Cross-sectional Study of PONV Risk Factors for Oral Surgery After Intubated General Anesthesia With Total Intravenous Anesthesia

Abstract

Objective:

The incidence of postoperative nausea and vomiting (PONV) after general anesthesia with total intravenous anesthesia (TIVA) was reported to be significantly lower than with volatile inhalational agents (13.3% vs 25%). However, no investigation of PONV risk factors associated with TIVA has ever been reported. This cross-sectional retrospective study aimed to investigate whether known risk factors influenced PONV in intubated general anesthetics utilizing TIVA for dental or oral and maxillofacial surgery.

Methods:

Subjects were 761 patients who underwent dental or oral and maxillofacial surgery under TIVA with propofol, fentanyl, and remifentanil. Univariate and multivariable logistic regression analyses were performed using PONV (within 24 hours) as the dependent variable and previously reported risk factors as independent variables.

Results:

Age (odds ratio [OR]: 1.020 per year decrease; 95% confidence interval [CI]: 1.0002–1.0418; P = .047) and female sex (OR: 2.73; 95% CI: 1.60–4.84; P < .001) were positively associated with PONV. Sagittal split ramus osteotomy (SSRO) (OR: 2.28; 95% CI: 1.21–4.33; P = .011) and bimaxillary osteotomy (OR: 5.69; 95% CI: 2.09–15.99; P < .001) were more likely to be associated with PONV than operations that were neither bimaxillary osteotomy nor SSRO. Late PONV (2–24 hours) had an ∼2.7 times higher incidence than early PONV (0–2 hours).

Conclusion:

These findings suggest further PONV countermeasures, aside from TIVA with propofol and prophylactic antiemetics for orthognathic surgeries especially bimaxillary osteotomy, are needed.

Postoperative nausea and vomiting (PONV) is one of the most frequent postoperative complications that can delay recovery after general anesthesia. The incidence of PONV after oral surgery has been reported as ∼24 to 46%^1,2 and is especially high after orthognathic surgery, approximating 40% to 73%.^1–3

Anesthesia-specific factors related to PONV include the use of general anesthesia, selection of specific anesthetic agents,^4,5 and prophylactic administration of antiemetics.^4,6 According to a meta-analysis that included 17 studies on the incidence of PONV after general anesthesia maintained with propofol vs volatile agents, propofol was significantly lower (median incidence 13.3% vs 25%).⁵ Although total intravenous anesthesia (TIVA) may have a lower incidence, PONV can still occur. Therefore, intervention to further reduce PONV risk may be needed even in TIVA cases. Although reports have identified patient-, surgery-, and anesthesia-specific factors related to PONV,^3,4,7–10 none have focused only on TIVA cases. The purpose of this study was to determine whether the above-mentioned factors identified by existing literature influenced PONV in general anesthetics utilizing TIVA for dental, oral, and maxillofacial surgery.

METHODS

This retrospective cross-sectional study included 791 patients identified as having undergone dental, oral, or maxillofacial surgery under intubated general anesthesia with TIVA using propofol, fentanyl, and remifentanil in the Hokkaido University Hospital between January 2011 and December 2018. Data were collected in August 2019. Approval from the Hokkaido University Hospital Institutional Review Board (clinical study code: 018-0412) was granted in May 2019 with waiver of written informed consent requirement. This study was conducted in accordance with the Helsinki Declaration guidelines, and this manuscript adheres to the applicable Strengthening the Reporting of Observational Studies in Epidemiology guidelines.

Prospective subject data were gathered based solely on the anesthesia ledger and anesthesia and nursing records from the study period. Of the initial 791 patients identified, only those without any of the following exclusion criteria were enrolled for further analysis. Exclusion criteria were as follows: (1) insufficient data regarding PONV in the anesthesia or nursing records; (2) American Society of Anesthesiologists physical status classification ≥3; and (3) patient age < 16 years. A total of 761 subjects matched the eligibility criteria.

