Skip to main content
Download PDF

Development and validation of a prediction model for post-induction hypotension in elderly patients undergoing non-cardiac surgery: a prospective cohort study

BMC Anesthesiology volume 25, Article number: 217 (2025) Cite this article

Abstract

Backgrounds

Post-induction hypotension (PIH) is prevalent in elderly surgical patients and associated with adverse outcomes; however, predicting PIH remains challenging. We aimed to develop a feasible and practical PIH prediction model for elderly patients undergoing non-cardiac surgery.

Methods

In this single-center prospective cohort study, 938 elderly patients undergoing non-cardiac surgery were enrolled from December 2022 to May 2023 (n = 657 in the development cohort) and from June 2023 to August 2023 (n = 281 in the temporal validation cohort), respectively. The study outcome was the occurrence of PIH, defined as hypotension during the first 15 min after anesthesia induction or until skin incision (whichever occurred first). Predictors were determined based on LASSO and logistic regression analyses. A nomogram and a dynamic application were used for model visualization. The internal and temporal validation were performed to evaluate the discriminability, calibration and clinical utility.

Results

The median age was 71 years in both cohorts. The incidence of PIH was 51.6% and 50.5% in the development and validation cohorts, respectively. Cardiac function, baseline mean arterial pressure in the ward, etomidate use, and pre-induction mean arterial pressure were determined as predictors. The PIH prediction model was visualized as a nomogram and a dynamic application. The area under the receiver operating characteristic curve was 0.680 (95% confidence interval [CI]: 0.639 to 0.720) in internal validation and 0.697 (95% CI: 0.635 to 0.759) in temporal validation. The mean absolute errors were 0.012 and 0.029 for the internal and temporal validation calibration curves, respectively. The Brier score was 0.223. The decision curve analysis indicated that the model had a gain in predicting PIH.

Conclusion

A PIH prediction model with four predictors was developed and validated for elderly patients undergoing non-cardiac surgery. This model provides a foundation for future refinements to enhance its value of assisting clinical decision-making across diverse healthcare settings.

Trial registration

This study was registered at the Chinese Clinical Trial Registry (ChiCTR2200066201).

Peer Review reports

Introduction

Post-induction hypotension (PIH) refers to hypotension occurring within 15 min after general anesthesia induction or until surgical incision (whichever occurs first) [1]. It is reported that PIH could account for approximately half of perioperative hypotension events in patients undergoing non-cardiac surgeries [2]. Hypotension is correlated with postoperative complications, prolonged length of hospitalization, and increased mortality, especially in elderly patients [3,4,5]. In comparison with intraoperative hypotension, PIH is more foreseeable because of its preoperative existing risk factors, such as advanced age, comorbidities, autonomic nervous system function, and volume status [6,7,8].

Several new predictors and models for PIH have emerged, including inferior vena cava ultrasound for intravascular volume, dynamic arterial elastance for arterial load, and dynamic pupillometry for autonomic function [9,10,11]. However, it may be difficult to use these tools in clinical practice due to the need for specific equipment or techniques. For the existing PIH prediction models, some were developed from retrospective studies with restricted data and factors, and some were derived from a specific type of surgery [12,13,14].

Therefore, we designed this prospective cohort study to establish a practical PIH prediction model for elderly patients undergoing non-cardiac surgery. We specifically focused on the elderly patients receiving general anesthesia with intravenous induction and tracheal intubation. With the use of this model, anesthesia providers can be more alert to PIH events and take appropriate precautions in advance.

Methods

Ethics statement

This single-center prospective cohort study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (N° 2022 − 443, Chairperson Prof. Yitao Xu) on 8 November 2022 and registered at the Chinese Clinical Trial Registry (http://www.chictr.org.cn, identifier: ChiCTR2200066201) on 28 November 2022. To revise the scientific title for a better description of this study, we obtained an ethical approval update (N° 2023-012, Chairperson Prof. Yitao Xu) on 10 January 2023. The registration identifier remained unchanged. Informed consent to participate was obtained from all of the participants in the study. This study was reported following the transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD) checklist [15]. The study protocol has been published previously [16].

Study population

Between December 2022 and August 2023, we enrolled elderly patients aged ≥ 65 years with the American Society of Anesthesiologists physical status classifications I to III and scheduled for elective non-cardiac surgery (estimated duration ≥ 30 min) under general anesthesia with tracheal intubation. The exclusion criteria were as follows: (1) severe heart, brain, liver, kidney, or lung dysfunction which may lead to hemodynamic instability; (2) induction with volatiles, use of nerve blocks, spinal or epidural anesthesia; (3) tracheotomy, or pre-existing tracheal tube; or (4) refusal for participation. Additionally, patients who underwent two or more attempts of intubations were excluded from analysis.

