Systematic review and meta-analysis of Trendelenburg position on intraocular pressure in adults undergoing surgery

Introduction

In the Trendelenburg position, the patient’s feet are higher than the patient’s head by 15 to 30 degrees (1). Many surgeons use a steep Trendelenburg position of 30 to 45 degrees, particularly during laparoscopic and robotic surgery. The benefit of the Trendelenburg position is that it moves the abdominal viscera cephalad to improve visibility and surgical access to the abdominal and pelvic organs. However, there are potential harms associated with the Trendelenburg position. The Trendelenburg position increases intraocular pressure (IOP) (2-10). According to the American Academy of Ophthalmology (11), normal IOP is 10 millimeters of mercury (mmHg) to 21 mmHg. IOPs higher than 21 mmHg pose a risk for glaucoma, detached retina, and postoperative vision loss (3,5,10,12-18).

Pathogenesis of postoperative vision loss

The specific pathogenesis of postoperative vision loss associated with increased IOP is unclear; however, it is known that elevated IOP can lead to optic nerve injury and decreased ocular perfusion pressure resulting in ischemic optic neuropathy (8,19). Ischemic optic neuropathy is the most common cause of postoperative vision loss (8,19). The ischemic process can occur as a direct result of decreased blood supply from the arteries of the optic nerve or by venous stasis that occurs as a result of decreased venous outflow (8,14,20). Periorbital swelling and venous congestion resulting from the Trendelenburg position can lead to a compartment syndrome in the orbital space that compromises blood flow to the eye, retina, and optic nerve (20,21). The amount of subsequent postoperative vision loss can range from temporary blurring to partial to complete blindness; however, once a loss of vision occurs, it is an irreversible complication (12,18).

Some researchers have found that increased IOP associated with the Trendelenburg position poses a greater risk for postoperative vision loss in patients who have existing ocular disease compared with patients without ocular disease (3,5,9,10,17,22). Older patients with elevated baseline IOPs are also at greater risk for ischemic optic neuropathy than younger patients with normal baseline IOPs (7-9,16,23). Likewise, patients with cardiovascular deficits may be at greater risk for postoperative vision loss than patients without cardiovascular deficits (5). The increase in IOP and subsequent risk for postoperative vision loss is related to the amount of time the patient is in the Trendelenburg position (21,24-27).

Incidence of postoperative vision loss

The incidence of postoperative vision loss following non-ocular surgery has been estimated to be as low as 0.0002% and as high as 0.2% (19,28); however, the incidence of postoperative vision loss specific to patients undergoing surgery in the Trendelenburg position remains unknown. To identify cases of ischemic optic neuropathy associated with prostatectomy procedures performed in the Trendelenburg position, Lee (29) reviewed the American Society of Anesthesiologists Postoperative Vision Loss Registry, a database of 175 cases of postoperative vision loss occurring between 1987 and 2010, and found six cases. A number of reports of postoperative vision loss following surgical procedures where the patient was in the Trendelenburg position have also been published (20,30-33).

Purpose

Although some researchers have studied the quantitative increase of IOP in surgical patients in the Trendelenburg position (3-5,9,10,17,20), there is a need for systematic review and meta-analyses of these studies to demonstrate the overall effect size and provide high-quality evidence supporting, or negating the need for, implementing intraoperative interventions designed to mitigate the increase of IOP and reduce the risk for postoperative vision loss. Meta-analysis research methods provide increased power compared to individual studies, improve estimates of effect size, and help resolve uncertainty when the results of individual studies disagree (34). Because all of the evidence pertaining to a particular phenomenon is included in the analysis, meta-analysis research provides a high level of objectivity, precision, and generalizability (35). Currently, there has been no quantitative meta-analytic synthesis of the existing studies examining the increase in IOP in adult patients undergoing surgery in the Trendelenburg position. The purpose of this systematic review and meta-analysis is to estimate the magnitude of the increase in IOP at selected perioperative time points in adult patients (i.e., individuals 18 years and older) undergoing any type of surgery in the Trendelenburg position.

Methods

To ensure rigorous and transparent presentation of the methods and results of this systematic review and meta-analysis, the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines have been followed (36).

Search strategies

An expert health sciences reference librarian was consulted to identify the most appropriate search terms and databases and to assist with refining search strategies for an exhaustive literature search. Keywords or medical subject headings included intraocular pressure or ocular tension, and Trendelenburg position or head-down tilt. Search strategies included online searching of the PubMed, Cumulative Index of Nursing and Allied Health Literature, and Scopus databases, and the Cochrane Database of Systematic Reviews for published and unpublished literature; ancestry searching of references from relevant reports to locate additional applicable references; author searching of individuals identified in the literature as experts in the field; and a dissertation search of the ProQuest database. The author reviewed abstracts for eligibility and obtained full-text copies of potentially eligible reports.

Inclusion and exclusion criteria

Criteria for inclusion in the meta-analysis were reports written in English; studies reported between January 01, 1990 and September 30, 2018; published or unpublished reports of primary studies that encompassed dissertations, conference abstracts, and presentations; studies that used either a one-group, pretest posttest design or a multiple-group, pretest posttest design; reports where the minimum age of the study participants was 18-year or older; studies that included a specific measured outcome of IOP using any type of tonometer; studies where the participants received any type of general anesthesia; and reports of studies that included sufficient data to calculate an effect size. The year 1990 was selected as the initial searching date because the first laparoscopic cholecystectomy was performed in 1987 (37). The use of steep Trendelenburg position has increased dramatically with the use of laparoscopic and robotic-assisted laparoscopic surgery. When reports did not include sufficient data to calculate an effect size, the author contacted the researchers on at least two separate occasions 2 to 3 weeks apart to obtain missing data.

Reports were excluded if the IOP was measured in adults not undergoing surgery. Reports were also excluded if data from only one time point of IOP measurement were present. Participant groups were additionally excluded if they were receiving an intervention specifically intended to mitigate IOP; however, participant groups representing control arms receiving placebos or no interventions were included in the systematic review and meta-analysis.

