Patient Satisfaction With Oral vs Intravenous Sedation for Vitrectomy Surgery: A Randomized, Noninferiority Clinical Trial

Abstract

Purpose:

This work aims to determine whether patient satisfaction with oral sedation is noninferior to intravenous (IV) sedation in vitrectomy surgery.

Methods:

This prospective, randomized, double-masked, noninferiority clinical trial measured patient satisfaction in 84 participants receiving oral or IV sedation during vitrectomy surgery under monitored anesthesia care. Patients were excluded if they were unable to receive benzodiazepines.

Results:

The primary outcome was patient satisfaction. Secondary outcomes included surgeon and anesthesia provider satisfaction, need for supplemental anesthesia, and surgical complications. Among the 84 patients (46 [54.8%] men; mean [SD] age, 57.0 [12.7 years]), mean patient satisfaction scores were 5.22 ± 0.81 (range, 3.08-6; scale 1-6) with oral and 5.25 ± 0.63 (range, 3.83-6; scale 1-6) with IV sedation. With an a priori noninferiority margin of 0.5 and a difference in mean scores between the groups of 0.03 (1-tailed 95% CI, infinity to 0.29), our results demonstrated the noninferiority of oral sedation (P = .002). There were no significant differences in surgeon or anesthesia satisfaction or major intraoperative complications. Five patients receiving oral (11.9%) and 3 receiving IV (7.1%) sedation required supplemental IV sedation (difference, 4.8%; P = .46).

Conclusions:

Patient satisfaction for oral sedation was noninferior to IV sedation for vitrectomy surgery.

Introduction

Over the last decade, the US health care system has seen a shift in emphasis to “value-based” care that incentivizes hospitals to prioritize low-cost, high-quality services with high patient satisfaction. ^1,2 In an effort to improve patient satisfaction, widespread research is being conducted across medical and surgical specialties, with thousands of related PubMed-indexed manuscripts published yearly. With the goal to improve patient care, it is tremendously important for ophthalmologists to carefully examine ways to deliver the highest-value care without sacrificing patients' experience or safety. The use of oral sedation in ophthalmic surgery offers the potential for improvement in efficiency over intravenous (IV) delivery without sacrificing patient safety or satisfaction.

For cataract surgery, the transition from IV to a less-invasive oral sedation has resulted in improvements in patient safety and reduction in costs. ^3,4 Several health care systems have retrospectively reported on their experience with oral anesthesia in cataract surgery following its large-scale adoption, and our ORVIS (The Oral versus Intravenous Sedation) study group demonstrated noninferior patient satisfaction following cataract surgery with oral vs IV sedation using benzodiazepines in a prospective, randomized clinical trial. ^{5 -8}

However, the data on patient satisfaction and safety with oral sedation for vitreoretinal surgery remain scant. In light of the recent emphasis on cost-containment in vitrectomy surgery, it is crucial to investigate new methods of delivering sedation without affecting patient safety. ⁹ The primary aim of our study was to determine whether patient satisfaction with vitrectomy surgery using oral sedation with benzodiazepines is noninferior to IV sedation with benzodiazepines in a prospective, randomized, double-masked fashion.

Methods

The study design and statistical methods, which we reported previously in our study comparing oral with IV sedation in cataract surgery, are summarized briefly. ⁸ The prospective, single-center, randomized, double-masked, parallel-group, noninferiority clinical trial (www.clinicaltrials.govidentifierNCT03246724) was conducted at an urban teaching hospital (Boston Medical Center, Boston, Massachusetts).

