Patient characteristics
There were total 3435 patients who underwent five types of surgical procedures questionnaired and interviewed, with 3223 patients remaining in this study after exclusion criteria applied as operation time less than one hour (n = 111), incomplete data (n = 71) and intraoperative bleeding more than 800 ml (n = 30)) (Fig. 1). They are initially divided into PONV (n = 726) and non-PONV (n = 2497) groups. To reduce bias, a matched patient group was established for propensity score matching analysis between the two groups (n = 700 equally in each group).
The demographic characteristics of the patients in PONV and non-PONV groups were showed in Table 1. Most of male patients (68.84%) did not experience PONV; Same as described in previous investigation, male patients were much less likely to have nausea and vomiting after surgery as compared with females (38.84% in male vs 61.16% in female). The operation time was slightly shorter in the PONV group. Other characteristics including ASA physical status, age and BMI were quite similar in both groups. In the whole PONV incidence, use of inhaled anesthetics and postoperative opioid account for 83.7% and 82.6% respectively, indicating both anesthetic agents highly involved in the occurrence of emetic episode. In non-PONV group, 65.4% patient with inhaled anesthetics did not experience PONV; however, when postoperative opioid was used, the number of non-PONV patient went up to 94.5%.
Table 1
Comparison of patient characteristics data between non-PONV group and PONV group. Data presents as mean±SD, n (%).
Variables | non-PONV group (n = 2497) (n = 2497) | PONV group (n = 726) | P-Value |
Female (n,%) | 778(31.2%) | 444 (61.2%) | < 0.001 |
Age(median-IQR, year) | 59.99 ± 11.12 | 59.10 ± 11.07 | 0.056 |
Age(year) | | | 0.034 |
\(\le\)60 | 1144(45.8%) | 365(50.3%) | |
\(>\)60 | 1353 (54.2 %) | 361 (49.7%) | |
BMI | 23.54 ± 3.08 | 23.27 ± 3.15 | 0.04 |
ASA physical status | | | 0.534 |
Ⅰ | 172(6.9 %) | 51(7 %) | |
Ⅱ | 2063(82.6 %) | 608(83.8 %) | |
≥Ⅲ | 262(10.5 %) | 67(9.2 %) | |
Duration of anesthesia | 2.55 ± 1.03 | 2.32 ± 0.98 | < 0.001 |
Postoperative opioid | 2360(94.5%) | 600(82.6%) | < 0.001 |
Inhaled anesthesia | 1633(65.4%) | 608(83.7%) | < 0.001 |
Vagus nerve trunk resection | 1046(41.9%) | 141(19.4%) | < 0.001 |
Non-vagus nerve trunk resection | 1451(58.1%) | 585(80.6%) | |
Notably, as compared with non-vagotomy surgeries (hepatectomy, pulmonary lobectomy and colorectomy), vagus nerve trunk resection performed in both esophagectomy and gastrectomy significantly reduced PONV incidence by approximately 4-fold, dropping from 80.6–19.4%. This result suggested that vagus nerve stimulation might play a predominant role in triggering PONV occurrence.
Vagotomy associated with PONV in the multivariate logistic regression model
To further examine the potential role of vagus nerve in PONV, we performed multivariate logistic regression analysis with factors associated with PONV in entire cohort (Table 2). There are four variables examined including gender, inhaled anesthetics,postoperative opioid and vagotomy. As expected, gender is one of risk factors of PONV as male patients had an OR value of 0.310 (95% CI, 0.258–0.373) as compared with females. In addition, the use of inhaled anesthetics was highly involved in PONV (OR = 3.946; 95%, CI, 3.135–4.966). Much lower incidence of PONV (OR = 0.325; 95% CI, 0.245–0.432) with postoperative opioid application was unexpectedly found in multivariate logistic regression analysis, in contrast to common point of view which consider postoperative opioid as a risk factor for PONV. This might be duo to: 1) dexamethasone being regularly used in premedication along with intra-operative 5HT3 antagonists; 2) sufficient anti-emetic (normally 9 mg ondensteron) accompanying opioid analgetics (butorphanol ± fentanyl at much lower dosage) during postoperative patient-controlled analgesia in our hospital. Butorphanol is less likely to cause PONV.[24]
Table 2
Multivariate logistic regression analysis with factors associated with PONV in entire cohort. CI,confidence interval.
Variables | OR(odds ratio) | 95% CI | P-value |
Sex(male vs female) | 0.310 | 0.258 ཞ 0.373 | < 0.001 |
Inhaled anesthesia | 3.958 | 3.145 ཞ 4.980 | < 0.001 |
Postoperative opioid | 0.327 | 0.247 ཞ 0.434 | < 0.001 |
Vagus nerve trunk resection | 0.321 | 0.259 ཞ 0.399 | < 0.001 |
Multivariate logistic regression result revealed that vagus nerve trunk resection served as one factor that significantly modulated the occurrence of PONV (OR = 0.311; 95% CI, 0.246–0.393), further providing the evidence that vagus nerve works as a primary afferent nerve to receive PONV-related signal inputs. These data also support the potential underlying mechanism by which intact gastric-vagal-brain reflex play a key role for PONV.
