Primary Retinal Detachment Outcomes

Study: Methodology and Overall

OutcomesdPrimary Retinal Detachment
Outcomes Study Report Number 1
Edwin H. Ryan, MD,1 Daniel P. Joseph, MD,2 Claire M. Ryan, BA,1 Nora J.K. Forbes, MS,1
Yoshihiro Yonekawa, MD,3 Robert A. Mittra, MD,1 D. Wilkin Parke, MD,1 Alex Ringeisen, MD,1
Geoffrey G. Emerson, MD,4 Gaurav K. Shah, MD,5 Kevin J. Blinder, MD,5 Antonio Capone, MD,5
George A. Williams, MD,5 Dean Eliott, MD,3 Omesh P. Gupta, MD,6 Jason Hsu, MD,6 Carl D. Regillo, MD6

Purpose: To detail the methodology for a large multicenter retrospective cohort study (RCS) of primary
rhegmatogenous retinal detachment (RRD) with detailed data collection and to present overall anatomic outcomes.
Design: This study used an RCS method.
Participants: All patients undergoing pars plana vitrectomy (PPV), scleral buckling (SB), and combined PPV/
SB for primary RRD in 2015 from 5 large retina groups were included in the database.
Methods: To ensure validity of the cohort method, a large and detailed database was generated. Double data
entry validation was conducted, and data audits were conducted. Anatomic and visual outcomes for all the cases in
the dataset will be described but not compared, because the cases were not matched. In future studies, comparable
cases of moderate-complexity RRD will be chosen naive to surgeon, surgery, and outcome for subgroup analysis.
Main Outcome Measures: Precision of data entry was confirmed by inter-rater reliability (IRR). Main surgical
outcome for each procedure type was single-surgery anatomic success (SSAS).
Results: Inter-rater reliability showed significant agreement among raters (P < 0.001). Of 2620 patients, 2335
had >90 days of follow-up. Of these, 320 eyes (13.7%) underwent SB, 1200 eyes (51.4%) underwent PPV, and
815 eyes (34.9%) underwent PPV/SB. The SSAS was 84.2% for PPV, 91.2% for SB, and 90.2% for PPV/SB.
Conclusions: We compiled a large, accurately documented database of primary RRD cases repaired by
PPV, SB, and PPV/SB from which cohort studies of moderately complex RRDs can be carried out. All 3 ap-
proaches had a high SSAS rate. Procedures chosen and their outcomes are described, but the broad case mix
makes comparisons not possible until future cohort studies are completed. Ophthalmology Retina 2020;4:814-
822 ª 2020 by the American Academy of Ophthalmology

Supplemental material available at www.ophthalmologyretina.org.

Scleral buckling (SB) was developed in the 1950s by Cus- PPV because of enhanced patient comfort, reduced surgi-
todis1 and Schepens et al2 and was essentially the only way cal time, and reduced postoperative discomfort relative to
to repair a rhegmatogenous retinal detachment (RRD) until PPV/SB or SB with reports of similar long-term
the introduction of pars plana vitrectomy (PPV) by outcomes.9,10
Machemer et al in 1971.3 Since that time, PPV has played A significant percentage of surgeons, however, continue
an increasing role in RRD repair, from being an adjunct to to use PPV/SB for repair of RRD,11,12 and primary SB still
SB in complex cases in the 1980s to being the primary has support, especially in phakic and younger patients.13,14
and only technique used in many cases more recently, Since evidence that the trend toward PPV alone improves
most often using small-gauge vitrectomy techniques.4,5 anatomic or visual outcomes is scant, the need for a more
The trend away from SB to PPV is well documented and comprehensive investigation is apparent. The last major
has increased with each generation of surgeons as shown study of surgical approaches for RRD repair was the
by Medicare data.6 The same trend has occurred in Scleral Buckling Versus Primary Vitrectomy in
Europe, with a large shift toward vitrectomy from 1999 to Rhegmatogenous Retinal Detachment (SPR), a large,
2006.7 A 2018 survey showed that a majority of surgeons prospective, randomized controlled trial (RCT) performed
placed an SB in less than 20% of RRD cases.8 Increasing in Europe with cases predominantly from the year 2000,
numbers of cases are done using primary small-gauge predating small-gauge PPV.15 The SPR study and a

814  2020 by the American Academy of Ophthalmology https://doi.org/10.1016/j.oret.2020.02.014

Published by Elsevier Inc. ISSN 2468-6530/20
 Ryan et al 
PRO Study Report Number 1, Methodology

