FormalPara Key Summary Points

Pneumatic retinopexy (PnR) has been suggested as an alternative procedure for rhegmatogenous retinal detachment (RRD) repair under certain circumstances, and the criteria for PnR use has been expanded and revised to optimize the treatment response.

There is no consensus on the superiority of pars plana vitrectomy (PPV) over PnR or vice versa. Herein, we compared the efficacy and safety of PnR and PPV for RRD repair.

PPV had a higher reattachment rate than PnR; however, PnR was superior to PPV in terms of retinal vessel printing, visual function, vertical metamorphopsia, and photoreceptor integrity.

While PnR patients had a better pre-op and post-op LogMAR, improvement in visual acuity after surgery was higher in PPV patients.

Cataract formation and cataract surgery were significantly higher in the PPV arm, while the occurrence of new retinal tears was more frequent in the PnR group.

Introduction

Optimal management of rhegmatogenous retinal detachment (RRD) remains a controversial issue despite the remarkable advances in ophthalmic surgeries [1]. There are three main surgical procedures for the treatment of RRD: scleral buckling (SB), pars plana vitrectomy (PPV), and pneumatic retinopexy (PnR). SB was the first treatment of RRD to show an acceptable success rate; however, PPV has become the most widely used surgical procedure for treating RRD [2, 3]. Recent advances in ophthalmic and vitreo-retinal instrumentation, including small-gauge surgeries, have made PPV an excellent choice for the treatment of RRD [4, 5].

Dominguez from Spain and Hilton and Grizzard from the United States were the first to describe PnR as an outpatient surgical procedure for the treatment of RRD [6, 7]. PnR use was limited to a superior single break or a group of breaks within 1 clock hour without evidence of proliferative vitreoretinopathy [6]. However, a growing body of literature has reevaluated the indications for PnR use [8,9,10,11,12]. While PnR is considered a less invasive procedure [13, 14], cost analysis of PnR and PPV revealed that PnR is more cost-effective [15]. In addition, PnR can be performed as an outpatient procedure, which obviates difficulties with a prolonged stay in the hospital [16] and is more satisfying for most patients. All of these advantages have made PnR an appealing choice for the treatment of RRD. However, its benefits should be judged alongside its success rate and adverse events.

Currently, there is no consensus on the superiority of PPV over PnR or vice versa. In this study, we conducted a systematic search protocol and performed a meta-analysis to compare the responses of RRD to PPV and PnR. We also assessed the frequency of adverse events in each procedure.

Methods

Search Strategy

We adhered to the recommendations provided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement in this study [17]. The study protocol has been registered in PROSPERO (ID: CRD42021256732).

We performed a computerized search of the MEDLINE (PubMed), EMBASE, Scopus, Web of Science, and Cochrane CENTRAL databases with the following search strategy: “retinopexy OR retinopexies AND Vitrectomy [Mesh] OR vitrectomy OR vitrectomies OR PPV AND retinal tear OR retina tear OR retinal break OR retina break OR retinal perforation OR retina perforation OR retinal detachment OR retina detachment OR Retinal Detachment” in March 2020. The search results were updated in July 2022 with the same keywords. Our search results were not restricted by study language, publication year, or any other criteria. We also hand-checked the reference lists of the included studies to find more relevant articles. The search strategy was checked and optimized with the help of an experienced librarian. We used EndNote X8 for importing and managing the references. The search protocol was performed and the articles were screened by two investigators, AR and SS.

Selection Criteria

Inclusion Criteria

We included both observational and interventional studies with the following criteria:

  1. (a)

    Sufficient data regarding the number of participants with an acceptable response or the mean and standard deviation (SD) of post-operative visual acuity (LogMAR) can be extracted or calculated

  2. (b)

    Sufficient data are presented for both the PnR and the PPV groups

  3. (c)

    Sufficient baseline data regarding country, design, number of eyes treated, mean age, and country are presented

  4. (d)

    Availability of the full texts.

Exclusion Criteria

  1. (a)

    Data regarding the number of participants with an acceptable response or the mean and SD of post-operative visual acuity (LogMAR) for both the PPV and the PnR groups cannot be extracted or calculated

  2. (b)

    Case reports, review articles, editorials, commentaries, and conference abstracts.

