Postoperative sore throat is common in patients who have undergone surgery under general anesthesia/endotracheal intubation, with a reported incidence of up to 68%.1,2,3,4,5,6,7 It negatively affects patient recovery and satisfaction,4,7,8,9,10 and considerable efforts have been undertaken to understand and treat this phenomenon. Mucosal trauma, erosion, and inflammation due to endotracheal intubation are proposed etiologies.11,12,13,14,15 Randomized-controlled trials have suggested that inhaled16 or topical17 corticosteroids, topical benzydamine hydrochloride,18 and topical licorice may help to prevent postoperative sore throat. Intravenous dexamethasone administration has also been proposed as a prophylactic measure. Two systematic reviews, based on data available in 2013,19,20 suggested that preoperative intravenous dexamethasone helps to prevent postoperative sore throat. Nevertheless, those reviews comprised a limited number of studies and the outcomes could not be appropriately pooled because the reporting methods used by the original studies were inconsistent.

Since those reviews, further studies have assessed the effect of dexamethasone treatment on postoperative sore throat and these are better suited for pooling of their findings. Therefore, we performed a systematic review and meta-analysis of the efficacy and safety of preoperative intravenous dexamethasone administration to prevent postoperative sore throat in adults subjected to tracheal intubation for general anesthesia. We focused on the incidence and severity of postoperative sore throat, and reported adverse events and postoperative cough and hoarseness.

Methods

We adhered to the Cochrane Collaboration methodology21 and PRISMA statement22 for the conduct and reporting of this systematic review. Our protocol is registered at PROSPERO (CRD42018086697).

Eligibility criteria

Type of studies

We included randomized-controlled studies that compared preoperative intravenous dexamethasone with non-analgesic or active control treatment in adults who underwent elective surgery under general anesthesia. We excluded observational studies and quasi-randomized and non-randomized controlled trials.

Type of participants

We included adult patients aged ≥ 18 yr who underwent elective surgery under general anesthesia and endotracheal intubation. We excluded patients whose general anesthesia care included a supraglottic airway device or those who underwent head and neck surgery that might induce confounding pain in the laryngopharynx.

Type of interventions

The intervention included dexamethasone that was intravenously administered to prevent sore throat or postoperative pain elsewhere. We placed no restrictions on the dose or the number of doses of dexamethasone, as long as it was administered preoperatively. We excluded studies that administered dexamethasone postoperatively.

Comparators included non-analgesic methods or active controls that were administered preoperatively. Non-analgesic strategies included no endotracheal tube lubrication, usual care, or the use of drugs without known analgesic potency, such as intravenous saline. Active controls involved agents that have known prophylactic effects against postoperative sore throat, such as topical ketamine or an analgesic combined with intravenous dexamethasone. We placed no restrictions on the dose or the number of doses of the comparators. We excluded studies that administered non-analgesic methods or active controls postoperatively.

Types of outcome measures

Our primary outcomes included 1) the incidence and 2) severity of postoperative sore throat at 24 hr after surgery/extubation and 3) adverse events. Because there is no established definition for postoperative sore throat, we accepted the investigators’ definitions. Studies on postoperative sore throat frequently use a four-level classification system: none, mild, moderate, and severe. When a study used this classification system, we calculated the incidence of sore throat from the sum of mild, moderate, and severe cases. We accepted each study author’s definition of adverse events.

Our secondary outcomes included 1) the incidence of moderate or severe sore throat, 2) cough, and 3) hoarseness at 24 hr after surgery/extubation. Since there are no established universal definitions for postoperative cough or hoarseness, we accepted the investigators’ definitions. We calculated the incidence of cough or hoarseness based on the sum of mild, moderate, and severe cases, according to the commonly-used classification system.

Search strategy

We searched PubMed, EMBASE, and the Cochrane Central Register of Controlled Trials for eligible studies. We reviewed the reference lists of the included publications and previous systematic reviews, and searched Google Scholar and ClinicaTrials.gov to identify further eligible studies. We did not search the grey literature. We placed no restrictions on publication status or language. Our search strategy is presented in Table 1. We updated the search on May 3, 2018.

Table 1 Search strategy
Full size table

Study selection

Two authors (A.K. and H.M.) independently, and in duplicate, reviewed the articles retrieved through the search and selected the eligible ones. We resolved disagreements by consensus (κ = 0.92).

Data extraction

These two authors then independently extracted the following data from each study: patient characteristics (age, sex, and American Society of Anesthesiologists [ASA] physical status), study characteristics (country and type of surgery), interventions used (dexamethasone dose, if reported, and comparators), and outcomes of interest.