second European prospective study, the Retina 1 project,7 postendophthalmitis, and others). Given the size and number of
showed better outcomes with SB in phakic patients and participating centers, 2500 to 3000 well-documented cases were
with PPV in pseudophakic patients. We seek to determine anticipated.
if these guidelines still hold true in the era of small-gauge
PPV and other modern advances in instrumentation and Data Collection
visualization.
Individuals trained by the study coordinator performed the data
Given the complexity and variability of anatomic pre-
collection and entry at each study center. A comprehensive data
sentation of primary RRD and the difficulties involved in entry manual was created by the study coordinator and used by
recruitment of both patients and surgeons for a randomized each of the data entry personnel at each site. The lead investigator
prospective trial, we believe that a comprehensive, high- at each site supervised the data collection. Further adjudication, if
quality, compulsively documented retrospective analysis of needed, was carried out with the coordinator and principal inves-
existing data would be the optimal method to study the tigator. Ten percent of cases underwent double entry as a method to
question of efficacy of modern methods for repair of pri- audit consistency of data entry. This was done by having 2
mary RRD. To this end, we proposed the compilation of a different data entry investigators enter 10% of charts at each site.
large database of primary retinal detachment surgeries in a Sources of patient data included the electronic medical record,
single year from several large retinal centers in the United operative notes, and communications with referring doctors. In
some cases, follow-up letters were requested from referring phy-
States from which we could perform cohort studies. From
sicians. All cases were included in the database, but minimum
these, we seek to determine if anatomic or visual outcomes follow-up to be included in outcomes analysis was 91 days. Data
are improved or harmed by the use of SB alone or in entry was halted on December 31, 2017, and the last eligible visit
combination with PPV for repair of RRD. In addition, we date was January 31, 2017.
believe the cohort study method described in this article Data entry was divided into 3 sections: preoperative, operative,
could serve as a model for future studies of RRD. and postoperative. A total of 256 variables were collected. This
In this article, we describe the methodology of the being a retrospective study, there were data points that were absent
multicenter study, including the demographics and pre- and could not be retrieved. Some data points were considered
senting anatomic characteristics of the cohort. We also critical (e.g., date of surgery and procedure), and if missing, the
present both the overall surgical outcomes and those of case was excluded. Cases containing this information but missing
other key data (e.g., adequate retinal detachment anatomy
various subgroups within the cohort. Detailed investigation
description, >90-day follow-up) were retained in the database, but
of specific variables and hypotheses will be described in not included in outcomes analysis. (A copy of the data entry
forthcoming publications. manual/data dictionary is available as a separate Excel [Microsoft
Corp, Redmond, WA] file in the Appendix [available at available
Methods at www.ophthalmologyretina.org]).
Retinal breaks were characterized and documented by size,
Surgeons and Study Centers type, and number on the basis of drawings or a clear description by
the examining or operating physician. Location of RD was docu-
Data were obtained from 61 surgeons at 5 participating centers, mented following a method created by the British and Eire Asso-
ranging from fully academic with fellows (Boston) to partially ciation of Vitreoretinal Surgeons (BEAVRS).16 A value of 0 to 3
academic with fellows (St. Louis, Detroit, and Philadelphia) to clock hours detachment (or unknown) was entered for each
fully private practice (Minneapolis). All centers reported managing quadrant. Values were based on drawing (preferred) or a clear
a large annual volume of primary retinal detachment cases. text description of the location (e.g., fluid from 12 to 3). If
neither of those was available but there was a quadrant
Patients mentioned (e.g., fluid superotemporally), then we instructed
people to enter a value of 3 clock hours for that quadrant. For
Institutional Review Board (IRB) approval and an informed con- both break findings and RD size and location, if intraoperative
sent waiver were obtained for this study using Salus IRB and the findings differed from preoperative findings, intraoperative values
IRBs of the respective institutions. Data were collected in accor- were considered more accurate and were used. If follow-up letter
dance with Health Insurance Portability and Accountability Act of findings differed from earlier postoperative findings, then the
1996 guidelines, and the study conformed to the tenets of the follow-up letter values were used.
Declaration of Helsinki. Patients who underwent surgery for RRD Whether the macula was attached or detached was stated
during the 2015 calendar year from 6 retina practices (5 centers) directly by the retina surgeon in the chart or inferred from chart
were identified by searching billing codes for office and operating information (i.e., drawings or OCT images). Macular status was
room procedures for that year. Patient data were de-identified and divided into 3 groups: “Macula off” was defined as the fovea being
stored securely on REDCap database. All patients were aged 5 fully separated, “macula on” was defined as the fovea being fully
years or older and underwent RRD repair using PPV, PPV/SB, or attached, and “fovea-split or threatened” comprising the others.
SB. Because of the relatively low numbers of RRD cases treated Cataracts were graded as 0 to 4þ, with <2þ considered not
with pneumatic retinopexy, laser alone, or cryopexy alone, these visually significant. Vitreous hemorrhage (VH) was graded as
cases were excluded. All nonprimary RRD surgeries were none, mild, moderate, or dense. Proliferative vitreoretinopathy was
excluded. In addition, the following RRD operating room cases graded as to type (A, B, or C) based on the examiner’s chart notes.
were excluded: RRD associated with penetrating injury or ruptured
globe; RRD causally associated with proliferative diabetic reti- Outcome Measures and Definitions
nopathy, sickle cell, retinopathy of prematurity, or other similar
conditions (i.e., familial exudative vitreoretinopathy or ischemic The primary clinical outcome measure in this study is anatomically
vein occlusion); and RRD associated with inflammation (i.e., acute successful single surgical retinal reattachment. Retinal reattach-
retinal necrosis, cytomegalovirus retinitis, pars planitis, ment is defined as attachment of the retina posteriorly with no

815
 Ophthalmology Retina Volume 4, Number 8, August 2020

Table 1. Overall Demographics perform an SB), comparison tests for outcomes across surgical
groups were not appropriate and thus not conducted. Rather, these
Characteristics No. or Mean % or SD or IQR data will be used in future studies on surgical outcome within
comparable subpopulations. Statistical comparisons of PPV, PPV/
Total cases 2620 100%
SB, and SB in the phakic subsets and PPV and PPV/SB in the
Male 1663 63.9%
Right eye 1323 51.1%
pseudophakic subsets will be carried out only on the cohorts of
Mean age (yrs) 59.8 13.2 moderately complex RRDs, which could be repaired with any of
Lens status the 3 procedures. Analysis of visual and anatomic outcomes will be
Phakic 1414 54.6% made using the proper associated tests for comparisons and con-
Pseudophakic 1158 44.7% trolling for covariates when appropriate.
Aphakic 16 0.6% Inter-rater reliability (IRR) is calculated for 15 key variables to
Surgery type determine consistency of doubly entered data. Simple percent
SB 353 13.5% agreement between the 2 raters is calculated for all 15 variables.
PPV 1355 51.7% Additionally, Cohen’s kappa statistic and interclass correlation
PPV/SB 912 34.8% coefficient and corresponding P values are calculated for categoric
Mean follow-up (days) 359.6 184.6 and quantitative variables, respectively. All IRR calculations are
Macular status preoperatively made using the statistical package “irr.”
Off 1370 52.5%
On 988 37.8%
Split/threatened 253 9.7% Results

