PONV is a common and distressing complication after thyroidectomy, with an incidence of 20–60% despite the use of prophylactic anti-emetics (1, 2). Beyond impairing comfort and satisfaction, PONV may precipitate serious consequences, including postoperative hemorrhage, delayed recovery, and increased healthcare costs (3, 4).

IONM is widely adopted during thyroid surgery to reduce the risk of RLN injury (5). However, IONM has been associated with a higher incidence of PONV (5, 6), likely reflecting anesthetic adjustments made to optimize monitoring—such as minimizing neuromuscular blockade, increasing opioid use to prevent movement, and prolonging operative time. Although guidelines for PONV prevention are available, their effectiveness in routine practice remains suboptimal (7). In addition, the limited efficacy and potential adverse effects of antiemetic drugs underscore the need for safer, non-pharmacological strategies.

Propofol-based TIVA, which has intrinsic antiemetic properties, is recommended for patients at high risk of PONV (7). Acupuncture has also attracted interest as a non-pharmacological intervention for postoperative gastrointestinal symptoms (8, 9). Common sensations during acupuncture—such as soreness, heaviness, or numbness at the stimulation site—are regarded by some practitioners as markers of physiological response and have been explored as behavioral indicators in research (8, 9). Stimulation at Neiguan (PC6), Zusanli (ST36), and Zhongwan (RN12) has been reported to alleviate nausea and gastrointestinal discomfort (10). These acupoints were selected according to traditional Chinese medicine theory; while prior trials support their clinical use, putative mechanisms remain largely theoretical and lack validation by modern biomedical research. Notably, clinical evidence on combining acupuncture with propofol-based anesthesia during thyroidectomy is limited.

This study aimed to evaluate whether propofol-based anesthesia combined with postoperative acupuncture reduces the incidence of PONV and improves early postoperative recovery in patients undergoing IONM-assisted thyroidectomy. We also explored whether the combination confers any additional benefit compared with either intervention alone.

Between April 9 and June 28, 2024, 160 patients were screened; 17 failed eligibility and 2 declined postoperative acupuncture. Six participants were lost to follow-up, leaving 135 participants who completed the trial (propofol + active n = 35; propofol + sham n = 35; sevoflurane + active n = 33; sevoflurane + sham n = 32) (Figure 1). Baseline characteristics were well balanced across groups (Table 1).

For the primary outcomes, the four-arm risks of any PONV were 78.1% (25/32) for sevoflurane + sham, 60.6% (20/33) for sevoflurane + active, 45.7% (16/35) for propofol + sham, and 28.6% (10/35) for propofol + active. In factorial models, propofol compared with sevoflurane reduced the risk of any PONV (aOR 0.25, 95% CI 0.12–0.52; p = 0.0002) and vomiting ≥1 episode (aOR 0.329, 95% CI 0.115–0.939; p = 0.038), and was associated with lower PONV severity (proportional-odds aOR 0.313, 95% CI 0.123–0.797; p = 0.015). Active compared with sham acupuncture also reduced any PONV (aOR 0.45, 95% CI 0.22–0.95; p = 0.035). No anesthesia × acupuncture interaction was detected for any PONV endpoints (Wald p = 0.897), indicating additive rather than synergistic effects (Table 2).

Early recovery at 24 h, measured by the 0–70 short-form score (higher = better), did not differ by anesthesia (mean difference [propofol–sevoflurane] + 1.39, 95% CI − 0.75 to +3.54; p = 0.204) or by acupuncture (active–sham +1.34, 95% CI − 0.80 to +3.48; p = 0.221); the interaction term was non-significant (p = 0.553). Medians [IQR] were 64 [60–67] vs. 62 [58–67] for propofol vs. sevoflurane and 62.5 [59–68] vs. 63 [58–66] for active vs. sham (Table 3).