Data Extraction

Data regarding the first author’s name, publication year, country of origin of the study, design, number of treated eyes in the PnR and PPV arms, mean age, duration of follow-up, status of the previous treatment (received previous treatment or treatment naive), lens (phakic vs. pseudophakic) and macula (macula-on vs. macula-off) status of the patients, and proportion of participants with severe RRD were extracted. To measure outcomes, the number of patients with an acceptable response (respondents) to PnR or PPV was obtained. In addition, the mean and SD of the pre-operative and post-operative LogMAR and their changes were extracted or calculated. We used the formulas proposed by the Cochrane handbook [18] and other sources [19] to calculate the mean and SD of the changes in outcome variables between before and after the procedure.

Two authors, AR and SS, performed the data extraction independently; disagreements were discussed with another author, SB, and consensus was achieved.

Risk of Bias Assessment

Two authors, AR and SS, independently assessed the risk of bias in the included studies. Retrospective studies were evaluated using Newcastle–Ottawa Scale (NOS) for non-randomized studies [20]. Randomized controlled trials were assessed with version 2 of the Cochrane risk-of-bias tool for randomized trials (RoB 2) [21].

Statistical Analysis

We used STATA version 14.0 software (STATA Corporation, College Station, TX, USA) to perform the meta-analysis in this study. The pooled proportion of respondents, pre-operative LogMAR, and post-operative LogMAR along with their 95% confidence intervals (CIs) were calculated separately for the PnR and PPV groups. The odds ratio (OR) and its 95% CI was applied to compare the response rate in the PnR and PPV groups. The standardized mean difference (SMD) between pre-operative and post-operative LogMAR was used to express changes in visual acuity after the operation in both groups. A p value of < 0.05 was considered statistically significant. The I2 statistic was used to measure heterogeneity between studies [22]. I2 > 50% indicates significant heterogeneity between studies. The DerSimonian and Laird random-effects model was used to perform the meta-analysis. When I2 exceeds 50%, subgroup analysis was used to investigate the possible etiology of the heterogeneity. Sensitivity analysis was performed to detect individual effects of any single study. Finally, we assessed the publication bias using funnel plots and Egger’s test.

Compliance with Ethics Guidelines

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

Results

Characteristics of the Included Studies

After an initial search, 2147 entries were identified. After an evaluation of titles and abstracts, 2095 articles were removed. Out of the 52 remaining papers, 11 were eligible to enter our study; these consisted of 6 retrospective studies, 1 non-randomized trial, and 4 randomized controlled trials (Fig. 1) [2, 23,24,25,26,27,28,29,30,31,32]. Only one of the studies compared the success rate of PnR and PPV in patients who had a previous history of failed PnR. In total, 11,346 patients with a mean age of 74.1 years were included in these studies. Overall, 1577 of the patients underwent PnR, while 4360 of them were operated on using PPV. The remaining included patients received different surgeries, e.g., SB or combined surgeries. In most of the included studies, the reattachment rate was used to define the success rate. Details of the included studies are presented in Table 1. The assessment of risk of bias revealed that the quality of the non-randomized studies was good. Details of the risk of bias assessment are provided in the Supplementary Information.

Fig. 1
figure 1

Flow diagram of the selection process for the included studies

Full size image
Table 1 Characteristics of the studies included in the meta-analysis
Full size table

Qualitative Synthesis

Comparisons of the efficacy (success rate and visual acuity) and safety (adverse events) of PPV and PnR are presented in the next section. Herein, we compare PPV and PnR based on some aspects which were not included in the quantitative analysis.

Retinal vessel printing was displayed in 7% of PnR patients and 44.4% of PPV patients [23]. The direction of retinal displacement was inferior in 94.5% of the PPV patients and 25% of the PnR patients. In the PIVOT study, Hillier et al. demonstrated that visual function questionnaire scores and vertical metamorphopsia scores were superior for PnR [24]. However, PnR patients had more office visits compared to the PPV group (10.8 vs. 9.6). Comparison of the vision-related functioning revealed that patients who underwent PnR had better visual acuity scores for distance activities, mental health, dependency, and peripheral vision at 6 months follow-up [26]. However, no significant difference was observed at 12 months. Muni et al. evaluated photoreceptor integrity using SD-OCT in a subset of patients of the PIVOT study [31]. In 3-mm (foveal) scans, ellipsoid zone (EZ) discontinuity was observed in 24% of PPV and 7% of PnR patients, and external limiting membrane (ELM) discontinuity was found in 20% of PPV and 6% of PnR patients. The superiority of PnR over PPV was also observed in 6-mm (foveal and non-foveal) scans.