IQR ¼ interquartile range; PPV ¼ pars plana vitrectomy; SB ¼ scleral Demographic Characteristics
buckle; SD ¼ standard deviation.
IQR is interquartile range, the distance between the 25th percentile and A total of 2620 primary RRD cases met inclusion criteria with
the 75th percentile. Note some data are missing lens and macular status. adequate documentation, and their characteristics are listed in
Table 1. In the group with greater than 90 days of follow-up, 320 of
2335 eyes (13.7%) underwent SB, 1200 of 2335 eyes (51.4%)
tamponade present and no subretinal fluid that could spread, underwent PPV, and 815 of 2335 eyes (34.9%) underwent
similar to that of BEAVRS.16 This includes those eyes noted to PPV/SB. Also, 315 of 320 eyes treated with SB were phakic.
have small traction detachments posterior to a circumferential or
encircling buckle. It also includes eyes noted to have anterior
Anatomic Characteristics
fluid walled off by 360-degree retinopexy. If the surgical eye
met these criteria and had greater than 90 days follow-up without a Retinal detachment extent distribution was documented in 2574 of
return trip to the operating room, it was considered to have had a 2620 eyes (98.2%). Of these eyes, the RD was limited to 1 to 3
single-surgery anatomic success (SSAS). Eyes with a silicone oil clock hours in 620 (24.1%), 4 to 6 clock hours in 1358 (52.7%), 7
(SO) tamponade at the end of follow-up were considered anatomic
to 9 clock hours in 382 (14.8%), and 10 to 12 clock hours in 204
failures, as per BEAVRS (e-mail with Dr. David Yorston,
November 26, 2019). We recognize that some eyes with primary (7.9%). The location of the detachment was documented in 2563 of
SO tamponade may achieve anatomic success but would be 2620 eyes (97.8%). Those with more clock hours represented
considered SSAS failures by this definition. Rates of both SSAS above the horizontal meridian than below were considered “supe-
and final attachment were tabulated. rior” and vice versa. Location of RD was mostly superior in 50.6%
The secondary outcome measure is final visual outcome. (1303), mostly inferior in 32.8% (843), or equally inferior and
Final Snellen best-corrected or pinhole visual acuity outcomes superior in 16.2% (417) of eyes.
were recorded and converted to logarithm of the minimum Any evidence of proliferative vitreoretinopathy was seen in 250
angle of resolution units for statistical analyses. In the phakic of 2620 cases (9.5%), and 9 underwent SB, 77 underwent PPV,
subset analysis, to control for the effects of PPV on post- and 164 underwent PPV/SB. The proliferative vitreoretinopathy
vitrectomy cataract and therefore final visual acuity, in patients
was graded as grade A in 27 cases, grade B in 50 cases, and grade
undergoing PPV or PPV/SB, only patients with mild or no
cataract, or those postecataract surgery were included in visual C in 110 cases, with 66 not specified. A macular hole was noted
outcomes. preoperatively in 26 of 2620 cases (1.0%). An epiretinal membrane
was noted in 124 cases (4.7%) preoperatively.
Statistical Analysis Presence or absence of VH was documented in 2580 of 2620
cases (98.4%). Vitreous hemorrhage was documented as mild in
Data are analyzed using R Statistical Software.17 A P value of 169 cases, moderate in 50, dense (red reflex only) in 43, and
< 0.05 is considered significant, and all tests are 2-tailed.
“degree not specified” in 152 cases. In patients with documented
Preoperative, intraoperative, and postoperative data are sum-
marized using descriptive statistics. Categoric variables are VH, SB was performed in 23 (only 1 having moderate or greater
described as count and percentage of subjects within groups, and VH), PPV in 251, and PPV/SB in 140.
continuous variables are given as mean  standard deviation.
Summaries are provided by age and site, in addition to other Outcomes
subpopulations of interest.
Given the wide range of anatomic RRD presentations in the A total of 2335 (89.1%) of the 2620 cases had >90 days of follow-
overall group, with many cases for which nearly all surgeons up and were considered eligible for anatomic and visual outcomes
would choose the same procedure (e.g., RRD, dense VH, multiple analysis. In the group undergoing primary SB, 292 of 320 cases
tears; all surgeons would likely perform PPV or PPV/SB; or in- (91.2%) had SSAS, with 315 of 320 (98.4%) having final success.
ferotemporal dialysis in a young patient; nearly all surgeons would For PPV, 1011 of 1200 cases (84.2%) had SSAS, and 1140 0f 1200

816
 Ryan et al 
PRO Study Report Number 1, Methodology

Table 2. Basic Retinal Detachment Demographics by Site

Site N Age (SD) Follow-up (SD) Right Eye/Left Eye Male/Female P/IOL/A SB/PPV/PPV-SB
1 574 61 (14) 386 (165) 264/278 338/219 285/252/5 (50/44/2) 79/343/152 (14/60/26)
2 462 61 (12) 363 (177) 235/227 293/169 234/227/1 (51/49/0) 83/266/113 (18/58/24)
3 303 60 (13) 341 (157) 135/168 192/111 166/135/2 (55/44/1) 28/144/131 (9/48/43)
4 154 60 (13) 354 (176) 77/77 104/50 84/66/4 (54/43/2) 6/55/93 (4/36/60)
5 284 55 (15)* 299 (186) 145/138 191/93 181/100/3 (64/35/1) 35/213/36 (12/75/13)
6 843 60 (13) 240 (189) 407/435 545/298 464/378/1 (55/45/0) 122/334/387 (14/40/46)

A ¼ aphakic; IOL ¼ intraocular lens; P ¼ phakic; PPV ¼ pars plana vitrectomy; SB ¼ scleral buckle; SD ¼ standard deviation.
N is the number of eyes from each site. Age and follow-up (days follow-up) are given as mean with standard deviation. Eye, gender lens status, and procedure
type are given in raw numbers. Lens status is defined as phakic (P), pseudophakic (IOL), aphakic (A). Procedure type is defined as SB, PPV, and PPV-SB.
Lens status and procedure data in parentheses are percentages rounded to the nearest whole integer. Numbers do not always add up to same totals because of
some missing data.
*P < 0.05.