Risk of bias within individual studies

To assess the risk of bias within individual studies, the Association of periOperative Registered Nurses Research Evidence Appraisal Tool – Study, available on the Association’s website (https://www.aorn.org/guidelines/about-aorn-guidelines/evidence-rating) was used by the author and an experienced evidence reviewer to independently evaluate and critically appraise each study for its level of strength and quality. When using the appraisal tool, strong study designs (e.g., randomized controlled trials) are assigned the highest level of strength (i.e., experimental). Non-experimental designs (e.g., observational studies) are assigned the lowest level of strength, and quasi-experimental designs (e.g., non-randomized) are assigned a moderate level of strength. Measures such as sample size, generalizability, bias, reliability, and validity are assessed to determine whether the study quality is high, good, or low. The author and evidence reviewer participated in conference calls to discuss their independent appraisals until 100% consensus was achieved on study design and quality levels for each of the included studies.

Risk of bias across studies

Risk of bias that may affect cumulative evidence was managed using several strategies. To avoid bias due to a narrow or limited search, a comprehensive and diverse literature search was conducted. Only research studies were included in the systematic review and meta-analysis to ensure the included studies were of sufficient strength and quality. As well, studies that included objective measurements of IOP at less than two perioperative time points were excluded. An analysis of publication bias was also conducted to determine whether unpublished research was unintentionally excluded.

Coding and data extraction

An iterative process that included studying codebooks used by experienced meta-analysts for data extraction and coding of their research studies was used to develop the codebook. The relevant literature was also reviewed. The codebook was pilot tested by the author using 10 randomly selected studies before being used to code and extract data from all eligible reports to identify missed coding categories and verify fit between coding categories and study characteristics.

Effect size data for each of the reports included in the systematic review and meta-analysis was independently coded by a trained researcher. The author and independent researcher discussed coding discrepancies until 100% consensus was achieved on effect size data for each of the eligible studies.

Data collected from each eligible study included study characteristics (i.e., authors, year of publication, publication status, geographic location, reported funding) and data related to study design (i.e., type of study, study quality, type of tonometer, inclusion of ophthalmologic exams by participants). When available, data related to participant and surgery characteristics [i.e., age; gender; American Society of Anesthesiologists Physical Status Classification (38); body mass index (BMI); comorbidities; type of surgery; degree of Trendelenburg; intra-abdominal pressure; type of anesthesia; duration of anesthesia, pneumoperitoneum, Trendelenburg position, and surgery; estimated blood loss] were collected. Data necessary to calculate effect sizes were extracted for the multiple time points recorded by the researchers during the perioperative phases of the procedures included in their studies.

Analyses

Meta-analyses were conducted using Comprehensive Meta-Analysis, Version 3, a statistical software developed specifically for meta-analysis research (39). Time points for meta-analysis were selected from the time points recorded by the researchers to achieve the greatest number of comparisons for analysis at each time point and to allow for similarity with the order of events as they occur during surgery. Analyses were conducted for the nine time points described in Box 1. Standardized mean difference effect sizes were calculated for each participant group at each measured time point.

To account for sample size and adjust for bias, effect size values were weighted by the inverse of the variance. To account for between- and within-study variation, and because heterogeneity was observed among study designs, sample attributes, and outcome measures, a random effects model was selected a priori to synthesize effect sizes. Using a random effects model assumes that the true effect size varies from one study to the next (40). A random effects model was used for seven analyses (T1, T2, T3, T5, T6, T7, T8). A fixed effect model was used to synthesize effect sizes for two analyses (T4, T9) because the number of included studies at these time points was limited. Using a fixed effect model assumes that the true effect size is the same for all studies (40). Borenstein et al. (40) suggest using a fixed effect analysis when the number of included studies is limited, even when heterogeneity among the studies is observed, because when using a random effects analysis, the estimate of between-studies variance may have poor precision. Relative to the interpretation of effect sizes, Cohen (41) recommended that 0.2 be considered a small effect size, 0.5 be considered a medium effect size, and 0.8 or greater be considered a large effect size. Following the procedures described by Lipsey and Wilson (42), the calculated effect sizes were converted to the metric used to measure IOP (i.e., mmHg) to facilitate interpretation of effect size findings.

The extent of publication bias for the meta-analysis was assessed by constructing a funnel plot. Notably, a funnel plot may suggest publication bias, but does not eliminate the bias (43). In order to create a funnel plot, there must be a minimum of three studies (39). An Egger’s test using linear regression was also conducted to measure asymmetry of the funnel plot (39,44,45). Using an Egger’s test is not advised when there are less than 10 studies included in the meta-analysis because the power of the test may be too low to distinguish true asymmetry from chance (44). Therefore, when the time point analysis included 10 or more participant groups (T1, T2, T6), a funnel plot was constructed and an Egger’s test conducted.

After deciding on the model and calculating effect sizes, the studies included in the meta-analysis were assessed for heterogeneity [i.e., variability among the studies (46)]. Heterogeneity testing explores the null hypothesis (i.e., that the same effect is being evaluated by all studies) (47). Heterogeneity among the included studies in a meta-analysis is very common and should be anticipated, not regarded as the exception (34).

Homogeneity of variance among effect sizes was tested using Cochrane’s Q, which estimates statistical significance; Tau-squared, which estimates the absolute value of the true variance between studies, but not the proportion of the variance; and I-squared, which estimates the proportion of true variance, but not the absolute value of the variance (48). Prediction intervals for each time point were also calculated to show the dispersion of true effect sizes around the mean (49).

Results

The flow of study selection is depicted in Figure 1. In total, 2,693 records were identified for possible inclusion, and of these, 18 studies were included in the systematic review and meta-analysis. Four non-experimental studies had multiple participant groups (50-53), resulting in a total of 24 participant groups and 762 participants for analysis. Table 1 contains a summary of the studies included in this review and meta-analysis.

Figure 1 Flow diagram of study selection. Adapted from (36). s, studies.

Study characteristics

All 18 studies included in the systematic review and meta-analysis were obtained from peer-reviewed journals. The researchers of six studies (33.3%) reported receiving some type of funding or donated supplies (53-56,60,61). Although the literature was searched from 1990 through 2018, studies included in the systematic review and meta-analysis were published between 2011 and 2018 (Figure 1 and Table 1). Some earlier studies were located during the literature search (4,63,64); however, these were excluded because of insufficient effect size data. As shown in Table 1, the greatest number of studies were published in 2015 (9,10,53,62) (s=4) and 2018 (51,56-58) (s=4), but the greatest number of participants occurred in 2013 (5,16,50) (n=147). The majority of studies (s=14) were conducted in Asia (10,50-53,56,57) (s=7) or North America (5,9,20,58-61) (s=7), with the majority of participants also from Asia (n=361) or North America (n=292).

