Machine Learning-based Prediction of Postoperative Nausea and Vomiting after Spinal Anesthesia: A Retrospective Observational Study

Machine Learning-based Prediction of Postoperative Nausea and Vomiting after Spinal Anesthesia: A Retrospective Observational Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Machine Learning-based Prediction of Postoperative Nausea and Vomiting after Spinal Anesthesia: A Retrospective Observational Study Hiroshi Hoshijima, Tomo Miyazaki, Shinichiro Omachi, Daisuke Konno, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4421679/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Purpose In this study, we apply analysis using artificial intelligence to identify risk factors for Postoperative nausea and vomiting (PONV) during surgery under spinal anesthesia. Methods This retrospective study used artificial intelligence to analyze data of adult patients (aged ≥ 20 years) who underwent surgery under spinal anesthesia. To evaluate PONV, patients who experienced nausea and/or vomiting or used antiemetics within 24 hours after surgery were extracted from postoperative medical records. We create a model that predicts probability of PONV using the gradient tree boosting model. The model implementation used the LightGBM framework. Results Data were available for 4,574 patients. The identified risk factors were duration of surgery, female, no blood transfusion, spinal level 3–4 puncture, no concomitant epidural anesthesia, use of propofol, and dexmedetomidine, postoperative fentanyl use, cesarean section, and not using phenylephrine, atropine, or oxytocin. Conclusions We used artificial intelligence to evaluate the extent to which risk factors for PONV contribute to the development of PONV. We identifies female and cesarean section, which are known risk factors for PONV after surgery under spinal anesthesia. Our findings also suggest that fluid volume, blood transfusion, and agents that normalize hemodynamics, such as phenylephrine and atropine, are important in preventing PONV. Trial registration number: UMIN000050012 artificial intelligence PONV risk factors Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Postoperative nausea and vomiting (PONV) is a frequent and serious complication in patients undergoing surgery.[ 1 ] This complication decreases patient satisfaction with surgery and is associated with increased healthcare costs due to prolonged hospitalization. Numerous methods have been proposed to avoid PONV but have not completely prevented it.[ 2 ] The incidence of PONV ranges from 5–42% in patients undergoing lower extremity surgery, lower abdominal surgery, and cesarean section under spinal anesthesia.[ 3 – 5 ] Previous studies have identified several risk factors for PONV after spinal anesthesia compared with general anesthesia, including female sex, puncture level, and intraoperative hypotension. However, these factors remain controversial because they are very different from the known risk factors after general anesthesia. Artificial intelligence (AI)-based technologies have evolved considerably and are starting to be applied in medicine.[ 6 , 7 ] The evolution of machine learning with AI, particularly deep learning methods such as neural networks and convolutional neural networks, is a driving force behind the development of AI. Deep learning differs from traditional machine learning in that AI learns differences between samples and selects the correct answer. Therefore, AI is uniquely capable of recognizing changes beyond human perception and establishing AI-specific identification methods. The objective of the present study was to identify risk factors for PONV using machine learning analysis of AI in patients undergoing surgery with spinal anesthesia. RESULTS We obtained data from the Tohoku University database on 5808 patients who underwent surgery under spinal anesthesia. Data for 1,234 patients were excluded (spinal puncture level unclear, n = 1,215; missing data, n = 12; no PONV adjudication, n = 3; no information on drugs used for spinal anesthesia, n = 3; concomitant general anesthesia, n = 3; and no surgery, n = 1) (Supplementary Fig. S1 ), leaving data for 4,574 patients available for analysis. The patient characteristics are shown in Table 1 . In total, 269 (5.9%) of the 4,574 patients developed PONV. Table 1 Patients characteristics Patients with PONV (269; 5.9%) Patients without PONV (4,305; 94.1%) Gender (Male/Female) 16 (5.9%)/269 (94.1%) 556 (12.9%)/3,749 (87.1%) Age (year) 40.8 ± 15.2 39.5 ± 14.4 Body Mass Index (BMI, kg/mg-2) 25.2 ± 4.76 26.4 ± 19.8 Duration of anesthesia (min) 97.4 ± 45.5 61.2 ± 27.3 Duration of surgery (min) 68.1 ± 38.2 90.3 ± 33.1 Total infusion volume (ml) 1366.6 ± 622.6 1189.4 ± 585.4 Total bleeding loss (ml) 720.5 ± 735.2 613.8 ± 660.7 Total blood transfusion volume (ml) 38.9 ± 175.9 18.7 ± 131.7 Total urine output (n) 253.3 ± 250.7 201.6 ± 247.8 Caesarean section (n*) 73 1,175 Anesthesic agents 0.5%Bupivacaine (High specific gravity) (n*) 218 3,651 0.5%Bupivacaine (Equal specific gravity) (n*) 51 654 Fentanyl Intrathecal (n*) 82 1,704 Fentanyl Post operative (0.25mg/1A) (n*) None 153 3127 1 (A) 112 1399 2 (A) 2 36 3 (A) 0 4 4 (A) 2 5 5 (A) 0 2 6 (A) 0 1 Morphine (n*) 14 44 Propofol (prefilled syringe) (mg**) 24 247 Midazolam (n*) 5 105 Dexmedetomidine (n*) 53 549 Puncture level of spinal anesthesia (n*) L2-3 20 515 L3-4 238 3554 L4-5 9 226 L5-S1 2 10 Puncture level of epidural anesthesia (n*) None 47 982 Th1-6 3 19 Th7-12 182 2988 L3 37 316 Cardiovascular agents Atropine (n*) 5 132 Ephedrine (n*) 77 1,479 Nicardipine (n*) 4 36 Phenylephrine (n*) 126 2,390 Atonin(Oxxytocin) (n*) 192 3,077 Steroid Hydrocortisone (n*) 0 17 Prednisolone (n*) 1 12 * Total number of ampules or prefilled syringe used in each group. **The amount of drug used (mg) is not the actual dose administered to the patients, but is calculated backwards from the number of ampules/vials of the drug administered to the patients. First, we evaluated the prediction model using three metrics, namely, the true positive rate, false positive rate, and area under the curve (AUC). We compared the model with a naive model (i.e., the k-nearest neighbor model [KNN]), with k set to 9. Supplementary Fig. S2 shows the receiver-operating characteristic curves. The model outperformed the KNN. Specifically, the AUC was 0.53 for the model and 0.54 for the KNN, demonstrating the ability of our model to recognize PONV and warranting further analysis. Next, we calculated the mean absolute Shapley Additive exPlanations (SHAP) value for each item by averaging the absolute SHAP values over the test data (Supplementary Fig. S3). If an item had a SHAP value greater than 0, it was judged to be a risk factor for PONV. Risk factors for PONV The total infusion volume was the item most strongly associated with PONV after surgery under spinal anesthesia. Duration of surgery, BMI, total urine output, and postoperative fentanyl were subsequently associated with PONV (Fig. 1 ). The analysis identified female sex as a patient-related risk factor for PONV (Fig. 2 A) and an operation time of 50–90 min as a time-related risk factor (Fig. 3 A). Fluid-related risk factors for PONV were need for blood transfusion (Fig. 3 E) and a total fluid infusion volume of over 1000 ml during anesthesia (Fig. 3 F). Anesthesia-related risk factors for PONV were puncture level for spinal anesthesia (spinal level 3–4) (Fig. 4 A) and no concomitant epidural anesthesia (Fig. 4 B). Use of propofol and dexmedetomidine during surgery was also a risk factor for PONV (Fig. 5 A and 5 B). Postoperative use of fentanyl was also a risk factor for PONV (Fig. 4 F), as was not using phenylephrine (Fig. 5 E), atropine (Fig. 5 F), or oxytocin (Fig. 5 C), and cesarean section (Fig. 3 C). In contrast, patient age (Fig. 2 B), BMI (Fig. 2 C), duration of anesthesia (Fig. 3 B), total blood loss (Fig. 3 D), total urine output (Fig. 3 G), intrathecal use of fentanyl (Fig. 4 E), use of bupivacaine (high or equal specific gravity) (Fig. 4 C, D), and use of ephedrine (Fig. 5 D) did not have any clear relationship with PONV. The associations of midazolam, nicardipine, hydrocortisone, prednisolone, and morphine with PONV could not be analyzed because of the small number of samples. DISCUSSION In this study, female sex was the only patient-related risk factor for PONV, and surgery-related risk factors included cesarean section, an operation time of 50–90 min, a total infusion volume of > 1000 ml, and need for blood transfusion. Previous studies found that PONV under spinal anesthesia was more common in women,[ 8 ] which is consistent with our results. Cesarean section also increased the risk of PONV in our study. Obstetric patients are more prone to nausea and vomiting due to physiological changes caused by pregnancy. PONV has been attributed to increased estrogen and progesterone levels during pregnancy, dysfunction of the lower esophageal sphincter, and hormonal changes during pregnancy affecting the neurovestibular system and emetic center in the brainstem.