(95%) had final success. For PPV/SB, 735 of 815 cases had SSAS Site and Age Analysis
(90.2%), and 768 of 815 (94.2%) had final success. For all 2335
cases, the overall SSAS was 87.3% and final success rate was The basic demographics of the RD population by site and the
95.2%. Silicone oil tamponade was used in the primary surgery in relative number of primary procedures used to repair RD at each
93 of 2335 cases, 14 using 5000 cS SO, and 79 using 1000 cS SO. site are shown in Table 2. The percentage of primary buckles
A total of 58 of 93 patients had no SO at final visit, and 52 patients varied between the sites and ranged from 4% to 18%. The
had retinal attachment. Silicone oil (used in primary or secondary percentage of PPV performed ranged from 36% to 75%, and the
surgery) was present at the last visit in 4 of 320 SB cases (1.3%), PPV/SB ranged from 13% to 60% between sites. The mean age
54 of 1200 PPV cases (4.5%), and 45 of 815 PPV/SB cases (5.5%). for the total cohort was 59 (13) years and was fairly similar
Dependence on SO was considered an anatomic failure. Average among the sites, except for 1 with a significantly lower age than
final visual acuities were 20/35 for SB (20/30 macula-on, 20/40 the other sites. This site had a higher percentage of phakic
macula-off), 20/53 for PPV (20/38 macula-on, 20/74 macula-off), patients compared with the other groups. A total of 61 surgeons’
and 20/61 for PPV/SB (20/37 macula-on, 20/86 macula-off). cases were included in this study, with a range of 2 to 192 cases
Note that the cases were not matched for preoperative character- per surgeon. Twelve surgeons contributed 20 or fewer cases, 40
istics and therefore cannot be used for comparison of visual had 21 to 60 cases, 5 had 61 to 100 cases, and 4 performed
outcomes. greater than 101 cases. Of the 57 surgeons who performed 10 or
more cases, 3 performed no SB cases, and 3 performed no PPV/
SB cases. Every surgeon performed at least 1 SB alone or in
Surgical Characteristics: Pars Plana Vitrectomy combination with PPV. Diplopia or need for prisms was noted at
Gauge Distribution final follow-up in 6 patients, 2 after PPV and 4 after PPV/SB.
Scleral buckle removal was performed in 4 patients, 3 after PPV/
Of 2268 eyes that underwent PPV or PPV/SB, 1463 (64.5%) were SB and 1 post-SB. Age distribution (decile) and age of patient
23-gauge, 677 (29.9%) were 25-gauge, 108 (4.8%) were 20-gauge, versus procedure chosen are detailed in Figure 1. The age range
8 (0.4%) were 27-gauge, and 11 were not documented. Outcomes was 5 to 91 years, and the majority of cases (1701/2620) were in
analysis was performed only on the 2023 cases with >90 days the 51 to 70 deciles. Scleral buckle was favored in the youngest
follow-up. Of the 23-gauge cases, 1330 had >90 days follow-up, patients, and PPV was favored in the eldest.
and the overall SSAS rate in this group was 86.2% (1146/1330)
for all eyes undergoing PPV or PPV/SB. Inter-rater Reliability
For the 761 eyes undergoing 23-gauge PPV alone, SSAS
was achieved in 632 (83.0%), whereas 514 of 569 eyes (90.3%) Reliability tests show a high level of agreement between data entry
that underwent 23-gauge PPV/SB achieved SSAS. A total of personnel. Average overall percent agreement was 96.3%. The
577 eyes undergoing 25-gauge PPV or PPV/SB had adequate average IRR, Cohen’s Kappa and interclass correlation coefficient,
follow-up, and combined SSAS was 502 of 577 (87.0%). Of was 0.98 (Table 3). The IRR was statistically significant for all 14
these, 407 underwent PPV alone and 170 underwent PPV/SB. individual key variables (P < 0.001).
The SSAS rates for PPV and PPV/SB were 351 of 407 (86.2%)
and 151 of 170 (88.8%), respectively. Of the 108 cases of Discussion
20-gauge PPV, 90 had >90 days follow-up, and 81 of 90
(90.0%) had SSAS. Of 20-gauge PPV/SB cases, 64 of 69 Intent
(92.8%) had SSAS, whereas 17 of 21 eyes (81.0%) repaired
with PPV had SSAS. Only 8 cases using 27-gauge had adequate The intent of this study was to examine the results of a large
follow-up, and 7 of these underwent PPV with 1 undergoing dataset of well-documented primary RRD cases managed
PPV/SB. The 27-gauge PPV/SB failed, whereas all of the PPV surgically by vitreoretinal surgeons skilled in SB, vitrec-
cases achieved SSAS. tomy, and combined vitrectomy and SB. The results show