Outcome measure

The study outcome was the occurrence of PIH, defined as hypotension during the first 15 min after anesthesia induction or until skin incision (whichever occurred first). The diagnosis of hypotension was based on a relative decrease in mean arterial pressure (MAP) ≥ 30% relative to baseline or an absolute MAP value ≤ 65 mmHg [4, 17, 18]. The baseline MAP was obtained in the ward preoperatively (using the average of multiple measurements to account for intra-individual variability). In the operating room, non-invasive cuff blood pressure was monitored on a single arm at a 1-min interval, aiming to ensure timely identification of PIH [16, 19]. During the data analysis phase, we utilized the traditional formula to calculate the MAP value at each time point: diastolic pressure + 0.33×pulse pressure [20].

Potential predictors

All candidate predictors were based on the published protocol (supplemental file 1) [16, 21]. On the day before surgery, a research assistant identified eligible patients and recorded preoperative potential predictors by reviewing electronic medical records and acquiring patients’ self-report data. Demographics, ASA classifications, comorbidities, autonomic function, cardiac function, vital signs, anxiety, and frailty status were collected. Demographic data included age, sex, and body mass index (BMI). Comorbidities included hypertension, diabetes mellitus, and age-adjusted Charlson Comorbidity index (aCCI) score. Autonomic and cardiac functions were assessed based on history, recent symptoms, and preoperative examination. Vital signs included baseline MAP and heart rate (HR) in the ward. Anxiety was assessed using the first four items of the Amsterdam Preoperative Anxiety and Information Scale (APAIS) [22]. Frailty status was evaluated using the FRAIL scale [23, 24].

On the day of surgery, the research assistant recorded the following potential predictors: duration of fasting (from 22:00 the day before surgery), preoperative volume of fluid infusion, regular MAP and HR values before induction, and anesthetics for induction (including propofol, ciprofol, etomidate, esketamine, fentanyl, and sufentanil).

Sample size Estimation

According to the existing literature [2, 25], the rate of PIH in elective non-cardiac surgery was approximately 35%. Based on the principle of at least 10 events per variable [26,27,28], 657 patients (multiply 23 by 10 and then divided by 35%) should be enrolled for the model construction and internal validation. In addition, 281 patients were continuously enrolled for temporal validation, in which the prediction model was evaluated in another time period [29]. Therefore, this study included a total of 938 patients.

Statistical analysis

Normally distributed variables were described as mean ± standard deviation (SD) and analyzed using the student t-test. Non-normally distributed variables are shown as median with interquartile range (IQR) and were analyzed using the Mann-Whitney U test. Categorical variables were presented as numbers (%) and were analyzed using the χ2 test or Fisher’s exact test. We used absolute standardized differences (ASD) to compare the baseline characteristics between the two cohorts, and an ASD ≥ 0.14 indicated imbalance (i.e., 1.96×\(\:\sqrt{(657+281)\:/\:(657\times\:281)}\)) [30]. Variables with ≤ 5% missing data were handled by multiple imputations, while variables with > 5% were excluded from the analysis.

For model development (n = 657), we applied least absolute shrinkage and selection operator (LASSO) regression with 10-fold cross-validation for variable shrinkage and selection. Based on the best lambda value, we performed multivariable logistic regression analysis to exam the potential interactions (variance inflation factor [VIF] ≥ 10 means severe interactions) and to determine the predictors with statistical significance (P < 0.05) for model construction. The final model was visualized as a nomogram, and a dynamic application was developed using the “Shiny” package. Internal validation was performed by bootstrapping method (n = 1000). Temporal validation was performed in another cohort (n = 281). Discriminative ability was assessed using the area under the receiver operating characteristic curve (AUC), and sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were also reported. The calibration was evaluated using the Hosmer-Lemeshow goodness-of-fit statistic, calibration curves and the Brier score. Calibration curves were used to measure the consistency between the predicted probability of the model and the actual observed results. In the calibration curves, the apparent curve represents the relationship between the predicted probability of the model and the actual observed probability, while the bias-corrected curve is the result of correcting the apparent curve [31, 32]. The Brier score reflects the mean squared difference between the observations and predictions, with lower values indicating higher accuracy (ranging from 0 to 100) [33]. Moreover, the clinical utility of the model was assessed using the decision curve analysis (DCA). DCA quantifies the net benefit of using the model and determines whether treating all or no patients is superior to the model, across the reasonable threshold probabilities [34, 35]. Statistical analyses were conducted using SPSS software (v 26.0, IBM SPSS, Chicago, Illinois, USA) and R software (v 4.2.2, www.R-project.org/).