Risk of bias assessment

The same authors independently and in duplicate evaluated the risk of bias with the Cochrane risk of bias assessment tool.21 We considered a study to be at a low risk of performance bias when the original study blinded participants to the intervention they received until all assessments were complete, given that sore throat is a subjective outcome. We also reviewed for conflicts of interest or industry sponsorship. We resolved any inconsistency through discussion (κ = 0.53–1.0). When an e-mail address was available, we contacted the authors of the original study for more information. Specifically, we attempted to retrieve information regarding the methodology of each trial and the outcomes of interest that were not presented in the articles. We deemed the authors to be unresponsive if they did not reply after three contact attempts.

Statistical analysis

We calculated the risk ratio (RR) and standardized mean difference (SMD) for dichotomous and continuous outcomes, respectively. When a trial had zero events in either arm, we applied continuity corrections by adding 0.5 to each cell of the 2 × 2 tables from the trial.23 When a study presented the data as the median with interquartile range, we converted the values to the mean and standard deviation using the method proposed by Wan et al.24 We pooled the data using the DerSimonian and Laird random-effects model.25 We assessed statistical heterogeneity using Q and I2 statistics.26 We deemed I2 ≥ 50% as substantial statistical heterogeneity and conducted subgroup analysis by the type of non-analgesic method. Since there is no evidence to suggest that non-analgesic methods work similarly, we conducted subgroup analysis by the type of comparator. We anticipated that there would be no difference in effect size between the subgroups by the type of comparator. Because the number of studies for each outcome was less than ten, we did not test for publication bias.21

We performed primary analyses by the type of comparator. We pooled the data into a single arm when a study examined groups based on different dexamethasone doses. We conducted sensitivity analyses by excluding trials at an unclear or high risk of bias with regard to sequence generation, allocation concealment, blinding of participants and outcome assessors, and conflicts of interest/industry sponsorship. We also conducted another sensitivity analysis by presenting the pooled outcomes of risk difference.

We conducted subgroup analysis and meta-regression to estimate the dose-response relationship for the incidence of postoperative sore throat, which had been a priori planned in case the number of studies was sufficient and a variety of regimens were used. Unless the mean body weight for the dexamethasone group was reported, we converted the dexamethasone dose to the equivalent value in mg·kg−1 based on the assumption that the mean body weight was 60 kg (most eligible studies were conducted outside North America and reported approximately 60 kg as the mean body weight of participants). We conducted subgroup analysis according to the following dexamethasone doses: low dose (0.1 mg·kg−1) and intermediate dose (0.11–0.20 mg·kg−1).27 We anticipated that the preventive effect of the intermediate dexamethasone dose might be larger than that of the lower dexamethasone dose.

To evaluate our findings’ robustness against type 1 and type 2 errors, we conducted trial sequential analysis (TSA) with the O’Brien-Fleming alpha-spending function method.28,29 We retained a 5% risk of type 1 errors and a power of 80%. Furthermore, we obtained the required information size according to relative risk reduction of 20% for postoperative sore throat, which is the most conservative value used for the power calculations employed in the included trials.30 The threshold for statistical significance was set at P < 0.05. We used Stata SE, version 15.1 (StataCorp., College Station, TX, USA) to perform conventional meta-analyses, while we used the TSA software, version 0.9 beta (Copenhagen Trial Unit, Copenhagen, Denmark) to conduct the TSA.

Results

Overview of included studies

Our database search initially produced 1,892 titles and abstracts, and an additional search found five articles. We ultimately included 15 randomized-controlled trials involving 1,849 study participants for the analysis after applying our inclusion and exclusion criteria30,31,32,33,34,35,36,37,38,39,40,41,42,43,44 (Fig. 1).

Fig. 1
figure 1

Flow diagram illustrating the literature screening process and selection of studies

Full size image

The mean age for the study participants ranged from 28–62 yr, and the proportion of females ranged from 16%–100% (Table 2). Twelve and three trials included patients with an ASA status of I–II and III, respectively. Eight trials listed the types of surgery as follows30,32,33,34,35,38,39,40,42,43: breast and abdominal surgery (including laparoscopic cholecystectomy), orthopedic surgery (including the limbs and spine), gynecologic surgery, middle ear surgery, and thoracic surgery (open pulmonary resection or video-assisted thoracoscopic surgery). The median sample size was 110 (range, 49–226).