Participant and surgery characteristics

Participant and surgery characteristics are shown in Table 2. Participant race was only reported by two researchers (5,54). Socioeconomic status was not reported by any researchers. Participants were slightly overweight with a mean BMI of 27.5 kg/m² (±2.3) (65). Notably, some researchers had exclusion criteria for participant age, BMI, ocular disease, and American Society of Anesthesiologist’s classification (Table 1).

Study design characteristics

As shown in Table 1, the 18 reports included in the systematic review and meta-analysis comprised four experimental (57,58,61,62) (n=100), two quasi-experimental (59,60) (n=168), and 12 non-experimental (5,9,10,16,20,50-56) (n=494). Nine were high quality (20,51,53,55,57-61) (n=461) and nine were good quality (5,9,10,16,50,52,54,56,62) (n=301). The researchers used five different types of tonometers to measure IOP, the Tono-Pen XL was used most frequently (5,10,20,52,53,56,59-61) (s=9; n=420). The researchers reported having 195 participants undergo preoperative ophthalmologic examinations (9,10,16,50,54,56,58) (s=7), and 100 participants undergo postoperative ophthalmologic examinations (9,10,56,58) (s=4). Notably, all participants who received postoperative ophthalmologic examinations also received preoperative examinations (n=100).

Effect sizes

Results of the synthesized effect sizes, prediction intervals, and meta-analyses for each time point of IOP measurement are shown in Table 3. A graphical representation of the magnitude of change in IOP for T0 through T9 is shown in Figure 2. In total, between abdominal insufflation in supine position (T2) and 5 minutes (T3), 60 minutes (T4), 150 minutes (T6), and 240 minutes (T8) of Trendelenburg position, IOP increases significantly by 13.6 mmHg (i.e., 3.5+4.4+2.6+1.5+1.6=13.6 mmHg). Based on the upper limits of the prediction intervals (Figure 3), in 95% of all populations IOP could increase by as much as 28.1 mmHg (i.e., 7.6+8.5+6.6+2.3+ 3.1=28.1 mmHg). The greatest increase in IOP occurs after the patient is placed into the Trendelenburg position (Figure 2; T3: +4.4 mmHg). The IOP continues to increase significantly while the patient is in Trendelenburg position, but to a lesser degree (T4: +2.6 mmHg; T6: +1.5 mmHg; T8: +1.6 mmHg). IOP decreases significantly after induction of anesthesia in supine position (T1: −5.2 mmHg) and after a return to supine position for arousal from anesthesia (T5: −7.5 mmHg; T7: −8.2 mmHg; T9: −6.0 mmHg). The forest plot of effect sizes for each participant group included in the meta-analysis for T6 is shown in Figure 4. The funnel plot for publication bias for T6 is shown in Figure 5.

Figure 2 Magnitude of mean change in IOP for T0 through T9 (Box 1). IOP increases significantly after abdominal insufflation in the supine position (T2: +3.5 mmHg, P<0.001), when the patient is placed in Trendelenburg position (T3: +4.4 mmHg, P<0.001), and with extended time in the Trendelenburg position (T4: +2.6 mmHg, P<0.001; T6: +1.5 mmHg, P<0.001; T8: +1.6 mmHg, P<0.001). IOP decreases significantly after induction of anesthesia (T1: −5.2 mmHg, P<0.001) and before arousal from anesthesia (T5: −7.5 mmHg, P<0.001; T7: −8.2 mmHg, P<0.001; T9: −6.0 mmHg, P<0.001) when the patient is in the supine position. *, Pooled mean at T0—Before induction of anesthesia. IOP, intraocular pressure; mmHg, millimeters of mercury; T, time point.

Figure 3 Magnitude of change in upper prediction intervals of IOP for T0 through T9 (Box 1). After 180 to 240 minutes in the Trendelenburg position, in 95% of all populations, IOP could increase to 35 mmHg (16.5–9.6+7.6+8.5+6.6+2.3+3.1=35 mmHg). *, pooled mean at T0—before induction of anesthesia. IOP, intraocular pressure; mmHg, millimeters of mercury; T, time point.

Figure 4 Forest plot of meta-analysis of T6. Abbreviations: min, minutes; Std, standard; T, time point. This analysis included 11 studies representing 12 participant groups (n=410). Effect sizes were calculated using a random effects model. The area of each square is proportional to study weight.

Figure 5 Funnel plot of publication bias for T6. Larger studies are shown at the top of the funnel. Positive smaller studies are shown at the right of the mean effect size (i.e., center line). The symmetrical distribution of studies (i.e., data points) around the mean effect size indicates a lack of publication bias. The non-significant Egger’s regression intercept for this funnel plot (bias =−0.05; P=0.47) is also indicative of the absence of apparent bias in the studies included in the meta-analysis for this time point. T, time point.

Discussion

The results of this systematic review and analysis have shown that IOP increases significantly for adult patients undergoing surgery in the Trendelenburg position. As shown in Figure 2, if an individual had a baseline IOP of 16.5 mmHg before induction of anesthesia (as indicated by the pooled mean calculated for T0), after 180 to 240 minutes in the Trendelenburg position, the patient’s IOP could increase to 24.9 mmHg (16.5–5.2+3.5+4.4+2.6+1.5+1.6=24.9 mmHg). Based on the upper limits of the prediction intervals (Figure 3), after 180 to 240 minutes in the Trendelenburg position, IOP could increase to 35 mmHg (16.5–9.6+7.6+8.5+6.6+2.3+3.1=35 mmHg). An IOP of 24.9 to 35 mmHg is above the highest parameter of normal IOP (i.e., 21 mmHg). As shown in Table 2, the mean duration of Trendelenburg position for the studies included in this systematic review and meta-analysis was 104.8 minutes (±58.2) with a range of 68 to 207 minutes. The greatest increases in IOP occur during abdominal insufflation and within the first 60 minutes after Trendelenburg position (Figure 2); however, based on the collective range of 68 to 207 minutes for duration of Trendelenburg position, a Trendelenburg time of 180 to 240 minutes is not implausible. Another important consideration regarding the findings of this systematic review and meta-analysis is that the mean degree of Trendelenburg position was 28.4 (Table 2). Steep Trendelenburg is generally considered to be a head-down tilt of 30 to 45 degrees (7,15,66); thus, it is likely that Trendelenburg positions greater than 28.4 degrees would have produced even greater increases in IOP.