Meta-Analysis

Comparison of PnR with PPV revealed that PPV had a higher success (reattachment) rate than PnR (OR = 3.39, 95% CI 2.25–5.11). The superiority of PPV was observed in both treatment-naïve patients (OR = 2.98, 95% CI 2.18–4.07) and those who had previously failed retinopexy (OR = 11.45, 95% CI 2.49–52.73); however, the OR of the success rate of PPV compared to that of PnR was much higher in previously operated patients (Fig. 2). Subgroup analysis based on lens and macula status showed that the advantage of PPV over PnR was more pronounced in studies with fewer phakic eyes and more macula-on patients (Fig. 3).

Fig. 2
figure 2

Comparison of the success rates of the two methods of retinal detachment surgery (pneumatic retinopexy vs. pars plana vitrectomy). The results are presented as the odds ratio (95% CI)

Full size image
Fig. 3
figure 3

Comparison of the success rates of the two methods of retinal detachment surgery (pneumatic retinopexy vs. pars plana vitrectomy) and subgroup analysis based on lens status (left figure) and macula status (right figure). The results are presented as the odds ratio (95% CI)

Full size image

In the next step, we compared the visual acuity (using LogMAR) between the PPV and PnR arms. PnR patients had better pre-op visual acuity (SMD = − 0.58, 95% CI − 1.16 to 0.00) (upper left figure in Fig. 4) and better post-op visual acuity (SMD = − 0.45, 95% CI − 0.60 to − 0.30) (upper right figure in Fig. 4); however, the improvement in visual acuity after surgery was more significant in PPV patients (SMD = 0.49, 95% CI − 0.15 to 1.13) (lower figure in Fig. 4). We also calculated the pooled success rates of PnR and PPV separately. The success rate of PPV for the treatment of retinal detachment was 91% for treatment-naïve and 85% for previously treated patients. The corresponding value for PnR was 69% for treatment-naïve and 33% for previously treated patients (Fig. 5).

Fig. 4
figure 4

Differences in pre-op visual acuity (upper left figure), post-op visual acuity (upper right figure), and the difference between pre-op and post-op visual acuity (lower figure) between pneumatic retinopexy and pars plana vitrectomy. The results are presented as the standardized mean differences (SMDs) and their 95% CIs

Full size image
Fig. 5
figure 5

Subgroup analysis of the success rates of the pneumatic retinopexy (left figure) and pars plana vitrectomy (right figure) arms, based on the history of previous surgery

Full size image

We performed another subgroup analysis to compare the success rates of PnR and PPV based on the study year. The success rate of PnR was significantly higher in studies published after 2015 compared to previous studies (82% vs. 59%). However, PPV showed quite similar success rates in studies published before and after 2015 (Fig. 6).

Fig. 6
figure 6

Subgroup analysis of the success rates of the pneumatic retinopexy (left figure) and pars plana vitrectomy (right figure) arms based on the publication year of the studies

Full size image

Further analysis revealed that patients who underwent PPV were at higher risk of developing retinal displacement as compared to those who underwent PnR (OR = 3.55, 95% CI 2.26–5.59) (Fig. S1 in the Supplementary Information). Finally, we compared the adverse events of these two procedures. Most of the adverse events, including endophthalmitis, cystoid macular edema, epiretinal membrane, hypotony, phthisis, and increased IOP, were not significantly different between PnR and PPV. Cataract formation and cataract surgery were significantly higher in the PPV arm, while the occurrence of new retinal tears was more frequent in the PnR group (Table 2).

Table 2 Prevalence of adverse events of pneumatic retinopexy vs. pars plana vitrectomy
Full size table

Sensitivity Analysis and Publication Bias

Sensitivity analysis revealed that none of the studies affected the pooled OR (Fig. S2 in the Supplementary Information). Publication bias was assessed using a funnel plot and Egger’s test. The funnel plot was asymmetrical; however, the p value of Egger’s test was 0.193, suggestive of a low possibility of publication bias (Fig. S3 in the Supplementary Information).