Results

Patient characteristics

Among 1225 elderly patients screened, we enrolled a consecutive sample of 938 eligible patients undergoing noncardiac surgery, with 657 (70%) in the development cohort from December 2022 to May 2023 and 281 (30%) in the validation cohort from June 2023 to August 2023 (Fig. 1). All data were prospectively collected, and there were no missing data. Baseline data and candidate predictors are listed in Table 1.

Fig. 1
figure 1

Study flow chart. Abbreviation: LASSO, least absolute shrinkage and selection operator

Full size image
Table 1 Comparison of characteristics between two cohorts
Full size table

The incidence of PIH

A total of 481 patients (51.3%) developed the PIH events, including 339 (51.6%) in the development cohort and 142 (50.5%) in the validation cohort (Table 2). In the post-hoc analysis, 42.2% of patients had MAP ≤ 65 mmHg, 44.8% of patients showed a relative decrease in MAP > 30% from baseline, and 44.8% of patients experienced PIH that lasted for 1 min. No significant differences were observed between the two cohorts in terms of the hypotension rates.

Table 2 Incidence of PIH
Full size table

Selection of predictors

Based on the LASSO regression (Fig. S1) and multivariate logistic regression analysis, we identified four variables (cardiac function, baseline MAP in the ward, etomidate, and pre-induction MAP) to establish the PIH prediction model (Table 3). All VIF values were approximately 1, representing low interactions.

Table 3 Final predictors for the model
Full size table

Development and validation of the model

Based on these four predictors, a nomogram was developed (Fig. 2A). In addition, we developed an app to facilitate the use of this PIH prediction model in clinical scenarios (Fig. 2B). For a male elderly patient who had cardiac dysfunction, MAP of 90 mmHg in the ward and 85 mmHg before induction without the use of etomidate, his risk of PIH would be 77.2% (95% confidence interval [CI]: 69.5%–83.4%).

Fig. 2
figure 2

Visualization for the prediction of post-induction hypotension. (A) Nomogram. Each predictor has a vertical line (“axis”) labeled with its values (e.g., for cardiac dysfunction, the axis ranges from 0 to 1). First, for a specific value (e.g., cardiac dysfunction = 1), draw a vertical line upward to the Points axis to find the corresponding score. Repeat this for all predictors. Next, add all the points from the individual predictors to get the total points. Finally, draw a vertical line from the Total Points axis to the Risk axis to read the predicted probability; (B) The interface of dynamic app with an example patient. Abbreviation: MAP, mean arterial pressure

Full size image

For internal validation, the AUC was 0.680 (95% CI: 0.639 to 0.720) (Fig. 3A), with sensitivity of 0.720 and specificity of 0.580 (PPV = 0.628 and NPV = 0.687). The value of the mean absolute error was 0.012 in the calibration curve (Fig. 3B), with a Hosmer-Lemeshaw goodness-of-fit statistic result of 0.652, suggesting a good agreement between the predicted and observed cases. The DCA curve showed that the prediction model had a gain in predicting PIH in most of the elderly patients (Fig. 3C).

Fig. 3
figure 3

Validation of the prediction model. (A) Receiver operating characteristic curve in the internal validation; (B) Calibration curve in the internal validation; (C) Decision curve analysis in the internal validation; (D) Receiver operating characteristic curve in the temporal validation; (E) Calibration curve in the temporal validation; (F) Decision curve analysis in the temporal validation

Full size image

For temporal validation, the AUC was 0.697 (95% CI: 0.635 to 0.759) (Fig. 3D), with sensitivity of 0.634 and specificity of 0.698 (PPV = 0.703 and NPV = 0.627). The value of the mean absolute error was 0.029 (Fig. 3E), with a Hosmer-Lemeshaw test result of 0.274 and a Brier score of 0.223. The DCA curve was similar to that of the internal validation (Fig. 3F).

Discussion

In this study, a preliminary prediction model for PIH was developed based on prospectively collected data, and this model was validated temporally. Cardiac dysfunction, high baseline MAP in the ward, etomidate not used for induction, and a relative low pre-induction MAP were risk factors for PIH. In both the internal and temporal validation cohorts, the model performance was consistent, suggesting there is no overfitting in the model training process. Both of the AUC values were approximately 0.7 with low mean absolute errors. A Brier score of 0.223 showed the model was well calibrated in that the predicted probabilities were close to the actual event rates. Furthermore, The DCA curves suggest that the application of this model can help in clinical decision-making.