817
 Ophthalmology Retina Volume 4, Number 8, August 2020

Advantages of Cohort Method

The present study provides outcome analysis of both anatomic
and functional outcomes after the repair of RD in the era of
small-gauge vitrectomy and wide-angle viewing systems.
Randomized controlled trials provide excellent level I
evidence in clinical research but can be difficult to conduct
for surgical investigations.18,19 A properly powered and
controlled RCT is nearly impossible to manage in a clinical
disorder of this complexity and variability, and because an
RCT has some inherent limitations, a large retrospective
cohort study (RCS) is an appropriate option to examine
outcomes of RRD repair.20 Observational studies are often
criticized as weaker for having results that are influenced
by confounding factors. However, comparable results
between observational studies and RCTs would be
anticipated with proper study planning and design.21,22
Other methods of testing for data accuracy, such as IRR,
can validate the quality of the data.23 To add confidence to
the validity of the results of the present study, IRR was
calculated for several variables to determine consistency of
data entry. Our results show that there is a high degree of
Figure 1. The bar graph shows the number of cases in each decile and agreement and therefore reliability of observations.
primary surgical procedure chosen. PPV ¼ pars plana vitrectomy; SB ¼ Another challenge in a surgical trial comparing differing
scleral buckling. approaches with similar anatomic problems is the setting
where a surgeon is skilled in one method and relatively un-
that a high degree of success can be obtained from all of the skilled in another. Forcing a surgeon to perform a procedure
surgical approaches. Moreover, despite the trend away from that he or she is not comfortable with or has a bias against
using SB, the results suggest that there is indeed a role for may skew the results. An observational study such as this one
SB in the modern era. examines the results of what the surgeons chose to do for the
From this dataset, we are able to select cases of patient. To mitigate this risk of wide variance in surgical
moderate-complexity primary RRD in both phakic and proficiency, particularly in light of decreasing use of SB and
pseudophakic groups for comparison of outcomes based on potentially lost skills in this technique, we invited participa-
surgical approach used, and these outcomes will be dis- tion from centers that are known to perform SB regularly.
cussed separately in companion articles. This article pre- Using an RCS method to study treatment outcomes has
sents overall anatomic and visual outcomes by procedure in its limitations but also has some advantages. For one, it does
the dataset from which the comparative analyses will be not have the same administrative time and cost consider-
performed. ations that an RCT has. With a disease as complex as RRD,
it is difficult to get or expect the participating surgeons to
adhere to prescribed treatments of buckle or vitrectomy.
Table 3. Inter-rater Reliability There are many reasons why a surgeon might deviate from
Percent Kappa
study protocol, including personal preference. For example,
Variable Agreement or ICC P Value the surgeon who is comfortable with vitrectomy but not SB
(or vice versa) will be forced at times to perform a surgical
Site 100 1.00 <0.0001
technique for which he or she is not as skilled and thus may
Preoperative date 97 0.99 <0.0001
Patient sex 98 0.96 <0.0001 jeopardize a patient’s outcome. (The patient’s own dispo-
Patient age 81 1.00 <0.0001 sition may also influence the technique best used to repair
Eye (right/left) 100 0.99 <0.0001 his or her RRD. For example, thin sclera or preexisting
Preoperative acuity 96 0.96 <0.0001 travel plans may alter a surgeon’s decision to recruit a pa-
Preoperative lens status 100 1.00 <0.0001 tient or perform a randomized surgery.) Surgeons may elect
Surgery date 95 0.99 <0.0001 to not enroll patients who meet study criteria but have an
Surgery type 100 0.99 <0.0001
Postoperative date 93 0.98 <0.0001
RRD anatomy that the surgeon thinks should be treated only
Postoperative acuity 94 0.99 <0.0001 with 1 of the 3 surgical methods. This occurred with some
Retina attached 98 0.93 <0.0001 surgeons during the SPR trial (e-mail communication with
Postoperative OR procedure 99 0.96 <0.0001 Heinrich Heimann, February 12, 2019). These and other
Postoperative OR procedure date 97 0.91 <0.0001 considerations make the practicality of recruiting surgeons
and patients to participate in a prospective surgical RCT for
ICC ¼ interclass correlation coefficient; OR ¼ operating room. RRD challenging. To adequately power such a study would
Inter-rater reliability is measured by Cohen’s Kappa for categoric data and require recruitment periods that are too lengthy and too
ICC for continuous data. Values >0.90 are considered excellent. costly to perform.

818
 Ryan et al 
PRO Study Report Number 1, Methodology

In this RCS, surgeons chose the procedure they thought cases of RRD treated by 176 surgeons that there was a
would work best for a specific case. We have clearly defined higher anatomic failure rate in cases treated with PPV or
boundaries in time (1 calendar year: 2015) and a large cohort PPV/SB than SB alone (1.2% vs. 0.5%, P ¼ 0.04).26 Azad
to work with. Because the study period is 1 year, it is unlikely et al27 also performed a small RCT in 2007 that showed in
that surgeons significantly varied their surgical techniques in phakic eyes with typical RRD that PPV and SB had
that short period. This is particularly important in a field such comparable anatomic outcomes, and PPV led to greater
as retina in which technological innovation often leads to incidence of cataract and delayed visual rehabilitation. The
changes of surgical technique. Because RRDs can be grouped intent of this study is to provide additional data to these
according to similar pathology, patient demographics, and investigations using a large retrospective cohort with
other ocular baseline characteristics, they lend themselves detailed data collection.
well to subgroup analysis. Part of this is because there are a The literature for treatment of primary RRD is extensive
limited number of surgical techniques to repair RRD, and and gives conflicting recommendations for its best treat-
most if not all surgeons have knowledge of how to perform ment. Studies with conflicting outcomes supporting the use
these techniques. Although we cannot control biases that may of primary PPV, small-gauge PPV, PPV/SB, and SB as
lead one surgeon to choose one technique over another, we the most effective treatment for primary RRD can be
can look at a large number of outcomes for a given technique found.10-15,26-34 Most of these studies were small retro-
and compare them with other techniques for similar de- spective reviews, and larger studies are likely to have more
tachments. We think that this study represents the most valid findings. The SPR study found SB to have similar
comprehensive examination of surgical outcomes in patients anatomic outcomes to PPV in phakic patients but with better
undergoing modern small-gauge vitrectomy or SB repair of visual outcomes and found similar visual outcomes but
RRD in the United States today. better anatomic outcomes in pseudophakic patients.15
Because all the vitrectomies performed in the SPR study
Review of Current Literature predated the introduction of small-gauge vitrectomy, it is
tempting to speculate whether the SPR findings would be
The SPR study was a large multicenter prospective RCT replicated in the current era of RRD repair. In addition to
carried out in Europe comparing SB with PPV in 681 smaller vitrectomy gauge, the viewing systems have
recruited participants, with cases performed from 1998 to improved since the SPR study. Wider-angle viewing sys-
2003 (before the introduction of small-gauge PPV and wide- tems and better preoperative diagnostic techniques such as
angle viewing systems), and found SB to have better visual OCT may also independently influence surgeon choice of
outcomes for phakic RD and PPV to have better anatomic one technique over another. These relatively recent in-
outcomes for pseudophakic RRD.15 The SPR studied novations would be expected to lead to improved outcomes
“moderate complexity” RRDs with the understanding that in the older studies had the same technology been available
simple cases with a single break and limited extent would and widely used when those studies were performed.
be repaired with SB or pneumatic retinopexy. This A number of recent studies have suggested that PPV
decision reflects standard practice at that time, but because (including small-gauge PPV) may be preferable to PPV/SB for
PPV is currently being performed on relatively simple repair of primary RRD.9,30,31 These studies cite comparable
RRDs routinely, the results of the SPR study do not anatomic outcomes and a higher complication rate with PPV/
necessarily apply to current practices because cases that SB. Other studies cite declining use of PPV/SB.4,33 We
were excluded in the SPR study design were included in found that the PRO cohort of surgeons continues to use PPV/
the present study. In addition, the SPR trial randomized SB at a fairly high rate, and this use is supported by greater
patients to SB or PPV, but left the option of a SASS than was achieved by PPV alone.
supplemental SB at the surgeon’s discretion in the PPV
cases.24 Subgroup analysis indicated that the addition of a Outcomes for Group as a Whole
buckle led to more favorable outcomes in the
pseudophakic patients but not the phakic subset.15 In the One of the goals of this observational study was to docu-
present study, PPV/SB was considered as a separate ment the procedure choices surgeons are making for surgical
procedure to be compared with both SB and PPV. The repair of primary RRD. We did not include cases treated in
subgroup analysis of phakic versus pseudophakic eyes is the office using laser or cryotherapy, or those treated with
provided in the companion articles, but in the group as a pneumatic retinopexy. Of the 2620 operating room cases
whole, PPV/SB had a success rate of 90%. with adequate documentation, 353 (13.5%) underwent pri-
Despite the trend away from SB alone for RRD, several mary SB, 1355 (51.7%) underwent primary PPV, and 912
studies have shown that there is a higher risk for severe (34.8%) underwent primary PPV/SB. This is consistent with
vision loss in eyes that underwent PPV for RRD compared recent trends toward use of PPV alone and away from SB
with SB. The Retina 1 project reported patients undergoing for management of primary RRD.4 The majority of the
PPV had a 25% rate of final vision <20/100 compared with primary PPV cases were performed using small-gauge
4% if no PPV.7 The UK National Ophthalmology Database PPV. In our study, 40.2% (912 of 2267) of cases treated
is an observational survey of RRD repair that showed SB to with a vitrectomy received a supplemental SB. This would
have a lower complication rate (3.6%) than PPV (6.1%) or suggest the use of supplemental SB is not declining, at least
PPV/SB (8.8%).25 Additionally, the European Vitreo- in this cohort of surgeons. A second goal of this study was
Retinal Society showed in self-reported surveys from 7678 to examine PPV/SB as a separate entity to compare

