual success rates in this setting remain poor compared with primary MH closure rates. ¹⁷ With that being said, an argument can be made for maximizing the first-surgery success rate through the use of longer-acting gas tamponade. Conversely, an air fill can achieve comparable success rates in certain settings with quicker visual recovery. Additionally, it is well established that longer-acting gas tamponade accelerates cataract formation in phakic patients following vitrectomy. ¹⁸ Approximately half the patients in our study were phakic at the time of MH surgery.

Limitations of this study are due to its retrospective nature. Multiple surgeons with differing protocols—including varying postoperative face-down positioning regimens—preclude broad application of the results of this study. Air was preferentially used by some surgeons and gas by others, creating unequally sized cohorts. Air may have been chosen for MHs that were perceived to be easier to close. Although there was no statistical difference in average preoperative MH size in the air and gas groups, there were significantly more holes in the gas group that were larger than 400 μm at the time of surgery. This can be a potential source of selection bias by which surgeons specifically used a longer-acting tamponade for larger holes. Other factors not mentioned in the medical record may have introduced bias when choosing endotamponade (eg, the patient’s ability to position postoperatively, difficulty with ILM peeling, size of bubble on day 1 after surgery).

A minimum of 3 months of postoperative follow-up was required for inclusion in the study. Given the low rate (4.8%) of late MH reopening after successful closure, ¹⁹ this was deemed to be adequate. OCT MH measurements were taken using the caliper function, although discrepancies may exist with progressive OCT generations or because of caliper placement.

In conclusion, the authors demonstrate air can serve as a viable alternative endotamponade to gas in the setting of PPV for idiopathic MH. Although both choices offer excellent hole closure rates, a randomized prospective trial is needed to confirm equal efficacy. Patient characteristics—including ability to comply with postoperative regimen and hole size—play an important role in endotamponade choice. An important strength of our paper is, despite its retrospective nature, it was a consecutive case series and holes were not excluded based on size or chronicity. In our study, most holes were smaller than 400 μm and of relatively short duration. Air proved to be an excellent endotamponade in this setting.