There were three hemodynamic factors for PIH prediction in this model. Among them, cardiac dysfunction and a high MAP in the ward were the risk factors, whereas a high pre-induction MAP was a protecting factor. Recent studies found hypertension history and abnormal preoperative echocardiographic parameters contributed to a higher PIH rate during noncardiac surgery [36, 37]. In contrast, perioperative fluid therapy helped to prevent PIH [38, 39]. In light of the pathophysiological mechanisms, cardiovascular abnormalities are predominant, including suppression of myocardial contractility, preload reduction caused by venous dilatation, reduced vascular resistance after arterial dilation, and baroreceptor inhibition [40,41,42]. A high MAP in the ward often indicated poor blood pressure control and abnormal cardiac function, while a relatively higher pre-induction MAP may reflect an adequate volume status and an overall good condition of sympathetic activity. For our patients, the baseline MAP was determined in the ward, reflecting a normal condition and avoiding possible impact by the operating room environment and preoperative fasting [18]. The interval of blood pressure measurement was set at every 1 min to increase the ability to detect PIH. The observed incidence of PIH in our study was 51.3%, which is in line with the results of recent cohort studies and randomized controlled trials [18, 19, 39, 43].

It is noteworthy that severe cardiac dysfunction with existing hemodynamic instability was an exclusion criterion in this study because such patients usually had very low left atrial ejection fraction (< 30%) and were already at high risk of PIH. However, the ‘cardiac dysfunction’ factor of the model was evaluated through relative anamneses and preoperative examinations, especially for the asymptomatic patients. The use of anesthetics is a strong factor that influences the occurrence of PIH. For patients included in this study, propofol, ciprofol, etomidate, and esketamine were used for anesthesia induction. We found that the use of etomidate was associated with a decreased occurrence of PIH. It has been confirmed that etomidate has a superior hemodynamic profile in comparison with propofol for patients undergoing either noncardiac or cardiac surgery [44]. A recent trial showed that a target-controlled infusion of etomidate was noninferior to propofol in terms of overall major in-hospital morbidity [45]. Two studies based on the National Emergency Airway Registry found that etomidate caused less postintubation hypotension compared with ketamine [46, 47]. Nevertheless, the comparison between ciprofol and etomidate is yet to be studied. In our clinical practice, most patients (~ 60%) received etomidate during induction. Regarding the analgesics, studies have shown that sufentanil is not associated with an increased risk of PIH [21, 42].

The predictors identified in our study (e.g., baseline blood pressure, cardiac function) are indeed recognized by clinicians as risk factors for PIH. However, the novelty of this work lies in synthesizing these factors into a validated, user-friendly tool tailored for elderly non-cardiac surgical patients. Compared with previous studies [12,13,14], our predictive model was based on prospectively collected data and a published protocol [16]. To our knowledge, this is the first prospective model to explicitly focus on elderly patients in non-cardiac settings, a population with unique physiological vulnerabilities.

However, this study has some limitations. First, 23 candidate predictors were selected based on the literature, and only four of them were finally selected for model development. On one hand, the potential interactions between predictors should be taken into consideration; on the other hand, it is also important to increase data diversity and precision (e.g. by including more laboratory examination results). Second, the endpoint event was only PIH. For elderly patients, intraoperative hypotension and related outcomes (such as acute kidney injury and myocardial injury) require further research. Third, the effects of anesthetics on blood pressure are dose-dependent. However, considering the various medication strategies, we only include the use of specific anesthetics as predictors. Fourth, the target population was elderly patients scheduled for general anesthesia with tracheal intubation, excluding the procedures with prior nerve blocks and the use of laryngeal mask. Finally, while our model provides a clinically feasible tool for PIH prediction, an AUC of ~ 0.7 underscores the need for further optimization. Future studies could integrate advanced machine learning algorithms (e.g., ensemble methods) and more preoperative data to improve performance. Multicenter collaborations may also help validate and refine the model across diverse populations, ensuring broader applicability.

Conclusion

PIH is prevalent among elderly patients undergoing noncardiac surgery. Our study identified several key risk factors for PIH, including cardiac dysfunction, elevated preoperative MAP in the ward, no use of etomidate, and low pre-induction MAP. By utilizing prospectively collected clinical data, we have developed and validated a preliminary model for the early prediction of PIH. This model provides a foundation for future refinements and improvements, with the ultimate goal of enhancing its value of assisting clinical judgment across diverse healthcare settings.