[ 9 – 11 ] Cesarean section has also been identified previously as a risk factor for PONV.[ 10 ] Therefore, female sex, pregnancy, and cesarean section can all be considered risk factors for PONV. Surgery-related risk factors for PONV were an operation time of 50–90 min, a total infusion volume of > 1000 ml, and need for blood transfusion. These risk factors suggest that surgical invasiveness and circulatory dynamics (especially intraoperative hypotension) increase the risk of PONV. A longer operation time increases the amount of time patients are exposed to blood loss, relative fluid deprivation, or hypotension. It is easy to imagine that the incidence of PONV is higher because of these factors. However, it is more likely that PONV is a secondary consequence of massive blood loss and the associated hypotension rather than a direct consequence of the transfusion procedure. Previous reports have also indicated that hypotension is a risk factor for PONV.[ 12 ] Our present findings are consistent with the possibility that lack of use of phenylephrine (a vasopressor agent) and lack of use of atropine (an anti-arrhythmic agent used to treat tachycardia) are risk factors for PONV and that maintenance of hemodynamics affects the occurrence of PONV. Furthermore, we found that the intraoperative infusion volume was most strongly associated with PONV. It is conceivable that this finding is associated with circulating dynamics including hypotension and operation time, but individual analyses using SHAP did not provide clear results on the association between infusion volume and PONV. Anesthesia-related risk factors for PONV included lack of use of epidural anesthesia, level of spinal anesthesia (L3-4), and postoperative use of fentanyl, propofol, or dexmedetomidine. In this study, the risk of PONV was lower when epidural anesthesia was used. This implies that postoperative pain increases the incidence of PONV. Previous reports have also indicated that severe postoperative pain increases PONV. [ 13 , 14 ] However, use of fentanyl for postoperative epidural anesthesia was a risk factor for PONV. There are numerous studies showing that use of intravenous fentanyl increases PONV. [ 15 – 17 ] In this study, administration of small amounts of fentanyl for spinal anesthesia was not a risk factor for PONV. Previous reports have shown that small doses of fentanyl administered for spinal anesthesia significantly prolonged the duration of analgesia postoperatively, even though the risk of PONV was comparable with that of local anesthesia alone.[ 18 ] [ 19 ] Based on these observations, use of fentanyl should be avoided postoperatively whenever possible, but if fentanyl must be used, it is better administered in small doses for spinal anesthesia. We found that puncture at L3-4 for spinal anesthesia was a risk factor for PONV. Previous studies have found that the incidence of PONV is greater when the puncture site for spinal anesthesia is at a higher level[ 8 ] and that the incidence of PONV is greater in the at higher levels of spinal anesthesia.[ 12 ] Hypotension and bradycardia were more common when spinal anesthesia was at a higher level than a lower level. Both hypotension and bradycardia may predispose to PONV. It has also been reported that development of PONV involves activation of vagal activity as a result of sympathetic blockade at a higher spinal level.[ 20 ] In this study, the sedatives propofol and dexmedetomidine were identified to be risk factors for PONV. In previous reports, these agents were not identified to increase the risk of PONV.[ 21 , 22 ] Considering that the operation time is usually longer when sedative medications are used, it is possible that surgery with propofol and dexmedetomidine was longer than that when other types of surgery were performed. Furthermore, inadvertent use of propofol and dexmedetomidine may have increased the incidence of PONV as a result of respiratory depression and hypoxia. There is a need for further research on the association of propofol and dexmedetomidine with PONV. Limitations A major limitation of this study was the small number of samples. As a result, the confidence level of the GBM model was comparable with that of the conventional KNN model. Moreover, there were some items that could not be analyzed adequately because of an insufficient number of samples. Another limitation was bias in the surgical procedures, with a large number of cesarean sections (27%, 1248/4574). Sedative medications used during surgery may have contributed to the risk of PONV, but these factors could not be excluded when performing the analysis. CONCLUSION In this study, we analyzed risk factors for PONV under spinal anesthesia using deep learning AI. We identifies female sex and cesarean section, which are known risk factors for PONV after surgery under spinal anesthesia. Our findings also suggest that fluid volume, blood transfusion, and agents that normalize hemodynamics, such as phenylephrine and atropine, are important in preventing PONV. However, a larger number of samples is necessary to improve the accuracy of analysis with AI. Methods This study was approved by the Ethics Committee of Tohoku University School of Medicine (#2023–31642, February 26, 2021). Before enrollment of any patients, the trial was registered in the UMIN Clinical Trials Registry (identifier UMIN000050012; principal investigator Hiroshi Hoshijima; registration date January 11, 2023). Due to the retrospective nature of the study, (Ethics Committee of Tohoku University School of Medicine) waived the need of obtaining informed consent. We performed the experiments following the STROBE statement. Research involving human research participants must have been performed in accordance with the Declaration of Helsinki. The study included patients aged older than 20 years who underwent surgery under general anesthesia at Tohoku University School of Medicine between 2010 and 2022. The patients were divided into two groups according to whether they developed PONV. Exclusion criteria for this study age younger than 20 years, emergency surgery, surgery under general anesthesia, severe intraoperative or postoperative complications (cardiac arrest, severe arrhythmia, myocardial infarction, massive bleeding, asthma, pulmonary embolus), and admission to an intensive care unit after surgery. Patient and surgical data were collected from the medical records, including the anesthetic record. We focused on data obtained within 24 h after surgery and collected the following information: sex, age, body mass index (BMI), duration of surgery, duration of anesthesia, puncture level for spinal anesthesia, puncture level for epidural anesthesia, type of local spinal anesthesia (equal or high specific gravity), drugs used during surgery (ephedrine [40 mg per ampoule], atropine [0.5 mg per ampoule], phenylephrine [1 mg per ampoule], nicardipine [2 mg per ampoule], hydrocortisone, prednisolone, oxytocin), administration of fentanyl into the spinal arachnoid, postoperative use of fentanyl as epidural anesthesia, total infusion volume, total urine output, total blood loss, total blood transfusion, and intraoperative use of sedative agents (propofol, midazolam, dexmedetomidine) (Table 1 ). The type of surgery was classified as cesarean section or other surgery. Machine learning modeling We constructed a model that predicts the probability of PONV and used it to determine the impact of items in the patient data. Next, we identified items that were risk factors for PONV according to their impact. The model predicts the probability of PONV by gradient tree boosting,[ 23 ] which is a widely used machine learning algorithm because of its accuracy and computational efficiency. The model consists of multiple decision trees and summarizes their prediction values, resulting in a final prediction probability. We implemented the model using the LightGBM framework.[ 24 ] We randomly divided the patient data into a training set (70%), a validation set (20%), and a test set (10%). We calculated the impact of each item using the Shapley value.