Effects of anesthetic depth on postoperative pain and delirium: a meta-analysis of randomized controlled trials with trial sequential analysis : Chinese Medical Journal

Secondary Logo

Journal Logo

Meta Analysis

Effects of anesthetic depth on postoperative pain and delirium: a meta-analysis of randomized controlled trials with trial sequential analysis

Long, Yuqin1,2; Feng, Xiaomei3; Liu, Hong4; Shan, Xisheng1,2; Ji, Fuhai1,2; Peng, Ke1,2

Editor(s): Ni, Jing

Author Information
Chinese Medical Journal ():10.1097/CM9.0000000000002449, January 3, 2023. | DOI: 10.1097/CM9.0000000000002449



Monitoring and maintaining brain function are important in daily anesthesia practice.[1,2] The brain functional indices derived from a processed electroencephalogram, such as bispectral index (BIS), auditory evoked potential index, and spectral entropy, have been utilized to evaluate the depth of anesthesia. The advantages of using these indices include prevention of intraoperative awareness, avoidance of excessive anesthetic depth, reduction of hypnotic agents used, and acceleration of postoperative recovery.[3-6]

Effective pain management is crucial to patients’ rehabilitation after surgery. Whether deep anesthesia alleviates postoperative pain remains unclear. Faiz et al[7] reported that deep anesthesia (BIS values of 35–44) vs. light anesthesia (BIS values of 45–55) led to better pain outcomes after laparoscopic cholecystectomy. However, other studies argued that deep anesthesia did not produce clinically useful analgesic effects.[8,9] There has not yet been a meta-analysis of postoperative pain in relationship with the depth of general anesthesia. Furthermore, postoperative pain and opioid-based analgesia are the risk factors for postoperative delirium (POD).[10]

Perioperative neurocognitive disorders are common and serious complications, particularly in elderly patients undergoing surgery. A consensus has been developed for perioperative cognitive changes, including acute events such as POD and cognitive decline up to 30 days postoperatively (delayed neurocognitive recovery [DNR]) and up to 12 months (postoperative neurocognitive disorder).[11] The effects of anesthesia depth on neurocognitive function are controversial in previous randomized controlled trials (RCTs).[8,12] Moreover, the results from meta-analyses are also conflicting,[13-15] without incorporating recently published trials.[16,17] Regarding postoperative mortality, observational studies and relevant meta-analyses showed that intraoperative low BIS was associated with increased postoperative mortality,[18-21] but a recent RCT did not demonstrate such a causal link.[22]

Therefore, we conducted this systematic review and meta-analysis to evaluate the effects of deep vs. light anesthesia on postoperative pain, cognitive function, recovery from anesthesia, complications, and mortality. We performed the trial sequential analysis (TSA) to assess the primary results and utilized the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach to evaluate the quality of evidence of this study.


Protocol and registration

We prospectively registered the review protocol at PROSPERO International Prospective Register of Systematic Reviews (identifier: CRD42019127973) on April 8, 2019. We conducted this systematic review and meta-analysis by following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement [Supplementary Table 1, https://links.lww.com/CM9/B335].[23]

Search strategy

Two review authors independently performed the literature search in PubMed, EMBASE, and Cochrane CENTRAL databases from inception to February 20, 2021, and the search results were updated on January 6, 2022. We used the following search strategy for PubMed: ((((((((bispectral index [Title/Abstract]) OR (bispectral index monitor [Title/Abstract])) OR (anesthesia depth [Title/Abstract])) OR (anesthetic depth [Title/Abstract])) OR (spectral entropy [Title/Abstract])) OR (depth of anesthesia [Title/Abstract])) OR (bis [Title/Abstract])) AND ((((((((((postoperative outcome) OR (postoperative complication)) OR (complications)) OR (pain)) OR (death)) OR (mortality)) OR (cognitive)) OR (cognition)) OR (delirium)) OR (POCD))) AND “Randomized Controlled Trial”[pt]. The search strategies for all databases are shown in Supplementary Table 2, https://links.lww.com/CM9/B335. We did not use language or other restrictions for the literature search. We manually checked the references of included studies to identify additional records.

Trial selection

We included studies that met the following criteria: (1) study design: RCT; (2) participants: adult patients undergoing cardiac or non-cardiac surgery; (3) intervention: light anesthesia vs. deep anesthesia (a mean between-group difference ≥5 in BIS [0–100] or ≥3 in auditory evoked potential index [0–60]); and (4) postoperative outcomes: pain intensity, cognitive function, postoperative nausea and vomiting (PONV), time to emergence from anesthesia, time to extubation from anesthesia length of stay, postoperative major complications, and mortality. The exclusion criteria were: (1) non-RCT, (2) duplicate publications, (3) surgical procedures under sedation other than general anesthesia, or (4) no specific results. Any discrepancy during the trial selection process was resolved by re-evaluation of the study and group discussion with other review authors.

Data extraction

Two review authors independently extracted data from each study, including the first author's name, publication year, region, type of surgery, type of anesthesia with anesthetic doses, intervention groups, mean age, number of patients, mean or median BIS values, and main outcomes reported. Any discrepancy during the data extraction process was resolved by re-checking the study data and group discussion.

Primary outcomes

The co-primary outcomes were postoperative pain scores at rest at 0–1 h postoperatively and the incidence of POD up to 1 week postoperatively or until discharge. Postoperative pain was measured using the visual analogue scale (VAS, 0–10). POD was assessed using the confusion assessment method. The definitions of perioperative neurocognitive disorders (NCDs, including POD, DNR, and postoperative NCD) are listed in Supplementary Table 3, https://links.lww.com/CM9/B335.

Secondary outcomes

The secondary outcomes were exploratory, including postoperative VAS pain scores at 8 h and 24 h postoperatively, intraoperative sufentanil consumption, postoperative rescue analgesia, persistent pain during 3 to 12 months postoperatively, DNR during 1 to 7 days postoperatively, NCD during 1 to 3 months postoperatively, Mini-mental State Examination (MMSE) scores, time to emergence from anesthesia, time to extubation from anesthesia, orientation recovery time, length of post-anesthesia care unit (PACU) stay, length of intensive care unit stay, length of hospital stay, quality of recovery on postoperative day 1, 90-day physical and mental recovery scores, clinically significant hypotension (necessitating fluid and/or drug intervention), PONV, any major complication, such as myocardial infarction, sepsis, stroke, and wound infection, intraoperative awareness, 1-year cancer recurrence, mortality within 30 to 90 days postoperatively, and 1-year mortality.