The primary objective of this study was to identify significant factors associated with PONV in cases only using TIVA. PONV was defined as the presence of any nausea and/or vomiting occurring within 24 hours after the end of anesthesia. This was determined from the anesthesia or nursing records detailing postoperative findings for at least 24 hours after anesthesia. Evaluation of factors associated with PONV was performed using the occurrence of PONV as the dependent variable and promotional or depressant PONV factors identified in the existing literature^1,3,4,6–12 as independent variables, which consisted of 4 patient-specific, 3 surgery-related, and 3 anesthesia-related factors (Table 1).

Table 1.

Univariate Analysis of Patient-Specific, Surgery-Related, and Anesthesia-Related PONV Risk Factors*

Variable	Total (n = 761)	PONV (n = 121)	No PONV (n = 640)	P value
Patient-specific factors
Sex
Males	228 (30.0%)	21 (17%)	207 (32.3%)	<.001
Females	533 (70.0%)	100 (83%)	433 (67.7%)
Age, y	26 (21–28)	22 (20–31)	27 (21–39)	.001
Body mass index, kg/m2	20.7 (19.0–23.0)	20.4 (19.0–22.6)	20.8 (18.9–23.2)	.232
Smoking
(Positive)	93 (12.2%)	10 (8%)	83 (13.0%)	.13
(Negative)	668 (87.8%)	111 (92%)	557 (87.0%)
Surgery-related factors
Type of surgery
SSRO	160 (21.0%)	33 (27%)	127 (19.8%)	<.001
Bimaxillary osteotomy	133 (17.5%)	47 (39%)	86 (13.4%)
Others	468 (61.5%)	41 (34%)	427 (66.7%)
Operation time, min	148 (87–245)	220 (146–386)	134 (182–229)	<.001
Blood loss, mL	20 (0–200)	150 (10–490)	0 (0–164)	<.001
Anesthesia-related factors
Fentanyl usage during anesthesia, mg	0.2 (0.1–0.4)	0.35 (0.2–0.55)	0.2 (0.1–0.4)	<.001
Fluid infusion volume during anesthesia, mL	800 (500–1350)	1350 (725–2210)	750 (500–1250)	<.001
Number of intraoperative antiemetic types used
0	207 (27.2%)	17 (14%)	190 (30%)	.001
1	200 (26.3%)	19 (16%)	181 (28%)
2 or more	354 (46.5%)	85 (70%)	269 (42.0%)

^*Values are presented as n (%) or median (interquartile range). PONV, postoperative nausea and vomiting; SSRO, sagittal split ramus osteotomy; bimaxillary osteotomy, Le Fort I osteotomy plus SSRO; others, operations that were neither bimaxillary osteotomy nor SSRO.

Secondary objectives were as follows: (1) assessing the overall incidence of PONV occurring within 24 hours after anesthesia maintained with TIVA; and (2) comparing the incidence of early (0–2 hours after anesthesia) and late PONV (2–24 hours after anesthesia).

In this study, a history of motion sickness or previous PONV^7,8 was excluded from the independent variables because of insufficient medical records. The use of postoperative analgesic opioids^3,8 was also excluded as there were no cases in which opioids were administered postoperatively for pain control.

Statistical Analysis

Mann-Whitney U test was used for univariate comparison of continuous variables for the occurrence of PONV, and χ² test was used for categorical variables. Binomial multivariable logistic regression analysis with adjusted odds ratios (aORs) and 95% confidence intervals (CIs) were used to evaluate significance of risk factors for PONV. The test for ratio-difference was used to compare the incidence of early and late PONV.

Statistical analysis and data calculations were made with statistical software (JMP Pro 14, SAS Institute, Inc). A P value < .05 was considered statistically significant. Data values are presented as median (interquartile range), n or n (%), aOR, and 95% CI as appropriate.

