Postoperative nausea and vomiting (PONV) is an important cause of distress for patients. Depending on the presence of known risk factors — female sex, non-smoking status, past history of PONV or motion sickness, or anticipated requirement for postoperative opioids (the Apfel criteria) — it can affect up to one-third of patients. ^{1, 2} Patients rank PONV ahead of pain in the list of undesirable outcomes after an operation. ³ Moreover, PONV can lead to complications such as wound dehiscence, aspiration, hypovolaemia, and even oesophageal rupture. ⁴

PONV is also common in the intensive care unit (ICU), particularly after cardiac surgery. In a study of 400 cardiac surgical patients, the baseline incidence of nausea was reported at 47% and retching/vomiting at 37%. ⁵ Both surgical-specific aspects (eg, longer operation duration, greater intraoperative opioid and volatile general anaesthetic agent exposure) as well as components of ICU management (ongoing opioid analgesia and sedation administration, as well as other emetic medications) combine to make ICU patients particularly prone to PONV. ^{6, 7}

Chewing gum stimulates gastrointestinal motility by vagal activation through effects of sham feeding and perhaps a reduction in systemic inflammation, ^{8, 9} with a good safety profile in resolving postoperative intestinal ileus in a number of surgical populations (including after gastrointestinal tract surgery, caesarean delivery and liver transplantation). ^{10, 11, 12, 13, 14} A meta-analysis of 26 randomised controlled trials involving 2214 colorectal surgical patients demonstrated a reduced time to recovery of intestinal function and shortened length of stay with chewing gum compared with standard postoperative care protocols. ¹⁵ More recently, chewing gum has been investigated as a novel treatment for PONV, demonstrating non-inferiority compared with ondansetron in a pilot randomised controlled study of 94 female patients undergoing breast and laparoscopic surgery. ¹⁶ No studies, however, have addressed its role in prophylaxis for PONV in postoperative patients in the ICU. It is unknown whether the unique factors to the ICU environment, such as the greater requirement for sedation and analgesia after major surgery, patient drowsiness and challenges relating to gum chewing with ongoing mechanical ventilation, may affect its efficacy in this patient group.

In patients admitted to the ICU after surgery, we therefore performed a pilot randomised controlled trial of the prophylactic effectiveness of chewing gum on PONV compared with a control group administered 20 mL of oral water. We hypothesised that chewing gum would decrease the incidence, duration and severity of PONV in this patient group.
 

Methods

Study design

This was a prospective, open label, pilot randomised controlled trial conducted in two metropolitan ICUs. Ethics approvals from Austin Health (LNR/17/Austin/205) and Epworth Health (EH2017-288) were obtained. The trial was prospectively registered with the Australian New Zealand Clinical Trial Registry (No. ACTRN12617001185358).

This trial was conducted in compliance with the Consolidated Standards of Reporting Trials (CONSORT) statement.

Eligible patients were those aged 18 years or more; admitted to the ICU within 16 hours after non-oropharyngeal, maxillary or oesophageal surgery; and anticipated to stay in the ICU for more than 12 hours. Patients who remained mechanically ventilated, who were deemed too sedated (based on the judgement of attending clinical staff), or who had partial or full dentures were excluded.

Patients were enrolled after written informed consent was obtained, either before or after the surgery. Patients were randomly allocated to either the intervention chewing gum group (Falim [Kent Gida], lightly mint flavoured, sugarless gum; ingredients: gum base, flavouring, colouring [E171], antioxidants [E321, E320]), which was chewed for at least 15 minutes every 4 hours, or to a control group, who were administered a 20 mL sip of water orally every 4 hours. Randomisation was achieved using a computerised random number generator; allocation concealment was done via sequentially numbered envelopes. Participants were enrolled by the two investigators (HA and NC). All patients were allowed other oral intake as tolerated and permitted by the surgical team. We pre-determined a cardiac surgery subgroup.

All enrolled patients had access to a standardised antiemetic rescue protocol:

• first line — ondansetron 8 mg intravenous;
• second line — droperidol 0.625 mg intravenous;
• third line — metoclopramide 10 mg intravenous;
• fourth line — prochlorperazine 12.5 mg intravenous.