[ 25 ] The Shapley value of an item represents its contribution to the prediction value. Eq. (1) defines the Shapley value \({\varphi }_{i}\) for item \(i\) in model \(f\) . \(\mathcal{R}\) is the set of all items in the patient data, \({P}^{R}\) is the set of features including item \(i\) , and \(M\) is the number of items ( M = 31 in this study). Thus, the Shapley value for item \(i\) is the expected value over sets of items with and without item \(i\) . We used an approximation framework to calculate the Shapley value.[ 26 ] $$\begin{array}{c}{\varphi }_{i}=\sum _{R\in \mathcal{R}}\frac{1}{M!}\left[f\left({P}^{R}\right)-f\left({P}^{R}\setminus \left\{i\right\}\right)\right]\#\left(1\right)\end{array}$$ Declarations Acknowledgments Assistance with the study: none. Authors contributions H.H. and T.M. had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: H.H., T.M., S.S., M.Y., K.M. Acquisition, analysis, or interpretation of data: H.H., T.M., K.D., T.S., K.M. Drafting of manuscript: H.H., T.M., D.K., S.S.,M.Y., K.M. Critical revision of the manuscript for important intellectual content: H.H., T.M., S.S., S.O., M.Y., K.M. Statistical analysis: T.M., S.O., T.S. Obtaining funding: H.H., K.M. Financial support and sponsorship This work was supported by a Grant-in-Aid from the Japan Society for the Promotion of Science (to H. Hoshijima; 22K10211, and to K. Mizuta; 21K19588). Competing interests None declared. Patient consent for publication Not required. Data availability statement: The datasets generated and/or analysed during the current study are not publicly available due clinical data obtained from patients but are available from the corresponding author on reasonable request. Ethics approval This study was approved by the Ethics Committee of Tohoku University School of Medicine (#2023-31642, February 26, 2021). Conflict of Interest Statement: We have no conflict interest. References Apfel CC, Korttila K, Abdalla M, Kerger H, Turan A, Vedder I, Zernak C, Danner K, Jokela R, Pocock SJ, Trenkler S, Kredel M, Biedler A, Sessler DI, Roewer N, Investigators I. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med. 2004;350(24):2441-51. doi: 10.1056/NEJMoa032196. Gan TJ, Belani KG, Bergese S, Chung F, Diemunsch P, Habib AS, Jin Z, Kovac AL, Meyer TA, Urman RD, Apfel CC, Ayad S, Beagley L, Candiotti K, Englesakis M, Hedrick TL, Kranke P, Lee S, Lipman D, Minkowitz HS, Morton J, Philip BK. Fourth Consensus Guidelines for the Management of Postoperative Nausea and Vomiting. Anesth Analg. 2020;131(2):411-48. doi: 10.1213/ANE.0000000000004833. Harmon D, Ryan M, Kelly A, Bowen M. Acupressure and prevention of nausea and vomiting during and after spinal anaesthesia for caesarean section. Br J Anaesth. 2000;84(4):463-7. doi: 10.1093/oxfordjournals.bja.a013471. Moore DC, Bridenbaugh LD. Spinal (subarachnoid) block. A review of 11,574 cases. JAMA. 1966;195(11):907-12. doi: 10.1001/jama.195.11.907. Ratra CK, Badola RP, Bhargava KP. A study of factors concerned in emesis during spinal anaesthesia. Br J Anaesth. 1972;44(11):1208-11. doi: 10.1093/bja/44.11.1208. Attia ZI, Noseworthy PA, Lopez-Jimenez F, Asirvatham SJ, Deshmukh AJ, Gersh BJ, Carter RE, Yao X, Rabinstein AA, Erickson BJ, Kapa S, Friedman PA. An artificial intelligence-enabled ECG algorithm for the identification of patients with atrial fibrillation during sinus rhythm: a retrospective analysis of outcome prediction. Lancet. 2019;394(10201):861-7. doi: 10.1016/S0140-6736(19)31721-0. Yamamoto Y, Tsuzuki T, Akatsuka J, Ueki M, Morikawa H, Numata Y, Takahara T, Tsuyuki T, Tsutsumi K, Nakazawa R, Shimizu A, Maeda I, Tsuchiya S, Kanno H, Kondo Y, Fukumoto M, Tamiya G, Ueda N, Kimura G. Automated acquisition of explainable knowledge from unannotated histopathology images. Nat Commun. 2019;10(1):5642. doi: 10.1038/s41467-019-13647-8. Carpenter RL, Caplan RA, Brown DL, Stephenson C, Wu R. Incidence and risk factors for side effects of spinal anesthesia. Anesthesiology. 1992;76(6):906-16. doi: 10.1097/00000542-199206000-00006. Broussard CN, Richter JE. Nausea and vomiting of pregnancy. Gastroenterol Clin North Am. 1998;27(1):123-51. doi: 10.1016/s0889-8553(05)70350-2. Balki M, Carvalho JC. Intraoperative nausea and vomiting during cesarean section under regional anesthesia. Int J Obstet Anesth. 2005;14(3):230-41. doi: 10.1016/j.ijoa.2004.12.004. Koch KL, Frissora CL. Nausea and vomiting during pregnancy. Gastroenterol Clin North Am. 2003;32:201–34. doi: https://doi.org/10.1016/S0889-8553(02)00070-5. Spelina KR, Gerber HR, Pagels IL. Nausea and vomiting during spinal anaesthesia. Effect of metoclopramide and domperidone: a double-blind trial. Anaesthesia. 1984;39(2):132-7. doi: 10.1111/j.1365-2044.1984.tb09500.x. Coughlin SM, Karanicolas PJ, Emmerton-Coughlin HM, Kanbur B, Kanbur S, Colquhoun PH. Better late than never? Impact of local analgesia timing on postoperative pain in laparoscopic surgery: a systematic review and metaanalysis. Surg Endosc. 2010;24(12):3167-76. doi: 10.1007/s00464-010-1111-1. Marks JL, Ata B, Tulandi T. Systematic review and metaanalysis of intraperitoneal instillation of local anesthetics for reduction of pain after gynecologic laparoscopy. J Minim Invasive Gynecol. 2012;19(5):545-53. doi: 10.1016/j.jmig.2012.04.002. Lee S, Woo S, Oh EJ, Park M. A randomized controlled trial of propofol-remifentanil total intravenous anesthesia and sevoflurane-fentanyl anesthesia on early postoperative fatigue in patients undergoing laparoscopic colorectal surgery. Qual Life Res. 2023. doi: 10.1007/s11136-023-03510-1. Apfel CC, Laara E, Koivuranta M, Greim CA, Roewer N. A simplified risk score for predicting postoperative nausea and vomiting: conclusions from cross-validations between two centers. Anesthesiology. 1999;91(3):693-700. doi: 10.1097/00000542-199909000-00022. Toleska M, Dimitrovski A, Dimitrovska NT. Postoperative Nausea and Vomiting in Opioid-Free Anesthesia Versus Opioid Based Anesthesia in Laparoscopic Cholecystectomy. Pril (Makedon Akad Nauk Umet Odd Med Nauki). 2022;43(3):101-8. doi: 10.2478/prilozi-2022-0042. Fonseca NM, Guimaraes GMN, Pontes JPJ, Azi L, de Avila Oliveira R. Safety and effectiveness of adding fentanyl or sufentanil to spinal anesthesia: systematic review and meta-analysis of randomized controlled trials. Braz J Anesthesiol. 2023;73(2):198-216. doi: 10.1016/j.bjane.2021.10.010. Uppal V, Retter S, Casey M, Sancheti S, Matheson K, McKeen DM. Efficacy of Intrathecal Fentanyl for Cesarean Delivery: A Systematic Review and Meta-analysis of Randomized Controlled Trials With Trial Sequential Analysis. Anesth Analg. 2020;130(1):111-25. doi: 10.1213/ANE.0000000000003975. Ward RJ, Kennedy WF, Bonica JJ, Martin WE, Tolas AG, Akamatsu T. Experimental evaluation of atropine and vasopressors for the treatment of hypotension of high subarachnoid anesthesia. Anesth Analg. 1966;45(5):621-9. Kang H, Lim T, Lee HJ, Kim TW, Kim W, Chang HW. Comparison of the effect of dexmedetomidine and midazolam under spinal anesthesia for cesarean delivery: a randomized controlled trial, single center study in South Korea. Anesth Pain Med (Seoul). 2023;18(2):159-68. doi: 10.17085/apm.22257. Kim H, Kim Y, Bae J, Yoo S, Lim YJ, Kim JT. Comparison of remimazolam and dexmedetomidine for intraoperative sedation in patients undergoing lower extremity surgery under spinal anesthesia: a randomized clinical trial. Reg Anesth Pain Med. 2023. doi: 10.1136/rapm-2023-104415. Friedman JH. Greedy function approximation: A gradient boosting machine. Ann Stat. 2001;29(5):1189-232. doi: DOI 10.1214/aos/1013203451. Ke GL, Meng Q, Finley T, Wang TF, Chen W, Ma WD, Ye QW, Liu TY. LightGBM: A Highly Efficient Gradient Boosting Decision Tree. Adv Neur In. 2017;30. Shapley LS. Stochastic Games. P Natl Acad Sci USA. 1953;39(10):1095-100. doi: DOI 10.1073/pnas.39.10.1095. Lundberg SM, Erion G, Chen H, DeGrave A, Prutkin JM, Nair B, Katz R, Himmelfarb J, Bansal N, Lee SI. From Local Explanations to Global Understanding with Explainable AI for Trees. Nat Mach Intell. 2020;2(1):56-67. doi: 10.1038/s42256-019-0138-9. Additional Declarations No competing interests reported. Supplementary Files SupplementaryOnlineContentSpinal.