Quality assessment

Two review authors independently conducted quality assessments using the Cochrane evaluation tool.[24,25] We evaluated the risk of bias for each study in seven domains: random sequence generation, allocation concealment, blinding of participants and personal information, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. After a judgment of low, high, or unclear risk of bias in each domain, we rated the study to be at a low risk of bias (if all domains were at low risk), a high risk of bias (if high risk in ≥1 domain), or unclear risk of bias (if unclear risk in ≥1 domain without any domain at a high risk). Furthermore, we assessed the quality of evidence for the main outcomes using the GRADE approach.[26,27] We assessed the certainty of evidence in six domains: study design, risk of bias, inconsistency, indirectness, imprecision, and other considerations. Based on the assessment, we rated the level of evidence as high, moderate, low, or very low. Any discrepancy during the quality assessment process was resolved by group discussion with other review authors.

Statistical analysis

We conducted the meta-analysis using the RevMan software (version 5.4, Cochrane Collaboration, Copenhagen, Denmark). For dichotomous outcomes, risk ratios (RRs) with 95% confidence intervals (CIs) were analyzed using the Mantel–Haenszel method. For continuous outcomes, weighted mean differences (WMDs) with 95% CIs were analyzed using the Inverse Variance method. Considering clinical heterogeneities, we applied a random-effects model for data pooling.[28] We used the I2 statistic test to evaluate heterogeneity among studies, with I2 > 50% indicating significant heterogeneity.[25,29] We assessed publication bias using the Begg's rank correlation test and Egger's linear regression test with the STATA software (version 14.0, Stata Corp, College Station, TX, USA).[27,30,31] Begg's funnel plot was also generated for visual inspection. For the two co-primary outcomes, we performed multiple testing using the Bonferroni method, with P < 0.025 indicating a statistical significance (i.e., 0.05/2). We conducted subgroup analyses for pain scores at different postoperative time points and POD according to cardiac or non-cardiac surgery. For the exploratory secondary outcomes, no multiple testing correction was applied.

We assessed the reliability of two primary results using the TSA viewer software (version beta, Copenhagen Trial Unit, Centre for Clinical Intervention Research, Rigshospitalet, Copenhagen, Denmark).[27,32] In a TSA diagram, a Z-curve crossing the trial sequential monitoring boundary or futility boundary suggests that the current evidence is sufficient for a conclusion and that further studies are unlikely to change the inference. On the contrary, a Z-curve not crossing any boundary suggests an insufficient level of evidence. To calculate the monitoring and futility boundaries, the following parameters were used: conventional test boundary (boundary type: two-sided; Type I error = 5%), Alpha-spending boundaries (hypothesis testing [boundary type: two-sided; Type I error = 5%; α-spending function: O’Brien-Fleming; Information axis: sample size], inner wedge [β-spending function: O’Brien-Fleming; Power = 80%], and required information size [information size: estimate; Power = 80%; Heterogeneity correction: model variance based]), and law of the Iterated logarithm (boundary type: two-sided; Type I error = 5%; Penalty = 2.0). We also reported the adjusted 95% CIs by TSA for each outcome.


Literature search

We initially identified a total of 2996 publications. After excluding duplicates and irrelevant articles, 99 studies were included for full-text review. Thereafter, we excluded 73 articles due to non-RCT, pediatric use, BIS not used in the control group, lack of specific outcomes, or surgery performed under sedation and spinal anesthesia. Finally, we included a total of 26 RCTs in this meta-analysis [Figure 1].[7-9,12,16,17,33-52]

Figure 1:
Flowchart of literature inclusion criteria of studies on effects of anesthetic depth on postoperative pain and delirium. BIS: Bispectral index; RCT: Randomized controlled trials.

Trial characteristics

Table 1 shows the trial characteristics. These RCTs were published between 1997 and 2021, involving 10,743 patients undergoing cardiac or non-cardiac surgery. Among 22 trials on non-cardiac surgery, 9 trials used volatile-based anesthesia (sevoflurane, isoflurane, or desflurane),[9,33-36,38-41] 8 trials used total intravenous anesthesia with propofol,[7,8,42,44-48] and 5 trials used propofol anesthesia combined with volatiles.[12,49-52] Two studies included both cardiac and non-cardiac surgeries using volatile-based anesthesia.[16,17] Two studies included patients undergoing cardiac surgery; isoflurane anesthesia was used in one study and propofol anesthesia was used in the other study.[37,43]

Table 1 - Trial characteristics of included studies comparing deep and light anesthesia.
Aneshesia depth