Implications for practice

Increased IOP puts the patient at risk for glaucoma, detached retina, or partial to complete vision loss (3,5,7,10,12,14,18). IOP increases of the magnitude found in this systematic review and meta-analysis clearly demonstrate the need for implementing intraoperative interventions to mitigate the increase in IOP and reduce the potential for serious ocular complications in patients undergoing surgery in the Trendelenburg position. These intraoperative interventions may include monitoring IOP at established intervals or continuously (17,67,68), reducing the degree of Trendelenburg position (7,62,69,70), implementing a modified Trendelenburg position (62), providing periodic position changes or rest periods (5,14,15,23,55,59) and administering specific medications or anesthetics (8,50,51,58,64,71-74).

Because IOP increases during abdominal insufflation and Trendelenburg position, intraoperative monitoring of IOP either continuously or at established intervals or time points (e.g., after abdominal insufflation, after initiation of Trendelenburg position, after 60 minutes of Trendelenburg position, etc.) seems prudent. Elevated IOPs can be an indication of ocular venous congestion and decreased perfusion of the optic nerve (21). Monitoring IOP can provide a baseline IOP and an objective measurement that can help the surgical team maintain awareness of the patient’s IOP, implement interventions to reduce IOP as needed, and thus reduce the potential for ocular complications and postoperative vision loss (17,67).

Steeper degrees of Trendelenburg increase the risk for postoperative complications because they place greater physiologic stress on the patient’s body (7,15,70). Ghomi et al. (7) found that robotic-assisted gynecologic surgery could be performed successfully with a modest head-down tilt of 16.4 degrees. In a study to determine the head-down tilt necessary to provide adequate surgical access and visibility, Gould et al. (15) found the mean head-down tilt most often selected by the participating surgeons was 28.1 degrees, which was much less than the 40-degree head-down tilt the surgeons were using.

Raz et al. (62) found that modifying the Trendelenburg position so that the patient’s head and shoulders remained level significantly decreased IOP and accelerated its return to baseline levels. Implementing periodic intraoperative position changes or rest periods in supine position (or positions where the ocular level is above the heart) can help to reduce IOP. In a quasi-experimental study, Molloy and Watson (59) implemented a 5-to-7-minute level supine intervention after 60 minutes of 32- to 40-degree Trendelenburg position and found there was a significant decrease in IOP after 120 minutes (Intervention: 18.7±5.22 mmHg; Control: 35.7±10.56 mmHg; P<0.001). The dramatic and significant decrease in IOP that occurs before arousal from anesthesia found in this systematic review and meta-analysis (T5: −7.5 mmHg, P<0.001; T7: −8.2 mmHg, P<0.001; T9: −6.0 mmHg, P<0.001) also supports the implementation of periodic intraoperative position changes or rest periods as a mechanism to help reduce IOP.

Administering specific medications or anesthetics may also be effective in reducing IOP or mitigating the intraoperative increase in IOP (50,51,58,64,67,71-74). Agrawal et al. (50) found that induction and maintenance of anesthesia with propofol was the most effective option for mitigating the increase in IOP in adult patients undergoing surgery in the Trendelenburg position. Likewise, Kaur et al. (51) found that propofol-based total intravenous anesthesia was more effective than inhalational anesthesia with sevoflurane in mitigating the increase in IOP in patients undergoing laparoscopic surgery in the Trendelenburg position. Kitamura et al. (57) found that continuous administration of dexmedetomidine in combination with propofol-based total intravenous anesthesia decreased IOP in patients undergoing robotic-assisted laparoscopic radical prostatectomy. Molloy and Cong (60) found that intraoperative treatment with dorzolamide-timolol eyedrops significantly reduced elevated IOP in patients undergoing lengthy laparoscopic procedures in the Trendelenburg position, while Molloy et al. (61) found that prophylactic therapy with dorzolamide-timolol eyedrops significantly reduced IOP in patients undergoing robotic-assisted laparoscopic prostate and gynecologic procedures.

Another important consideration for practice is the need to evaluate whether patients undergoing surgery in the Trendelenburg position should receive a preoperative ophthalmologic examination to reduce the risk for ocular injury (5,67). Preoperative ophthalmologic examinations may be helpful in identifying patients at risk for postoperative vision loss or other ocular complications. Increases in IOP may be more harmful in older patients or patients who are predisposed to developing glaucoma than in younger, healthier patients (5,9,10,16).

Grosso et al. (75), recommend developing a risk assessment model to help identify individuals who may be at risk for ocular complications following prolonged laparoscopic or robotic-assisted laparoscopic procedures. In selected patients (e.g., individuals with a diagnosis of glaucoma or those who are currently using glaucoma medications, individuals with a history of ocular trauma or surgery, individuals undergoing minimally invasive or robotic-assisted laparoscopic surgery anticipated to last longer than 1 hour), collaboration with an ophthalmologist to monitor intraoperative IOP and recommend treatment for increases in IOP is prudent.

Implications for future research

Further research relative to the magnitude of IOP increases in patients undergoing surgery in the Trendelenburg position is warranted. To allow for consistent data collection, comparison, meta-analysis, and reporting, researchers of future studies should use standardized time points for measurement (i.e., before arousal, after arousal, after abdominal insufflation, after change to Trendelenburg position and every 30 to 60 minutes thereafter, after return to supine position, before arousal, and postoperatively). Further, researchers should present data in a consistent format for each time point (i.e., sample size, mean, standard deviation). Additionally, to determine whether certain variables affect the strength of the relationship between Trendelenburg position and IOP, researchers should include patients of all ages (e.g., children, older adults), without restricting patients based on BMI or comorbidities.