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4421679","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":307480820,"identity":"e8f4b5d0-531f-45e1-8549-c350e584bbc9","order_by":0,"name":"Hiroshi Hoshijima","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAvklEQVRIiWNgGAWjYBACAwkGBmYGAxs5EOfAAxK0pBmDtSQQr4XhUGIDiEeUFnPp5sOfCwoOpM8PO/wQaIudnG4DAS2Wc46lSc8wuJO78XaaAVBLsrHZAUIOu5Fjxsxj8Cx34+wEkJYDiduI0GL8mcfgcLrh7PQPRGsxkAZqSZCXziHalrQ0oJY0ww3SOQUHEgyI8kvy4c88f2zk5Wenb/7wocJOjqAWhF6wSgNilYOAfAMpqkfBKBgFo2BEAQAImUahhqRo6gAAAABJRU5ErkJggg==","orcid":"","institution":"Tohoku University Graduate School of Dentistry","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Hiroshi","middleName":"","lastName":"Hoshijima","suffix":""},{"id":307480821,"identity":"f6874879-1e57-4dd3-89a6-1264ea1bbf82","order_by":1,"name":"Tomo Miyazaki","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Tomo","middleName":"","lastName":"Miyazaki","suffix":""},{"id":307480822,"identity":"7e3273d0-9d11-46ce-b075-284ddd3cf793","order_by":2,"name":"Shinichiro Omachi","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Shinichiro","middleName":"","lastName":"Omachi","suffix":""},{"id":307480823,"identity":"a8ea26f3-5713-4074-90e0-95cb96870a3f","order_by":3,"name":"Daisuke Konno","email":"","orcid":"","institution":"Tohoku University School of Medicine","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Daisuke","middleName":"","lastName":"Konno","suffix":""},{"id":307480824,"identity":"250f373e-d6bd-4614-a0d4-6ba3abe4c488","order_by":4,"name":"Shigekazu Sugino","email":"","orcid":"","institution":"Tohoku University School of Medicine","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Shigekazu","middleName":"","lastName":"Sugino","suffix":""},{"id":307480825,"identity":"b1ff1602-5e2a-4068-85df-564faa557567","order_by":5,"name":"Masanori Yamauchi","email":"","orcid":"","institution":"Tohoku University School of Medicine","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Masanori","middleName":"","lastName":"Yamauchi","suffix":""},{"id":307480826,"identity":"5bd37330-59e8-4066-87e6-40e579b34d08","order_by":6,"name":"Toshiya Shiga","email":"","orcid":"","institution":"International University of Health and Welfare Ichikawa Hospital","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Toshiya","middleName":"","lastName":"Shiga","suffix":""},{"id":307480827,"identity":"b201b731-ef38-4ee7-a892-3f94a7a49e80","order_by":7,"name":"Kentaro Mizuta","email":"","orcid":"","institution":"Tohoku University Graduate School of Dentistry","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Kentaro","middleName":"","lastName":"Mizuta","suffix":""}],"badges":[],"createdAt":"2024-05-14 23:53:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4421679/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4421679/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":57727582,"identity":"2d9b26a2-d3e1-4aa3-9820-9cc28573167b","added_by":"auto","created_at":"2024-06-04 21:39:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":69442,"visible":true,"origin":"","legend":"\u003cp\u003eSummary of plot. The risk factors for PONV are plotted in descending order of relevance.\u003c/p\u003e","description":"","filename":"SpinalFigure151.png","url":"https://assets-eu.researchsquare.com/files/rs-4421679/v1/ef555b58b69f0ee2e617c177.png"},{"id":57728601,"identity":"568efa13-bb36-4311-8032-3590e375258d","added_by":"auto","created_at":"2024-06-04 21:47:04","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":31711,"visible":true,"origin":"","legend":"\u003cp\u003eRelationship between patient-related and time-related risk factors and PONV in SHAP value. The X-axis represents (A) sex (0: female, 1: male), (B) age, and (C) BMI (kg m\u003csup\u003e2 -1\u003c/sup\u003e), The Y-axis represents the risk of PONV. SHAP value above 0 is related to the risk of PONV.\u003c/p\u003e","description":"","filename":"SpinalFigure152.png","url":"https://assets-eu.researchsquare.com/files/rs-4421679/v1/87efca098eeed4a920fe9887.png"},{"id":57727584,"identity":"3796afe6-cf3c-4254-b2a1-44472588fa24","added_by":"auto","created_at":"2024-06-04 21:39:04","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":63972,"visible":true,"origin":"","legend":"\u003cp\u003eRelationship between fluid-related risk factors and PONV in SHAP value. The X-axis represents (A) duration of surgery (min), (B) duration of anesthesia (min), (C) Caesarean section, (D) total blood loss (ml), (E) total blood transfusion volume (ml), (F) total infusion volume (ml), and (G) total urine volume. The Y-axis represents the risk of PONV. SHAP value above 0 is related to the risk of PONV.\u003c/p\u003e","description":"","filename":"SpinalFigure153.png","url":"https://assets-eu.researchsquare.com/files/rs-4421679/v1/0161639cab839d79a861b567.png"},{"id":57727587,"identity":"c20703fb-6844-43ce-9438-919a1accde94","added_by":"auto","created_at":"2024-06-04 21:39:04","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":57651,"visible":true,"origin":"","legend":"\u003cp\u003eRelationship between anesthetic agent-related risk factors and PONV in SHAP value. The X-axis represents (A) Puncture level of spinal anesthesia (0: L 2-3, 1: L 3-4, and 2: L 4-5), (B) Puncture level of epidural anesthesia (0: No use of epidural, 1: Th1-6, and 2: Th 7-12), (C) Bupivacaine (high specific gravity, (D) Bupivacaine (equal specific gravity), (E) Administration of fentanyl to spinal arachnoid, and (F) fentanyl in epidural anesthesia. The Y-axis represents the risk of PONV. SHAP value above 0 is related to the risk of PONV.\u003c/p\u003e","description":"","filename":"SpinalFigure154.png","url":"https://assets-eu.researchsquare.com/files/rs-4421679/v1/43b5fa3f36d243ef7997f429.png"},{"id":57727585,"identity":"2b10009a-b299-407a-bdf1-4539303b15a3","added_by":"auto","created_at":"2024-06-04 21:39:04","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":45299,"visible":true,"origin":"","legend":"\u003cp\u003eRelationship between anesthetic agent-related risk factors and PONV in SHAP value. The X-axis represents (A) Propofol, (B) Dexmedetomidine, (C) Atonin, (D) Ephedrine, (E) Phenylephrine, and (F) Atropine. The Y-axis represents the risk of PONV. SHAP value above 0 is related to the risk of PONV.\u003c/p\u003e","description":"","filename":"SpinalFigure155.png","url":"https://assets-eu.researchsquare.com/files/rs-4421679/v1/c4c0cee8cf60d72f720d8d97.png"},{"id":62639728,"identity":"620868ed-15df-47e8-83d9-586c75e58ade","added_by":"auto","created_at":"2024-08-16 18:26:08","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":745926,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4421679/v1/96eba474-bc13-4b1e-b834-d387db6e4213.pdf"},{"id":57728600,"identity":"7433078c-6123-46c5-bcc6-b9d3a76c0d89","added_by":"auto","created_at":"2024-06-04 21:47:04","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":262289,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryOnlineContentSpinal.docx","url":"https://assets-eu.researchsquare.com/files/rs-4421679/v1/6ea04a9e4a82aed9e38a0f5b.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine Learning-based Prediction of Postoperative Nausea and Vomiting after Spinal Anesthesia: A Retrospective Observational Study","fulltext":[{"header":"Introduction","content":"\u003cp\u003ePostoperative nausea and vomiting (PONV) is a frequent and serious complication in patients undergoing surgery.[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] This complication decreases patient satisfaction with surgery and is associated with increased healthcare costs due to prolonged hospitalization. Numerous methods have been proposed to avoid PONV but have not completely prevented it.[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eThe incidence of PONV ranges from 5\u0026ndash;42% in patients undergoing lower extremity surgery, lower abdominal surgery, and cesarean section under spinal anesthesia.[\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e] Previous studies have identified several risk factors for PONV after spinal anesthesia compared with general anesthesia, including female sex, puncture level, and intraoperative hypotension. However, these factors remain controversial because they are very different from the known risk factors after general anesthesia.\u003c/p\u003e \u003cp\u003eArtificial intelligence (AI)-based technologies have evolved considerably and are starting to be applied in medicine.[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e] The evolution of machine learning with AI, particularly deep learning methods such as neural networks and convolutional neural networks, is a driving force behind the development of AI. Deep learning differs from traditional machine learning in that AI learns differences between samples and selects the correct answer. Therefore, AI is uniquely capable of recognizing changes beyond human perception and establishing AI-specific identification methods.\u003c/p\u003e \u003cp\u003eThe objective of the present study was to identify risk factors for PONV using machine learning analysis of AI in patients undergoing surgery with spinal anesthesia.