Reference Country Surgery Drug (dose, MAC) Deep group Light group BIS value (deep vs. light) Main outcomes
Abdelmalak et al [33] USA Major non-cardiac Sevoflurane (N/A) 187 (65) 194 (63) 44 vs. 50 Any major complication, myocardial infarction, infection, sepsis, stroke, and death (30 days, 1 year)
Abdelmalak et al [34] USA Major non-cardiac Sevoflurane (N/A) 159 (64) 167 (64) 35 vs. 55 SF-12 physical/mental scores (30 days)
An et al [8] China Microvascular decompression Propofol (1100 mg vs. 655 mg) 40 (45) 40 (49) 38 vs. 58 VAS pain scores (1 day, 2 days) and DNR (5 days)
Chan et al [12] China Major non-cardiac Propofol + volatile (138 mg vs. 136 mg, 0.93% vs. 0.57%) 452 (67.6) 450 (68.1) 39 vs. 53 POD, DNR (7 days), NCDs (3 months), time to emergence, time to extubation, PACU stay, hospital stay, any complication, infection, QoR-9 scores, hypotension, and SF-36 physical/mental scores (3 months)
Cotoia et al [42] Italy Urologic surgery Propofol (7.7 mg · kg−1 · h−1 vs. 5.07 mg · kg−1 · h−1) 32 (60) 32 (65) 38 vs. 45 MMSE scores (15 min) and time to extubation
Evered et al [16] USA Cardiac and major non-cardiac Volatile (0.79% vs. 0.59%) 262 (71.1) 253 (70.8) 38 vs. 51 POD, NCDs (30 days/1 year), PACU stay, hospital stay, and MMSE scores (at discharge)
Faiz et al [7] Iran Laparoscopic cholecystectomy Propofol (627 mg vs. 624 mg) 30 (44.1) 30 (44.7) 35–44 vs. 45–55 VAS pain scores (0 h/8 h/16 h/24 h at rest and on movement), rescue analgesia, and PONV
Farag et al [35] USA Abdominal, spine, and pelvic Isoflurane (N/A) 36 (63.8) 38 (63.9) 39 vs. 51 PACU stay and NCDs (4–6 weeks)
Hou et al [49] China Total knee replacement Propofol + sevoflurane (4.56 mg · kg−1 · h−1 vs. 2.88 mg · kg−1 · h−1) 30 (67.9) 30 (68.5) 42 vs. 63 Time to emergence, time to extubation, DNR (1 day/3 days/7 days), and VAS pain scores (1 day/3 days/7 days)
Jildenstål et al [36] Sweden Ophthalmic Desflurane (3.3% vs. 2.5%) 226 (60.5) 224 (60.0) 12 vs. 18 (AAI) DNR (1 day, 7 days), NCDs (1 month), mortality (1 year), and hypotension
Kunst et al [37] United Kingdom Coronary artery bypass grafting Isoflurane (N/A) 40 (72.0) 42 (71.6) 35 vs. 41 POD, MMSE scores (3–5 days/6 weeks/1 year), infection, ICU stay, and hospital stay
Law et al [9] New Zealand Non-emergent Desflurane (N/A) 66 (42.1) 69 (43.2) 33 vs. 42 VAS pain scores (at PACU, 1 day on movement), hypotension, and morphine consumption (PACU, 24 h)
Lehmann et al [43] Germany Coronary artery bypass grafting Midazolam + propofol (N/A) 33 (65) 33 (65) 35–44 vs. 45–55 Time to extubation, PONV, and hypotension
Quan et al [44] China Abdominal Propofol (1308 mg vs. 1024 mg) 52 (65.6) 53 (63.9) 39 vs. 53 Time to extubation, hospital stay, VAS pain scores (1–2 h), DNR (7 days), NCDs (3 months), PONV, infection, intraoperative awareness, rescue analgesia, death (7 days, 3 months), cancer recurrence (1 year), and persistent pain
Sahni et al [52] India Laparoscopic cholecystectomy Propofol + isoflurane (0.93% vs. 0.90%) 40 (39.5) 40 (38.4) 45 vs. 63 Time to emergence, VAS pain scores (0 h/8 h/16 h/24 h at rest and on movement), PONV, and rescue analgesia
Short et al [50] New Zealand, Australia Major non-cardiac Propofol or volatile (4.0 μg/mL vs. 3.1 μg/mL, 0.98% vs. 0.64%) 61 (74) 64 (72) 39 vs. 48 VAS pain scores (at PACU), PACU stay, hospital stay, QoR-9 scores (1 day/2 days/3 days), any major complication, infection, death (1 year), cancer recurrence (1 year), and hypotension
Short et al [17] International multicenter Cardiac and major non-cardiac Volatile (0.88% vs. 0.62%) 3328 (72) 3316 (72) 39 vs. 47 PACU stay, hospital stay, intraoperative awareness, PONV, myocardial infarction, infection, sepsis, stroke, persistent pain, cancer recurrence (1 year), and death (1 year)
Shu et al [38] China Gynecologic laparoscopic Sevoflurane (N/A) 64 (41) 64 (41.5) 30–40 vs. 50–60 MMSE scores (1 day)
Song et al [39] USA Laparoscopic tubal ligation Sevoflurane (1.2% vs. 0.8%); 15 (27); 15 (28); 44 vs. 60; 42 vs. 60 Time to emergence, time to extubation, PACU stay, and intraoperative awareness
Desflurane (1.5% vs. 0.7%) 15 (26) 15 (26)
Soumpasis et al [40] Greece Major urological Sevoflurane (3.2% vs. 0.9%) 30 (62) 30 (60) 20–30 vs. 50–60 Time to emergence, VAS pain scores (8 h, 24 h at rest and on movement), and rescue analgesia
Valentin et al [45] Brazil Non-cardiac and non-neurological Propofol (1051–1093 mg vs. 855–931 mg) 40 (67.2) 32 (68.7) 38 vs. 49 DNR (3 days, 7 days, 21 days), NCDs (90 days, 180 days), and SF-36 physical/mental scores (21 days, 180 days)
36 (68.0) 32 (69.2) 36 vs. 46
Wong et al [41] Canada Orthopedic Isoflurane (7.7 mL vs. 5.6 mL) 31 (70) 29 (71) 44 vs. 51 Time to emergence, PACU stay, intraoperative awareness, and MMSE scores (30 min/60 min/90 min/120 min, 1 day/2 days/3 days)
Xu et al [51] China Hip arthroplasty Propofol + sevoflurane (412 mg vs. 360 mg) 40 (72.2) 41 (74.3) 40–49 vs. 50–59 Time to extubation, PACU stay, DNR (3 h), MMSE score (3 h), and hypotension
Yang et al [46] China Laparoscopic nephrectomy Propofol (N/A) 32 (49.7) 33 (51.4) 30–40 vs. 50–60 MMSE score (1 day)
Zhang and Nie [47] China Gynecologic laparoscopic Propofol (N/A) 51 (36.6) 48 (37.0) 30–40 vs. 50–60 MMSE score (1 day)
Zhou et al [48] China Colon radical surgery Propofol (N/A) 40 (68.9) 41 (68.3) 41 vs. 51 POD, DNR (1–5 days), and hypotension
Data are shown as n (average age, years). mean; median; range. AAI: Auditory evoked potential index; BIS: Bispectral index; DNR: Delayed neurocognitive recovery; MAC: Minimum alveolar concentration; MMSE: Mini-mental State Examination; N/A: not available; NCDs: Neurocognitive disorders; PACU: Post-anesthesia care unit; POD: Postoperative delirium; PONV: Postoperative nausea and vomiting; QoR: Quality of recovery; SF-12: Health Survey Short Form-12; SF-36: Health Survey Short Form-36; VAS: Visual analogue scale.