Data availability

The datasets and the preliminary dynamic app during the study are available from the corresponding authors on reasonable request.

Abbreviations

aCCI:

Age-adjusted Charlson Comorbidity index

APAIS:

Amsterdam Preoperative Anxiety and Information Scale

ASA-PS:

American Society of Anesthesiologists Physical Status Classification

ASD:

Absolute standardized differences

AUC:

Area under the receiver operating characteristic curve

BMI:

Body mass index

CI:

Confidence interval

DCA:

Decision curve analysis

LASSO:

Least absolute shrinkage and selection operator

MAP:

Mean arterial pressure

OR:

Odds ratio

PIH:

Post-induction hypotension

VIF:

Variance inflation factor

References

  1. Fechner J, El-Boghdadly K, Spahn DR, Motsch J, Struys M, Duranteau O, Ganter MT, Richter T, Hollmann MW, Rossaint R, et al. Anaesthetic efficacy and postinduction hypotension with remimazolam compared with Propofol: a multicentre randomised controlled trial. Anaesthesia. 2024;79(4):410–22.

    Article  CAS  PubMed  Google Scholar 

  2. Maheshwari K, Turan A, Mao G, Yang D, Niazi AK, Agarwal D, Sessler DI, Kurz A. The association of hypotension during non-cardiac surgery, before and after skin incision, with postoperative acute kidney injury: a retrospective cohort analysis. Anaesthesia. 2018;73(10):1223–8.

    Article  CAS  PubMed  Google Scholar 

  3. Green RS, Butler MB. Postintubation hypotension in general anesthesia: A retrospective analysis. J Intensive Care Med. 2016;31(10):667–75.

    Article  PubMed  Google Scholar 

  4. Sessler DI, Bloomstone JA, Aronson S, Berry C, Gan TJ, Kellum JA, Plumb J, Mythen MG, Grocott MPW, Edwards MR, et al. Perioperative quality initiative consensus statement on intraoperative blood pressure, risk and outcomes for elective surgery. Br J Anaesth. 2019;122(5):563–74.

    Article  PubMed  Google Scholar 

  5. Czajka S, Putowski Z, Krzych LJ. Post-induction hypotension and intraoperative hypotension as potential separate risk factors for the adverse outcome: a cohort study. J Anesth. 2023;37(3):442–50.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Frandsen MN, Mehlsen J, Foss NB, Kehlet H. Pre-operative autonomic nervous system function - a missing link for post-induction hypotension? Anaesthesia. 2022;77(2):139–42.

    Article  CAS  PubMed  Google Scholar 

  7. Chen B, Pang QY, An R, Liu HL. A systematic review of risk factors for postinduction hypotension in surgical patients undergoing general anesthesia. Eur Rev Med Pharmacol Sci. 2021;25(22):7044–50.

    CAS  PubMed  Google Scholar 

  8. Lee HJ, Kim YJ, Woo JH, Oh HW. Preoperative frailty is an independent risk factor for postinduction hypotension in older patients undergoing noncardiac surgery: A retrospective cohort study. J Gerontol Biol Sci Med Sci 2024, 79(1).

  9. Chowdhury SR, Datta PK, Maitra S, Rawat D, Baidya DK, Roy A, Nath S. The use of preoperative inferior Vena Cava ultrasound to predict anaesthesia-induced hypotension: a systematic review. Anaesthesiol Intensive Ther. 2023;55(1):18–31.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Shao L, Zhou Y, Yue Z, Gu Z, Zhang J, Hui K, Xiong J, Xu M, Duan M. Pupil maximum constriction velocity predicts post-induction hypotension in patients with lower ASA status: a prospective observational study. BMC Anesthesiol. 2022;22(1):274.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Oh EJ, Min JJ, Kwon E, Choi EA, Lee JH. Evaluation of pre-induction dynamic arterial elastance as an adjustable predictor of post-induction hypotension: A prospective observational study. J Clin Anesth. 2023;87:111092.

    Article  PubMed  Google Scholar 

  12. Noto K, Uchida S, Kinoshita H, Takekawa D, Kushikata T, Hirota K. Predictive model for post-induction hypotension in patients undergoing transcatheter aortic valve implantation: a retrospective observational study. JA Clin Rep. 2024;10(1):33.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Kang AR, Lee J, Jung W, Lee M, Park SY, Woo J, Kim SH. Development of a prediction model for hypotension after induction of anesthesia using machine learning. PLoS ONE. 2020;15(4):e0231172.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Kendale S, Kulkarni P, Rosenberg AD, Wang J. Supervised Machine-learning predictive analytics for prediction of postinduction hypotension. Anesthesiology. 2018;129(4):675–88.