Patients' medical records were evaluated for key eligibility criteria by the study team prior to enrollment in the study. All patients who were scheduled to undergo a pars plana vitrectomy (PPV) under monitored anesthesia care sedation for any indication between October 2017 and October 2018 were approached for participation preoperatively by the research team. Inclusion criteria included: age 18 years or older; ability to speak and read in English, Spanish, or Haitian Creole; and ability to independently participate in the informed consent process. Key exclusion criteria included a plan for general anesthesia at the time of surgery, contraindication to receive benzodiazepine, women currently pregnant or nursing, current use of a medication inhibiting cytochrome P450 3A, or enrollment in this study for the fellow eye or in another study with an investigational drug within the prior 3 months. All patients undergoing PPV who met the inclusion criteria and were able to give informed consent were eligible and were not excluded based on their ophthalmic diagnosis or potential complexity of the surgery. Patients undergoing scleral buckle procedures were excluded because these are performed under general anesthesia at our institution.

Prior to surgery, patients aged 70 years or older were administered a delirium screening questionnaire to determine their safety to receive benzodiazepine and thus, their eligibility in the study. The day of surgery, eligibility was reassessed by the anesthesia and study teams to ensure there were no interval changes to the patient’s eligibility.

Study participants underwent 1:1 random assignment to either oral triazolam with IV placebo or IV midazolam with oral placebo. The majority of individuals were randomly assigned 1 to 2 days prior to surgery if written consent was on file, with the remainder randomly assigned the day of surgery if telephone-based consent was on file and written consent was obtained the day of surgery. The randomization tables, which were created by a biostatistician blinded to patient characteristics and surgical plans, used a blocking scheme with a block size of 6. The participants and all clinical personnel were masked to the random assignment throughout the study. The Investigational Pharmacy Services was not masked to treatment assignment to allow for correct study drug administration but they were masked to all outcome data and had no direct contact with any study participants. Masked study personnel recorded study data. At completion of the study, the statistician unmasked the treatment assignments and analyzed the data.

Sedative medication dosages were standardized and based on the participant’s body mass index (BMI). Individuals with a BMI of less than 35 received 0.125 mg of oral triazolam with IV placebo or 1.0 mg/mL of IV midazolam with oral placebo. Individuals with a BMI of 35 or greater received 0.25 mg of oral triazolam with IV placebo and 2.0 mg/mL of IV midazolam with oral placebo. Each participant received the assigned oral medication or placebo approximately 30 minutes prior to surgery and the assigned IV medication or placebo approximately 5 minutes prior to surgery. To satisfy the internal review board's and surgeons' concerns about adequate patient comfort and safety, an anesthesiologist or certified registered nurse anesthetist (CRNA) monitored all patients throughout the procedure, and additional IV anesthesia was administered intraoperatively, regardless of study arm, if deemed necessary by the surgeon or anesthesia provider.

At the start of each surgical case, the 3 participating vitreoretinal surgeons (S.N., N.H.S., M.L.S.) administered topical tetracaine drops and either a sub-Tenon or retrobulbar block injection, based on the surgeon's preference. All but 2 of the patients received a standard retrobulbar block in the inferotemporal quadrant using a 25-gauge, 1.5-inch Atkinson needle (Beaver-Visitec International, Inc) along with a limited van Lint facial nerve block that was injected subcutaneously 1 cm behind the lateral margin of the orbit. The remaining 2 patients received an inferonasal sub-Tenon block using a 19-gauge hydrodissection cannula (Beaver-Visitec International, Inc) instead of a retrobulbar block because of an increased risk of bleeding from systemic anticoagulation.

Local anesthetic injections consisted of a 1:1 mixture of 2% lidocaine and 0.5% bupivacaine, with or without hyaluronic acid, which were prepared by the hospital pharmacy. All surgical procedures were completed using standard-of-care techniques and 25-gauge vitrectomy instruments. Ophthalmology trainees, including residents and vitreoretinal fellows, performed some aspects of the anesthetic injections and surgical procedures at the discretion of the primary vitreoretinal surgeon. Operative times were recorded. All complications (perioperative and intraoperative) were documented and reviewed on a bimonthly basis by an independent data safety monitoring board.