Supplementary Figure 1, https://links.lww.com/CM9/B334 depicts the risk of bias in the included studies. Of these, 12 RCTs had a low risk of bias, 13 had an unclear risk of bias, and one had a high risk of bias.

Primary outcomes

The VAS pain score at rest at 0–1 h postoperatively was significantly lower in the deep anesthesia group than that in the light anesthesia group (WMD = −0.72, 95% CI = −1.25 to −0.18, P = 0.009, I2 = 33%; Supplementary Table 4, https://links.lww.com/CM9/B335 and Figure 2A), with a moderate level of GRADE evidence [Supplementary Table 5, https://links.lww.com/CM9/B335]. The mean difference of BIS values between groups was −12 in the included studies. There was no publication bias with the Begg's funnel plot (P = 1.000; Figure 2B) or Egger's test (P = 0.894). In the TSA diagram, the Z-curve (blue) crossed the trial sequential monitoring boundary (red) and conventional benefit boundary (brown), suggesting sufficient evidence for this result [Figure 2C]. For this pain outcome, the adjusted 95% CI by TSA was from −1.32 to −0.12.

Figure 2:
Deep vs. light anesthesia on postoperative pain at rest at 0–1 h postoperatively. (A) forest plot; (B) Begg's funnel plot; and (C) TSA. Pain intensity was assessed using the VAS (0–10). Red lines indicate the trial sequential monitoring boundary; green lines indicate the futility boundary; brown lines indicate the conventional benefit boundary; blue line is the Z-curve; CI: Confidence interval; IV: Inverse variance; RIS: Required information size; SD: Standard deviation; TSA: Trial sequential analysis; VAS: Visual analogue scale; WMD: Weighted mean difference.

The deep anesthesia group had a significantly higher incidence of POD (24.95%) compared with the light anesthesia group (15.92%; risk ratio [RR] = 1.57, 95% CI = 1.28–1.91, P < 0.0001, I2 = 0%; Supplementary Table 4, https://links.lww.com/CM9/B335 and Figure 3A), with a high level of GRADE evidence [Supplementary Table 5, https://links.lww.com/CM9/B335]. The mean difference of BIS values between groups was −11. We did not detect significant publication bias in the Begg's funnel plot (P = 0.308; Figure 3B) or Egger's test (P = 0.196). In the TSA diagram, the Z-curve crossed the monitoring boundary and conventional benefit boundary, suggesting that the current evidence is sufficient [Figure 3C]. For the POD outcome, the adjusted 95% CI by TSA was 1.17–2.09.

Figure 3:
Deep vs. light anesthesia on the incidence of POD. (A) forest plot; (B) Begg's funnel plot; and (C) TSA. Red lines indicate the trial sequential monitoring boundary; green lines indicate the futility boundary; brown lines indicate the conventional benefit boundary; blue line is the Z-curve. CI: Confidence interval; M-H: Mantel-Haenszel; POD: Postoperative delirium; RIS: Required information size; RR: Risk ratio; TSA: Trial sequential analysis.

Secondary outcomes

For the secondary pain outcomes [Supplementary Table 4, https://links.lww.com/CM9/B335], the deep anesthesia group had lower VAS pain scores at rest at 8 h (WMD = −1.16, 95% CI = −1.74 to −0.57, P = 0.0001) and 24 h postoperatively (WMD = −0.50, 95% CI = −0.94 to −0.06, P = 0.03) and on movement at 8 h postoperatively (WMD = −1.25, 95% CI = −1.88 to −0.61, P = 0.0001). There were no between-group differences in VAS pain scores on movement at 24 h postoperatively, intraoperative sufentanil consumption, need for rescue analgesia, and persistent pain during 3–12 months postoperatively. For the secondary cognitive function outcomes [Supplementary Table 4, https://links.lww.com/CM9/B335], the two anesthesia groups were comparable in terms of the incidence of DNR during 1–7 days postoperatively (very low-quality evidence; Supplementary Table 5, https://links.lww.com/CM9/B335), NCD during 1–3 months postoperatively (moderate-quality evidence; Supplementary Table 5, https://links.lww.com/CM9/B335), MMSE scores on postoperative day 1, and MMSE scores during 3–5 days postoperatively.

Regarding postoperative recovery [Supplementary Table 4, https://links.lww.com/CM9/B335], the deep anesthesia group had prolonged time to emergence from anesthesia (WMD = 3.65 min, 95% CI = 1.94–5.36 min, P < 0.0001), time to extubation (WMD = 3.64 min, 95% CI = 1.39–5.90 min, P = 0.002), and orientation recovery time (WMD = 4.51 min, 95% CI = 1.61–7.40 min, P = 0.002). In addition, the deep anesthesia group had prolonged length of PACU stay (WMD = 5.85 min, 95% CI = 2.30–9.41 min P = 0.001; very low-quality evidence; Supplementary Table 5, https://links.lww.com/CM9/B335) and length of hospital stay (WMD = 1.00 day, 95% CI = 0.14–1.86 days, P = 0.02; low-quality evidence; Supplementary Table 5, https://links.lww.com/CM9/B335).

As for postoperative complications and mortality [Supplementary Table 4, https://links.lww.com/CM9/B335], there were no between-group differences in clinically significant hypotension, PONV, any major complication, myocardial infarction, sepsis, stroke, wound infection, intraoperative awareness, 1-year cancer recurrence, mortality within 30–90 days postoperatively, and 1-year mortality. Clinically significant hypotension which necessitated fluid and/or drug intervention was recorded in 23.8% (209/878) and 19.5% (172/881) of patients in the deep and light anesthesia groups, respectively. Two patients in the light anesthesia group experienced intraoperative awareness.