819
 Ophthalmology Retina Volume 4, Number 8, August 2020

outcomes with both SB and PPV. In the SPR study, the use 2. Schepens CL, Okamura ID, Brockhurst RJ. The scleral
of an additional SB was at the surgeons’ discretion and not buckling procedures. I. Surgical techniques and management.
part of the randomization. Choice of additional SB (i.e., AMA Arch Ophthalmol. 1957;58:797e811.
PPV/SB) in this study was also at the surgeons’ discretion, 3. Machemer R, Buettner H, Norton EW, et al. Vitrectomy: a pars
but the comparison of similar cases selected without plana approach. Trans Am Acad Ophthalmol Otolaryngol.
1971;75:813e820.
knowledge of surgeon, surgery, and outcome could mitigate 4. Chong DY, Fuller DG. The declining use of scleral buckling
this confounding bias. These outcomes will be compared in with vitrectomy for primary retinal detachments. Arch Oph-
separate sub-study articles. thalmol. 2010;128:1206e1207.
There was a strong preference for the use of SB in pa- 5. Orlin A, Hewing NJ, Nissen M, et al. Pars plana vitrectomy
tients aged less than 40 years and a stronger preference for compared with pars plana vitrectomy combined with scleral
the use of PPV in patients aged more than 60 years in the buckle in the primary management of noncomplex rhegma-
present study cohort. The highest use of PPV/SB was in the togenous retinal detachment. Retina. 2014;34:1069e1075.
cohort aged 40 to 60 years. This distribution of cases in 6. McLaughlin MD, Hwang JC. Trends in vitreoretinal proced-
young patients is not surprising, given that numerous pub- ures for Medicare beneficiaries, 2000 to 2014. Ophthalmology.
lications support SB in young patients.10,35 That the trend in 2017;124:667e673.
7. de la Rua ER, Pastor JC, Fernandez I, et al. Non-complicated
general is away from primary SB is supported by the present retinal detachment management: variations in 4 years. Retina 1
study with only 13.5% (353 of 2620) of patients undergoing project; report 1. Br J Ophthalmol. 2008;92:523e525.
primary SB. This finding is particularly noted in 8. Stone TW, ed. ASRS 2018 Preferences and Trends Member-
pseudophakic patients, in whom only 6 of 1100 cases ship Survey. Chicago, IL: American Society of Retina Spe-
were treated with primary SB. Supplemental SB was used cialists; 2018.
in 40% of patients who underwent PPV, so the trend 9. Von Fricken MA, Kunjukunju N, Weber C, et al. 25-Gauge
toward use of PPV and away from use of PPV/SB was sutureless vitrectomy versus 20-gauge vitrectomy for the
not confirmed in this cohort of patients. repair of primary rhegmatogenous retinal detachment. Retina.
2009;29:444e450.
10. Mohamed YH, Ono K, Kinoshita H, et al. Success rates of
Study Limitations vitrectomy in treatment of rhegmatogenous retinal detachment.
This study has the usual limitations of a retrospective J Ophthalmol. 2016;2016:2193518.
11. Storey P, Alshareef R, Khuthaila M, et al. Pars plana vitrectomy
analysis. Data were not always easily found, record keeping
and scleral buckle versus pars plana vitrectomy alone for pa-
was imperfect, and retinal drawings were not always well tients with rhegmatogenous retinal detachment at high risk for
done. There was an effective system for adjudicating data proliferative vitreoretinopathy. Retina. 2014;34:1945e1951.
uncertainties, however, and we think this dataset will allow 12. Chbat E, Morel C, Conrath J, et al. PPV and SB versus PPV
for cohort studies of the different surgical procedures. There alone for patients with retinal detachment. EC Ophthalmology.
was a range of surgeon bias in procedure choice, as well as 2018;9:425e429.
(likely) variability in proficiency with each procedure. 13. Schaal S, Sherman MP, Barr CC, et al. Primary retinal
Conclusions about the superiority of one procedure over the detachment repair: comparison of 1-year outcomes of four
other cannot be made from the data presented in this article surgical techniques. Retina. 2011;31:1500e1504.
but are anticipated with the matched cases to be considered 14. Wong CW, Yeo IY, Loh BK, et al. Scleral buckling versus
vitrectomy in the management of macula-off primary rheg-
in the cohort studies.
matogenous retinal detachment: a comparison of visual out-
comes. Retina. 2015;35:2552e2557.
15. Heimann H, Bartz-Schmidt KU, Bornfeld N, et al. Scleral
Conclusions buckling versus primary vitrectomy in rhegmatogenous retinal
detachment: a prospective randomized multicenter clinical
Treatment of RRD has evolved away from SB and toward study. Ophthalmology. 2007;114:2142e2154.
small-gauge vitrectomy techniques, and most evidence used 16. Aylward GW, Laidlaw A, Patton N, et al. Royal College of
to defend this trend is from small retrospective studies. Ophthalmologists Retinal Detachment Data Set. V0.97. 2011.
Larger studies such as the SPR still found a role for SB, and 17. R: A Language and Environment for Statistical Computing
we show with this cohort study that outcome results support [computer program]. Vienna, Austria: R Foundation for Sta-
the continued role for SB in the repair of RRD. This article tistical Computing; 2018.
presents an overview of the methodology and design of this 18. Burns PB, Rohrich RJ, Chung KC. The levels of evidence and
large cohort study of primary RRD with overall primary and their role in evidence-based medicine. Plast Reconstr Surg.
2011;128:305e310.
secondary outcomes, whereas the companion articles look at 19. Chung KC, Burns PB. A guide to planning and executing a
the specific subgroups of moderately complex phakic and surgical randomized controlled trial. J Hand Surg Am.
pseudophakic detachments, the techniques used to repair 2008;33:407e412.
them, and the anatomic and functional outcomes. 20. Song JW, Chung KC. Observational studies: cohort and case-
control studies. Plast Reconstr Surg. 2010;126:2234e2242.
References 21. Benson K, Hartz AJ. A comparison of observational studies
and randomized, controlled trials. N Engl J Med. 2000;342:
1. Custodis E. [Treatment of retinal detachment by circumscribed 1878e1886.
diathermal coagulation and by scleral depression in the area of 22. Concato J, Shah N, Horwitz RI. Randomized, controlled trials,
tear caused by imbedding of a plastic implant]. Klin Monbl observational studies, and the hierarchy of research designs.
Augenheilkd Augenarztl Fortbildung. 1956;129:476e495. N Engl J Med. 2000;342:1887e1892.