Limitations

This systematic review and meta-analyses have several limitations. The literature search yielded 107 potentially eligible studies. Studies were excluded for a variety of reasons (Figure 1); however, 25 studies were excluded from the analyses because of a lack of data necessary to calculate an effect size. The researchers were contacted a minimum of two times to obtain missing data, but most did not respond. Some of the researchers excluded participants based on age, BMI, and comorbidities; therefore, the mean values for these variables may not fully reflect the characteristics of all adult surgical patients. Because researchers measured IOP at different intraoperative time points, all studies could not be included at all time points examined in the meta-analysis. Likewise, there were not enough studies included at each time point to allow for moderator analyses. With the exception of T6 and T8, heterogeneity was significant, indicating that variation across studies was substantial, potentially limiting generalizability. The non-significant Egger’s regression intercept (bias = −0.05; P=0.47) is indicative of the absence of bias in the studies included in the meta-analysis for T6 (s=12); however, the Egger’s test has low power for meta-analyses containing small to moderate numbers of studies (43).

IOP can be affected by a number of factors (e.g., age, gender, blood pressure, BMI, total cholesterol, low-density lipoprotein, high-density lipoprotein, blood glucose, diabetes, smoking) (76). Although the researchers conducting each study included in the meta-analysis attempted to control for factors other than the effect of Trendelenburg position on IOP, it is possible that the increase in IOP was affected by factors other than patient position.

There is a potential for IOP to be affected by anesthetic agents and positive end-expiratory pressure (PEEP) greater than 15 cmH₂O (77). In all of the studies except one (20), a standardized protocol for anesthetic administration was used as a mechanism to help control potentially confounding variables. Succinylcholine is an anesthetic agent known to increase IOP (77). This agent was used for induction by only one researcher (54). Propofol, pentothal, opiates, and volatile anesthetic agents are known to reduce IOP (77). Rocuronium has no effect on IOP (78). Therefore, the primary confounding effect of these anesthetic agents, if present, would be to either lower or have no effect on IOP. This means that the actual increase in IOP could be even greater than the increase shown in the meta-analysis. As well, the variety of anesthetic agents used by the researchers reduces the likelihood of the anesthetic agents being a confounding factor in the meta-analysis (Table 4).

None of the studies included in the meta-analysis discuss the PEEP levels used during general anesthesia, so there is no way to know if these levels were greater than 15 cmH₂O. Notably, Blecha et al. (55) found that IOP doubled and the optic nerve sheath diameter increased to values indicative of increased intracranial pressure when the patient was in a 45-degree Trendelenburg position during robotic-assisted laparoscopic prostatectomy. The researchers concluded that this nonsignificant increase in IOP could be due to the effect of PEEP.

Pneumoperitoneum can also cause an increase in IOP (75,79). Uno et al. (79) found there was a profound increase in IOP during abdominal insufflation with Trendelenburg position. The researchers speculated this was due to the effect of increased central venous pressure, abdominal insufflation, and position change. In the studies included in this meta-analysis, abdominal insufflation occurred while the patient was in the supine position. After abdominal insufflation, the trocars were inserted and the patient was moved to a Trendelenburg position.

Grosso et al. (75) conducted a prospective study to determine the effects of pneumoperitoneum used during colorectal procedures performed in Trendelenburg and reverse Trendelenburg positions (degree of Trendelenburg not specified). The researchers measured IOP at eight different intraoperative timepoints and used optimal coherence tomography to measure the thickness of the peripapillary retinal nerve fiber layer (RNFL) before and after surgery. They found that pneumoperitoneum (12 to 14 mmHg CO₂) resulted in a mean increase of 4 mmHg in IOP. Notably, there was no statistically significant difference in the increase in IOP during surgery between the two groups (P>0.05), and no statistically significant difference in RNFL thickness before or after surgery (P>0.05). Additionally, there were no statistically significant associations between time in the Trendelenburg position and an increase in IOP (right eye P=0.786; left eye P=0.668). The researchers emphasized the importance of this finding because it demonstrates that the increase in IOP caused by pneumoperitoneum is a temporary, reversible occurrence.

The results of this meta-analysis showed there was a significant increase in IOP after insufflation in the supine position (+3.5 mmHg, P<0.001). There is no way to separate the effect of maintaining pneumoperitoneum from the effect of Trendelenburg position on IOP and it is likely that some of the increase in IOP was due to maintaining abdominal insufflation. However, the greatest increase in IOP occurred immediately after the patient was placed into the Trendelenburg position (Figure 2; T3: +4.4 mmHg). The IOP continues to increase significantly while the patient is in Trendelenburg position, but to a lesser degree (T4: +2.6 mmHg; T6: +1.5 mmHg; T8: +1.6 mmHg). This supports the concept that the increase in IOP is primarily due to the Trendelenburg position. The meta-analysis included both laparoscopic and robotic-assisted laparoscopic procedures. Some surgeons may use much steeper degrees of Trendelenburg during robotic-assisted laparoscopic procedures, leading to a greater increase in IOP for these procedures than for laparoscopic procedures. Only one study of robotic-assisted laparoscopic procedures included in the meta-analysis used a steep 45-degree Trendelenburg position (55). The mean degree of Trendelenburg for all studies was 28.4 (±6.5; range, 17.5–45). The mean degree of Trendelenburg for laparoscopic procedures was 28.1 (±6.3; range, 17.5–36). The mean degree of Trendelenburg for robotic-assisted laparoscopic procedures was only minimally greater at 31.2 (±6.3; range, 23–45), and therefore, the difference in the degree of Trendelenburg used for laparoscopic compared with robotic-assisted laparoscopic procedures was not likely to substantively influence IOP levels.

Conclusions

IOP increases significantly between abdominal insufflation in supine position and 240 minutes of Trendelenburg position. The greatest increase in IOP occurs within 5 minutes of placing the patient into the Trendelenburg position. The IOP continues to increase significantly while the patient is in Trendelenburg position, but to a lesser degree. IOP increases of the magnitude found in this systematic review and meta-analysis demonstrate the need for implementing intraoperative interventions to mitigate the increase in IOP and reduce the risk for postoperative vision loss and other ocular complications in patients undergoing surgery in the Trendelenburg position.

