Postoperative vision loss (POVL) in the setting of non-ophthalmic surgery is a rare but disastrous event that can be permanently debilitating for the patient. This complication has been reported in patients undergoing surgery in the prone position but also in the steep Trendelenburg position.1–7 The most common type of POVL after non-ocular surgery is known to be ischemic optic neuropathy (ION),5,8 which is thought to be caused by venous congestion of the orbit and decreased ocular perfusion pressure (OPP).3 It is known that increased venous pressure within the eye can possibly lead to an increase in intraocular pressure (IOP), and OPP is estimated as the difference between mean arterial blood pressure and IOP.3 In this sense, elevated IOPs may serve as a surrogate marker of ocular venous congestion and decreased perfusion of the optic nerve. The existing literature indicates that laparoscopic surgery performed with intraperitoneal carbon dioxide (CO2) insufflation results in a typical state of increased IOP.9,10 This increase in IOP is known to be a time-dependent phenomenon which is considerably worsened when combined with a steep Trendelenburg position.1–3,7

Previous studies have reported a decrease in IOP with intravenous hypnotic agents, inhalation anesthetics, and opioids;9,11–15 however, there is controversy regarding the superiority of propofol over inhalation anesthetics in suppressing the increase in IOP. While some reported that propofol was more effective than sevoflurane13 and isoflurane,9 others reported otherwise.14–16 Robot-assisted laparoscopic radical prostatectomy (RALRP) is performed in the steep Trendelenburg position and often requires a prolonged duration of CO2 pneumoperitoneum, which may place the patient at risk for ocular complications due to increased IOP and decreased OPP. This randomized controlled trial was conducted to evaluate and compare the changes in IOP in patients receiving RALRP under general anesthesia with either propofol-based total intravenous anesthesia (TIVA) or inhalation anesthesia with sevoflurane.

Methods

The study protocol was approved by the Institutional Review Board and Hospital Research Ethics Committee of Severance Hospital, Yonsei University Health System, Seoul, Korea on April 28, 2011. All participants were recruited at the anesthesiology preoperative evaluation clinic, and all provided written informed consent. Sixty-six adult patients who were over the age of 50 yr, American Society of Anesthesiologists physical status I or II, and scheduled for RALRP using the da Vinci™ Robot System (Intuitive Surgical, Inc., Mountain View, CA, USA) under general anesthesia were enrolled from May 2011 to March 2012. An ophthalmologist not involved in the study conducted a preoperative evaluation on all patients to measure preoperative IOP, examine the optic fundus, and screen for any corneal abnormalities or ophthalmologic disorders, such as glaucoma, cataract, macular degeneration, and retinopathies. Exclusion criteria included patients with baseline IOPs > 30 mmHg, diabetic retinopathy, cataract, known allergies to anesthetic drugs, uncontrolled hypertension, unstable angina, or congestive heart failure. Patients who had received surgery or medications for treatment of previously diagnosed glaucoma were also excluded. A computer-generated table of random numbers placed in sealed envelopes was used to assign patients immediately and randomly in a 1:1 ratio for parallel treatment arms to either the sevoflurane-based inhalation anesthesia (Sevoflurane Group, n = 33) or the propofol-based TIVA (Propofol Group, n = 33). No stratification or block randomization was done. All study investigators, except the anesthesiologist who took part in the operation, were blinded to group assignment.