    Article  PubMed  Google Scholar 

  15. Collins GS, Reitsma JB, Altman DG, Moons KGM. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD). Ann Intern Med. 2015;162(10):735–6.

    Article  PubMed  Google Scholar 

  16. Hu JH, Xu N, Bian Z, Shi HJ, Ji FH, Peng K. Protocol for development and validation of a prediction model for post-induction hypotension in elderly patients undergoing non-cardiac surgery: a prospective cohort study. BMJ Open. 2023;13(9):e074181.

    Article  PubMed  PubMed Central  Google Scholar 

  17. Jor O, Maca J, Koutna J, Gemrotova M, Vymazal T, Litschmannova M, Sevcik P, Reimer P, Mikulova V, Trlicova M, et al. Hypotension after induction of general anesthesia: occurrence, risk factors, and therapy. A prospective multicentre observational study. J Anesth. 2018;32(5):673–80.

    Article  PubMed  Google Scholar 

  18. Khan AI, Fischer M, Pedoto AC, Seier K, Tan KS, Dalbagni G, Donat SM, Arslan-Carlon V. The impact of fluid optimisation before induction of anaesthesia on hypotension after induction. Anaesthesia. 2020;75(5):634–41.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Takaki R, Yokose M, Mihara T, Saigusa Y, Tanaka H, Yamamoto N, Masui K, Goto T. Hypotension after general anaesthesia induction using remimazolam or Propofol in geriatric patients undergoing Sevoflurane anaesthesia with remifentanil: a single-centre, double-blind, randomised controlled trial. Br J Anaesth. 2024;133(1):24–32.

    Article  CAS  PubMed  Google Scholar 

  20. Papaioannou TG, Protogerou AD, Vrachatis D, Konstantonis G, Aissopou E, Argyris A, Nasothimiou E, Gialafos EJ, Karamanou M, Tousoulis D, et al. Mean arterial pressure values calculated using seven different methods and their associations with target organ deterioration in a single-center study of 1878 individuals. Hypertens Res. 2016;39(9):640–7.

    Article  PubMed  Google Scholar 

  21. Reich DL, Hossain S, Krol M, Baez B, Patel P, Bernstein A, Bodian CA. Predictors of hypotension after induction of general anesthesia. Anesth Analg. 2005;101(3):622–8.

    Article  PubMed  Google Scholar 

  22. Moerman N, van Dam FS, Muller MJ, Oosting H. The Amsterdam preoperative anxiety and information scale (APAIS). Anesth Analg. 1996;82(3):445–51.

    CAS  PubMed  Google Scholar 

  23. van Abellan G, Rolland Y, Bergman H, Morley JE, Kritchevsky SB, Vellas B. The I.A.N.A task force on frailty assessment of older people in clinical practice. J Nutr Health Aging. 2008;12(1):29–37.

    Article  Google Scholar 

  24. Kojima G. Quick and simple FRAIL scale predicts incident activities of daily living (ADL) and instrumental ADL (IADL) disabilities: A systematic review and Meta-analysis. J Am Med Dir Assoc. 2018;19(12):1063–8.

    Article  PubMed  Google Scholar 

  25. Schonberger RB, Dai F, Michel G, Vaughn MT, Burg MM, Mathis M, Kheterpal S, Akhtar S, Shah N, Bardia A. Association of Propofol induction dose and severe pre-incision hypotension among surgical patients over age 65. J Clin Anesth. 2022;80:110846.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Peduzzi P, Concato J, Feinstein AR, Holford TR. Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates. J Clin Epidemiol. 1995;48(12):1503–10.

    Article  CAS  PubMed  Google Scholar 

  27. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49(12):1373–9.

    Article  CAS  PubMed  Google Scholar 

  28. Li J, Tian Y, Wang M, Fang J, Zhang H, Yue F, Xu M, Wang J, Li M, Guo X, et al. Radiological indicators and a novel combined predictive model for anticipating difficult laryngoscopy in cervical spondylosis patients: a prospective cohort study. BMC Anesthesiol. 2024;24(1):446.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Efthimiou O, Seo M, Chalkou K, Debray T, Egger M, Salanti G. Developing clinical prediction models: a step-by-step guide. BMJ. 2024;386:e078276.

    Article  PubMed  PubMed Central  Google Scholar 

  30. Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28(25):3083–107.

