Is ‘anxiety sensitivity’ predictive of postoperative nausea and vomiting?

A prospective observational study

Abstract

BACKGROUND 

Postoperative nausea and vomiting (PONV) is an extremely distressing side effect for patients. Despite PONV prophylaxis guided by well established scoring systems, the incidence of PONV is still high.

OBJECTIVE 

The aim of the current study was to investigate the predictive value of anxiety sensitivity as an additional independent risk factor for PONV in patients with an increased risk of PONV.

DESIGN 

A noninterventional, observational study.

SETTING 

A tertiary care university hospital.

PATIENTS 

Patients with an increased risk of PONV (i.e. female, nonsmoking) undergoing elective surgery (general, gynaecological, urological, musculoskeletal or neurosurgical) under general anaesthesia.

MAIN OUTCOME MEASURES 

The number of patients with anxiety sensitivity assessed pre-operatively with the Anxiety Sensitivity Index-3 questionnaire, the number of patients experiencing PONV, predictive value of anxiety sensitivity compared with other established risk factors for PONV.

RESULTS 

Some 41.5% of the patients experienced PONV within the first 24 h after surgery. In these patients increased anxiety sensitivity (Anxiety Sensitivity Index-3 score higher than seven points) was associated with a five-fold increase in the odds ratio (OR) for PONV. From the regression model, the risk of PONV was increased by lack of PONV prophylaxis (OR, 3.68), the postoperative administration of opioids (OR, 3.60) and patient age (OR, 1.03), but laparoscopic surgery did not increase the risk.

CONCLUSION 

In addition to the well established risk factors, anxiety sensitivity can help to predict the risk of PONV. It seems justifiable to add psychological factors such as anxiety sensitivity to PONV risk-scores. PONV prophylaxis should be considered when anxiety sensitivity is high.

TRIAL REGISTRATION 

ClinicalTrials.gov identifier: NCT01875120.

Introduction

Postoperative nausea and vomiting (PONV) is extremely unpleasant for patients after general anaesthesia and deserves the same attention as the treatment of postoperative pain. The burden of this distressing symptom impairs the quality of recovery.¹ The incidence of PONV is influenced by pre-existing risk factors as well as prophylactic antiemetic treatment before or during surgery and the anaesthetic technique.^2–4

The cause of PONV is multifactorial and includes factors related to the patient, the anaesthetic, and the surgery. However, the understanding of risk factors is still incomplete.² There are a number of independent risk factors for PONV such as female sex, history of PONV or motion sickness, nonsmoking status and age. Anaesthesia-related predictors, such as the use of volatile anaesthetic agents or the postoperative use of opioids, are also established risk factors, whereas conflicting data exist for additional factors such as pre-operative fasting, the stage of the menstrual cycle, the type of surgery and psychological factors.^5,6

Several PONV scoring systems have been developed for adults and children based on these risk factors.^3,4,7 The frequently used Apfel-Score³ includes four risk factors (female sex, nonsmoking status, motion sickness or history of PONV and the expected postoperative administration of opioids). If 0, 1, 2, 3 or 4 of these risk factors are present, the probability for PONV is about 10, 20, 40, 60 or 80%,^7–11 respectively. However, although the Apfel score and other scoring systems are well established and often used to define standardised procedures for a risk-based antiemetic strategy for patients with an increased risk for PONV, about 30% of patients will still suffer from PONV. Therefore, PONV remains the most frequent side effect after general anaesthesia.¹² As PONV is associated with distress and a poorer quality of recovery after surgery it seems justified to characterise additional influencing factors to further reduce the number of patients suffering from PONV.¹³