Regarding secondary outcomes within 24 h, sore throat (any vs. none) showed no significant differences (propofol vs. sevoflurane aOR 0.56, 95% CI 0.18–1.78; p = 0.330; active vs. sham aOR 1.01, 95% CI 0.33–3.07; p = 0.983). Similarly, headache did not differ significantly (propofol aOR 0.90, 95% CI 0.42–1.94; p = 0.847; active aOR 1.33, 95% CI 0.62–2.86; p = 0.560), nor did dizziness (propofol aOR 0.49, 95% CI 0.20–1.22; p = 0.125; active aOR 1.83, 95% CI 0.74–4.58; p = 0.194). Severity distributions for sore throat and headache were comparable across main effects (both p > 0.40, Mann–Whitney), and severe headache (NRS ≥ 7) was rare (Table 4). Intra-operative opioid consumption, converted to morphine milligram equivalents, differed significantly across the four groups (Kruskal–Wallis H = 14.8, p = 0.002). Pairwise Mann–Whitney tests showed no difference between active and sham acupuncture within either anesthetic regimen (propofol p = 1.000; sevoflurane p = 1.000), whereas propofol-based TIVA required significantly higher MME than sevoflurane across both acupuncture strata (all Bonferroni-adjusted p < 0.001) (Table 5).

Omnibus four-arm tests confirmed these patterns. The omnibus χ² test for any PONV was significant (χ²(3) = 16.01, p = 0.00113); after Holm adjustment, the only significant pairwise contrast was sevoflurane + sham vs. propofol + active (OR 7.50, p_adj = 0.0013). For vomiting ≥1, the omnibus χ² was significant (χ²(3) = 12.98, p = 0.00469), but no pairwise contrast remained significant after adjustment. Omnibus tests for sore throat, headache, and dizziness were not significant (all p ≥ 0.06). Full results are provided in Supplementary Tables S1–S3.

Post-hoc power analyses indicated that, for the primary nausea NRS outcome, power was 68% for the anesthesia main effect, 76% for the acupuncture main effect, and only 5% for the anesthesia × acupuncture interaction (β = 0.041, SE = 0.739), showing that the trial was underpowered to detect modest synergy; similarly low interaction power was observed for other outcomes (Supplementary Tables S4). No severe adverse events occurred, and no participant discontinued because of an adverse event.

This randomized controlled trial demonstrated that both propofol-based anesthesia and postoperative acupuncture significantly reduced the incidence of PONV within 24 h after thyroidectomy with intraoperative neuromonitoring. However, the combination of the two interventions did not confer synergistic benefits, indicating additive rather than interactive effects.

The observed incidence of PONV in the sevoflurane groups exceeded that reported in previous thyroidectomy studies (≈44%) (2, 22, 23), reaching up to 60% without acupuncture. This higher rate may reflect the anesthetic adjustments required for IONM, including low-dose cisatracurium and higher opioid administration (24, 25). In this trial, intra-operative opioid consumption, expressed as morphine milligram equivalents, was significantly higher with propofol-based TIVA than with sevoflurane anesthesia, whereas acupuncture did not alter opioid requirements. This suggests that the antiemetic benefit of acupuncture is independent of opioid sparing, and may instead relate to modulation of vagal activity, serotonergic signaling, β-endorphin release or central emetic pathways, consistent with mechanisms proposed in previous experimental studies. These multi-target mechanisms help explain why acupuncture remained effective in reducing PONV even under propofol anesthesia, despite higher opioid consumption. Moreover, despite higher opioid use, propofol anesthesia still reduced the incidence of PONV, indicating that its intrinsic antiemetic effect outweighed the potential emetogenic impact of increased opioid exposure.

As a safe non-pharmacological intervention, acupuncture has increasingly been integrated into multimodal therapy (26–28). Previous studies have shown that bilateral Zusanli (ST36) acupuncture combined with ondansetron significantly reduced PONV in patients undergoing laparoscopic surgery (29). Yan et al. also found that combined acupuncture at Neiguan (PC6), Hegu, Zusanli (ST36), and Sanyinjiao was more effective than ondansetron alone in preventing postoperative nausea in high-risk patients (30). Our findings extend this evidence by confirming that acupuncture remains beneficial even alongside propofol-based anesthesia, although without synergy.

Propofol’s antiemetic effects are well recognized and may involve multiple mechanisms, including enhancement of GABA_A receptor activity, suppression of neuronal excitability in the nucleus tractus solitarius and chemoreceptor trigger zone, modulation of 5-HT3 and dopamine D2 signaling, and anti-inflammatory actions (31). Our finding that propofol reduced postoperative dizziness aligns with prior studies (32).