On the day of the operation, patients were premedicated with intramuscular midazolam 0.05 mg·kg−1 and glycopyrrolate 0.2 mg at one hour and just before the induction of anesthesia, respectively. The patients arrived at the operating room with a secured intravenous access through which they received lactated Ringer’s solution at a rate of 5 mL·kg−1·hr−1. Infusion rates were tailored to each patient according to intraoperative urine output and blood loss. Standard monitoring devices were applied, and a 20G radial artery catheter was inserted under local anesthesia before induction of anesthesia for continuous blood pressure monitoring and repeated arterial blood gas analysis. Anesthesia was induced with propofol and remifentanil. After loss of consciousness, rocuronium 0.6 mg·kg−1 was administered to facilitate tracheal intubation. Controlled ventilation was performed with 40% oxygen in air to maintain end-tidal CO2 (ETCO2) at 35-40 mmHg, and a positive end-expiratory pressure of 5 cmH2O was applied. Body temperature was maintained at 36-37°C using a forced-air warming system. After induction of anesthesia, a central venous catheter was inserted via the right internal jugular vein for monitoring of central venous pressure (CVP).

A remifentanil infusion was administered in all patients using a target-controlled infusion system (Orchestra® Base Primea, Fresenius Vial, France) according to the Minto model,17 and the effect-site concentration was adjusted in the range of 2-5 ng·mL−1. In the Sevoflurane Group, end-tidal sevoflurane concentrations were maintained at 1.5-2.5% after induction with propofol 1.5 mg·kg−1. In the Propofol Group, propofol was infused using a target-controlled infusion system (Orchestra® Base Primea, Fresenius Vial, France) for induction and maintenance of anesthesia according to the Marsh model,18 and the effect-site concentration was kept in the range of 2-5 μg·mL−1. The anesthetics administered in each group were titrated to maintain mean arterial pressure (MAP) and heart rate within 20% of baseline values and to provide adequate depth of anesthesia. A bispectral index score (BIS) monitor (Aspect A-2000®, Aspect Medical System Inc., Newton, MA, USA) was used, and the BIS was maintained in the range of 40-60. Additional boluses of rocuronium 0.15 mg·kg−1 were given as needed to maintain adequate intraoperative neuromuscular blockade. At the end of surgery, reversal of neuromuscular blockade was performed with neostigmine 50 μg·kg−1 and glycopyrrolate 10 μg·kg−1intravenously. During the operation, CO2 pneumoperitoneum was induced to maintain a mean (SD) intra-abdominal pressure of 15 (5) mmHg using a CO2 insufflator in a 30° Trendelenburg position. The Trendelenburg angle was accurately adjusted by the operating table controller which showed the tilted angle in numbers.

The arterial line transducer was placed at patient eye level and moved as appropriate whenever the patient’s position was changed. Central venous pressure was measured with respect to heart level. Intraocular pressure was measured by a single ophthalmologist using a handheld applanation tonometer (Tono-Pen® XL, Medtronic, Jacksonville, FL, USA); the ophthalmologist was blinded to the anesthetic given. Both MAP and IOP were measured at nine predefined time points (Table 1). The IOP was measured three times at each time point, and the median value was retained for analysis. The primary outcome measurement was the change in IOP from baseline (T0) to 30 min after being positioned in the steep Trendelenburg position with CO2 pneumoperitoneum (T3). For each patient, we recorded whether they experienced an abnormally high IOP at any time. Normal IOP was defined as being < 21 mmHg,19 while IOP ≥ 24 mmHg was defined as the treatment indication of glaucoma.20 Ocular perfusion pressure was calculated as the difference between IOP and MAP measured at eye level.21 Peak inspiratory pressures (PIP), CVP, and end-tidal CO2 (ETCO2) were measured at T1 to T5. Blood was drawn for arterial blood gas analysis at T1, T3, and T5. Recording of vital signs and data analyses were performed by a person who was blinded to patient allocation.

Table 1 Time points of intraocular pressure measurements
Full size table

Dichotomous variables are shown as numbers (percentages) and continuous variables are shown as mean (SD). Dichotomous variables were compared using the Chi square test. Repeated measured variables, such as IOP and OPP, were analyzed using a linear mixed model with patient indicator as a random effect and group, time, and group-by-time as fixed effects. The group-by-time interaction assesses whether the change over time differs between groups. Between-group comparisons of continuous variables other than those previously mentioned were performed by Student’s two-sample t test. The changes in IOP at T3 compared with T0 were analyzed with the linear mixed model. Post-hoc analysis with Bonferroni correction was done for multiple comparisons when repeated measured variables showed a significant difference between the two groups. Linear mixed models were used to assess correlation between IOP and PIP, CVP and ETCO2 over the time period T1-T5, and P values were calculated using Fisher’s z transformation. All statistical tests were two-tailed, and P values < 0.05 were considered statistically significant.