The influence of psychological factors such as anxiety as a cause of PONV has not been explored sufficiently so far. Studies investigating anxiety as a potential risk factor have shown inconsistent results.^14,15 Anxiety sensitivity is a psychological construct which describes the tendency of a person to fear the sensations caused by anxiety (e.g. those sensations following activation of the sympathetic nervous system: tachycardia, palpitations, dyspnoea, sweating or trembling) because of a belief that such sensations are harmful. Anxiety sensitivity is believed to intensify states of fear and anxiety. Higher levels of anxiety sensitivity are positively correlated with awareness of bodily symptoms (e.g. palpitations¹⁶) increased subjective complaints,¹⁷ and are associated with general distress.¹⁸ Thus, we hypothesised that anxiety sensitivity would increases the probability of PONV in predisposed patients, and may be an additional predictor for PONV. The aim of the current study was to explore the predictive value of anxiety sensitivity as an independent risk factor for PONV in a sample of patients with an increased risk of PONV – female nonsmokers.³

Methods

Ethics

The study was approved by the Regional Ethics Committee of Rhineland-Palatinate, State Board of Physicians, Deutschhausplatz 3, 55116 Mainz, Germany, protocol number 837.327.12 (8428), Prof Dr Krämer, the date of approval was 2 August 2012. The study was registered in clinicalTrials.gov (Identifier: NCT01875120).

Inpatients, aged between 18 and 75 years, undergoing elective interventions in general surgery, gynaecology, urology, musculoskeletal surgery or neurosurgery (except intracranial interventions), were recruited at the Department of Anaesthesiology of the University Medical Center of the Johannes Gutenberg University, Mainz, Germany. Nonsmoking patients of the female sex (i.e. at least two risk factors for PONV according to the Apfel-Score³) were enrolled in this study if they were to receive general anaesthesia which included inhalational agents (sevoflurane or desflurane). Written informed consent was obtained from all patients. Exclusion criteria were emergency surgery, regional anaesthetic techniques, planned total intravenous anaesthesia (TIVA) or participation in another clinical study.

From 192 eligible patients, complete data sets from 135 patients were analysed. Data were collected between October 2012 and July 2013. The CONSORT diagram of recruitment is shown in Fig. 1.

Fig. 1: Consort flow diagram. Gen, general surgery; Gyn, gynaecological surgery; Musc, musculoskeletal surgery; Neuro, neurosurgery; TIVA, total intravenous anaesthesia; Uro, urological surgery.

Pre-operatively, all patients completed the Anxiety Sensitivity Index-3 questionnaire (ASI-3, supplemental data file, https://links.lww.com/EJA/A192).^19,20 The ASI-3 is an internationally used 18-item self-reporting questionnaire, assessing the patient's concerns about the harmfulness of bodily perceptions associated with anxiety arousal. The ASI-3 contains three subscales with six items each. First, the Somatic Concerns subscale, which assesses the fear of somatic consequences from anxiety symptoms (sample item: ‘When I feel pain in my chest, I worry that I am going to have a heart attack’). Second, the Cognitive Concerns subscale COG, which assesses catastrophic cognitive consequences of anxiety symptoms (sample item: ‘When I cannot keep my mind on a task, I worry that I might be going crazy’). Third, the Social Concerns subscale, which assesses fear of social consequences of observable anxiety-related reactions, for example, blushing and embarrassment (sample item: ‘It is important for me not to appear nervous’). All items have to be answered on a five-point Likert-scale, indicating a patient's agreement, ranging from ‘very little’ (0 points) to ‘very much’ (four points). The ASI-3 total score ranges from 0 to 72 points, with higher scores representing a higher manifestation of anxiety sensitivity. Excellent reliability and validity of the ASI-3 has been shown in various studies.^21–23

According to the Apfel³ score, we assessed the risk of four factors for PONV: female sex, nonsmoking status, expected postoperative administration of opioids, and a history of PONV or motion sickness. These, were assessed pre-operatively by questionnaire. Data regarding anaesthesia, actual intra-operatively and postoperatively administered opioids, surgical technique and duration of surgery were collected from electronic and article-based medical records.

PONV prophylaxis, consisting of dexamethasone (5 to 10 mg intravenously before general anaesthesia) and ondansetrone (4 mg intravenously at the end of surgery) was administered according to clinical guidelines. The incidence of PONV within the first 24 h postsurgery was evaluated. Nausea was assessed with a numerical rating scale ranging from 0 to 10, with 0 representing no nausea, and 10 representing maximum nausea.^24,25 Any value greater than 0 was defined as PONV. Retching and/or vomiting were also rated as PONV.