Discussion

The results of the previous studies which were included in our meta-analysis indicated that PnR was superior to PPV in terms of retinal vessel printing [23], visual function, vertical metamorphopsia [24, 26], and photoreceptor integrity (EZ and ELM) [31]. However, the results of our meta-analysis revealed that PPV showed a higher reattachment rate than PnR in both treatment-naïve patients and those who had previously failed retinopexy. We also found that the higher rate of reattachment in the PPV arm was more pronounced in studies with fewer phakic eyes and more macula-on patients. In terms of the visual acuity, while PnR patients had a better pre-op and post-op LogMAR, improvement in visual acuity after surgery was higher in PPV patients. Success rates of PPV for the treatment of retinal detachment were 91% and 85% for treatment-naïve and previously treated patients, respectively. The corresponding values for PnR were 69% and 33%. It was also found that patients who underwent PPV were at higher risk of developing retinal displacement. Comparison of the safety of PnR and PPV revealed that cataract formation and cataract surgery were significantly higher in the PPV arm, while the occurrence of new retinal tears was more frequent in the PnR group.

We observed that the success rate of PnR varies greatly—from 51% to 82%—while for PPV, the success rate varied in a narrower range, between 79% to 93%. This high heterogeneity of the success rate of PnR can be attributed to the variable pre-operative characteristics of the patients. PnR was indicated for patients with a single break or multiple breaks within 1 clock-hour, providing that the breaks involve the retina superior to eight and four o'clock meridians [34]. Goldman et al. developed expanded patient inclusion criteria for PnR [12, 35]. They concluded that while mild vitreous hemorrhage and lattice degeneration did not decrease the efficacy of PnR, PPV or SB should be considered for patients with visible traction and inferior retinal breaks. Rahat et al. stated that the delay to surgery and PVR (proliferative vitreoretinopathy) are the most important predictors of PnR failure [36]. Interestingly, they found that even grade A PVR was accompanied by poorer PnR outcomes compared to individuals without PVR (42% vs. 97%). Therefore, we suggest that the absence of PVR should remain a solid inclusion criterion for PnR and that the eligibility of patients with grade A PVR is at the surgeon's discretion, based on other pre-operative characteristics. Besides, it seems that the experience of the surgeon [37] and the technique adopted for the surgery can impact the outcomes of PnR. Adding adjunctive techniques such as cryotherapy or laser treatment, the types of gas being used (perfluoropropane, sulfur hexafluoride), the volume of gas being injected, and properly maneuvering the patients could have affected the surgical outcomes [35, 36, 38]. We think that better selection of eligible patients for PnR and improvements in the surgical techniques are the reasons for the observed increasing pattern in the PnR success rate, from 59% in older studies to 82% in more recent studies.

We should be aware of the differences between best practice and real-world outcomes [39]. In a study by Hillier et al., time to surgery was 7.5 h for macula-on and 18.1 h for macula-off cases [24]. These numbers might be much higher in a real-world situation compared to the controlled conditions in a prospective trial, especially in resource-limited areas. Assessment of retrospective studies, although performed in uncontrolled conditions, might provide success rates closer to real practice. We identified two retrospective studies evaluating the success rates of PPV and PnR in patients eligible to enter the PIVOT trial [40, 41]. The primary surgery success rate was near 94% for PPV and 74–79% for PnR. However, we should consider the difficulties accompanying the immediate provision of operation room facilities and experienced staff. Taking the risk of delaying the surgery to prepare the PPV operation conditions, especially in resource-limited areas, is accompanied by poorer PPV outcomes. In these circumstances, PnR could obviate the need for urgent preparation of the operation room conditions.

We have shown that the success rate of primary RRD repair is lower in PnR compared to PPV. Our subgroup analysis based on the publication year revealed that the success rates of PnR and PPV were 82% and 90% in more recent studies. Considering the success rates presented in the previous paragraph and our meta-analysis (PnR success rate: 74–82%, PPV success rate: 88–94%), we assume that the absolute risk reduction of PPV is about 12–14%. Therefore, the number needed to treat to observe one different outcome between PnR and PPV is 7–8. This means that from every 7–8 patients who are treated using PnR instead of PPV, only one of them might require an additional secondary RRD repair. This approach not only deprived the patients of some superior outcomes of PnR (visual function, photoreceptor integrity, retinal displacement), but also the patients may have experienced more complications of PPV. Besides, a cost-utility analysis by Elhusseiny et al. demonstrated that PnR was accompanied by lower total imputed costs, higher estimated lifetime quality-adjusted life years (QALYs), and lower cost per QALY [15], favoring the use of PnR in selected cases.