    Article  PubMed  PubMed Central  Google Scholar 

  31. Van Calster B, Nieboer D, Vergouwe Y, De Cock B, Pencina MJ, Steyerberg EW. A calibration hierarchy for risk models was defined: from utopia to empirical data. J Clin Epidemiol. 2016;74:167–76.

    Article  PubMed  Google Scholar 

  32. Van Calster B, McLernon DJ, van Smeden M, Wynants L, Steyerberg EW. Topic group ‘evaluating diagnostic t, prediction models’ of the Si: calibration: the Achilles heel of predictive analytics. BMC Med. 2019;17(1):230.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Angraal S, Mortazavi BJ, Gupta A, Khera R, Ahmad T, Desai NR, Jacoby DL, Masoudi FA, Spertus JA, Krumholz HM. Machine learning prediction of mortality and hospitalization in heart failure with preserved ejection fraction. JACC Heart Fail. 2020;8(1):12–21.

    Article  PubMed  Google Scholar 

  34. Vickers AJ, Van Calster B, Steyerberg EW. Net benefit approaches to the evaluation of prediction models, molecular markers, and diagnostic tests. BMJ. 2016;352:i6.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Van Calster B, Wynants L, Verbeek JFM, Verbakel JY, Christodoulou E, Vickers AJ, Roobol MJ, Steyerberg EW. Reporting and interpreting decision curve analysis: A guide for investigators. Eur Urol. 2018;74(6):796–804.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Zhang H, Gao H, Xiang Y, Li J. Maximum inferior Vena Cava diameter predicts post-induction hypotension in hypertensive patients undergoing non-cardiac surgery under general anesthesia: A prospective cohort study. Front Cardiovasc Med. 2022;9:958259.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Tarao K, Daimon M, Son K, Nakanishi K, Nakao T, Suwazono Y, Isono S. Risk factors including preoperative echocardiographic parameters for post-induction hypotension in general anesthesia. J Cardiol. 2021;78(3):230–6.

    Article  PubMed  Google Scholar 

  38. Zhang Y, Wang X, Sang X, Zhou Z, Dai G, Zhang X. Effect of fluid therapy in early morning on the incidence of Post-Induction hypotension during Non-Cardiac surgery after noon: A Single-Center retrospective study. Drug Des Devel Ther. 2024;18:1339–47.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Huang S, Liao Z, Chen A, Wang J, Xu X, Zhang L. Effect of carotid corrected flow time combined with perioperative fluid therapy on preventing hypotension after general anesthesia induction in elderly patients: a prospective cohort study. Int J Surg. 2024;110(2):799–809.

    PubMed  Google Scholar 

  40. Dorantes Mendez G, Aletti F, Toschi N, Canichella A, Dauri M, Coniglione F, Guerrisi M, Signorini MG, Cerutti S, Ferrario M. Baroreflex sensitivity variations in response to Propofol anesthesia: comparison between normotensive and hypertensive patients. J Clin Monit Comput. 2013;27(4):417–26.

    Article  PubMed  Google Scholar 

  41. Wong GTC, Irwin MG. Post-induction hypotension: a fluid relationship? Anaesthesia 2021, 76(1):15–8.

  42. Saugel B, Bebert EJ, Briesenick L, Hoppe P, Greiwe G, Yang D, Ma C, Mascha EJ, Sessler DI, Rogge DE. Mechanisms contributing to hypotension after anesthetic induction with sufentanil, Propofol, and Rocuronium: a prospective observational study. J Clin Monit Comput. 2022;36(2):341–7.

    Article  PubMed  Google Scholar 

  43. Park JY, Yu J, Kim CS, Baek JW, Jo Y, Kim YK. Effect of pneumatic leg compression on post-induction hypotension in elderly patients undergoing robot-assisted laparoscopic prostatectomy: a double-blind randomised controlled trial. Anaesthesia. 2023;78(6):730–8.

    Article  CAS  PubMed  Google Scholar 

  44. Hannam JA, Mitchell SJ, Cumin D, Frampton C, Merry AF, Moore MR, Kruger CJ. Haemodynamic profiles of etomidate vs Propofol for induction of anaesthesia: a randomised controlled trial in patients undergoing cardiac surgery. Br J Anaesth. 2019;122(2):198–205.

    Article  CAS  PubMed  Google Scholar 

  45. Lu Z, Zheng H, Chen Z, Xu S, Chen S, Mi W, Wang T, Chai X, Guo Q, Zhou H, et al. Effect of etomidate vs Propofol for total intravenous anesthesia on major postoperative complications in older patients: A randomized clinical trial. JAMA Surg. 2022;157(10):888–95.