Descriptive statistics are expressed as number (relative frequencies, %) or mean (SD). Distribution of continuous variables was checked using the Kolmogorov–Smirnov test of normality and Spearman's rank correlation (r) was used to examine the association between variables. Differences between groups were examined using Student's t test, Mann–Whitney U test or Chi-square test, according to the type of variables and their distribution. The joint influence of different variables was assessed in logistic regression models, therefore, the ASI-3 total score was quartile cut. Variables showing significant differences between patients with or without PONV were entered into the regression model. Thus, ASI-3 total score, age, PONV prophylaxis, laparoscopic technique and postoperative opioids constituted the set of independent variables.

All analyses were performed using IBM SPSS Statistics for Windows, Version 22 (Released 2013. IBM SPSS Statistics for Windows; IBM Corp., Armonk, New York, USA). A P value less than 0.05 was considered statistically significant.

Results

Patients

Of 192 patients who fulfilled the inclusion criteria, 182 (95%) agreed to participate. Reasons for participation refusal were lack of interest or lack of time, thus a screening bias is not obvious. Complete data sets for 135 patients were entered into the analysis. The majority of surgical interventions were in the three specialities of general surgery (n = 77), gynaecological (n = 39) and urological surgery (n = 19) (Fig. 1).

Demographic and clinical data for the whole sample and for the two groups (PONV and No PONV) are shown in Table 1.

Table 1: Demographic and clinical data for the patient groups with or without postoperative nausea and vomiting

Anxiety sensitivity

ASI-3 mean scores show levels similar to samples of normal healthy individuals. In the present sample the mean ASI-3 total score was 16.1 (12.5), corresponding to a moderate degree of anxiety sensitivity compared with other studies.^21,22,26 Comparing patients with PONV and those without PONV, higher ASI-3 scores were found in the PONV group, with statistically significant differences in the ASI-3 total score and the cognitive concerns (COG) subscale (Table 1).

Postoperative nausea and vomiting

In the whole sample 56 (41.5%) patients were found to have PONV within the first 24 h after surgery. Overall, 33 (24.4%) received PONV prophylaxis pre-operatively of whom eight experienced PONV symptoms.

Demographic and clinical parameters were compared between the two groups (with or without PONV). Patients with PONV were younger, were less likely to have received PONV prophylaxis, had more often received opioids for postoperative analgesia, and underwent laparoscopic surgery more frequently. No differences could be found with regard to a history of PONV or motion sickness and duration of anaesthesia (Table 1). In the group of patients with PONV significantly more patients reached higher ASI-3 scores (i.e. above the first quartile) than in the group of patients without PONV (87.5 vs. 65.8%, Chi² = 8.173; P = 0.004).

Prediction of postoperative nausea and vomiting

First we estimated the association between each candidate predictor and the outcome PONV or No PONV. Potential predictors were all variables showing a significant difference (P < 0.05) between the groups and these were entered into a logistic regression model (Table 1). Thus, patient age, PONV prophylaxis, laparoscopic surgery, postoperative administration of opioids and the sum score from the ASI-3 questionnaire were included. The results of the regression analysis are shown in Table 2 and reveal an odds ratio (OR) for PONV of 4.95 in patients with ASI-3 scores of eight or more, compared with patients with ASI-3 scores of seven or less. The ORs were also statistically significant for patient's age, lack of PONV prophylaxis and postoperative administration of opioids. Laparoscopic surgery was not statistically significant in the model (Table 2).

Table 2: Odds ratio for postoperative nausea and vomiting in the patient cohort

Discussion

In the current study, we tested the predictive value of anxiety sensitivity as an independent risk factor for PONV in a cohort of patients at increased risk of PONV according to the Apfel Score,³ that is nonsmoking patients of the female sex. In these patients, anxiety sensitivity (ASI-3 score of eight or more) increased the OR for the occurrence of PONV almost five-fold, suggesting that anxiety sensitivity can contribute to the estimation of an individual's risk of experiencing PONV.