Immediately following completion of the surgery, both the anesthesiologist/CRNA and the attending surgeon completed a satisfaction survey. To avoid potential confounding effects of residual anesthesia in the early postoperative period, narcotics were not prescribed on discharge, as per our standard protocol, and the patient satisfaction survey was completed independently at the postoperative day 1 visit. The survey design (Supplementary Table 1) was based on the Iowa Satisfaction with Anesthesia Scale, which has previously been validated for use in intraocular surgery and was used with permission from Franklin Dexter and the University of Iowa Research Foundation. ^{10 -13}

Noninferiority Margin and Sample Size Justification

The null hypothesis was defined as a mean patient satisfaction score in the oral sedation group less than that of the IV sedation group minus the noninferiority margin (0.5). The alternative hypothesis was mean patient satisfaction in the oral group greater than or equal to the IV group minus the noninferiority margin. The mean scores for the oral and IV sedation groups’ satisfaction were compared using a 1-tailed t test for noninferiority with a 95% CI.

The a priori noninferiority margin was determined to be 0.5, with an expected mean satisfaction score of 5.5 using the Iowa Satisfaction with Anesthesia Scale (scale, 1-6) and an SD of 0.75, based on an internal pilot study of 20 patients. The mean score of 5.5 in the pilot study meant that respondents fell between “agree” (score, 5) and “agree very much” (score, 6) when asked if they were satisfied with the sedation received during surgery. The noninferiority margin of 0.5 ensured that the oral sedation cohort would at least “agree” that it was satisfied with the level of sedation provided, which we felt to be a clinically acceptable level of satisfaction. The total sample size estimated to obtain statistical significance with a power of 90% and 1:1 randomization of oral triazolam to IV midazolam was 80, with approximately 40 individuals in each group.

Primary Outcome Measure: Patient Satisfaction

Participants' responses to the 13-question satisfaction surgery were scored on a scale of 1 to 6. The average of these scored questions was used to determine the mean patient satisfaction score, with 6 being the highest possible average score.

Secondary Outcomes

Surgeon's and Anesthesiologist's/Certified Registered Nurse Anesthetist's Satisfaction

The surgeon and either the attending anesthesiologist or CRNA present in the operating room were administered a satisfaction survey with 6 scored questions (scale, 1-6). The survey had a highest possible mean score of 6.

Need for Supplemental Intravenous Anesthesia

Additional IV anesthesia was made optional for patients after initiation of the vitrectomy, if deemed necessary by either the surgeon or anesthesia provider to ensure the patient's comfort, and the medication name, dosage, and reason for administration were recorded.

Complications

All intraoperative and perioperative complications were recorded.

Statistical Analysis for Secondary Outcome Measures

Analysis of surgeon's satisfaction, anesthesiologist's/CRNA's satisfaction, supplemental anesthesia interventions, intraoperative complications, and operative times was carried out using independent-sample t tests for continuous variables and the χ² test for categorical variables. All analyses were performed using SAS software, version 9.4 (SAS Institute).

Results

During the time of open enrollment, 88 patients were randomly assigned. Four patients (4.5%) were lost to the study after random assignment, either because of withdrawal of consent (2 individuals) or ineligibility the day of the surgery, as determined by the anesthesia team secondary to concern for administration of benzodiazepines (2 individuals). There was an equal distribution of participants analyzed in each treatment arm: 42 received oral triazolam with IV placebo and 42 received the IV midazolam with oral placebo (Figure 1). Patient demographics are shown in Table 1.

Figure 1.

Consolidated Standards of Reporting Trials flowchart. IV indicates intravenous.

Table 1.

Patient Characteristics.