Subgroup analyses

Data on postoperative pain were reported from studies in non-cardiac surgery only. We conducted subgroup analyses for pain scores at different postoperative time points. The VAS pain scores at rest during 24 h postoperatively were significantly lower in the deep anesthesia group than that in the light anesthesia group (WMD = −0.69, 95% CI = −0.97 to −0.40, P < 0.0001, I2 = 31%), without statistically significant subgroup differences (P = 0.210; Supplementary Figure 2, https://links.lww.com/CM9/B334). The VAS pain scores on movement during 24 h postoperatively were also significantly lower in the deep anesthesia group than that in the light anesthesia group (WMD = −0.78, 95% CI = −1.26 to −0.30, P = 0.002, I2 = 43%), without statistically significant subgroup differences (P = 0.110; Supplementary Figure 3, https://links.lww.com/CM9/B334).

For the outcome of POD, the subgroup analysis according to cardiac and non-cardiac surgeries showed that deep anesthesia was associated with a higher incidence of POD after non-cardiac surgery (RR = 1.54, 95% CI = 1.26–1.89, P < 0.0001, I2 = 0%) and cardiac surgery (RR = 8.20, 95% CI = 1.07–62.6, P = 0.04), without significant subgroup differences (P = 0.11; Supplementary Figure 4, https://links.lww.com/CM9/B334).


This meta-analysis included 26 RCTs with 10,743 patients to demonstrate the effects of deep vs. light anesthesia on postoperative pain, cognitive function, postoperative recovery, complications, and mortality. We found that deep anesthesia led to lower postoperative pain but a higher incidence of POD when compared with light anesthesia. The TSA results suggest that the current evidence is sufficient for the two primary outcomes. Based on the GRADE methodology, the level of evidence was moderate for the VAS pain scores and was high for POD. For the secondary outcomes, the deep anesthesia group had reduced pain up to 24 h postoperatively, prolonged recovery from anesthesia, and prolonged hospital stay, without between-group differences in the incidence of DNR, NCD during 1 to 3 months postoperatively, other major complications, or mortality.

Surgical patients experience a peak of acute postoperative pain during the first 24 h after surgery.[53,54] In our meta-analysis, deep anesthesia provided maximum pain relief at 8 h postoperatively, both at rest (a mean reduction of 1.16 points on the 0–10 VAS) and on movement (a mean reduction of 1.25 points). For these pain outcomes, we did not detect significant heterogeneity among the studies. Generally, these effects are not that large, but the comparable intraoperative sufentanil consumption suggested the differences were mainly attributable to the depth of anesthesia. A possible explanation is that general anesthetics such as sevoflurane and propofol attenuate noxious stimuli.[55-59] In addition, our previous meta-analysis did not support that propofol-based anesthesia significantly reduced postoperative pain than volatile-based anesthesia.[60] Therefore, it is the depth of anesthesia, other than the choice of general anesthetics, that plays a part in postoperative analgesic effects.

Perioperative NCD are often characterized by impairment in attention, memory, mental status, and psychomotor function in patients who are undergoing surgical procedures.[11,61] As a form of the acute event, POD typically occurs from hours to days after surgery, increasing the risks of morbidities and reducing the quality of daily living.[62,63] As for the effects of deep vs. light anesthesia on neurocognitive function after surgery, previous meta-analyses have yielded conflicting results.[14,64,65] Lu et al[14] investigated the association between anesthetic depth and postoperative cognitive impairment based on four studies. However, only one study was included for POD, comparing depth of sedation (BIS value of 50 vs. BIS values ≥80) during spinal anesthesia.[66] In our meta-analysis, we excluded studies on different depths of sedation, because we believed that pooling studies with sedation and those with general anesthesia would introduce significant heterogeneities. Miao et al[64] included nine RCTs to suggest that the use of BIS monitoring was not associated with reduced incidences of POD, DNR, and postoperative NCD in older patients. In their study, two trials had a mean difference of BIS values <5 between the BIS-guided and usual care groups.[67,68] In contrast, our meta-analysis included studies with clinically significant separation between BIS values between groups. Regarding long-term neurocognitive outcomes, a recent meta-analysis of 10 RCTs suggested that light vs. deep anesthesia was associated with a reduction in postoperative NCD at 90 days after surgery.[15] As the authors mentioned, their results should be treated with caution due to heterogeneity of outcome measures. In our present meta-analysis, we found no between-group differences in the incidences of DNR during 1–7 days postoperatively and NCD during 1 to 3 months postoperatively. We noted that the included studies used different neurocognitive tests, which introduced significant heterogeneities. Thus, more studies are required to investigate the impact of anesthesia depth on long-term postoperative neurocognitive function.

This meta-analysis has several limitations. First, the BIS targets in the deep and light anesthesia groups were not uniform among the included studies, which may have introduced heterogeneities. To better discriminate the deep and light anesthesia groups, we emphasized a mean between-group difference ≥5 in BIS values in our eligibility criteria. Second, the diagnosis of DNR or NCD during 1–3 months postoperatively was based on different neuropsychological tests, which may have confounded these results. Third, while the VAS pain outcomes and POD have a low heterogeneity, several outcomes (including DNR, postoperative recovery, length of PACU and hospital stays, hypotension, and any major complication) are significantly heterogenous, possibly due to different intravenous or inhalational anesthetics used, varied surgical procedures, and different patient populations. Fourth, individual patient data were not available for our meta-analysis. Finally, although there are moderate to high level of evidence for our primary outcomes, the numbers of included studies and patients are relatively small. Hence, we encourage further studies with a large sample size to ascertain these findings.

In conclusion, deep anesthesia compared with light anesthesia was associated with a moderately reduced pain during the early postoperative period but led to an increased incidence of POD. From the current evidence, the risks of maintaining deep anesthesia outweigh its benefits for patients undergoing surgical procedures.


The authors are grateful to Dr. Talmage D. Egan and Dr. Nathan L. Pace (Department of Anesthesiology, University of Utah Health) for their kind support to this study.


This study was supported in part by grants from the Jiangsu Government Scholarship for Overseas Studies (No. JS-2018–178) and the Six Talent Peaks Project in Jiangsu Province (No. WSN-022).