Other factors influencing the risk of PONV in our patient population were younger age, postoperative administration of opioids and lack of pre-operative PONV prophylaxis. Neither laparoscopic surgery, duration of anaesthesia, nor a history of PONV or motion sickness was found to be statistically significant in the regression analysis.

In our patients, 41.5% experienced PONV within the first 24 h after surgery. This high percentage can be explained by the fact that PONV was interpreted in a restrictive fashion – any nausea, retching or vomiting was defined as PONV. However, although PONV episodes in some patients may have been transient or mild and included only retching or nausea without vomiting, it can be assumed that they were bothersome for most patients.

Previous studies examining an association between common anxiety measures and PONV found varying results. In a study evaluating the relationship between pre-operative anxiety and PONV while accounting for established predictors, the authors found only a weak correlation between pre-operative anxiety and PONV and concluded that pre-operative anxiety did not improve the prediction of PONV.¹⁴ In contrast, in a detailed review of PONV,⁵ pre-operative anxiety was found to be a potential risk factor for PONV, although it was investigated only in two of the studies included in this review. Roh et al.¹⁵ demonstrated that pre-operative anxiety, measured with the Amsterdam Preoperative Anxiety and Information Scale, was positively correlated with PONV in patients aged 16 to 77 years undergoing ambulatory hand surgery under general anaesthesia. However, other studies with the same focus found no evidence for a relationship between anxiety and nausea after anaesthesia.^27,28 Comparison of these results is difficult, because the studies differed not only with regard to the definition of PONV and the instruments used to measure anxiety, but also the anaesthetic techniques, medication, surgical procedure, and the studied population, for example children.²⁹

In contrast to previous studies, we investigated inpatients with a pre-existing risk for PONV (nonsmoking and female sex) assuming that anxiety sensitivity might have an additive predictive value for PONV in this selected patient group. In line with this hypothesis, anxiety sensitivity increased significantly the OR for PONV in this cohort of patients.

The results of our study also suggest that anxiety sensitivity might be more predictive of PONV than a history of PONV or motion sickness. Although a history of PONV or motion sickness is part of all commonly used risk scoring systems and such a history often triggers the administration of pre-operative PONV prophylaxis, the relevance of PONV history as a risk factor could not be confirmed in our patient population. Nevertheless, a thorough assessment of previous PONV episodes is recommended, because it may help to differentiate if PONV is patient related (e.g. history of migraine, hormonal factors), surgery or anaesthesia related (e.g. laparoscopic surgery, volatile anaesthetics, opioids), or caused by other influencing factors. In his review article, Gan⁵ mentioned the problem of confusing association and causality by differentiating between ‘true’ and ‘surrogate’ risk factors thus, although the postoperative administration of opioids was confirmed as an independent risk factor in our study, we could not establish if opioids or pain triggered the PONV.

Significantly, the risk of experiencing PONV was comparable if patients received opioids postoperatively or if they had not received any PONV prophylaxis prior to the induction of anaesthesia.

Some limitations of this study should understood. Only patients receiving inhalational anaesthesia were included to standardise the anaesthetic technique and it would have been interesting to evaluate the influence of the anaesthetic technique (e.g. TIVA) on the occurrence of PONV in our population. Furthermore, in our study nausea was assessed with a numeric rating scale ranging from 0 to 10, with any value greater than 0 defined as PONV. Due to this liberal interpretation of PONV and due to the inclusion of only female nonsmoking patients, the incidence of PONV requiring antiemetic medication may be overestimated.

Conclusion

The current study reveals that psychological factors such as anxiety sensitivity help to predict the risk of PONV in a selected group of nonsmoking women. PONV prophylaxis should be considered in patients who meet these criteria, especially if postoperative opioid administration is likely.

Acknowledgements relating to this article

Assistance with the study: none.

Financial support and sponsorship: none.

Conflicts of interest: none.

Presentation: preliminary data for this study were presented as a poster presentation at the German Society of Anaesthesiology and Intensive Care Medicine (DGAI) Deutscher Anästhesiekongress DAC Düsseldorf, 7 to 9 May 2015, Düsseldorf.