Table 2 Characteristics of included studies
Full size table

Intravenous dexamethasone was administered as a single-dose infusion prior to surgery in all studies. The dose was fixed in eight studies (4 mg in one study36; 8 mg in five studies,32,36,39,42,44 and 10 mg in three studies34,40,41) and was determined according to patient’s body weight in five studies (0.05 mg·kg−1 in one study33; 0.1 mg·kg−1 in three studies,30,33,35 and 0.2 mg·kg−1 in five studies).30,31,35,37,43 The remaining study did not specify the dose. Intravenous saline was used as a non-analgesic control in 12 studies. Comparator agents with known analgesic effects included betamethasone38 and triamcinolone gel41 and were applied over endotracheal tubes (one study each); ketamine gargles37,43 or ketamine gargles combined with intravenous dexamethasone37,43 (two studies); and intravenous lidocaine32 or acetaminophen34 combined with intravenous dexamethasone (one study each). Four trials were performed in South Korea,30,32,34,35 three in India,31,39,43 two each in Iran37,38 and Thailand,36,42 and one each in China,40 Indonesia,41 Nepal,44 and the United States of America.33 All studies were published as full text between 2007 and 2018. All studies but two were reported in English, with the remaining ones published in Chinese40 and Indonesian.41

Eight studies used a four-level classification system to rate the severity of postoperative sore throat and confirm its presence.30,31,34,37,40,41,43,44 Two studies used a visual analog scale35,39 while two used a numeric rating scale to rate the severity.33,42 Five studies reported the incidence of postoperative sore throat.30,32,35,36,38 We contacted 11 study authors with a valid e-mail address for further information, and six responded.31,32,34,36,41,44 Most studies were at relatively low risk of bias, as shown in Table 3. Nevertheless, in one study, the duration of patient intubation in the dexamethasone group was significantly longer compared with the control group (142 vs 121 min).42 As duration of intubation is a known risk factor for postoperative sore throat,8 we considered this study to have a high risk of bias.

Table 3 Risk of bias in included studies
Full size table

Dexamethasone vs non-analgesic controls

Incidence of postoperative sore throat

Nine studies comprising 983 participants provided data on the incidence of sore throat at 24 hr after surgery/extubation.30,31,35,36,37,38,40,43,44 Intravenous administration of dexamethasone was associated with a lower incidence of postoperative sore throat (RR, 0.62; 95% CI, 0.51 to 0.75; P < 0.001; df = 7; I2 = 0%; Fig. 2).

Fig. 2
figure 2

Forest plot for the incidence of postoperative sore throat at 24 hr after surgery/extubation. The plot shows decreased incidence in patients treated with iv dexamethasone compared with non-analgesic control. RR = risk ratio

Full size image

Severity of postoperative sore throat

Four studies including 431 participants provided data on the severity of sore throat at 24 hr after surgery/extubation.33,35,39,42 Intravenous administration of dexamethasone was associated with a lower severity of postoperative sore throat (SMD, −1.06; 95% CI, −1.80 to −0.33; P = 0.005; df = 3; I2 = 91%; Fig. 3).

Fig. 3
figure 3

Forest plot for the severity of postoperative sore throat at 24 hr after surgery/extubation. The plot shows decreased severity in patients treated with iv dexamethasone compared with non-analgesic controls. SD = standard deviation; SMD = standardized mean difference

Full size image

Adverse events

Four studies reported on adverse events. One study reported that the frequency of adverse events during 24 hr after surgery was comparable between the dexamethasone and control groups,39 another noted that no adverse event was associated with dexamethasone,35 the third reported that there were no differences in the incidence of local or systemic side effects between the dexamethasone and control groups,37 and the fourth reported that there were no local or systemic adverse events, except one episode of vomiting in the non-analgesic control group.43

Secondary outcomes

Moderate and severe postoperative sore throat

Six studies involving 562 participants provided data on the incidence of moderate or severe sore throat at 24 hr after surgery/extubation.30,31,36,40,43,44 Intravenous administration of dexamethasone was associated with a lower incidence of moderate or severe postoperative sore throat (RR, 0.10; 95% CI, 0.03 to 0.39; P = 0.001; df = 4; I2 = 0%; Fig. 4).

Fig. 4
figure 4

Forest plot for the incidence of moderate or severe postoperative sore throat at 24 hr after surgery/extubation. The plot shows decreased incidence in patients treated with iv dexamethasone compared with non-analgesic controls. RR = risk ratio

Full size image

Postoperative cough

Three studies comprising 349 participants provided data on the incidence of cough at 24 hr after surgery/extubation.30,31,33 Intravenous administration of dexamethasone was associated with a reduced incidence of postoperative cough (RR, 0.66; 95% CI, 0.55 to 0.80; P < 0.001; df = 2; I2 = 0%; Fig. 5).