Conflicts of interest



1. Scheeren TWL, Kuizenga MH, Maurer H, Struys MMRF, Heringlake M. Electroencephalography and brain oxygenation monitoring in the perioperative period. Anesth Analg 2019;128:265–277. doi: 10.1213/ANE.0000000000002812.
2. Constant I, Sabourdin N. The EEG signal: a window on the cortical brain activity. Paediatr Anaesth 2012;22:539–552. doi: 10.1111/j.1460-9592.2012.03883.x.
3. Bruhn J, Myles PS, Sneyd R, Struys MMRF. Depth of anaesthesia monitoring: What's available, what's validated and what's next? Br J Anaesth 2006;97:85–94. doi: 10.1093/bja/ael120.
4. Liu SS. Effects of bispectral index monitoring on ambulatory anesthesia: a meta-analysis of randomized controlled trials and a cost analysis. Anesthesiology 2004;101:311–315. doi: 10.1097/00000542-200408000-00010.
5. Punjasawadwong Y, Phongchiewboon A, Bunchungmongkol N. Bispectral index for improving anaesthetic delivery and postoperative recovery. Cochrane Database Syst Rev 2014;2014:Cd003843. doi: 10.1002/14651858.CD003843.pub3.
6. Kreuer S, Bruhn J, Stracke C, Aniset L, Silomon M, Larsen R, et al. Narcotrend or bispectral index monitoring during desflurane-remifentanil anesthesia: a comparison with a standard practice protocol. Anesth Analg 2005;101:427–434. doi: 10.1213/01.ANE.0000157565.00359.E2.
7. Faiz SHR, Siamdoust SAS, Rahimzadeh P, Houshmand L. An investigation into the effect of depth of anesthesia on postoperative pain in laparoscopic cholecystectomy surgery: a double-blind clinical trial. J Pain Res 2017;10:2311–2317. doi: 10.2147/JPR.S142186.
8. An J, Fang Q, Huang C, Qian X, Fan T, Lin Y, et al. Deeper total intravenous anesthesia reduced the incidence of early postoperative cognitive dysfunction after microvascular decompression for facial spasm. J Neurosurg Anesthesiol 2011;23:12–17. doi: 10.1097/ANA.0b013e3181f59db4.
9. Law CJ, Jacobson GM, Kluger M, Chaddock M, Scott M, Sleigh JW. Randomized controlled trial of the effect of depth of anaesthesia on postoperative pain. Br J Anaesth 2014;112:675–680. doi: 10.1093/bja/aet419.
10. Subramaniam B, Shankar P, Shaefi S, Mueller A, O’Gara B, Banner-Goodspeed V, et al. Effect of intravenous acetaminophen vs placebo combined with propofol or dexmedetomidine on postoperative delirium among older patients following cardiac surgery: the DEXACET randomized clinical trial. JAMA 2019;321:686–696. doi: 10.1001/jama.2019.0234.
11. Evered L, Silbert B, Knopman DS, Scott DA, DeKosky ST, Rasmussen LS, et al. Recommendations for the nomenclature of cognitive change associated with anaesthesia and surgery-2018. Anesthesiology 2018;129:872–879. doi: 10.1097/ALN.0000000000002334.
12. Chan MTV, Cheng BCP, Lee TMC, Gin T. BIS-guided anesthesia decreases postoperative delirium and cognitive decline. J Neurosurg Anesthesiol 2013;25:33–42. doi: 10.1097/ANA.0b013e3182712fba.
13. Bocskai T, Kovács M, Szakács Z, Gede N, Hegyi P, Varga G, et al. Is the bispectral index monitoring protective against postoperative cognitive decline? A systematic review with meta-analysis. PLoS One 2020;15:e0229018. doi: 10.1371/journal.pone.0229018.
14. Lu X, Jin X, Yang S, Xia Y. The correlation of the depth of anesthesia and postoperative cognitive impairment: a meta-analysis based on randomized controlled trials. J Clin Anesth 2018;45:55–59. doi: 10.1016/j.jclinane.2017.12.002.
15. Li Y, Zhang B. Effects of anesthesia depth on postoperative cognitive function and inflammation: a systematic review and meta-analysis. Minerva Anestesiol 2020;86:965–973. doi: 10.23736/S0375-9393.20.14251-2.
16. Evered LA, Chan MTV, Han R, Chu MHM, Cheng BP, Scott DA, et al. Anaesthetic depth and delirium after major surgery: a randomised clinical trial. Br J Anaesth 2021;127:704–712. doi: 10.1016/j.bja.2021.07.021.
17. Short TG, Campbell D, Frampton C, Chan MTV, Myles PS, Corcoran TB, et al. Anaesthetic depth and complications after major surgery: an international, randomised controlled trial. Lancet 2019;394:1907–1914. doi: 10.1016/S0140-6736(19)32315-3.
18. Sessler DI, Sigl JC, Kelley SD, Chamoun NG, Manberg PJ, Saager L, et al. Hospital stay and mortality are increased in patients having a “triple low” of low blood pressure, low bispectral index, and low minimum alveolar concentration of volatile anesthesia. Anesthesiology 2012;116:1195–1203. doi: 10.1097/ALN.0b013e31825683dc.
19. Lindholm ML, Traff S, Granath F, Greenwald SD, Ekbom A, Lennmarken C, et al. Mortality within 2 years after surgery in relation to low intraoperative bispectral index values and preexisting malignant disease. Anesth Analg 2009;108:508–512. doi: 10.1213/ane.0b013e31818f603c.
20. Zorrilla-Vaca A, Healy RJ, Wu CL, Grant MC. Relation between bispectral index measurements of anesthetic depth and postoperative mortality: a meta-analysis of observational studies. Can J Anaesth 2017;64:597–607. doi: 10.1007/s12630-017-0872-6.
21. Liu YH, Qiu DJ, Jia L, Tan JT, Kang JM, Xie T, et al. Depth of anesthesia measured by bispectral index and postoperative mortality: a meta-analysis of observational studies. J Clin Anesth 2019;56:119–125. doi: 10.1016/j.jclinane.2019.01.046.
22. Sessler DI, Turan A, Stapelfeldt WH, Mascha EJ, Yang D, Farag E, et al. Triple-low alerts do not reduce mortality: a real-time randomized trial. Anesthesiology 2019;130:72–82. doi: 10.1097/ALN.0000000000002480.
23. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009;339:b2535. doi: 10.1136/bmj.b2535.