Acknowledgments

The author wishes to thank Jo-Ana D. Chase, PhD, APRN-BC, Vicki S. Conn, PhD, RN, FAAN, Deidre D. Wipke-Tevis, PhD, RN, Sinclair School of Nursing, and Christine Proulx, PhD, Department of Human Development and Family Science, University of Missouri, Columbia, for their support and guidance in developing this systematic review and meta-analysis; Marie Bashaw, DNP, RN, NEA-BC, Department of Nursing, Hartwick College, Oneonta, New York, for her assistance in providing independent review and appraisal of the studies included in this systematic review and meta-analysis; Danielle Karasko, MSN, CRNP, PhD candidate, Sinclair School of Nursing, University of Missouri, Columbia, and Philadelphia Children’s Hospital, Pennsylvania, for her assistance in providing independent coding of the studies included in this systematic review and meta-analysis; and Rebecca Graves, Health Science Library, University of Missouri, Columbia, for her assistance with developing effective search terms and strategies.

Funding: None.

Footnote

Conflicts of Interest: The author has completed the ICMJE uniform disclosure form (available at http://dx.doi.org/10.21037/ales.2019.07.09). The author has no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. MacDonald JJ, Washington SJ. Positioning the surgical patient. Anaesth Intensive Care Med 2012;13:528532 [Crossref]
2. Akhavan A, Gainsburg DM, Stock JA. Complications associated with patient positioning in urologic surgery. Urology 2010;76:1309-16. [Crossref] [PubMed]
3. Astuto M, Minardi C, Uva MG, et al. Intraocular pressure during laparoscopic surgery in paediatric patients. Br J Ophthalmol 2011;95:294-5. [Crossref] [PubMed]
4. Awad H, Santilli S, Ohr M, et al. The effects of steep Trendelenburg positioning on intraocular pressure during robotic radical prostatectomy. Anesth Analg 2009;109:473-8. [Crossref] [PubMed]
5. Borahay MA, Patel PR, Walsh TM, et al. Intraocular pressure and steep Trendelenburg during minimally invasive gynecologic surgery: is there a risk? J Minim Invasive Gynecol 2013;20:819-24. [Crossref] [PubMed]
6. Cullen A, Ferguson A. Perioperative management of the severely obese patient: a selective pathophysiological review. Can J Anaesth 2012;59:974-96. [Crossref] [PubMed]
7. Ghomi A. Robotics in practice: new angles on safer positioning. Contemp OB/GYN 2012;57:26-37.
8. Kan KM, Brown SE, Gainsburg DM. Ocular complications in robotic-assisted prostatectomy: a review of pathophysiology and prevention. Minerva Anestesiol 2015;81:557-66. [PubMed]
9. Mondzelewski TJ, Schmitz JW, Christman MS, et al. Intraocular pressure during robotic-assisted laparoscopic procedures utilizing steep Trendelenburg positioning. J Glaucoma 2015;24:399-404. [Crossref] [PubMed]
10. Taketani Y, Mayama C, Suzuki N, et al. Transient but significant visual field defects after robot-assisted laparoscopic radical prostatectomy in deep Trendelenburg position. PLoS One 2015;10:e0123361 [Crossref] [PubMed]
11. American Academy of Ophthalmology. Intraocular pressure. Available online: https://www.aao.org/bcscsnippetdetail.aspx?id=f010bbf6-3f3e-486b-b5cd-0ad86ddb9d74. Updated 2018. Accessed November 10, 2018.
12. Emery SE, Daffner SD, France JC, et al. Effect of head position on intraocular pressure during lumbar spine fusion: a randomized, prospective study. J Bone Joint Surg (Am) 2015;97:1817-23. [Crossref] [PubMed]
13. Ghomi A, Kramer C, Askari R, et al. Trendelenburg position in gynecologic robotic-assisted surgery. J Minim Invasive Gynecol 2012;19:485-9. [Crossref] [PubMed]
14. Gkegkes ID, Karydis A, Tyritzis SI, et al. Ocular complications in robotic surgery. Int J Med Robot 2015;11:269-74. [Crossref] [PubMed]
15. Gould C, Cull T, Wu YX, et al. Blinded measure of Trendelenburg angle in pelvic robotic surgery. J Minim Invasive Gynecol 2012;19:465-8. [Crossref] [PubMed]
16. Grosso A, Scozzari G, Bert F, et al. Intraocular pressure variation during colorectal laparoscopic surgery: standard pneumoperitoneum leads to reversible elevation in intraocular pressure. Surg Endosc 2013;27:3370-6. [Crossref] [PubMed]
17. Hoshikawa Y, Tsutsumi N, Ohkoshi K, et al. The effect of steep Trendelenburg positioning on intraocular pressure and visual function during robotic-assisted radical prostatectomy. Br J Ophthalmol 2014;98:305-8. [Crossref] [PubMed]
18. Lee LA, Newman N. Postoperative visual loss after anesthesia for nonocular surgery. UpToDate. Available online: https://www.uptodate.com/contents/postoperative-visual-loss-after-anesthesia-for-nonocular-surgery. Updated January 16, 2018. Accessed November 10, 2018.
19. Newman NJ. Perioperative visual loss after nonocular surgeries. Am J Ophthalmol 2008;145:604-10. [Crossref] [PubMed]
20. Molloy BL. Implications for postoperative visual loss; steep Trendelenburg position and effects on intraocular pressure. AANA J 2011;79:115-21. [PubMed]
21. Yoo YC, Shin S, Choi EK, et al. Increase in intraocular pressure is less with propofol than with sevoflurane during laparoscopic surgery in the steep Trendelenburg position. Can J Anaesth 2014;61:322-9. [Crossref] [PubMed]
22. Grosso A, Ceruti P, Morino M, et al. Comment on the paper by Mondzelewski and colleagues: “Intraocular pressure during robotic assisted laparoscopic procedures utilizing steep Trendelenburg positioning.” J Glaucoma 2015;24:399-404. J Glaucoma 2017;26:e166-7. [Crossref] [PubMed]