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cp\u003e We obtained data from the Tohoku University database on 5808 patients who underwent surgery under spinal anesthesia. Data for 1,234 patients were excluded (spinal puncture level unclear, n\u0026thinsp;=\u0026thinsp;1,215; missing data, n\u0026thinsp;=\u0026thinsp;12; no PONV adjudication, n\u0026thinsp;=\u0026thinsp;3; no information on drugs used for spinal anesthesia, n\u0026thinsp;=\u0026thinsp;3; concomitant general anesthesia, n\u0026thinsp;=\u0026thinsp;3; and no surgery, n\u0026thinsp;=\u0026thinsp;1) (Supplementary Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e), leaving data for 4,574 patients available for analysis. The patient characteristics are shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. In total, 269 (5.9%) of the 4,574 patients developed PONV.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePatients characteristics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePatients with PONV (269; 5.9%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePatients without PONV (4,305; 94.1%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eGender (Male/Female)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16 (5.9%)/269 (94.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e556 (12.9%)/3,749 (87.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eAge (year)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e40.8\u0026thinsp;\u0026plusmn;\u0026thinsp;15.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e39.5\u0026thinsp;\u0026plusmn;\u0026thinsp;14.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eBody Mass Index (BMI, kg/mg-2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25.2\u0026thinsp;\u0026plusmn;\u0026thinsp;4.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e26.4\u0026thinsp;\u0026plusmn;\u0026thinsp;19.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eDuration of anesthesia (min)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e97.4\u0026thinsp;\u0026plusmn;\u0026thinsp;45.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e61.2\u0026thinsp;\u0026plusmn;\u0026thinsp;27.3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eDuration of surgery (min)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e68.1\u0026thinsp;\u0026plusmn;\u0026thinsp;38.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e90.3\u0026thinsp;\u0026plusmn;\u0026thinsp;33.1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eTotal infusion volume (ml)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1366.6\u0026thinsp;\u0026plusmn;\u0026thinsp;622.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1189.4\u0026thinsp;\u0026plusmn;\u0026thinsp;585.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eTotal bleeding loss (ml)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e720.5\u0026thinsp;\u0026plusmn;\u0026thinsp;735.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e613.8\u0026thinsp;\u0026plusmn;\u0026thinsp;660.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eTotal blood transfusion volume (ml)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38.9\u0026thinsp;\u0026plusmn;\u0026thinsp;175.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e18.7\u0026thinsp;\u0026plusmn;\u0026thinsp;131.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eTotal urine output (n)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e253.3\u0026thinsp;\u0026plusmn;\u0026thinsp;250.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e201.6\u0026thinsp;\u0026plusmn;\u0026thinsp;247.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eCaesarean section (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,175\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eAnesthesic agents\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.5%Bupivacaine (High specific gravity) (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e218\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3,651\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.5%Bupivacaine (Equal specific gravity) (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e654\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFentanyl Intrathecal (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,704\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFentanyl Post operative (0.25mg/1A) (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNone\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e153\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3127\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (A)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e112\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1399\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (A)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e36\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 (A)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4 (A)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5 (A)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6 (A)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMorphine (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e44\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePropofol (prefilled syringe) (mg**)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e247\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMidazolam (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e105\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDexmedetomidine (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e549\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003ePuncture level of spinal anesthesia (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eL2-3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e515\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eL3-4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e238\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3554\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eL4-5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e226\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eL5-S1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003ePuncture level of epidural anesthesia (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNone\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e982\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTh1-6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTh7-12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e182\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2988\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eL3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e316\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eCardiovascular agents\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAtropine (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e132\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEphedrine (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,479\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNicardipine (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e36\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePhenylephrine (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e126\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2,390\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAtonin(Oxxytocin) (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e192\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3,077\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eSteroid\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHydrocortisone (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePrednisolone (n*)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003e* Total number of ampules or prefilled syringe used in each group.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003e**The amount of drug used (mg) is not the actual dose administered to the patients, but is calculated backwards from the number\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e \u003cp\u003eof ampules/vials of the drug administered to the patients.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eFirst, we evaluated the prediction model using three metrics, namely, the true positive rate, false positive rate, and area under the curve (AUC). We compared the model with a naive model (i.e., the k-nearest neighbor model [KNN]), with k set to 9. Supplementary Fig. S2 shows the receiver-operating characteristic curves. The model outperformed the KNN. Specifically, the AUC was 0.53 for the model and 0.54 for the KNN, demonstrating the ability of our model to recognize PONV and warranting further analysis.