It was shown in a study that near 53% of the patients attending a tertiary retina center were eligible for PnR according to the PIVOT criteria [40]. However, PnR constituted less than 20% of all RDD repair procedures from 2005 to 2016, with a declining pattern observed in the United States [42]. One justification for treating cases eligible for PnR with PPV is the higher chance of failure in the secondary RRD repair. We showed that the success rate of PnR for secondary RRD repair falls to 33%. However, the success rate of secondary PPV after primary PnR failure (90% to 99%) is approximately equal to the success rate of primary PPV [24, 29]. Similarly, visual acuity after the secondary PPV following failed PnR was comparable to that of primary PPV [29]. Assuming that the success rate was 80% for primary PnR and 90% for secondary PPV, only 2% of the patients will fail to achieve reattachment after the secondary PPV following primary PnR failure. It is worth giving the patients one chance to be treated by PnR, and then, if it fails, PPV can be performed with an acceptable success rate.

The results of our study indicated that the superiority of PPV over PnR is greater in pseudophakic eyes and macula-on patients; however, the results were not significant. In the literature, there is remarkable controversy about the role of lens and macula status in surgical outcomes. Hillier et al. and Rahat et al. stated that neither lens nor macula status affected surgical outcomes [24, 36]. On the other hand, some studies showed that phakia was associated with more favorable outcomes after surgery [40]. The most concordant results with our meta-analysis were observed in a study by Jung et al. indicating that pseudophakia and macula-on detachments were correlated with lower best corrected visual acuity (BCVA) [14]. Retinal detachment in pseudophakic patients is usually diagnosed at more advanced stages. This is mainly because of the delayed diagnosis of small peripheral retinal tears in pseudophakic patients, in whom the opacification of the implanted lens may impair the view of the peripheral retina [25, 38]. Later diagnosis of retinal tears is accompanied by a higher chance of PVR development. As noted previously, high-grade PVR is an indication for using PPV instead of PnR.

In summary, PPV is accompanied by a higher primary reattachment rate and lower retinal tears; however, patients who underwent PnR experienced more favorable visual acuity and visual function. The observed discrepancy in reattachment rate and visual acuity between PPV and PnR can be attributed to the presence of retinal displacement and the low integrity of the attached retina in the PPV arm [32]. As discussed earlier, PnR was superior to PPV in terms of retinal displacement and photoreceptor integrity. Also, PnR is less invasive, has fewer complications, is less expensive, and it is easier to provide the required conditions for PnR as an outpatient procedure. PnR can be used as the primary procedure for RRD repair in selected cases according to the PIVOT trial with favorable outcomes. Still, we believe that some modifications could be applied to the PIVOT criteria. As noted earlier, the roles of some criteria, including lens/macula status and mild PVR, in the outcomes of PnR are still controversial. We recommend that patients with several of these possibly negative predictive factors should be excluded at the surgeon’s discretion. With these modifications, we can expect higher success rates for PnR. Note that, although we have advocated the use of PnR for uncomplicated RDD, PPV is still the treatment of choice for most cases which do not meet the recommended criteria.

Our study is the first meta-analysis which compares the efficacy and safety of PnR and PPV. We performed a comprehensive search strategy and robust statistical analysis. However, our study has certain limitations. First, several of the included studies did not report some of the factors which might affect the success rate, including lens, macula, and PVR status. Thus, we have a low number of studies in some parts of the subgroup analysis. Therefore, we encourage other researchers to conduct more studies to elucidate the role of lens, macula, and mild PVR in PnR outcomes. Secondly, the details of the surgery and inclusion criteria for patient selection were not presented in some of the studies. This precludes an exact comparison between different studies to find some differences between the outcomes. However, the results of our study could be justified according to the previous literature, as discussed earlier. We also tried to present a holistic view of the issue, encompassing not only the clinical issues surrounding the efficacy and safety of the procedures but also considering economic aspects, patient comfortability, and the accessibility of the procedures.

Conclusion

PnR is accompanied by a lower primary success, but it has a more favorable visual function. PnR was also superior to PPV in terms of photoreceptor integrity and absence of retinal displacement. Considering its fewer complications, PnR can be used as the primary procedure for RRD repair in selected cases. However, we propose some modifications to the PIVOT criteria, e.g., the exclusion of cases presenting with several risk factors of poor outcomes (pseudophakia, macula-on RRD, mild PVR), although none of them alone is considered an exclusion criterion.