Vagus nerve trunk resection is associated with the incidence of PONV
To further investigate the role of vagus nerve in PONV, we performed propensity score matching to adjust the imbalance of concomitant variable in order to avoid the bias. Propensity score matching is an effective method to analyze the observational data when randomized trials are not feasible.[25] After propensity score matching, 700 patients remained in the PONV group and 700 remained in the non-PONV group. Gender, age, the use of volatile anesthetics, BMI, postoperative opioid, ASA physical, and duration of anesthesia were similar between two groups. In PONV group, 35.3% patients underwent surgeries (esophagectomy and gastrectomy) with vagus nerve trunk resected, while 60.9% patients in non-PONV group had surgeries (hepatectomy, pulmonary lobectomy and colorectomy) with intact vagus nerve (P < 0.001) (Table 3).
Table 3
Prospensity score matching analysis on the association of vagus nerve trunk resection with PONV
Variables | Before propensity score matching | P-Value | After propensity score matching | P-Value |
| non-PONV group (n = 2497) | PONV group (n = 726) | | non-PONV group (n = 700) | PONV group (n = 700) | |
Female (n,%) | 778(31.2%) | 444 (61.2%) | < 0.001 | 285(40.7%) | 282(40.3%) | 0.870 |
Age(median-IQR, year) | 59.99 ± 11.12 | 59.10 ± 11.07 | 0.056 | 58.99 ± 11.8 | 59.01 ± 11.0 | 0.963 |
Age(year) | | | 0.034 | | | |
≦ 60 | 1144(45.8%) | 365(50.3%) | | 340(48.6%) | 349(49.9%) | |
> 60 | 1353 (54.2 %) | 361 (49.7%) | | 360 (51.1 %) | 351 (50.1%) | |
BMI | 23.54 ± 3.08 | 23.27 ± 3.15 | 0.04 | 23.13 ± 3.15 | 23.29 ± 3.16 | 0.380 |
ASA physical status | | | 0.534 | | | 0.929 |
Ⅰ | 172(6.9 %) | 51(7 %) | | 46(6.6%) | 49(7.0%) | |
Ⅱ | 2063(82.6 %) | 608(83.8 %) | | 590(84.3%) | 585(83.6%) | |
≥Ⅲ | 262(10.5 %) | 67(9.2 %) | | 64(9. 1%) | 66(9.4%) | |
Duration of anesthesia | 2.55 ± 1.03 | 2.32 ± 0.98 | < 0.001 | 2.35± 0.96 | 2.34 ± 0.99 | 0.848 |
Postoperative opioid | 2360(94.5%) | 600(82.6%) | < 0.001 | 605(86.4%) | 600(85.7%) | 0.700 |
Inhaled anesthesia | 1633(65.4%) | 608(83.7%) | < 0.001 | 590(84.3.4%) | 582(83.1%) | 0.848 |
Vagus nerve trunk resection | 1046(41.9%) | 141(19.4%) | < 0.001 | 426(60.9%) | 247(35.3%) | < 0.001 |
Emetic outcomes related to vagotomy
As crucial role of vagus nerve trunk resection was confirmed through multiple statistical analysis, we further evaluated the detailed percentage of nausea and the intensity of vomiting (Table 4). Most of patients with esophagectomy and gastrectomy in vagus nerve trunk resection group did not report nausea and vomiting (90.9% and 86.5% respectively), while the percentage of non-PONV was reduced to approximately 70% in patients without vagus nerve trunk resection. The overall incidence of PONV were much lower in vagotomy patients with 9.1% in esophagectomy and 13.5% in gastrectomy. By contrast, when intact vagus nerve was maintained, the highest incidence of PONV (ཞ30%) was reported in pulmonary lobectomy and hepatectomy, followed by 23% of PONV incidence in colorectomy patients. Consistently, nausea was reported in 5.9% esophagectomy patients and 8.6% gastrectomy patients as compared with that in 12ཞ17% patients without vagotomy. Most vomiting episodes were reported less than 3 times, being around 3% in vagotomy patients including esophagectomy and gastrectomy and 8ཞ13% in non-vagotomy patients. Sever vomiting graded as more than 3 times were much less experienced in patients undergoing vagotomy (ཞ1%) or non-vagotomy (ཞ3%).
Table 4
Details of PONV incidence between two groups with and without vagus nerve trunk resection.
| Non-PONV | PONV |
Vagus nerve trunk resection group (n,%) | | nausea | Vomiting < 3 times | Vomiting > 3times | total |
Esophagectomy | 401(90.9%) | 26(5.9%) | 10(2.3%) | 4(0.9%) | 40(9.1%) |
Gastrectomy | 645(86.5%) | 64(8.6%) | 29(3.9%) | 8(1.1%) | 101(13.5%) |
Non-vagus nerve trunk resection group (n,%) | | |
Colorectomy | 442(77%) | 71(12.4%) | 46(8.0%) | 15(2.6%) | 132(23%) |
Hepatectomy | 366(69.8%) | 87(16.6%) | 57(10.9%) | 14(2.7%) | 158(30.2%) |
Pulmonary lobectomy | 643(68.6%) | 144(15.4%) | 122(13.0%) | 29(3.1%) | 295(31.4%) |