	Oral sedation		Intravenous sedation
Parameter	No. of participants	% of cohort	No. of participants	% of cohort
Overall (N = 84)	42	50.0	42	50.0
Age, mean ± SD, y	56.76 ± 11.73		57.29 ± 13.65
Sex
Female	21	50.0	17	40.5
Male	21	50.0	25	59.5
BMI, mean ± SD	28.88 ± 4.56		30.83 ± 7.26
<35, lower dose	38	90.5	30	71.4
≥35, higher dose	4	9.5	12	28.6
Preoperative diagnosis
Tractional retinal detachment	17	40.5	6	14.3
Nonclearing vitreous hemorrhage	9	21.4	13	31.0
Rhegmatogenous retinal detachment	7	16.7	11	26.2
Epiretinal membrane	5	11.9	5	11.9
Macular hole	1	2.4	7	16.7
Vitreous opacity	2	4.8	0	0.0
Exudative retinal detachment	1	2.4	0	0.0
Ethnicity
Hispanic/Latino	8	19.0	13	31.0
Non-Hispanic/Latino	33	78.6	27	64.3
Unknown	1	2.4	2	4.7
Race
Black/African American	14	33.3	14	33.3
Declined	16	38.1	13	31.0
White	10	23.8	15	35.7
Asian	1	2.4	0	0.0
Native Hawaiian/Pacific Islander	1	2.4	0	0.0
Language
English	30	71.4	26	61.9
Spanish	7	16.7	13	31.0
Haitian-Creole	5	11.9	3	7.1
Medical conditions
Systemic hypertension	27	64.3	28	66.7
Tobacco use	7	16.7	7	16.7
Hypercholesterolemia	7	16.7	10	23.8
Heart disease	10	23.8	8	19.0
Diabetes mellitus	27	64.3	24	57.1
Hyperlipidemia	18	42.9	12	28.6
Asthma/bronchitis/COPD	7	16.7	2	4.8
None	4	9.5	7	16.7

Abbreviations: BMI, body mass index; COPD, chronic obstructive pulmonary disease.

Satisfaction Surveys

The mean patient satisfaction score (scale, 1-6) was 5.22 ± 0.81 (range, 3.08-6.00) for the oral sedation group and 5.25 ± 0.63 (range, 3.83-6.00) for the IV sedation group. With a difference in mean patient satisfaction scores between the oral and IV sedation groups of 0.03 (1-tailed 95% CI, infinity-0.29) and an a priori noninferiority margin of 0.5, our results demonstrated noninferiority of oral sedation (P = .002).

The mean surgeon satisfaction score (scale 1-6) was 5.44 ± 0.99 (range, 1.50-6.00) for the oral sedation group and 5.69 ± 0.64 (range, 2.17-6.00) for the IV sedation group (P = .16). The mean anesthesiologist satisfaction score was 5.02 ± 1.00 (range, 1.17-6.00) for the oral group and 5.25 ± 0.88 (range, 1.67-6.00) for the IV group (P = .26).

Supplemental Intravenous Anesthesia Intervention

Five patients in the oral group (11.9%) and 3 in the IV group (7.1%) received additional IV sedation (difference, 4.8%; P = .46). When the 8 patients who received additional IV medications were excluded from the analysis, the mean estimates of patient satisfaction scores did not appreciably change (5.27 ± 0.77 [range, 3.25-6.00] for the oral group and 5.24 ± 0.62 [range, 3.83-6] for the IV group) and noninferiority was maintained (P = .001). When these participants were excluded, the surgeons' mean satisfaction scores increased to 5.62 ± 0.70 (range, 2.83-6.00) for the oral sedation group and 5.82 ± 0.27 (range, 5.00-6.00) for the IV group, but no significant difference existed between the groups (P = .10). The anesthesia group's mean satisfaction scores increased to 5.25 ± 0.70 (range, 3.5-6.00) for the oral and 5.45 ± 0.47 (range, 4.17-6.00) for the IV group, but analysis also failed to detect a significant difference (P = .14).

Safety Outcomes

Adverse Events

Seven patients (16.7%) from the oral group and 9 (21.4%) from the IV group experienced a perioperative adverse event (AE; difference, 4.7%, P = .58). AEs were defined as any untoward or unfavorable medical occurrence, including any abnormal sign, symptom, or disease, that occurred during the patient’s participation in the research, whether or not it was related to the study intervention. The most common AE was postoperative pain around the eye by 14 participants (16.7%). All AEs were determined to be expected perioperative events that were not related to the study interventions (Table 2).