\u003c/p\u003e \u003cp\u003eNext, we calculated the mean absolute Shapley Additive exPlanations (SHAP) value for each item by averaging the absolute SHAP values over the test data (Supplementary Fig. S3). If an item had a SHAP value greater than 0, it was judged to be a risk factor for PONV.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eRisk factors for PONV\u003c/h2\u003e \u003cp\u003eThe total infusion volume was the item most strongly associated with PONV after surgery under spinal anesthesia. Duration of surgery, BMI, total urine output, and postoperative fentanyl were subsequently associated with PONV (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe analysis identified female sex as a patient-related risk factor for PONV (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA) and an operation time of 50\u0026ndash;90 min as a time-related risk factor (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Fluid-related risk factors for PONV were need for blood transfusion (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE) and a total fluid infusion volume of over 1000 ml during anesthesia (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eF).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAnesthesia-related risk factors for PONV were puncture level for spinal anesthesia (spinal level 3\u0026ndash;4) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA) and no concomitant epidural anesthesia (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). Use of propofol and dexmedetomidine during surgery was also a risk factor for PONV (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). Postoperative use of fentanyl was also a risk factor for PONV (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eF), as was not using phenylephrine (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE), atropine (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF), or oxytocin (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC), and cesarean section (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn contrast, patient age (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB), BMI (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC), duration of anesthesia (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB), total blood loss (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD), total urine output (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eG), intrathecal use of fentanyl (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE), use of bupivacaine (high or equal specific gravity) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC, D), and use of ephedrine (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD) did not have any clear relationship with PONV. The associations of midazolam, nicardipine, hydrocortisone, prednisolone, and morphine with PONV could not be analyzed because of the small number of samples.\u003c/p\u003e \u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eIn this study, female sex was the only patient-related risk factor for PONV, and surgery-related risk factors included cesarean section, an operation time of 50\u0026ndash;90 min, a total infusion volume of \u0026gt;\u0026thinsp;1000 ml, and need for blood transfusion.\u003c/p\u003e \u003cp\u003ePrevious studies found that PONV under spinal anesthesia was more common in women,[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] which is consistent with our results. Cesarean section also increased the risk of PONV in our study. Obstetric patients are more prone to nausea and vomiting due to physiological changes caused by pregnancy. PONV has been attributed to increased estrogen and progesterone levels during pregnancy, dysfunction of the lower esophageal sphincter, and hormonal changes during pregnancy affecting the neurovestibular system and emetic center in the brainstem.[\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e] Cesarean section has also been identified previously as a risk factor for PONV.[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] Therefore, female sex, pregnancy, and cesarean section can all be considered risk factors for PONV.\u003c/p\u003e \u003cp\u003eSurgery-related risk factors for PONV were an operation time of 50\u0026ndash;90 min, a total infusion volume of \u0026gt;\u0026thinsp;1000 ml, and need for blood transfusion. These risk factors suggest that surgical invasiveness and circulatory dynamics (especially intraoperative hypotension) increase the risk of PONV. A longer operation time increases the amount of time patients are exposed to blood loss, relative fluid deprivation, or hypotension. It is easy to imagine that the incidence of PONV is higher because of these factors. However, it is more likely that PONV is a secondary consequence of massive blood loss and the associated hypotension rather than a direct consequence of the transfusion procedure. Previous reports have also indicated that hypotension is a risk factor for PONV.[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] Our present findings are consistent with the possibility that lack of use of phenylephrine (a vasopressor agent) and lack of use of atropine (an anti-arrhythmic agent used to treat tachycardia) are risk factors for PONV and that maintenance of hemodynamics affects the occurrence of PONV. Furthermore, we found that the intraoperative infusion volume was most strongly associated with PONV. It is conceivable that this finding is associated with circulating dynamics including hypotension and operation time, but individual analyses using SHAP did not provide clear results on the association between infusion volume and PONV.\u003c/p\u003e \u003cp\u003eAnesthesia-related risk factors for PONV included lack of use of epidural anesthesia, level of spinal anesthesia (L3-4), and postoperative use of fentanyl, propofol, or dexmedetomidine. In this study, the risk of PONV was lower when epidural anesthesia was used. This implies that postoperative pain increases the incidence of PONV. Previous reports have also indicated that severe postoperative pain increases PONV. [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e] However, use of fentanyl for postoperative epidural anesthesia was a risk factor for PONV. There are numerous studies showing that use of intravenous fentanyl increases PONV. [\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eIn this study, administration of small amounts of fentanyl for spinal anesthesia was not a risk factor for PONV. Previous reports have shown that small doses of fentanyl administered for spinal anesthesia significantly prolonged the duration of analgesia postoperatively, even though the risk of PONV was comparable with that of local anesthesia alone.[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e] [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e] Based on these observations, use of fentanyl should be avoided postoperatively whenever possible, but if fentanyl must be used, it is better administered in small doses for spinal anesthesia.\u003c/p\u003e \u003cp\u003eWe found that puncture at L3-4 for spinal anesthesia was a risk factor for PONV. Previous studies have found that the incidence of PONV is greater when the puncture site for spinal anesthesia is at a higher level[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] and that the incidence of PONV is greater in the at higher levels of spinal anesthesia.[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] Hypotension and bradycardia were more common when spinal anesthesia was at a higher level than a lower level. Both hypotension and bradycardia may predispose to PONV. It has also been reported that development of PONV involves activation of vagal activity as a result of sympathetic blockade at a higher spinal level.[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eIn this study, the sedatives propofol and dexmedetomidine were identified to be risk factors for PONV. In previous reports, these agents were not identified to increase the risk of PONV.[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] Considering that the operation time is usually longer when sedative medications are used, it is possible that surgery with propofol and dexmedetomidine was longer than that when other types of surgery were performed. Furthermore, inadvertent use of propofol and dexmedetomidine may have increased the incidence of PONV as a result of respiratory depression and hypoxia. There is a need for further research on the association of propofol and dexmedetomidine with PONV.