23. Freshcoln M, Diehl MR. Repositioning during robotic procedures to prevent postoperative visual loss. OR Nurse 2014;8:36-41. [Crossref]
24. Mizumoto K, Gosho M, Iwaki M, et al. Ocular parameters before and after steep Trendelenburg positioning for robotic-assisted laparoscopic radical prostatectomy. Clin Ophthalmol 2017;11:1643-50. [Crossref] [PubMed]
25. Nuzzi R, Tridico F. Ocular complications in laparoscopic surgery: review of existing literature and possible prevention and treatment. Semin Ophthalmol 2016;31:584-92. [PubMed]
26. Stang-Veldhouse KN, Yeu E, Rothenberg DM, et al. Unusual presentation of perioperative ischemic optic neuropathy following major spine surgery. J Clin Anesth 2010;22:52-5. [Crossref] [PubMed]
27. Pinkney TD, King AJ, Walter C, et al. Raised intraocular pressure (IOP) and perioperative visual loss in laparoscopic colorectal surgery: a catastrophe waiting to happen? A systematic review of evidence from other surgical specialities. Tech Coloproctol 2012;16:331-5. [Crossref] [PubMed]
28. Berg KT, Harrison AR, Lee MS. Perioperative visual loss in ocular and nonocular surgery. Clin Ophthalmol 2010;4:531-46. [PubMed]
29. Lee LA. Visual loss, venous congestion, and robotic prostatectomies. ASA Mon 2011;75:26-7.
30. Williams GC, Lee AG, Adler HL, et al. Bilateral anterior ischemic optic neuropathy and branch retinal artery occlusion after radical prostatectomy. J Urol 1999;162:1384-5. [Crossref] [PubMed]
31. Weber ED, Colyer MH, Lesser RL, et al. Posterior ischemic optic neuropathy after minimally invasive prostatectomy. J Neuroophthalmol 2007;27:285-7. [Crossref] [PubMed]
32. Mizrahi H, Hugkulston CE, Vyakarnam P, et al. Bilateral ischaemic optic neuropathy following laparoscopic proctocolectomy: a case report. Ann R Coll Surg Engl 2011;93:e53-4. [Crossref] [PubMed]
33. Kumar G, Vyakarnam P. Postoperative vision loss after colorectal laparoscopic surgery. Surg Laparosc Endosc Percutan Tech 2013;23:e87-8. [Crossref] [PubMed]
34. Berlin JA. Invited commentary: benefits of heterogeneity in meta-analysis of data from epidemiological studies. Am J Epidemiol 1995;142:383-7. [Crossref] [PubMed]
35. Thompson SG. Why sources of heterogeneity in meta-analysis should be investigated. BMJ 1994;309:1351-5. [Crossref] [PubMed]
36. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6:e1000097 [Crossref] [PubMed]
37. Vecchio R, MacFayden BV, Palazzo F. History of laparoscopic surgery. Panminerva Med 2000;42:87-90. [PubMed]
38. American Society of Anesthesiologists. ASA Physical Status Classification System. Updated 2018. Available online: https://www.asahq.org/standards-and-guidelines/asa-physical-status-classification-system. Accessed November 10, 2018.
39. Borenstein M, Hedges LV, Higgins J, et al. Comprehensive Meta-Analysis. 2018. Available online: https://www.meta-analysis.com/. Updated 2018. Accessed November 10, 2018.
40. Borenstein M, Hedges LV, Higgins JP, et al. Introduction to Meta-Analysis. West Sussex, UK: John Wiley & Sons, Ltd., 2009:77-86.
41. Cohen J. Power primer. Psychol Bull 1992;112:155-9. [Crossref] [PubMed]
42. Lipsey M, Wilson D. Practical Meta-Analysis: Applied Social Research. 1st edition. Thousand Oaks, CA: Sage, 2001:146-68.
43. Sutton AJ. Publication bias. In: Cooper H, Hedges LV, Valentine JC. editors. The Handbook of Research Synthesis and Meta-Analysis. 2nd edition. New York, NY: Russell Sage Foundation, 2009:435-52.
44. Sterne JA, Egger M, Moher D. Chapter 10: Addressing reporting biases. Available online: https://handbook-5-1.cochrane.org/
45. Egger M, Smith GD, Schneider M, et al. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629-634. [Crossref] [PubMed]
46. Deeks JJ, Higgins JP, Atlman, DG. Chapter 9: Analyzing data and undertaking meta-analysis. In: Deeks JJ, Higgins JPT, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [Updated March 2011]. The Cochrane Collaboration;2011:9.1-9.9. www.handbook.cochrane.org. Accessed November 10, 2018.
47. Higgins JPT, Thompson SG, Decks JJ, Altman DG. Measuring inconsistency in meta-analysis. BMJ 2003;327:557-560. [Crossref] [PubMed]
48. Borenstein M, Hedges LV, Higgins JP, et al. Introduction to Meta-analysis. West Sussex, UK: John Wiley & Sons, Ltd, 2009:107-25.
49. Borenstein M, Hedges LV, Higgins JP, et al. Introduction to Meta-analysis. West Sussex, UK: John Wiley & Sons, Ltd, 2009:127-33.
50. Agrawal M, Dureja V, Verma AP, et al. A comparative study of four combinations of anesthetic drugs for assessing the intraocular pressure changes during gynaecological laparoscopic procedures. Anesth Essays Res 2013;7:319-25. [Crossref] [PubMed]
51. Kaur G, Sharma M, Kalra P, et al. Intraocular pressure changes during laparoscopic surgery in Trendelenburg position in patients anesthetized with propofol-based total intravenous anesthesia compared to sevoflurane anesthesia: a comparative study. Anesth Essays Res 2018;12:67-72. [Crossref] [PubMed]
52. Nishikawa M, Watanabe H, Kurahashi T. Effects of 25- and 30-degree Trendelenburg positions on intraocular pressure changes during robot-assisted radical prostatectomy. Prostate Int 2017;5:135-8. [Crossref] [PubMed]
53. Yoo YC, Kim NY, Shin S, et al. The intraocular pressure under deep versus moderate neuromuscular blockade during low-pressure robot assisted laparoscopic radical prostatectomy in a randomized trial. PLoS One 2015;10:e0135412 [Crossref] [PubMed]
54. Adisa AO, Onakpoya OH, Adenekan AT, et al. Intraocular pressure changes with positioning during laparoscopy. JSLS 2016;20:e2016.00078.