Table 2.

Summary of Perioperative Adverse Events.^a

AE	No. of participants (N = 84)
	Oral sedation	IV sedation
No AEs experienced (n = 68)	35	33
AEs experienced (n = 16)	7	9
	No. of AE occurrences (N = 23)
Type of AE	Oral sedation	IV sedation
Dizziness	0	0
Agitation	0	0
Headache	1	1
Nausea	2	5
Pain	6	8
Malaise	0	0

Abbreviations: AE, adverse event; IV, intravenous.

^aThere was no significant difference in the incidence of perioperative AEs between the 2 groups (P = .58).

Intraoperative Complications

Two patients (4.8%) in the oral group and 6 (14.3%) in the IV group experienced minor surgical or anesthesia-related intraoperative complications not affecting the final surgical outcome (difference 9.5%, P = .14) (Table 3).

Table 3.

Summary of Intraoperative Complications.^a

Complication	No. of participants (N = 84)
	Oral sedation	IV sedation
No complication (n = 76)	40	36
Complication (n = 8)	2	6
	No. of complication occurrences (N = 11)
Complication	Oral sedation	IV sedation
Somnolence	1	0
Iris prolapse	1	0
Dropped nucleus	1	0
Anterior capsule tear	1	1
Posterior capsule tear	1	1
Respiratory obstruction	0	1
Descemet detachment	0	1
Retinal break/detachment	0	2

Abbreviation: IV, intravenous.

^a There were no significant differences in the incidence of intraoperative complications between the 2 groups (P = .13).

Surgical Time

The mean surgical time, from incision to closure, was 50.07 ± 19.56 minutes (range, 15-108 minutes) for the oral group and 47.88 ± 22.17 minutes (range, 15-110 minutes) for the IV group (P = .63). In both groups, most cases lasted 30 to 60 minutes (57.1% and 52.4% in the oral and IV sedation groups, respectively). However, a higher proportion of the surgical procedures longer than 60 minutes were in the oral group (31.0%) compared with the IV group (23.8%), whereas the IV group had more cases shorter than 30 minutes (Table 4).

Table 4.

Comparison of Surgical Case Times.

Surgical case times, min	Oral sedation		IV sedation
	No. of participants	% of cohort	No. of participants	% of cohort
<30	5	11.9	10	23.8
30-60	24	57.1	22	52.4
60-90	12	28.6	8	19.0
>90	1	2.4	2	4.8

Abbreviation: IV, intravenous.

Conclusions

The 21st century has seen continued improvements in the instrumentation for vitreoretinal surgery, including the initial introduction of microincision vitrectomy surgery (MIVS) with 25-gauge instruments in 2002. ¹⁴ The advent of 23-, 25-, and 27-gauge MIVS instruments with several-fold increases in cut rates, compared with the older 20-gauge instruments, has resulted in improvements in operative efficiency and surgical time for PPV surgery, resulting in its broad implementation both nationally and internationally within 10 years of its availability. ^14,15

In conjunction, there has been a trend from general anesthesia to local anesthesia for vitrectomy surgery. However, regional variations exist in the preference for different methods of anesthesia. Evaluating the data from the United Kingdom at the turn of the 21st century, the use of general anesthesia, which accounted for greater than 95% of the anesthesia in 2001, drastically diminished, and the use of local anesthesia for routine vitreoretinal surgery, most commonly with a sub-Tenon injection, increased from 5% in 2001 to nearly 60% in 2010. ¹⁶ Interestingly, only 2% of the local anesthesia cases in this UK study incorporated IV systemic sedation. In contrast, in the United States, where vitreoretinal surgeons adopted local anesthesia techniques earlier than their UK counterparts, ¹⁷ the use of IV sedation remains the standard of care for most vitreoretinal surgical cases. ¹⁶

There is a lack of prospective randomized controlled trials in the current literature that compare different methods of anesthesia for PPV surgery. Licina et al were unable to complete a Cochrane review in 2016 that compared general vs local anesthesia for PPV as there were no eligible randomized controlled trials. ¹⁸ We believe that our trial can serve as a pilot study to demonstrate the maintained patient safety and satisfaction associated with oral sedation for PPV surgery.