24. Higgins JPT, Altman DG, G⊘tzsche PC, Jüni P, Moher D, Oxman AD, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011;343:d5928. doi: 10.1136/bmj.d5928.
25. Cumpston M, Li T, Page MJ, Chandler J, Welch VA, Higgins JP, et al. Updated guidance for trusted systematic reviews: a new edition of the Cochrane Handbook for Systematic Reviews of Interventions. Cochrane Database Syst Rev 2019;10:Ed000142. doi: 10.1002/14651858.ED000142.
26. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008;336:924–926. doi: 10.1136/bmj.39489.470347.AD.
27. Peng K, Li D, Applegate RL, 2nd, Lubarsky DA, Ji FH, Liu H. Effect of dexmedetomidine on cardiac surgery-associated acute kidney injury: a meta-analysis with trial sequential analysis of randomized controlled trials. J Cardiothorac Vasc Anesth 2020;34:603–613. doi: 10.1053/j.jvca.2019.09.011.
28. Subramani Y, Nagappa M, Kumar K, Mortuza R, Fochesato LA, Chohan MBY, et al. Medications for the prevention of pruritus in women undergoing cesarean delivery with Intrathecal morphine: a systematic review and bayesian network meta-analysis of randomized controlled trials. J Clin Anesth 2021;68:110102. doi: 10.1016/j.jclinane.2020.110102.
29. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–560. doi: 10.1136/bmj.327.7414.557.
30. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–1101. doi: 10.2307/2533446.
31. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–634. doi: 10.1136/bmj.315.7109.629.
32. Brok J, Thorlund K, Wetterslev J, Gluud C. Apparently conclusive meta-analyses may be inconclusive – Trial sequential analysis adjustment of random error risk due to repetitive testing of accumulating data in apparently conclusive neonatal meta-analyses. Int J Epidemiol 2009;38:287–298. doi: 10.1093/ije/dyn188.
33. Abdelmalak BB, Bonilla A, Mascha EJ, Maheshwari A, Wilson Tang WH, You J, et al. Dexamethasone, light anaesthesia, and tight glucose control (DeLiT) randomized controlled trial. Br J Anaesth 2013;111:209–221. doi: 10.1093/bja/aet050.
34. Abdelmalak BB, You J, Kurz A, Kot M, Bralliar T, Remzi FH, et al. The effects of dexamethasone, light anesthesia, and tight glucose control on postoperative fatigue and quality of life after major noncardiac surgery: a randomized trial. J Clin Anesth 2019;55:83–91. doi: 10.1016/j.jclinane.2018.12.038.
35. Farag E, Chelune GJ, Schubert A, Mascha EJ. Is depth of anesthesia, as assessed by the bispectral index, related to postoperative cognitive dysfunction and recovery? Anesth Analg 2006;103:633–640. doi: 10.1213/01.ane.0000228870.48028.b5.
36. Jildenstål PK, Hallén JL, Rawal N, Gupta A, Berggren L. Effect of auditory evoked potential-guided anaesthesia on consumption of anaesthetics and early postoperative cognitive dysfunction: A randomised controlled trial. Eur J Anaesthesiol 2011;28:213–219. doi: 10.1097/EJA.0b013e328340dbb9.
37. Kunst G, Gauge N, Salaunkey K, Spazzapan M, Amoako D, Ferreira N, et al. Intraoperative optimization of both depth of anesthesia and cerebral oxygenation in elderly patients undergoing coronary artery bypass graft surgery – A randomized controlled pilot trial. J Cardiothorac Vasc Anesth 2019;34:1172–1181. doi: 10.1053/j.jvca.2019.10.054.
38. Shu AH, Wang Q, Chen XB. Effect of different depths of anesthesia on postoperative cognitive function in laparoscopic patients: A randomized clinical trial. Curr Med Res Opin 2015;31:1883–1887. doi: 10.1185/03007995.2015.1075968.
39. Song D, Joshi GP, White PF. Titration of volatile anesthetics using bispectral index facilitates recovery after ambulatory anesthesia. Anesthesiology 1997;87:842–848. doi: 10.1097/00000542-199710000-00018.
40. Soumpasis I, Kanakoudis F, Vretzakis G, Arnaoutoglou E, Stamatiou G, Iatrou C. Deep anaesthesia reduces postoperative analgesic requirements after major urological procedures. Eur J Anaesthesiol 2010;27:801–806. doi: 10.1097/EJA.0b013e328337cbf4.
41. Wong J, Song D, Blanshard H, Grady D, Chung F. Titration of isoflurane using BIS index improves early recovery of elderly patients undergoing orthopedic surgeries. Can J Anaesth 2002;49:13–18. doi: 10.1007/BF03020413.
42. Cotoia A, Mirabella L, Beck R, Matrella P, Assenzo V, Chazot T, et al. Effects of closed-loop intravenous anesthesia guided by bispectral index in adult patients on emergence delirium: a randomized controlled study. Minerva Anestesiol 2018;84:437–446. doi: 10.23736/S0375-9393.17.11915-2.
43. Lehmann A, Schmidt M, Zeitler C, Kiessling AH, Isgro F, Boldt J. Bispectral index and electroencephalographic entropy in patients undergoing aortocoronary bypass grafting. Eur J Anaesthesiol 2007;24:751–760. doi: 10.1097/00003643-200706003-00023.
44. Quan C, Chen J, Luo Y, Zhou L, He X, Liao Y, et al. BIS-guided deep anesthesia decreases short-term postoperative cognitive dysfunction and peripheral inflammation in elderly patients undergoing abdominal surgery. Brain Behav 2019;9:e01238. doi: 10.1002/brb3.1238.
45. Valentin LSS, Pereira VFA, Pietrobon RS, Schmidt AP, Oses JP, Portela LV, et al. Effects of single low dose of dexamethasone before noncardiac and nonneurologic surgery and general anesthesia on postoperative cognitive dysfunction – A phase III double blind, randomized clinical trial. PLoS One 2016;11:e0152308. doi: 10.1371/journal.pone.0152308.