820
 Ryan et al 
PRO Study Report Number 1, Methodology

23. Bland JM, Altman DG. Statistical methods for assessing 29. Escoffery RF, Olk RJ, Grand MG, et al. Vitrectomy without
agreement between two methods of clinical measurement. scleral buckling for primary rhegmatogenous retinal detach-
Lancet. 1986;1:307e310. ment. Am J Ophthalmol. 1985;99:275e281.
24. Heimann H, Hellmich M, Bornfeld N, et al. Scleral buckling 30. Kinori M, Moisseiev E, Shoshany N, et al. Comparison of pars
versus primary vitrectomy in rhegmatogenous retinal detach- plana vitrectomy with and without scleral buckle for the repair
ment (SPR Study): design issues and implications. SPR Study of primary rhegmatogenous retinal detachment. Am J Oph-
report no. 1. Graefes Arch Clin Exp Ophthalmol. 2001;239: thalmol. 2011;152:291e297.e292.
567e574. 31. Kessner R, Barak A. Pseudophakic rhegmatogenous retinal
25. Jackson TL, Donachie PH, Sparrow JM, et al. United detachment: combined pars plana vitrectomy and scleral
Kingdom National Ophthalmology Database Study of Vitre- buckle versus pars plana vitrectomy alone. Graefes Arch Clin
oretinal Surgery: report 1; case mix, complications, and cata- Exp Ophthalmol. 2016;254:2183e2189.
ract. Eye. 2013;27:644e651. 32. Miller DM, Riemann CD, Foster RE, et al. Primary repair of
26. Adelman RA, Parnes AJ, Ducournau D. Strategy for the retinal detachment with 25-gauge pars plana vitrectomy.
management of uncomplicated retinal detachments: the Euro- Retina. 2008;28:931e936.
pean vitreo-retinal society retinal detachment study report 1. 33. Lai MM, Ruby AJ, Sarrafizadeh R, et al. Repair of
Ophthalmology. 2013;120:1804e1808. primary rhegmatogenous retinal detachment using 25-gauge
27. Azad RV, Chanana B, Sharma YR, Vohra R. Primary vitrec- transconjunctival sutureless vitrectomy. Retina. 2008;28:
tomy versus conventional retinal detachment surgery in phakic 729e734.
rhegmatogenous retinal detachment. Acta Ophthalmol Scand. 34. Mehta S, Blinder KJ, Shah GK, Grand MG. Pars plana
2007;85:540e545. vitrectomy versus combined pars plana vitrectomy and
28. Weichel ED, Martidis A, Fineman MS, et al. Pars plana vit- scleral buckle for primary repair of rhegmatogenous
rectomy versus combined pars plana vitrectomy-scleral buckle retinal detachment. Can J Ophthalmol. 2011;46:237e241.
for primary repair of pseudophakic retinal detachment. 35. Park SW, Kwon HJ, Byon IS, et al. Impact of age on scleral
Ophthalmology. 2006;113:2033e2040. buckling surgery for rhegmatogenous retinal detachment.
Korean J Ophthalmol. 2017;31:328e335.