Fig. 5
figure 5

Forest plot for the incidence of postoperative cough at 24 hr after surgery/extubation. The plot shows decreased incidence in patients treated with iv dexamethasone compared with non-analgesic controls. RR = risk ratio

Full size image

Postoperative hoarseness

Six studies including 621 participants provided information on the incidence of hoarseness at 24 hr after surgery/extubation.30,31,33,35,37,43 Intravenous administration of dexamethasone was associated with a reduced incidence of postoperative hoarseness (RR, 0.47; 95% CI, 0.31 to 0.74; P = 0.001; df = 5; I2 = 61.0%; Fig. 6).

Fig. 6
figure 6

Forest plot for the incidence of postoperative hoarseness at 24 hr after surgery/extubation. The plot shows decreased incidence in patients treated with iv dexamethasone compared with non-analgesic controls. RR = risk ratio

Full size image

Subgroup, sensitivity, and trial sequential analyses

There was no significant difference in the incidence of postoperative sore throat between subgroups created according to the type of comparator (P = 0.46; eTable, available as Electronic Supplementary Material). The results of sensitivity analyses excluding studies at an unclear or high risk of bias were consistent with those of primary analyses for all outcomes (eTable, available as Electronic Supplementary Material [ESM]). Trial sequential analysis suggested that the cumulative z curve crossed both the conventional and trial sequential monitoring boundaries for benefit before reaching the required information size (2,020 participants), which therefore supports the true positive efficacy of intravenous dexamethasone for preventing postoperative sore throat (Fig. 7).

Fig. 7
figure 7

Trial sequential analysis for the incidence of postoperative sore throat at 24 hr after surgery/extubation in patients treated with iv dexamethasone or non-analgesic controls. The cumulative z curve (blue line) crosses the conventional boundary for benefit (z score of 1.96; horizontal brown line) and the trial sequential monitoring boundary for benefit (red line), demonstrating true positive evidence to suggest that no further studies are required. A diversity-adjusted required information size of 2,020 patients (red vertical line) was calculated, using α = 0.05 (two-sided) and β = 0.20 (power of 80%) and an anticipated relative risk reduction of 20%

Full size image

We examined the relationship between the effect size (incidence of sore throat) and dexamethasone dose. Subgroup analysis found a similar effect size between groups that reported the dexamethasone dose (P = 0.61): low dose (RR, 0.62; 95% CI, 0.35 to 1.09; df = 2; I2 = 42%) and intermediate dose (RR, 0.60; 95 % CI, 0.47 to 0.76; df = 6; I2 = 0%). Meta-regression (P = 0.19) failed to find a dose-response relationship (eFigure, available as ESM).

Dexamethasone vs active controls

Five trials employed analgesic agents as comparators.32,34,37,38,41 The comparators were corticosteroids applied to endotracheal tube cuffs in two trials,38,41 ketamine gargles and intravenous dexamethasone in two trials,37,43 ketamine gargles37,43 in two trials, intravenous dexamethasone and intravenous lidocaine in one trial,32 and intravenous dexamethasone and acetaminophen in one trial.34

With regard to the prevention of postoperative sore throat, intravenous dexamethasone administration was similar to corticosteroids applied to endotracheal tube cuffs38,41 or ketamine gargles alone (RR, 0.80; 95% CI, 0.41 to 1.52),37,43 or intravenous dexamethasone and acetaminophen,34 while it was inferior to ketamine gargles and intravenous dexamethasone (RR, 2.25; 95% CI, 1.13 to 4.48)37,43 or intravenous dexamethasone and intravenous lidocaine.32 With regard to the prevention of postoperative hoarseness, intravenous dexamethasone was similar to ketamine gargles and intravenous dexamethasone (RR, 1.17; 95% CI, 0.42 to 3.26),37,43 ketamine gargles alone (RR, 0.70; 95% CI, 0.29 to 1.71),37,43 intravenous dexamethasone and intravenous lidocaine,32 or intravenous dexamethasone and acetaminophen.34

Discussion

Our review suggests that, compared with non-analgesic controls, preoperative intravenous administration of dexamethasone was associated with a reduced incidence and severity of postoperative sore throat in adults undergoing elective surgery under anesthesia/endotracheal intubation. Our analysis indicates that the number of patients needed to treat with intravenous dexamethasone (to prevent one case of postoperative sore throat) is eight (95% CI, 5 to 20), which indicates a considerable prophylactic benefit. Intravenous dexamethasone administration also decreased the incidence of moderate or severe sore throat as well as hoarseness and cough after surgery. Limited available evidence suggests that intravenous dexamethasone administration is not associated with significant adverse events. Our findings were robust throughout the sensitivity analysis and TSA, providing strong evidence that preoperative intravenous dexamethasone administration effectively alleviates postoperative sore throat.