The use of oral, instead of IV, sedation in vitrectomy surgery offers the potential for preserved patient satisfaction and efficiency without sacrificing patient safety. In our diverse patient population, we determined that patient satisfaction for oral sedation with benzodiazepines was noninferior to IV sedation with benzodiazepines for vitrectomy surgery. Further study could be undertaken to compare oral benzodiazepine use with nonbenzodiazepines, namely IV propofol, at the time of the retrobulbar block or during vitrectomy surgery.

In addition to maintaining patient satisfaction and safety, the use of oral rather than IV sedation represents a potential opportunity to decrease the costs associated with vitrectomy surgery. The American health care system continues to experience a tremendous rise in costs, ^{19 -21} and it is critical to investigate ways to improve cost-efficiency while maintaining high-quality surgery. ^{22 -25} Although we did not calculate the associated costs for IV vs oral sedation, given the large number of vitrectomy procedures performed yearly in the United States, even modest per-case cost reductions could translate into significant savings. As more than 100 000 vitrectomies were performed in 2014 in the Medicare population alone, the potential for cost reduction by substituting oral for IV sedation could translate into millions of dollars. ²⁶

The use of oral sedation could potentially allow some ophthalmic procedures to be performed outside an operating room. Since 2006, Kaiser Permanente Colorado has been performing anterior segment surgery using oral sedation in an office-based minor procedure room (MPR) setting. The MPRs are either connected to an ambulatory surgery center or in a private office with varying degrees of support from an anesthesia provider. Ianchulev et al published their 10-year experience of more than 21 000 cases, and they demonstrated both the safety and effectiveness of oral sedation. ⁷ More recently, our group showed oral sedation in cataract surgery to be noninferior to IV for patient satisfaction in a randomized clinical trial. ⁸

However, there are multiple differences between anterior segment and vitreoretinal surgical procedures. A recent survey-based study evaluating the response of 556 retina specialists worldwide found that regional block anesthesia with monitored anesthesia care remains the overwhelming preference of surgeons. ²⁷ Historically, prior to the introduction of MIVS, vitreoretinal procedures tended to be significantly longer in duration than cataract surgery. ^14,15 Additionally, given that PPVs are frequently performed for complications related to diabetic retinopathy, patients undergoing PPVs often have more medical comorbidities, including poorly controlled diabetes, hypertension, renal failure requiring dialysis, or vascular disease requiring anticoagulation use. ²⁸

There are also aspects of vitreoretinal surgery—including the use of endolaser, which was performed in 55 of the surgical cases in our study, trocar placement, scleral depression, and conjunctival manipulation—that could cause more discomfort than the maneuvers performed in a routine cataract surgery. However, in our study, the more complex diagnoses were in the oral group (40.5% of the tractional retinal detachment cases were in the oral vs 14.3% in the IV group) and the surgical procedures requiring less manipulation were greater in the IV group, namely macular hole procedures, which comprised 2.4% of the oral and 16.7% of the IV group. It would be expected that the complexity of the surgery requiring more manipulation could decrease the patient’s satisfaction, which gives more credence to our findings of noninferiority with oral vs IV sedation.