If nausea was rated greater than 5 points on a visual analogue scale from 0 to 10 points, ¹⁷ actual vomiting occurred or, at any time, if requested by the patient, rescue antiemetic was administered. After 30 minutes, PONV assessment was made by the attending ICU nurse to determine if further rescue antiemetic was needed. After randomisation, the attending ICU nurse collected data for 24 hours, divided into continuous 4-hour observation periods. Data collected included presence, duration and severity of nausea (measured on a ten-point Likert scale), number of retching/vomiting episodes, and antiemetic medication administered. If nausea was not self-reported, the frequency of querying the patient for nausea within each 4-hour period by the ICU nurse was not standardised. The number of episodes of gum chewing was recorded, with the underlying reason if not able to be chewed (eg, patient refusal or drowsiness). The ICU nurse attending the patient reconciled all gum administered and discarded, with any instances of swallowed or aspirated gum recorded.

The primary outcome was the number of patient-reported nausea episodes in the first 24 hours after the operation, with secondary outcomes including the combination of vomiting or dry retching episodes, duration of nausea episodes, use of rescue antiemetics in the two groups, and the severity of nausea as reported in a visual analogue scale from 0 to 10 every 4 hours.
 

Sample size calculation

We based our power calculation on a pilot study comparing chewing gum with ondansetron, with a full resolution rate of PONV in the chewing gum group of 75% and in the ondansetron group of 39%. ¹⁶ We assumed that, in our prophylaxis study, a smaller effect would apply, with a decrease in combined nausea and vomiting/dry retching episode from 75% in the intervention group to 50% in the control group, which, at an 80% power and an α = 0.05, would require 90 patients in total.
 

Statistical analysis

Data are expressed as means with standard deviation (SD), when normally distributed, or median and interquartile range (IQR), when non-parametric. The χ² or Fisher exact test was applied to categorical variables (including the primary outcome), with normally distributed continuous variables compared using the Student t test, and the Mann– Whitney U test used for non-parametric continuous variables. A P < 0.05 indicated statistical significance. All statistical analyses were performed using R, version 3.4.2 (R Foundation for Statistical Computing, Vienna, Austria).
 

Results

Between October 2017 and January 2019, we screened a total of 171 patients, after exclusions (28 patients with dentures, 25 patients remaining mechanically ventilated, 19 refusals, and five patients too drowsy), 90 patients were enrolled. Forty-six patients were randomised to the chewing gum group and 44 patients to the control group (Figure 1).

The demographic features of enrolled patients are outlined in Table 1. Groups were similar in baseline characteristics, including Apfel PONV risk scores. Thirty-seven patients (41%) were admitted after cardiac surgery, 14 (15%) after bariatric surgery, and 16 (17%) after other gastrointestinal surgery, with the remainder undergoing other procedures.

Gum was chewed on 148 occasions by all patients randomised to the chewing gum group (median, 3 occasions; IQR, 2–4). It was not administered on 93 occasions (38%): 72 occasions because the patient was sleeping, 15 occasions because the patient refused, and six occasions due to patient drowsiness. No gum was swallowed or aspirated.

There was no difference between groups in the number of patients with nausea (10 [22%] chewing gum patients [95% CI, 4–15; 9–34%] ν 12 [27%] control patients [95% CI, 5–18; 14–41%]; P = 0.72) or vomiting (2 [4%] chewing gum ν 6 [14%] control patients; P = 0.24) with a combined incidence of these outcomes of 12 (16%) versus 18 (41%) (Table 2). There was no difference in the number of patient-reported nausea episodes (total 22 nausea episodes per patient in the chewing gum group [median, 0; IQR, 0–0] ν 21 in the control group [median, 0; IQR, 0–1]; P = 0.706). There was also no difference in the degree, duration or severity of nausea; number of episodes of vomiting or retching; or the use of rescue antiemetics (Table 2). Thirty-seven patients were enrolled in the pre-determined cardiac surgical subgroup: 19 in the chewing gum group and 18 in the water group, with no differences in outcomes between groups.
 

Discussion

Key findings

In this prospective, open label, two-centre randomised controlled trial, we assessed the use of chewing gum for the prevention of postoperative nausea and vomiting in the ICU. We found chewing gum treatment feasible in the ICU, but we also found no difference between chewing gum and water administered orally every 4 hours in the number of patients with, or in the episodes of, nausea, vomiting or retching, nor in the severity or duration or in the use of rescue antiemetics. There was also no difference between the two groups in the subgroup of patients after cardiac surgery.

Relationship to previous studies

Implications of the study findings

Study strengths and limitations

Conclusion