46. Yang C, Sun Z, Feng Y, Guo K, Chen W, Gao X, et al. Effect of different depth of anesthesia on postoperative cognitive function after retroperitoneal laparoscopic radical nephrectomy. Int J Clin Exp Med 2018;11:12381–12386. doi: NODOI.
47. Zhang D, Nie A. Assessment of different anesthesia depth under total intravenous anesthesia on postoperative cognitive function in laparoscopic patients. J Res Med Sci 2016;21:73. 10.4103/1735-1995.189679.
48. Zhou Y, Li Y, Wang K. Bispectral index monitoring during anesthesia promotes early postoperative recovery of cognitive function and reduces acute delirium in elderly patients with colon carcinoma: a prospective controlled study using the attention network test. Med Sci Monit 2018;24:7785–7793. doi: 10.12659/MSM.910124.
49. Hou R, Wang H, Chen L, Qiu Y, Li S. POCD in patients receiving total knee replacement under deep vs light anesthesia: a randomized controlled trial. Brain Behav 2018;8:e00910. doi: 10.1002/brb3.910.
50. Short TG, Leslie K, Campbell D, Chan MTV, Corcoran T, O’Loughlin E, et al. A pilot study for a prospective, randomized, double-blind trial of the influence of anesthetic depth on long-term outcome. Anesth Analg 2014;118:981–986. doi: 10.1213/ANE.0000000000000209.
51. Xu G, Huang YL, Li PL, Li LL, Shang XD, Sun ZT, et al. Effect of bispectral index values on hip arthroplasty in elderly patients under general anaesthesia combined with lumbar plexus nerve block. Int J Clin Exp Med 2020;13:8447–8454.
52. Sahni N, Anand LK, Gombar KK, Gombar S. Effect of intraoperative depth of anesthesia on postoperative pain and analgesic requirement: a randomized prospective observer blinded study. J Anaesthesiol Clin Pharmacol 2012;28:266–267. doi: 10.4103/0970-9185.86595.
53. Kuhn S, Cooke K, Collins M, Jones JM, Mucklow JC. Perceptions of pain relief after surgery. BMJ 1990;300:1687–1690. doi: 10.1136/bmj.300.6741.1687.
54. Jensen MP, Mardekian J, Lakshminarayanan M, Boye ME. Validity of 24-h recall ratings of pain severity: biasing effects of “Peak” and “End” pain. Pain 2008;137:422–427. doi: 10.1016/j.pain.2007.10.006.
55. Hao S, Ogawa H. Sevoflurane suppresses behavioral response in the rat formalin test: Combination with intrathecal lidocaine produced profound suppression of the response. Neurosci Lett 1998;248:124–126. doi: 10.1016/S0304-3940(98)00282-1.
56. Hao S, Takahata O, Mamiya K, Iwasaki H. Sevoflurane suppresses noxious stimulus-evoked expression of Fos-like immunoreactivity in the rat spinal cord via activation of endogenous opioid systems. Life Sci 2002;71:571–580. doi: 10.1016/S0024-3205(02)01704-6.
57. Bandschapp O, Filitz J, Ihmsen H, Berset A, Urwyler A, Koppert W, et al. Analgesic and antihyperalgesic properties of propofol in a human pain model. Anesthesiology 2010;113:421–428. doi: 10.1097/ALN.0b013e3181e33ac8.
58. Antognini JF, Wang XW, Piercy M, Carstens E. Propofol directly depresses lumbar dorsal horn neuronal responses to noxious stimulation in goats. Can J Anaesth 2000;47:273–279. doi: 10.1007/BF03018926.
59. Cheng SS, Yeh J, Flood P. Anesthesia matters: patients anesthetized with propofol have less postoperative pain than those anesthetized with isoflurane. Anesth Analg 2008;106:264–269. doi: 10.1213/01.ane.0000287653.77372.d9.
60. Peng K, Liu HY, Wu SR, Liu H, Zhang ZC, Ji FH. Does propofol anesthesia lead to less postoperative pain compared with inhalational anesthesia? : A systematic review and meta-analysis. Anesth Analg 2016;123:846–858. doi: 10.1213/ANE.0000000000001504.
61. Evered LA, Silbert BS. Postoperative cognitive dysfunction and noncardiac surgery. Anesth Analg 2018;127:496–505. doi: 10.1213/ANE.0000000000003514.
62. Marcantonio ER. Delirium in hospitalized older adults. N Engl J Med 2017;377:1456–1466. doi: 10.1056/NEJMcp1605501.
63. Rasmussen LS. ISPOCD2 investigators. Post-operative cognitive dysfunction in the elderly. Acta Anaesthesiol Scand 2005;49:1573. doi: 10.1111/j.1399-6576.2005.00860.x.
64. Miao M, Xu Y, Sun M, Chang E, Cong X, Zhang J. BIS index monitoring and perioperative neurocognitive disorders in older adults: a systematic review and meta-analysis. Aging Clin Exp Res 2020;32:2449–2458. doi: 10.1007/s40520-019-01433-x.
65. Orena EF, King AB, Hughes CG. The role of anesthesia in the prevention of postoperative delirium: a systematic review. Minerva Anestesiol 2016;82:669–683. doi: NODOI.
66. Sieber FE, Zakriya KJ, Gottschalk A, Blute MR, Lee HB, Rosenberg PB, et al. Sedation depth during spinal anesthesia and the development of postoperative delirium in elderly patients undergoing hip fracture repair. Mayo Clin Proc 2010;85:18–26. doi: 10.4065/mcp.2009.0469.
67. Radtke FM, Franck M, Lendner J, Krüger S, Wernecke KD, Spies CD. Monitoring depth of anaesthesia in a randomized trial decreases the rate of postoperative delirium but not postoperative cognitive dysfunction. Br J Anaesth 2013;110 (Suppl 1):i98–i105. doi: 10.1093/bja/aet055.
68. Wildes TS, Mickle AM, Ben Abdallah A, Maybrier HR, Oberhaus J, Budelier TP, et al. Effect of electroencephalography-guided anesthetic administration on postoperative delirium among older adults undergoing major surgery: the ENGAGES randomized clinical trial. JAMA 2019;321:473–483. doi: 10.1001/jama.2018.22005.

Anesthetic depth; GRADE level of evidence; Postoperative delirium; Postoperative pain; Trial sequential analysis

Supplemental Digital Content

Copyright © 2023 The Chinese Medical Association, produced by Wolters Kluwer, Inc. under the CC-BY-NC-ND license.