Footnotes and Financial Disclosures

Originally received: September 20, 2019. J.H.: Grant – Roche/Genentech, Ophthotech, Santen, Novartis, Adverum,
Final revision: February 20, 2020. Chengdu Kanghong Biotechnology, Apellis, Regeneron.
Accepted: February 24, 2020. C.R.: Grant e Genentech, Regeneron, Novartis, Allergan, Astellis, Notal,
Available online: March 4, 2020. ORET_D_19_00177. Chengdu Kanghong, Opthea, Ophthotech, Adverum; Personal fees e
1
VitreoRetinal Surgery PA, Minneapolis, Minnesota. Genentech, Novartis, Allergan, Chengdu Kanghong, Opthea, Adverum,
2 Notal, Shire-Takeda, Kodiak, Allegro, Graybug, Biotime, Merck, Santen,
The Retina Institute, St. Louis, Missouri.
Aldeyra.
3
Massachusetts Eye and Ear, Boston, Massachusetts. HUMAN SUBJECTS: Human subjects were included in this study.
4 Institutional review board approval and an informed consent waiver
Retina Center, Minneapolis, Minnesota.
5 were obtained for this study using Salus IRB and the IRBs of the
The Retina Institute, St. Louis, Missouri.
respective institutions. Data were collected in accordance with Health
6
Mid-Atlantic Retina, Philadelphia, Pennsylvania. Insurance Portability and Accountability Act of 1996 guidelines, and the
Presented in part or in whole at multiple meetings: American Society of study conformed to the tenets of the Declaration of Helsinki. Informed
Retinal Specialists 2018, Vancouver, Canada; Retina Society 2017, Boston, consent was not required or obtained due to the retrospective nature of
MA and 2018, San Francisco, CA; Macula Society 2018, Beverly Hills, the study.
CA; Club Jules Gonin 2018, Jersey, UK; and Vermont Ophthalmologic No animal subjects were used in this study.
Society 2018, Burlington, VT. Author Contributions:
Financial support: Phillips Eye Institute Foundation (Minneapolis, MN) and Conception and design: C.M.Ryan, Joseph, Parke, E.H.Ryan, Shah, Hsu,
VitreoRetinal Surgery Foundation (Edina, MN). The sponsor or funding Forbes, Gupta, Mittra, Yonekawa, Ringeisen
organization had no role in the design or conduct of this research. Data collection: Capone, C.M.Ryan, Regillo, Elliot, Joseph, Parke,
Financial Disclosure(s): E.H.Ryan, Emerson, Shah, Williams, Hsu, Blinder, Forbes, Gupta, Mittra,
The author(s) have made the following disclosure(s): E.R.: Grants – Phillips Yonekawa, Ringeisen
Eye Institute Foundation, VitreoRetinal Surgery Foundation; Personal fees Analysis and interpretation: C.M.Ryan, Joseph, Parke, E.H.Ryan, Shah,
– Alcon Surgical. Hsu, Forbes, Mittra, Yonekawa, Ringeisen
C.R.: Grants – Philips Eye Institute Foundation, VRS Foundation. Obtained funding: E.H.Ryan
Y.Y.: Consultant e Alcon. Overall responsibility: Capone, C.M.Ryan, Regillo, Elliot, Joseph, Parke,
G.E.: Stock e Novartis, Allergan, Celgene, Glaukos, Regeneron, Pfizer, E.H.Ryan, Emerson, Shah, Williams, Hsu, Blinder, Forbe, Gupta, Mittra,
Ocular Therapeutix, Regenexbio. Yonekawa, Ringeisen
KB: Personal fees – Bausch & Lomb, Regeneron, Novartis, Allergan, Abbreviations and Acronyms:
Genentech. BEAVRS ¼ British and Eire Association of Vitreoretinal Surgeons;
AC: Consultant – Alcon Laboratories, Allergan, ThromboGenics; Equity IRB ¼ Institutional Review Board; IRR ¼ inter-rater reliability;
owner – Retinal Solutions, LLC; DSMC Chair, RAINBOW Study e Novartis. PPV ¼ pars plana vitrectomy; RCS ¼ retrospective cohort study;
D.E.: Consultant e Alcon, Alimera, Allergan, Dutch Ophthalmic, Genentech, RCT ¼ randomized controlled trial; RRD ¼ rhegmatogenous retinal
Regenxbio; Grants (institution) e Neurtotech; Stock – Aldeyra Therapeutics, detachment; SB ¼ scleral buckling; SO ¼ silicone oil; SPR ¼ Scleral
Pykus Therapeutics.

821
 Ophthalmology Retina Volume 4, Number 8, August 2020

Buckling Versus Primary Vitrectomy in Rhegmatogenous Retinal Detach- Correspondence:

ment; SSAS ¼ single-surgery anatomic success; VH ¼ vitreous Edwin H. Ryan, MD, VitreoRetinal Surgery, PA, 7760 France Avenue
hemorrhage. South, Suite 310, Edina, MN 55435. E-mail: [email protected].

Pictures & Perspectives

Swept Source OCT Angiography of Optic Nerve Head Retinal Capillary Hemangioma
An 86-year-old woman with nonexudative macular degeneration reported a new floater and examination revealed a peripapillary
hemorrhage in the right eye. After hemorrhage resolution, a vascular elevation was visualized on the optic nerve head (ONH) (A); the left
ONH was unremarkable (B). Swept source OCT angiography revealed dilated vessels with associated flow (C), confirming the presence of
a retinal capillary hemangioma. The left ONH was normal (D). Vision was 20/20 and observation was recommended. Given the negative
personal and family history and a solitary lesion, genetic screening for Von Hippel-Lindau disease was deferred after discussion with a
genetic counselor.
BRITTNEY STATLER, MD
GREGORY KOSMORSKY, DO
ALEKSANDRA RACHITSKAYA, MD
Cole Eye Institute, Cleveland Clinic, Cleveland, Ohio

822