Acupuncture’s anti-PONV effects may involve the modulation of central and peripheral 5-HT levels, enhancement of vagal tone, and promotion of β-endorphin release in cerebrospinal fluid. Stimulation at PC6 has been shown to modulate vagal activity and gastric slow waves, regulate serotonergic signaling, and promote β-endorphin release in cerebrospinal fluid, thereby attenuating nausea and improving gastrointestinal motility (15, 17). ST36 complements PC6 by enhancing vagal tone and HRV, while GB20 targets vestibular–occipital circuits implicated in postoperative dizziness and headache. Prior studies combining acupuncture with antiemetics such as ondansetron demonstrated additive efficacy (10, 29). These multi-target mechanisms explain why acupuncture remained effective even in the presence of propofol-based anesthesia.

However, in this factorial RCT, the anesthesia × acupuncture interaction for the PONV endpoints was not significant (e.g., any PONV, Wald p = 0.897), indicating additive rather than synergistic effects. This pattern is biologically plausible: propofol-based TIVA already lowers baseline emetogenicity through overlapping antiemetic pathways and by avoiding volatile agents, creating a ceiling once risk is suppressed (7). Acupuncture can reduce PONV and modulate vagal/gastrointestinal signaling (e.g., PC6 stimulation), but the added value on top of effective anesthesia strategies is variable (15, 17). Moreover, propofol may attenuate stimulus-evoked central responses, potentially blunting acupuncture-evoked neural effects (33). Finally, acupuncture is likely dose-dependent (intensity, duration, frequency) (34). Consistent with these mechanisms, post-hoc power for the interaction was low across outcomes (~5–15%; peak 34% for dizziness), implying that only very large interaction effects (interaction OR ≈ 0.07–0.16 or 6–15) would have been detectable; our single, brief perioperative session may have provided insufficient activation. Future trials should test pre-operative initiation, longer or repeated sessions, and standardized stimulation parameters, and be a priori powered for an interaction (35).

Early recovery, assessed with a 0–70 short-form score, did not differ between groups (36, 37). The results revealed no significant differences between acupuncture and propofol in improving sleep quality or alleviating sore throat. This could be related to the specific nature of thyroid surgery, including use of large electrophysiological monitoring tubes and required neck hyperextension, which could exacerbate postoperative throat discomfort. In addition, acupuncture was performed only once, and the limited stimulation intensity and frequency may not have been sufficient to yield therapeutic effects within such a short period. The choice of recovery assessment tools may also influence outcomes.

This study has several unique strengths. The active acupuncture procedures were consistently performed by a single experienced acupuncturist, who has extensive clinical experience in managing postoperative gastrointestinal dysfunction (38). Standardized manipulation techniques were applied to ensure consistent stimulation intensity across participants. Additionally, a validated sham acupuncture control was implemented to minimize performance and detection bias.

Several limitations should be acknowledged. First, this was a single-center factorial RCT conducted under uniform IONM workflows with one acupuncturist, which may limit external validity. Routine antiemetic prophylaxis was omitted primarily to isolate the effects of the interventions; local practice constraints, including limited insurance coverage and cost, also contributed. This approach did not fully align with guideline recommendations (7) and represents an important limitation of the trial. Second, the trial was powered for main effects rather than interaction, and the anesthesia × acupuncture interaction estimate was imprecise, so modest synergistic effects cannot be excluded. Third, early recovery was assessed using a 0–70 short form that showed a 24-h ceiling, potentially attenuating differences; future studies could employ more comprehensive tools such as the QoR-40 or Pittsburgh Sleep Quality Index. Fourth, outcomes were assessed to 24 h only, whereas a substantial proportion of PONV occurs within 48 h (39). Fifth, despite rigorous blinding, placebo and expectancy effects cannot be entirely ruled out (15). Finally, the acupuncture protocol involved a single, brief perioperative session; as efficacy is dose- and timing-dependent, preoperative initiation, longer duration, or repeated stimulation may yield stronger effects (34).

In conclusion, this study demonstrated that both acupuncture and propofol anesthesia are effective in reducing PONV. However, their combined application did not demonstrate synergistic advantages, potentially due to differences in intervention mechanisms, limited stimulation frequency, or individual variability. Future research should focus on optimizing acupuncture timing (pre- vs. postoperative), frequency, and intensity, while integrating serological, physiological, and neuroimaging studies to elucidate mechanisms. Additionally, multicenter large-sample randomized trials are needed to confirm the applicability and cost-effectiveness of this approach across different surgeries and populations.

Acupuncture and propofol-based anesthesia each reduced the incidence of PONV in patients undergoing thyroidectomy with intraoperative neuromonitoring. However, their combination did not demonstrate superiority over either intervention alone.