The present study was designed to validate the superiority of propofol-based TIVA over sevoflurane anesthesia. In a previous study, the mean (SD) IOP in the propofol-based TIVA group after induction of anesthesia was 6.0 (3.2) mmHg compared with 8.9 (3.4) mmHg in the sevoflurane inhalational anesthesia group.13 To detect a mean (SD) difference in IOP of 2.9 (3.4) mmHg, power estimation analysis suggested that 30 patients per group would be required to obtain a power of 90%, considering a type I error of 0.05. Considering a drop-out rate of 10%, we recruited 33 patients in each group. All statistical analyses were performed using SAS® version 9.2 (SAS Institute Inc., Cary, NC, USA).

Results

Sixty-seven patients were assessed for eligibility, and one patient was excluded after ophthalmologic evaluation due to senile cataract. The remaining 66 patients were randomized into two groups, and RALRP was performed as planned in all patients. Complete data sets from all 66 patients were analyzed without any missing data. One patient in the Sevoflurane Group and one patient in the Propofol Group were found to have normotensive glaucoma at the preoperative visit to the ophthalmologist. The two patients were advised regarding postoperative follow up; however, they were not excluded from the study. There were no clinically important differences between the two groups in demographic data, duration of anesthesia, pneumoperitoneum and Trendelenburg time, and intraoperative fluid input and output (Table 2). Intraoperative arterial blood gas data were also comparable between the two groups at time points T1, T3, and T5. Mean arterial pressure during the perioperative period, intraoperative CVP, respiratory settings, ETCO2, BIS, and effect-site concentration of remifentanil were similar between the two groups (Table 3).

Table 2 Patient characteristics and intraoperative variables
Full size table
Table 3 Vital signs, respiratory settings, end-tidal CO2, bispectral index scores, and anesthetic concentrations
Full size table

At baseline, IOP was similar in both groups, with mean (SD) values being 17.5 (3.6) mmHg in the Sevoflurane Group and 17.9 (3.7) mmHg in the Propofol Group. Six patients in the Sevoflurane Group and eight patients in the Propofol Group were found to have baseline IOPs > 20 mmHg. After induction of anesthesia, IOP decreased in both groups but increased with the establishment of pneumoperitoneum and Trendelenburg position. Values of IOP returned to baseline at one and 24 hr after surgery (Figure A). During maintenance of anesthesia (from T2 to T6), IOP was greater in the Sevoflurane Group than in the Propofol Group (linear mixed model analysis; P = 0.0039).

Figure 1
figure 1

Mean intraocular pressure (IOP) (A) and ocular perfusion pressure (OPP) (B) in the Sevoflurane and Propofol Groups at each time point, as defined in Table 1. Error bars denote the standard deviation. Time points are defined in Table 1. *P < 0.05 compared with the Sevoflurane Group

Full size image

In the Sevoflurane Group, IOP was increased significantly at T3 compared with baseline; the difference was 6.0 (5.0) mmHg (P < 0.001). In the Propofol Group, the increase was 2.1 (5.1) mmHg and did not reach statistical significance (P = 0.136). The IOP values at T3 were 23.5 (4.3) mmHg in the Sevoflurane Group and 19.9 (3.8) mmHg in the Propofol Group (difference = 3.5 mmHg; 95% confidence interval [CI]: 1.5 to 5.5; P = 0.0019). The IOP was also greater in the Sevoflurane Group than in the Propofol Group at T2 [18.4 (5.6) mmHg vs 15 (3.4) mmHg, respectively; P = 0.038] and at T6 [21.4 (5.4) vs 17 (4.7) mmHg, respectively; P = 0.009]. Figure A shows the mean IOP values at each time point in each group. Post-hoc analysis with Bonferonni correction revealed that IOP values at T2, T3, and T6 were significantly lower in the Propofol Group compared with the Sevoflurane Group.