The presumed etiology underlying postoperative sore throat is mucosal inflammation around the tracheal tube cuff.12 Previous studies suggest that topical and intravenous anti-inflammatory agents prevent postoperative sore throat in surgical patients,16,17,18,45 thus supporting this explanation. Furthermore, intravenous corticosteroids administered to critical care patients prior to elective extubation decrease laryngeal edema and reduce post-extubation airway complications.46 In line with this evidence, our study suggests that preoperative intravenous dexamethasone administration is associated with a reduced risk of postoperative sore throat.

Some studies have attempted to examine a dose-response relationship by testing two doses of dexamethasone. Our subgroup and meta-regression analyses suggest no association between the dose of dexamethasone and the effect size. We assumed a mean body weight of 60 kg because most studies were conducted outside North America. Re-analysis with an assumed body weight of 70 kg yielded similar findings. Although our findings do not elucidate the optimal dexamethasone dose for preventing postoperative sore throat, we believe that low-dose dexamethasone may be effective.

Previous studies have suggested that topical application of ketamine and corticosteroids help to prevent postoperative sore throat.17,47 Our study suggests that intravenous dexamethasone is superior to topical ketamine, while topical corticosteroids and most analgesic agents combined with intravenous dexamethasone are no better than intravenous dexamethasone alone. Nevertheless, most of these comparisons were performed in single, small-sized trials and we believe that additional studies of this nature are warranted.

Our review could not adequately assess adverse events associated with intravenous dexamethasone. In 1996, The Consolidated Standards of Reporting Trials (CONSORT) statement recommended that trial investigators should report unintended effects related to interventions.48 All studies that are included in our review were published after this statement. Nevertheless, only four studies reported adverse events and few details were provided. There is literature indicating that short-term, high-dose corticosteroids and single-dose intravenous dexamethasone are not associated with significant adverse events.27,46 By analogy, we speculate that the prophylactic intravenous administration of dexamethasone to prevent sore throat is unlikely to be related to significant adverse events.

Two previous systematic reviews suggested that intravenous administration of dexamethasone helps to prevent postoperative sore throat.19,20 Our findings are in agreement with these studies. Nevertheless, there are substantial differences between those reviews and the present one. First, the previous reviews included four and seven trials, respectively, whereas this report includes 15 studies, 13 of which included non-analgesic controls. Second, we pooled efficacy outcomes strictly at 24 hr after surgery/extubation. In contrast, the previous reviews pooled outcomes recorded over a wide range of time intervals because the original studies were few and inconsistent in reporting their outcomes; we excluded two such trials, which were included in the previous reviews. Third, we included two non-English language reports and studies involving comparisons with other analgesic agents; thus, our review was more comprehensive. Fourth, our TSA suggested true positive efficacy of intravenous dexamethasone administration to help prevent postoperative sore throat.

Additional strengths of our study include the comprehensive nature of the literature search, which identified 15 publications thus permitting sensitivity and meta-regression analyses. Moreover, we compared intravenous dexamethasone with topical corticosteroids and ketamine, which are known prophylactic measures to prevent postoperative sore throat. Finally, we complied with the Cochrane methodology. We contacted the original study authors to more accurately evaluate the risk of bias and requested necessary unpublished data. We also confirmed the robustness of our findings with sensitivity analysis and TSA.

Nevertheless, our review is not without limitations. First, the number of pooled studies for each outcome was relatively small even though the TSA provided strong evidence for the efficacy of preoperative intravenous dexamethasone. Second, there is a possibility of publication bias for each outcome. We were unable to test for publication bias given the limited number of studies available for each outcome, according to the recommendation of the Cochrane Collaboration. Third, there were potentially diverse, un-controlled confounding variables of the included trials. For example, operator inexperience, elevated cuff pressure, larger tracheal tubes, and a prolonged duration of intubation are known risk factors for postoperative sore throat. Unfortunately, we could not adjust for these variables because they were inconsistently reported.

In conclusion, our study suggests that preoperative intravenous dexamethasone administration alleviates postoperative sore throat more effectively than non-analgesic strategies.