\u003c/p\u003e \u003cp\u003eLimitations\u003c/p\u003e \u003cp\u003eA major limitation of this study was the small number of samples. As a result, the confidence level of the GBM model was comparable with that of the conventional KNN model. Moreover, there were some items that could not be analyzed adequately because of an insufficient number of samples. Another limitation was bias in the surgical procedures, with a large number of cesarean sections (27%, 1248/4574). Sedative medications used during surgery may have contributed to the risk of PONV, but these factors could not be excluded when performing the analysis.\u003c/p\u003e"},{"header":"CONCLUSION","content":"\u003cp\u003eIn this study, we analyzed risk factors for PONV under spinal anesthesia using deep learning AI. We identifies female sex and cesarean section, which are known risk factors for PONV after surgery under spinal anesthesia. Our findings also suggest that fluid volume, blood transfusion, and agents that normalize hemodynamics, such as phenylephrine and atropine, are important in preventing PONV. However, a larger number of samples is necessary to improve the accuracy of analysis with AI.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e This study was approved by the Ethics Committee of Tohoku University School of Medicine (#2023\u0026ndash;31642, February 26, 2021). Before enrollment of any patients, the trial was registered in the UMIN Clinical Trials Registry (identifier UMIN000050012; principal investigator Hiroshi Hoshijima; registration date January 11, 2023). Due to the retrospective nature of the study, (Ethics Committee of Tohoku University School of Medicine) waived the need of obtaining informed consent. We performed the experiments following the STROBE statement. Research involving human research participants must have been performed in accordance with the Declaration of Helsinki. The study included patients aged older than 20 years who underwent surgery under general anesthesia at Tohoku University School of Medicine between 2010 and 2022. The patients were divided into two groups according to whether they developed PONV. Exclusion criteria for this study age younger than 20 years, emergency surgery, surgery under general anesthesia, severe intraoperative or postoperative complications (cardiac arrest, severe arrhythmia, myocardial infarction, massive bleeding, asthma, pulmonary embolus), and admission to an intensive care unit after surgery.\u003c/p\u003e \u003cp\u003ePatient and surgical data were collected from the medical records, including the anesthetic record. We focused on data obtained within 24 h after surgery and collected the following information: sex, age, body mass index (BMI), duration of surgery, duration of anesthesia, puncture level for spinal anesthesia, puncture level for epidural anesthesia, type of local spinal anesthesia (equal or high specific gravity), drugs used during surgery (ephedrine [40 mg per ampoule], atropine [0.5 mg per ampoule], phenylephrine [1 mg per ampoule], nicardipine [2 mg per ampoule], hydrocortisone, prednisolone, oxytocin), administration of fentanyl into the spinal arachnoid, postoperative use of fentanyl as epidural anesthesia, total infusion volume, total urine output, total blood loss, total blood transfusion, and intraoperative use of sedative agents (propofol, midazolam, dexmedetomidine) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The type of surgery was classified as cesarean section or other surgery.\u003c/p\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eMachine learning modeling\u003c/h2\u003e \u003cp\u003eWe constructed a model that predicts the probability of PONV and used it to determine the impact of items in the patient data. Next, we identified items that were risk factors for PONV according to their impact. The model predicts the probability of PONV by gradient tree boosting,[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e] which is a widely used machine learning algorithm because of its accuracy and computational efficiency. The model consists of multiple decision trees and summarizes their prediction values, resulting in a final prediction probability. We implemented the model using the LightGBM framework.[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e] We randomly divided the patient data into a training set (70%), a validation set (20%), and a test set (10%).\u003c/p\u003e \u003cp\u003eWe calculated the impact of each item using the Shapley value.[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] The Shapley value of an item represents its contribution to the prediction value. Eq.\u0026nbsp;(1) defines the Shapley value \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({\\varphi }_{i}\\)\u003c/span\u003e\u003c/span\u003e for item \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e in model \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(f\\)\u003c/span\u003e\u003c/span\u003e. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\mathcal{R}\\)\u003c/span\u003e\u003c/span\u003e is the set of all items in the patient data, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({P}^{R}\\)\u003c/span\u003e\u003c/span\u003e is the set of features including item \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(M\\)\u003c/span\u003e\u003c/span\u003e is the number of items (\u003cem\u003eM\u003c/em\u003e\u0026thinsp;=\u0026thinsp;31 in this study). Thus, the Shapley value for item \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e is the expected value over sets of items with and without item \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\)\u003c/span\u003e\u003c/span\u003e. We used an approximation framework to calculate the Shapley value.[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\begin{array}{c}{\\varphi }_{i}=\\sum _{R\\in \\mathcal{R}}\\frac{1}{M!}\\left[f\\left({P}^{R}\\right)-f\\left({P}^{R}\\setminus \\left\\{i\\right\\}\\right)\\right]\\#\\left(1\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e Assistance with the study: none.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors contributions\u003c/strong\u003e H.H. and T.M. had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.\u003c/p\u003e\n\u003cp\u003eStudy concept and design: H.H., T.M., S.S., M.Y., K.M.\u003c/p\u003e\n\u003cp\u003eAcquisition, analysis, or interpretation of data: H.H., T.M., K.D., T.S., K.M.\u003c/p\u003e\n\u003cp\u003eDrafting of manuscript: H.H., T.M., D.K., S.S.,M.Y., K.M.\u003c/p\u003e\n\u003cp\u003eCritical revision of the manuscript for important intellectual content: H.H., T.M., S.S., S.O., M.Y., K.M.\u003c/p\u003e\n\u003cp\u003eStatistical analysis: T.M., S.O., T.S.\u003c/p\u003e\n\u003cp\u003eObtaining funding: H.H., K.M.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFinancial support and sponsorship\u003c/strong\u003e This work was supported by a Grant-in-Aid from the Japan Society for the Promotion of Science (to H. Hoshijima; 22K10211, and to K. Mizuta; 21K19588).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e None declared.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePatient consent for publication\u0026nbsp;\u003c/strong\u003eNot required.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement:\u003c/strong\u003e The datasets generated and/or analysed during the current study are not publicly available due clinical data obtained from patients but are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e This study was approved by the Ethics Committee of Tohoku University School of Medicine (#2023-31642, February 26, 2021).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest Statement:\u0026nbsp;\u003c/strong\u003eWe have no conflict interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eApfel CC, Korttila K, Abdalla M, Kerger H, Turan A, Vedder I, Zernak C, Danner K, Jokela R, Pocock SJ, Trenkler S, Kredel M, Biedler A, Sessler DI, Roewer N, Investigators I. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med. 2004;350(24):2441-51. doi: 10.1056/NEJMoa032196.\u003c/li\u003e\n\u003cli\u003eGan TJ, Belani KG, Bergese S, Chung F, Diemunsch P, Habib AS, Jin Z, Kovac AL, Meyer TA, Urman RD, Apfel CC, Ayad S, Beagley L, Candiotti K, Englesakis M, Hedrick TL, Kranke P, Lee S, Lipman D, Minkowitz HS, Morton J, Philip BK. Fourth Consensus Guidelines for the Management of Postoperative Nausea and Vomiting. Anesth Analg. 2020;131(2):411-48. doi: 10.1213/ANE.0000000000004833.\u003c/li\u003e\n\u003cli\u003eHarmon D, Ryan M, Kelly A, Bowen M. Acupressure and prevention of nausea and vomiting during and after spinal anaesthesia for caesarean section. Br J Anaesth. 2000;84(4):463-7. doi: 10.1093/oxfordjournals.bja.a013471.