Beyond the efficiency and cost-saving benefits already discussed, one must also consider the potential for improved patient safety associated with using oral in lieu of IV sedation. Known risks associated with IV sedation, namely apnea, respiratory depression, and hypotension, could be lessened with the use of oral sedation. ⁴ Oral sedation may reduce the risk of aspiration and could eliminate the need for patients to fast preoperatively. Reduced fasting requirements may not only further improve patient satisfaction but may also prevent severe glucose fluctuations in patients with diabetes who are forced to interrupt their normal dietary and medication routines during the preoperative period. ²⁹

However, safety considerations of oral sedation include an unanticipated, emergent conversion to general anesthesia intraoperatively, in which case having both IV access and a fasting patient remains important. Further study is needed to risk stratify the preoperative or intraoperative factors that could predict the need for supplemental anesthesia or other required IV medications. Additionally, an implementation protocol could include IV access on all patients, regardless of their route of sedation, in case unanticipated medication use is necessary. It will also be important to evaluate the feasibility for vitreoretinal surgery to be performed with oral sedation in an MPR setting. ³⁰

As far as we know, our study was the first to compare patient satisfaction outcomes for oral vs intravenous sedation for PPV. We feel that this study has several strengths. With a prospective, double-masked, randomized design, we sought to limit patients', surgeons', and anesthesiologists' bias. Recent research suggests that racial bias may affect the assessment of another person’s pain. ³¹ Given the possibility that ethnic and racial differences may influence one’s satisfaction, and thus affect our data, we collected each participant’s race and ethnicity. Both of our patient cohorts were ethnically (Hispanic/Latino, 19.0% and 31.0% in the oral and IV groups, respectively) and racially (Black/African American, 33.0% in both groups and White, 23.8% and 35.7% in the oral and IV groups, respectively) diverse, which should make our results more generalizable.

Other than scleral buckle surgery, which is routinely performed with general anesthesia at our facility, we did not exclude any surgical indication from trial inclusion. While mean operative times (incision to closure) in both groups were less than an hour, we feel that these surgical times are representative of most cases performed in the MIVS era. Additionally, both groups included significantly longer cases, including combined phaco-vitrectomy operations and diabetic traction retinal detachment cases.

There are limitations to this study as well. Noninferiority trials are inherently complex and may miss subtle differences between groups, particularly if a poor outcome measure is used. We are confident in our satisfaction survey methodology (the Iowa Satisfaction with Anesthesia Scale) as it has previously been validated for use in ophthalmic surgery. We also selected a noninferiority margin to ensure that any subtle differences between the 2 groups would not be clinically relevant. Although the secondary outcomes—specifically surgeon and anesthesia-provider satisfaction—were not directly accounted for in our a priori power calculations, they provide useful corroboration of the main study’s conclusions. The provider perceptions, based on observations of undesired movement or complaints of pain or anxiety by the patient, did not differ significantly between the 2 treatment arms. Given that benzodiazepine medications can act as anxiolytics as well as cause amnesia, the high surgeon satisfaction scores reassure us that patients were truly comfortable during surgery and not simply unable to recall a traumatic episode.

A final limitation of our study involved the availability of supplemental IV anesthesia to both study arms, which served as a means to ensure patient comfort. While 9.5% of participants required supplemental anesthesia, we found no significant difference between treatment arms, and removing those patients who received additional medication from the analysis did not affect the conclusion. Regardless, the small subset of patients requiring additional medication echoed a conclusion from our cataract sedation study—specifically that “anxious outliers” exist whose satisfaction certainly would have suffered without additional anesthesia. This highlights the need for standardized screening tools to identify patients who require a higher level of sedation and potentially need continued anesthesia monitoring during intraocular surgery.

While the potential change from IV to oral sedation may seem unconventional for retina surgery, we present this study as a proof of concept of the feasibility of oral sedation for posterior segment surgery. More evaluation is required before widespread adoption could be considered. Further study needs to be conducted to determine the patient characteristics, length of the surgery, preoperative diagnoses, or other risk factors that could predict optimal candidates to receive oral sedation.

Over the last 20 years there has been widespread advancement in the instrumentation and efficiency of vitreoretinal surgery as well as a shift from general to local anesthesia. This study demonstrates that in the age of MIVS, oral sedation is noninferior to IV sedation for patient satisfaction and safety during vitrectomy surgery. However, further study is needed to determine the feasibility and safety of performing vitreoretinal surgery under oral anesthesia without the presence of an anesthesia provider.