RESULTS

The selection criterion and breakdown of present cases are shown in Figure 1. Of the 5076 general anesthetics performed during the study period, 791 cases were identified as only using TIVA. After applying the exclusion criteria, a total of 761 cases were selected for inclusion with no instances of data unavailability. Of those cases, 121 were positive and 640 were negative for PONV for an overall PONV (0–24 hours) incidence rate of 15.9% (121/761 patients).

Figure 1.

Selection criterion and breakdown of study cases. Of the 791 patients who underwent general anesthesia with total intravenous anesthesia (TIVA) using propofol, fentanyl, and remifentanil, 761 patients without any of the exclusion criteria were enrolled. A total of 121 patients had postoperative nausea and vomiting (PONV), whereas 640 did not.

PONV-Related Factors

Univariate analysis identified multiple significant factors correlated with PONV. Comparison of the PONV positive and negative groups demonstrated that of all the patient-specific factors, only body mass index and smoking were found to lack significance while all surgery- and anesthesia-specific factors were significant (Table 1). However, further analysis using binomial multivariable logistic regression only identified a few patient- and surgery-specific factors as being significant risks for PONV (0–24 hours; Table 2). Those factors in descending order, based on the calculated ORs (95% CI), were bimaxillary osteotomy (Le Fort I osteotomy plus sagittal split ramus osteotomy [SSRO]) 5.69 (2.09–15.99), female sex 2.73 (1.60–4.84), and SSRO 2.28 (1.21–4.33). Younger age was also a significant risk factor for PONV based on the unit OR (95% confidence limit [CL]) 1.02 (1.0002–1.0418).

Table 2.

The Results of Adjusted Odds Ratio for PONV (0–24 h) in Binomial Multivariable Logistic Regression Analysis

Variable	Unit aOR	aOR	95% CI	P value
Patient-specific factors
Females (vs males)	—	2.726	1.600–4.836	<.001
Age (y) per year decrease	1.02	—	1.0002–1.0418	.047
Body mass index, kg/m2	0.998	—	0.932–1.065	.989
Smoker (vs nonsmoker)	—	0.667	0.306–1.325	.258
Surgery-related factors
Type of surgery
Bimaxillary osteotomy (vs SSRO)	—	2.495	1.099–5.723	.029
Bimaxillary osteotomy (vs others*)	—	5.693	2.085–15.987	<.001
SSRO (vs others)	—	2.282	1.207–4.331	.011
Operation time, min	0.999	—	0.995–1.002	.401
Blood loss, mL	0.9997	—	0.999–1.001	.624
Anesthesia-related factors
Fentanyl usage during anesthesia, mg	1.059	—	0.937–1.199	.36
Fluid infusion volume during anesthesia, mL	1	—	0.999–1.001	.745
Number of intraoperative antiemetic types used
2 or more (vs 1)	—	1.34	0.778–2.365	.295
2 or more (vs 0)	—	1.205	0.690–2.152	.516
1 (vs 0)	—	0.9	0.468–1.726	.749

^R2 = 0.172.

^*Tooth extraction, n = 186 (24.4%); removal of locking plate, n = 84 (11%). aOR, adjusted odds ratio; CI, confidence interval; PONV, postoperative nausea and vomiting; SSRO, sagittal split ramus osteotomy; bimaxillary osteotomy, Le Fort I osteotomy plus SSRO; others, operations that were neither bimaxillary osteotomy nor SSRO.

PONV Timing

Of the 15.9% (121/761 patients) that had PONV, the timing was distributed throughout both the early (0–2 hours) and late (2–24 hours) postoperative periods. There was a significantly higher incidence rate of late PONV compared with early PONV (70.2% vs 25.6%; P < .001; Table 3). Additionally, there was a small subset of cases (n = 5) with PONV in both time periods.

Table 3.