Among patients with a normal IOP at baseline (≤ 20 mmHg), IOP was ≥ 24 mmHg at any time point in 20/27 (74%) patients in the Sevoflurane Group and 8/25 (32%) patients in the Propofol Group (difference 42%; 95% CI: 22 to 62; P = 0.002).

A significant difference in OPP was found between the two groups in the linear mixed model analysis (P = 0.047) (Figure B); however, post-hoc analysis with Bonferroni correction revealed no significant difference at any specific time point. Both CVP and PIP were found to be significantly correlated with IOP between T1 and T5 (r = 0.44; 95% CI: 0.21 to 0.62; P = 0.0003, and r = 0.34; 95% CI: 0.1 to 0.54; P = 0.0064, respectively), while ETCO2 was not (r = 0.0816; P = 0.5264). None of the patients experienced ocular complications.

Discussion

In the present study, we found that propofol-based TIVA is more effective than sevoflurane-based inhalation anesthesia in attenuating IOP increase during RALRP with pneumoperitoneum and steep Trendelenberg. Moreover, propofol-based TIVA may be more helpful in maintaining higher OPP during RALRP compared with sevoflurane-based inhalation anesthesia.

In several previous studies, propofol-based TIVA has been reported to be superior to inhalation anesthesia for lowering IOP. Schafer et al. 13 reported that the decrease in IOP during cataract surgery was significantly greater with propofol-based TIVA than with sevoflurane inhalation anesthesia. Moreover, Mowafi et al. 9 found that propofol-based TIVA was superior to isoflurane inhalation anesthesia in attenuating the increase in IOP during laparoscopic surgery in the Trendelenburg position. Nevertheless, despite the significant increase in IOPs observed with isoflurane anesthesia in their study, all absolute values of IOP were found to be within the normal range (< 20 mmHg). In contrast, the mean IOP levels in the present study were > 20 mmHg in the Sevoflurane group during pneumoperitoneum in the Trendelenburg position. This discrepancy between the previous study and the present study seems to be due to the difference in study population. While Mowafi et al. 9 conducted their study in young female patients with a mean age of 31 yr, most patients in the present study were older than 60 yr. In another study by Awad et al.,7 results showed that IOPs increased beyond 20 mmHg in patients undergoing RALRP in the steep Trendeleburg position with desflurane anesthesia. Patients undergoing RALRP are usually elderly, and therefore, they often present with medical comorbidities. Evidence suggests that vascular diseases, such as atherosclerosis, vasospastic diseases, and diabetes, are related to dysfunctional autoregulation of ocular blood flow, likely contributing to glaucomatous optic neuropathy.22 It is well known that even transient episodes of increased IOP may jeopardize retinal perfusion and cause retinal ischemia in older patients with arteriosclerotic involvement of the retinal artery.23 The incidence of undiagnosed glaucoma is also known to increase with age; therefore, the overall risk of damage to the optic nerve due to elevated IOP may be higher in this patient population.2,6,21,24 Ocular hypertension is reported to be found in 4 ~ 10% of the population over 40 yr of age,25 and the incidence of primary open-angle glaucoma is known to be increased by ten to 15 times in these patients.26 Patients with chronic open-angle glaucoma have reduced capacity for drainage of the aqueous humour; consequently, they have a reduced capacity to compensate for an acute rise in IOP. We found that 14/66 patients in this study actually had IOPs > 20 mmHg at baseline evaluation, and two patients were suspected as having normotensive glaucoma.