55. Blecha S, Harth M, Schlachetzki F, et al. Changes in intraocular pressure and optic nerve sheath diameter in patients undergoing robotic-assisted laparoscopic prostatectomy in steep 45° Trendelenburg position. BMC Anesthesiol 2017;17:40. [Crossref] [PubMed]
56. Hirooka K, Ukegawa K, Nitta E, et al. The effect of steep Trendelenburg positioning on retinal structure and function during robotic-assisted laparoscopic procedures. J Ophthalmol 2018;2018:1027397 [Crossref] [PubMed]
57. Kitamura S, Takechi K, Nishihara T, et al. Effect of dexmedetomidine on intraocular pressure in patients undergoing robot-assisted laparoscopic radical prostatectomy under total intravenous anesthesia: a randomized, double blinded placebo controlled clinical trial. J Clin Anesth 2018;49:30-5. [Crossref] [PubMed]
58. Mathew DJ, Greene RA, Mahsood YJ, et al. Preoperative brimonidine tartrate 0.2% does not prevent an intraocular pressure rise during prostatectomy in steep Trendelenburg position. J Glaucoma 2018;27:965-70. [PubMed]
59. Molloy BL, Watson C. A comparative assessment of intraocular pressure in prolonged steep Trendelenburg position versus level supine position intervention. J Anesth Clin Sci. 2012. Available online: http://www.hoajonline.com/journals/pdf/2049-9752-1-9.pdf. Accessed November 10, 2018.
60. Molloy B, Cong X. Perioperative dorzolamide-timolol intervention for rising intraocular pressure during steep Trendelenburg positioned surgery. AANA J 2014;82:203-11. [PubMed]
61. Molloy BL, Cong X, Watson C. Preventive dorzolamide-timolol for rising intraocular pressure during steep Trendelenburg position surgery. AANA J 2016;84:189-96. [PubMed]
62. Raz O, Boesel TW, Arianayagam M, et al. The effect of the modified Z Trendelenburg position on intraocular pressure during robotic assisted laparoscopic radical prostatectomy: a randomized, controlled study. J Urol 2015;193:1213-9. [Crossref] [PubMed]
63. Lentschener C, Benhamou D, Niessen F, et al. Intra-ocular pressure changes during gynaecological laparoscopy. Anaesthesia 1996;51:1106-8. [Crossref] [PubMed]
64. Mowafi HA, Al-Ghamdi A, Rushood A. Intraocular pressure changes during laparoscopy in patients anesthetized with propofol total intravenous anesthesia versus isoflurane inhaled anesthesia. Anesth Analg 2003;97:471-4. [Crossref] [PubMed]
65. Centers for Disease Control and Prevention. Defining adult overweight and obesity. Updated 2016. Available online: https://www.cdc.gov/obesity/adult/defining.html. Accessed November 10, 2018.
66. Demasi CL, Porpiglia F, Tempia A, et al. Ocular blood flow in steep Trendelenburg positioning during robotic-assisted radical prostatectomy. Eur J Ophthalmol 2018;28:333-8. [Crossref] [PubMed]
67. Lee M, Dallas R, Daniel C, et al. Intraoperative management of increased intraocular pressure in a patient with glaucoma undergoing robotic prostatectomy in the Trendelenburg position. A A Case Rep 2016;6:19-21. [Crossref] [PubMed]
68. Vitish-Sharma P, Acheson AG, Stead R, et al. Can the SENSIMED Triggerfish® lens data be used as an accurate measure of intraocular pressure? Acta Ophthalmol 2018;96:e242-6. [Crossref] [PubMed]
69. Ozcan MF, Akbulut Z, Gurdal C, et al. Does steep Trendelenburg positioning effect the ocular hemodynamics and intraocular pressure in patients undergoing robotic cystectomy and robotic prostatectomy? Int Urol Nephrol 2017;49:55-60. [Crossref] [PubMed]
70. Kadono Y, Yaegashi H, Machioka K, et al. Cardiovascular and respiratory effects of the degree of head-down angle during robot-assisted laparoscopic radical prostatectomy. Int J Med Robot 2013;9:17-22. [Crossref] [PubMed]
71. Hwang JW, Oh AY, Hwang DW, et al. Does intraocular pressure increase during laparoscopic surgeries? It depends on anesthetic drugs and the surgical position. Surg Laparosc Endosc Percutan Tech 2013;23:229-32. [Crossref] [PubMed]
72. Joo J, Koh H, Lee K, et al. Effects of systemic administration of dexmedetomidine on intraocular pressure and ocular perfusion pressure during laparoscopic surgery in a steep Trendelenburg position: prospective, randomized, double-blinded study. J Korean Med Sci 2016;31:989-96. [Crossref] [PubMed]
73. Joo J, Kim J, Lee J. Effect of continuous systemic administration of esmolol on intraocular pressure during surgery in a sustained steep Trendelenburg position. J Glaucoma 2017;26:1068-71. [Crossref] [PubMed]
74. Kim NY, Yoo YC, Park HJ, et al. The effect of dexmedetomidine on intraocular pressure increase in patients during robot-assisted laparoscopic radical prostatectomy in the steep Trendelenburg position. J Endourol 2015;29:310-6. [Crossref] [PubMed]
75. Grosso A, Ceruti P, Crowston J, et al. Short-term intraocular pressure spikes induced by pneumoperitoneum are safe for the optic nerve in a healthy eye. Eur Ophthal Rev 2018;12:60-5. [Crossref]
76. Kim YJ, Chun YS, Lee MY, et al. Association of IOP with systemic factors in a Korean cohort. Optom Vis Sci 2015;92:1182-8. [Crossref] [PubMed]
77. Nathan N. The eyes have it: factors that influence intraocular pressure (IOP). Anesthesia and analgesia 2018;126:1431. [Crossref] [PubMed]
78. Arunakirinathan M, Papmichael E, Hacking R. Use of suxamethonium to open eye injuries: a dilemma with explosive consequences. Br J Hosp Med (Lond) 2014;75:418. [Crossref] [PubMed]
79. Uno T, Hattori S, Itol K, et al. Intra-ocular pressure changes during laparoscopic cholecystectomy Masui 1994;43:1899-902. [abstract]. [PubMed]