Patients With PONV by Period of Occurrence*

Period of occurrence	Number of PONV present, n (%)
Early PONV (0–2 h)	31 (25.6)
Late PONV (2–24 h)	85 (70.2)†
Both of early and late PONV	5 (4.1)

^*The incidence of overall PONV 0–24 hours was 15.9% (121/761 patients). PONV, postoperative nausea and vomiting.

^†P < .001 (vs early PONV).

Prophylactic Antiemetic Usage

Although not included for detailed analysis in this study, the rates of antiemetics used were as follows: dexamethasone 70.2% (534/761), droperidol 44.2% (336/761), and metoclopramide 29.7% (226/761). Intraoperatively, 2 or more types of antiemetics were prophylactically administered in 87.5% (140/160 patients) of SSRO and 97.7% (130/133 patients) of bimaxillary osteotomy surgical cases.

DISCUSSION

The existing literature has identified multiple patient-, surgery-, and anesthesia-related factors associated with increased PONV risks after general anesthesia that were subsequently evaluated in this study that assessed PONV after intubated general anesthetics using only TIVA (Tables 1 and 2). Of those previously identified PONV risk factors, only young age, female sex, bimaxillary osteotomy, and SSRO were determined to be significant risks for PONV according to the binomial multivariable logistic regression analysis (Figure 2).

Figure 2.

Risk factors for postoperative nausea and vomiting (PONV; 0–24 h) after intubated general anesthesia using total intravenous anesthesia (TIVA). Arrows illustrate factors with increased PONV risk. Arrow thickness illustrates size of the factor's adjusted odds ratio (OR) per study data. The strongest PONV risk factor was bimaxillary osteotomy (OR 5.69) followed by female sex (OR 2.73) and sagittal split ramus osteotomy (SSRO; OR 2.28). Factors lacking arrows were not significantly associated with PONV.

PONV Risk Factors

According to the systematic review by Apfel et al,⁸ female sex was the strongest overall independent predictor of PONV compared with other independent predictors (aOR: 1.39–2.09) among patient-specific, surgery-related, and anesthesia-related risk factors. The aOR for female sex in our study (2.73) was similar to that same review (2.57).⁸ However, the strongest predicted risk factor in this study was bimaxillary osteotomy (aOR: 5.69).

In general, bimaxillary osteotomy is a lengthy operation with the possibility of heavy blood loss. Moreover, surgery is usually targeted toward young female patients because it is associated with cosmetic concerns. However, PONV was significantly associated with bimaxillary osteotomy after correcting for confounding factors in this study. Therefore, this surgery type was likely to directly influence the occurrence of PONV. One of the vomiting-inducing pathways associated with the emetic center located in the parvicellular reticular formation includes an afferent pathway from the pharynx and gastrointestinal tract.¹³ Factors for PONV induction associated with bimaxillary osteotomy may be surgical stimulation, bleeding, postoperative edema in the pharynx or surrounding soft tissues, and blood being swallowed postoperatively.^3,14,15 Brookes et al¹⁵ recommended subbacteriostatic oral doses of erythromycin (250 mg) administered 8 hours preoperatively to capitalize on erythromycin's established properties as a motilin agonist for preventing PONV as motilin promotes gastrointestinal peristalsis. Other approaches include controlling intraoperative bleeding (ie, suctioning) and ensuring adequate hemostasis postoperatively.

Sinclair et al¹⁰ reported that a 30-minute increase in the duration of anesthesia increased the likelihood of PONV by 59%. However, logistic regression analysis in this study showed operation time was not a significant independent predictor of PONV when TIVA is used. This suggests the pro-PONV action of time-dependent factors (ie, prolonged operation time or anesthesia time) may be offset by antiemetic potential of a propofol continuous infusion during TIVA management. The increased emetic potential of blood loss and fentanyl dosage during general anesthesia may also be offset via continuous infusion of propofol. Results of this study also suggest that when investigating multifactorial complications like PONV, analytical methods like logistic regression analysis that can exclude the effects of confounding factors should be used rather than methods strongly influenced by confounding factors such as univariate analysis.