\u003c/li\u003e\n\u003cli\u003eMoore DC, Bridenbaugh LD. Spinal (subarachnoid) block. A review of 11,574 cases. JAMA. 1966;195(11):907-12. doi: 10.1001/jama.195.11.907.\u003c/li\u003e\n\u003cli\u003eRatra CK, Badola RP, Bhargava KP. A study of factors concerned in emesis during spinal anaesthesia. Br J Anaesth. 1972;44(11):1208-11. doi: 10.1093/bja/44.11.1208.\u003c/li\u003e\n\u003cli\u003eAttia ZI, Noseworthy PA, Lopez-Jimenez F, Asirvatham SJ, Deshmukh AJ, Gersh BJ, Carter RE, Yao X, Rabinstein AA, Erickson BJ, Kapa S, Friedman PA. An artificial intelligence-enabled ECG algorithm for the identification of patients with atrial fibrillation during sinus rhythm: a retrospective analysis of outcome prediction. Lancet. 2019;394(10201):861-7. doi: 10.1016/S0140-6736(19)31721-0.\u003c/li\u003e\n\u003cli\u003eYamamoto Y, Tsuzuki T, Akatsuka J, Ueki M, Morikawa H, Numata Y, Takahara T, Tsuyuki T, Tsutsumi K, Nakazawa R, Shimizu A, Maeda I, Tsuchiya S, Kanno H, Kondo Y, Fukumoto M, Tamiya G, Ueda N, Kimura G. Automated acquisition of explainable knowledge from unannotated histopathology images. Nat Commun. 2019;10(1):5642. doi: 10.1038/s41467-019-13647-8.\u003c/li\u003e\n\u003cli\u003eCarpenter RL, Caplan RA, Brown DL, Stephenson C, Wu R. Incidence and risk factors for side effects of spinal anesthesia. Anesthesiology. 1992;76(6):906-16. doi: 10.1097/00000542-199206000-00006.\u003c/li\u003e\n\u003cli\u003eBroussard CN, Richter JE. Nausea and vomiting of pregnancy. Gastroenterol Clin North Am. 1998;27(1):123-51. doi: 10.1016/s0889-8553(05)70350-2.\u003c/li\u003e\n\u003cli\u003eBalki M, Carvalho JC. Intraoperative nausea and vomiting during cesarean section under regional anesthesia. Int J Obstet Anesth. 2005;14(3):230-41. doi: 10.1016/j.ijoa.2004.12.004.\u003c/li\u003e\n\u003cli\u003eKoch KL, Frissora CL. Nausea and vomiting during pregnancy. Gastroenterol Clin North Am. 2003;32:201\u0026ndash;34. doi: https://doi.org/10.1016/S0889-8553(02)00070-5.\u003c/li\u003e\n\u003cli\u003eSpelina KR, Gerber HR, Pagels IL. Nausea and vomiting during spinal anaesthesia. Effect of metoclopramide and domperidone: a double-blind trial. Anaesthesia. 1984;39(2):132-7. doi: 10.1111/j.1365-2044.1984.tb09500.x.\u003c/li\u003e\n\u003cli\u003eCoughlin SM, Karanicolas PJ, Emmerton-Coughlin HM, Kanbur B, Kanbur S, Colquhoun PH. Better late than never? Impact of local analgesia timing on postoperative pain in laparoscopic surgery: a systematic review and metaanalysis. Surg Endosc. 2010;24(12):3167-76. doi: 10.1007/s00464-010-1111-1.\u003c/li\u003e\n\u003cli\u003eMarks JL, Ata B, Tulandi T. Systematic review and metaanalysis of intraperitoneal instillation of local anesthetics for reduction of pain after gynecologic laparoscopy. J Minim Invasive Gynecol. 2012;19(5):545-53. doi: 10.1016/j.jmig.2012.04.002.\u003c/li\u003e\n\u003cli\u003eLee S, Woo S, Oh EJ, Park M. A randomized controlled trial of propofol-remifentanil total intravenous anesthesia and sevoflurane-fentanyl anesthesia on early postoperative fatigue in patients undergoing laparoscopic colorectal surgery. Qual Life Res. 2023. doi: 10.1007/s11136-023-03510-1.\u003c/li\u003e\n\u003cli\u003eApfel CC, Laara E, Koivuranta M, Greim CA, Roewer N. A simplified risk score for predicting postoperative nausea and vomiting: conclusions from cross-validations between two centers. Anesthesiology. 1999;91(3):693-700. doi: 10.1097/00000542-199909000-00022.\u003c/li\u003e\n\u003cli\u003eToleska M, Dimitrovski A, Dimitrovska NT. Postoperative Nausea and Vomiting in Opioid-Free Anesthesia Versus Opioid Based Anesthesia in Laparoscopic Cholecystectomy. Pril (Makedon Akad Nauk Umet Odd Med Nauki). 2022;43(3):101-8. doi: 10.2478/prilozi-2022-0042.\u003c/li\u003e\n\u003cli\u003eFonseca NM, Guimaraes GMN, Pontes JPJ, Azi L, de Avila Oliveira R. Safety and effectiveness of adding fentanyl or sufentanil to spinal anesthesia: systematic review and meta-analysis of randomized controlled trials. Braz J Anesthesiol. 2023;73(2):198-216. doi: 10.1016/j.bjane.2021.10.010.\u003c/li\u003e\n\u003cli\u003eUppal V, Retter S, Casey M, Sancheti S, Matheson K, McKeen DM. Efficacy of Intrathecal Fentanyl for Cesarean Delivery: A Systematic Review and Meta-analysis of Randomized Controlled Trials With Trial Sequential Analysis. Anesth Analg. 2020;130(1):111-25. doi: 10.1213/ANE.0000000000003975.\u003c/li\u003e\n\u003cli\u003eWard RJ, Kennedy WF, Bonica JJ, Martin WE, Tolas AG, Akamatsu T. Experimental evaluation of atropine and vasopressors for the treatment of hypotension of high subarachnoid anesthesia. Anesth Analg. 1966;45(5):621-9. \u003c/li\u003e\n\u003cli\u003eKang H, Lim T, Lee HJ, Kim TW, Kim W, Chang HW. Comparison of the effect of dexmedetomidine and midazolam under spinal anesthesia for cesarean delivery: a randomized controlled trial, single center study in South Korea. Anesth Pain Med (Seoul). 2023;18(2):159-68. doi: 10.17085/apm.22257.\u003c/li\u003e\n\u003cli\u003eKim H, Kim Y, Bae J, Yoo S, Lim YJ, Kim JT. Comparison of remimazolam and dexmedetomidine for intraoperative sedation in patients undergoing lower extremity surgery under spinal anesthesia: a randomized clinical trial. Reg Anesth Pain Med. 2023. doi: 10.1136/rapm-2023-104415.\u003c/li\u003e\n\u003cli\u003eFriedman JH. Greedy function approximation: A gradient boosting machine. Ann Stat. 2001;29(5):1189-232. doi: DOI 10.1214/aos/1013203451.\u003c/li\u003e\n\u003cli\u003eKe GL, Meng Q, Finley T, Wang TF, Chen W, Ma WD, Ye QW, Liu TY. LightGBM: A Highly Efficient Gradient Boosting Decision Tree. Adv Neur In. 2017;30. \u003c/li\u003e\n\u003cli\u003eShapley LS. Stochastic Games. P Natl Acad Sci USA. 1953;39(10):1095-100. doi: DOI 10.1073/pnas.39.10.1095.\u003c/li\u003e\n\u003cli\u003eLundberg SM, Erion G, Chen H, DeGrave A, Prutkin JM, Nair B, Katz R, Himmelfarb J, Bansal N, Lee SI. From Local Explanations to Global Understanding with Explainable AI for Trees. Nat Mach Intell. 2020;2(1):56-67. doi: 10.1038/s42256-019-0138-9.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"artificial intelligence, PONV, risk factors","lastPublishedDoi":"10.21203/rs.3.rs-4421679/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4421679/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003ePurpose\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn this study, we apply analysis using artificial intelligence to identify risk factors for Postoperative nausea and vomiting (PONV) during surgery under spinal anesthesia.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis retrospective study used artificial intelligence to analyze data of adult patients (aged ≥ 20 years) who underwent surgery under spinal anesthesia. To evaluate PONV, patients who experienced nausea and/or vomiting or used antiemetics within 24 hours after surgery were extracted from postoperative medical records. We create a model that predicts probability of PONV using the gradient tree boosting model. The model implementation used the LightGBM framework.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData were available for 4,574 patients. The identified risk factors were duration of surgery, female, no blood transfusion, spinal level 3–4 puncture, no concomitant epidural anesthesia, use of propofol, and dexmedetomidine, postoperative fentanyl use, cesarean section, and not using phenylephrine, atropine, or oxytocin.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe used artificial intelligence to evaluate the extent to which risk factors for PONV contribute to the development of PONV. We identifies female and cesarean section, which are known risk factors for PONV after surgery under spinal anesthesia. Our findings also suggest that fluid volume, blood transfusion, and agents that normalize hemodynamics, such as phenylephrine and atropine, are important in preventing PONV.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTrial registration number: \u003c/strong\u003eUMIN000050012\u003c/p\u003e","manuscriptTitle":"Machine Learning-based Prediction of Postoperative Nausea and Vomiting after Spinal Anesthesia: A Retrospective Observational Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-06-04 21:38:59","doi":"10.21203/rs.3.rs-4421679/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"b6668003-342a-41ff-bdb5-407d990265ea","owner":[],"postedDate":"June 4th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-08-16T18:18:00+00:00","versionOfRecord":[],"versionCreatedAt":"2024-06-04 21:38:59","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4421679","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4421679","identity":"rs-4421679","version":["v1"]},"buildId":"cBFmMYwuxLRRLfASyISRj","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}