The decrease in OPP together with mechanical compression on ganglion cell axons is an important mechanism of optic neuropathy that is caused by increased IOP.2,21,27 The results of the present study have clinical significance given that propofol may decrease the risk of damage to the optic nerve during laparoscopic surgery by attenuating the increase in IOP and, therefore, decreasing the risk of ocular hypoperfusion.

The duration of pneumoperitoneum in the Trendelenburg position, ETCO2, and PIP are all known to contribute to increased IOP.2,7,28 One key determinant of IOP is the volume of the aqueous humour, which depends on the pressure gradient between IOP and episcleral venous pressure for drainage.29 Episcleral pressure increases when CVP increases, eventually causing IOP elevation. Although we found an increase in CVP in both groups following pneumoperitoneum and the Trendelenburg position, CVP was similar in both groups. Mechanical ventilation at high PIP is also known to cause an increase in IOP by increasing intrapulmonary pressure,28 and indeed, the relatively high PIP observed in our study seems to have contributed to the increase in IOP. On the other hand, while hyperventilation is known to decrease IOP,30 we were not able to find any significant relationship between ETCO2 and IOP in the present study. Overall, although the results of our study partially support those of previous reports, we were not able to identify any of the aforementioned factors as the cause of differences in IOP between the two groups as there were no intergroup differences in CVP, PIP, or ETCO2.

Adrenergic stimulation is known to increase IOP by increasing the resistance to the outflow of aqueous humour at the trabecular meshwork between the anterior chamber and Schlemm’s canal by causing vasoconstriction and venoconstriction.31 Patients usually experience high postoperative pain and anxiety, which leads to increased sympathetic tone during the acute postoperative period. This may explain the increased IOP at T6 in the Sevoflurane Group. Interestingly, this was not observed in the Propofol Group, a finding also reported by Sator et al.14 who measured increased IOPs after emergence from anesthesia with desflurane and isoflurane, but not after anesthesia with propofol. Since there is no guarantee that the depth of intraoperative anesthetic can be maintained at the level of MAC-BAR,32 it can be said that adrenergic stimulation exists to a variable degree throughout the operation. Based on the results of the present study, we can only assume that propofol-based TIVA may have a greater ability to alleviate adrenergic stimulation than sevoflurane and thus maintain IOP at a lower level. The underlying interaction between propofol and the activity of the autonomic nervous system during laparoscopic surgery in the steep Trendelenburg position remains to be studied.

This study has limitations. First, OPP was calculated from MAP and IOP, and blood flow was not measured at the optic nerve head. Second, the duration of surgery in this study was relatively short. Longer operation times have been shown to be correlated with ION in spine surgeries in the prone position33 as well as in surgeries performed in the steep Trendelenburg position.7 In this study, the mean duration of pneumoperitoneum in the Trendelenburg position was approximately 90 min; thus, the clinical effect of propofol-based TIVA to reduce IOP in patients undergoing procedures of a prolonged duration is not clear. Further studies conducted in longer robotic procedures are needed to explore this aspect. Lastly, several patients were found to present with extremely elevated IOP, which may not have been due simply to choice of anesthetic. While this may represent a confounding factor, exclusion of these patients would have been a possible cause of bias; therefore, patients with elevated IOP were included for analysis.

In conclusion, we found that propofol-based TIVA was more effective than inhalation anesthesia with sevoflurane in attenuating the increase in IOP during laparoscopic surgery requiring CO2 pneumoperitoneum in the steep Trendelenburg position. Although the primary outcome of this study was not the change in the actual incidence of postoperative ocular complications, IOP seems appropriate as a surrogate marker of ocular venous congestion and can therefore be considered to reflect the risk of ION. Overall, our results seem to have clinical significance given that this study was conducted in patients who are relatively vulnerable to the damaging effects of an elevation in IOP. The use of propofol-based TIVA may be of benefit in this patient population.