Administration of opioid agonists for postoperative pain control is a powerful promoter of PONV.^3,8 However, local anesthetics, plus remifentanil and a relatively small amount of fentanyl as components of TIVA were used for intraoperative pain control in this retrospective study. Nonsteroidal anti-inflammatory drugs and acetaminophen were administered immediately before the end of surgery for prophylactic postoperative pain control, and additional nonopioid analgesics were administered (IV, suppository, oral) in the ward according to each patient's needs. Unlike in Europe and the United States, opioids are not administered in the ward postoperatively for analgesia after oral surgery. Therefore, opioids were not involved in the development of PONV in this study.

Antiemetics

Habib et al¹⁶ reported in a randomized controlled trial study that the incidence of “no PONV and no rescue antiemetic” in a multimodal group (TIVA with propofol, droperidol, and ondansetron) was significantly higher than the incidence of “no PONV and no rescue antiemetic” in the TIVA with propofol alone group (80% vs 43%; P < .05). Multiple antiemetics (dexamethasone, droperidol, and/or metoclopramide) were used during general anesthesia for cases included in this study when moderate-severe PONV was predicted to occur in advance. Unfortunately, health insurance in Japan does not cover ondansetron for PONV. Two or more antiemetics were used for 87.5% of all patients who underwent SSRO, 97.7% of all patients who underwent bimaxillary osteotomy, and for 18% of other operations. Interestingly, a paradoxical result was seen in the univariate analysis such that PONV was more likely to occur despite more antiemetics being administered.

Incidence of PONV

The overall incidence of PONV (0–24 hours) in all patients was 15.9% in this study. This value was well in accordance with the reported incidence of 13.3% in anesthesia patients maintained with propofol in a meta-analysis.⁵ Details of the various mechanisms involved with PONV are still relatively unclear,¹⁷ as are the mechanistic details of propofol's antiemetic actions.^13,18 However, it is obvious that propofol has a clinically inhibitory effect on PONV.^4,5,9,19 In this study, the incidence of late PONV was significantly higher (∼2.7 times) than early PONV, similar to the report of Apfel et al.²⁰ Consideration should be given to utilizing an antiemetic for the later period following the use of TIVA for oral surgery. Meanwhile, this result may be consistent with those of 2 RCTs,^21,22 which reported that gastric aspiration at the end of an operation is not recommended for preventing PONV despite being previously recommended. In summary, the data from this study suggest that further PONV countermeasures, aside from TIVA with propofol and prophylactic antiemetics for orthognathic surgeries especially bimaxillary osteotomy, are needed.

Limitations

This study had several limitations. First, these results cannot be applied to the general population because this study was limited to 1 institution. Second, the external validity of this study may be considered low as opioids were not administered for postoperative analgesia. Third, a potential confounding factor in the statistical analysis due to selection bias was the intraoperative administration of antiemetics, which was done for cases where the occurrence of moderate-severe PONV was anticipated. In addition, the type, dose, and timing of intraoperative antiemetic administration were at the discretion of each dental anesthesiologist; thus, it cannot be ruled out as a confounding factor influencing the relationship between PONV and the number of intraoperative antiemetics administered.

CONCLUSION

Of the previously established risk factors for PONV after general anesthesia, only young age, female sex, and surgical type, specifically bimaxillary osteotomy and SSRO, were identified in this study as being significant for PONV (0–24 hours) following intubated general anesthesia with TIVA. Bimaxillary osteotomy was the strongest predictor of PONV. Interestingly, 2 or more types of antiemetics were prophylactically administered in 97.7% of bimaxillary osteotomy and 87.5% of SSRO cases. The incidence of late PONV (2–24 hours) was almost 3 times higher than early PONV (0–2 hours). These data suggest that further PONV countermeasures, aside from TIVA with propofol and prophylactic antiemetics for orthognathic surgeries especially bimaxillary osteotomy, are likely needed.