    Article  PubMed  PubMed Central  Google Scholar 

  46. April MD, Arana A, Schauer SG, Davis WT, Oliver JJ, Fantegrossi A, Summers SM, Maddry JK, Walls RM, Brown CA. Ketamine versus etomidate and Peri-intubation hypotension: A National emergency airway registry study. Acad Emerg Med. 2020;27(11):1106–15. 3rd et al.

    Article  PubMed  Google Scholar 

  47. Driver BE, Trent SA, Prekker ME, Reardon RF, Brown CA. 3rd: sedative dose for rapid sequence intubation and postintubation hypotension: is there an association?? Ann Emerg Med. 2023;82(4):417–24.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

We thank all participants, the nurses at the Department of Anesthesiology from the First Affiliated Hospital of Soochow University, and the research assistants (Ning Xu, Weifan Wu, and Ruyi Li) for their support in this study.

Funding

This study was funded by the College Students’ Extracurricular Academic Research Project of Soochow University (KY2023101A to JHH), Pre-research Fund of the Second Affiliated Hospital of Soochow University (SDFEYGR2258 to YYY), National Natural Science Foundation of China (82471290 to KP), Suzhou Medical Health Science and Technology Innovation Project (SKY2022136 to KP), and Suzhou Basic Research Pilot Project (SSD2024082 to KP). The funders were not involved in the design or execution of the study, or in the analysis and interpretation of the data.

Author information

Author notes

  1. Zhen Bian, Wei Dou and Yaoyu Ying contributed equally to this work.

Authors and Affiliations

  1. Department of Anesthesiology, First Affiliated Hospital of Soochow University, Suzhou, Jiangsu, 215006, China

    Zhen Bian, Wei Dou, Haijing Shi, Fuhai Ji, Jinghui Hu & Ke Peng

  2. Institute of Anesthesiology, Soochow University, Suzhou, Jiangsu, 215006, China

    Wei Dou, Haijing Shi, Fuhai Ji, Jinghui Hu & Ke Peng

  3. Department of Medical Affairs, The Second Affiliated Hospital of Soochow University, Suzhou, 215006, China

    Yaoyu Ying

Authors
  1. Zhen Bian
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Wei Dou
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Yaoyu Ying
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Haijing Shi
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Fuhai Ji
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Jinghui Hu
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Ke Peng
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Z. Bian, J. Hu and K. Peng: Study design; Z.Bian, W. Dou and H. Shi: Execution, acquisition of data; W. Dou and Y. Ying: Analysis and interpretation of data; Y. Ying: Preparing figures and tables; Z. Bian and J. Hu: Drafting the manuscript; F. Ji and K. Peng: Revising the manuscript and giving final approval of the version of the manuscript. All authors reviewed the manuscript.

Corresponding authors

Correspondence to Jinghui Hu or Ke Peng.

Ethics declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki and clinical study regulations from the Ethics Committee of the First Affiliated Hospital of Soochow University. This ethical approval of this study was obtained from the Ethics Committee of the First Affiliated Hospital of Soochow University (N° 2022 − 443, Chairperson Prof. Yitao Xu) on 8 November 2022. Then, this study was registered at the Chinese Clinical Trial Registry (http://www.chictr.org.cn, identifier: ChiCTR2200066201) on 28 November 2022. To revise the scientific title for a better description of this study, an ethical update was approved (N° 2023-012, Chairperson Prof. Yitao Xu) on 10 January 2023, but the registry identifier remained unchanged. Informed consent to participate was obtained from all of the participants in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

12871_2025_3090_MOESM3_ESM.tif

Supplementary Material 3: Fig. S1 LASSO selection path diagram. (A) LASSO selection path plot: the left vertical dashed line indicates the minimum error (lambda.min), and the right dashed line indicates one standard error away from the minimum error (lambda.lse). (B) LASSO path plot: the curve of regression coefficients versus Log(lambda) as the coefficient scores are gradually decreasing. Abbreviation: LASSO, least absolute shrinkage and selection operator.

Supplementary Material 4

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Bian, Z., Dou, W., Ying, Y. et al. Development and validation of a prediction model for post-induction hypotension in elderly patients undergoing non-cardiac surgery: a prospective cohort study. BMC Anesthesiol 25, 217 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12871-025-03090-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12871-025-03090-2

Keywords

BMC Anesthesiology

ISSN: 1471-2253

Contact us