A meta-analysis of randomized controlled trials: efficiency and safety of ondansetron in preventing post-anesthesia shivering during cesarean section

Although ondansetron was considered to prevent post-anesthesia shivering during cesarean section, its efficiency remained controversial. Our review was conducted to estimate the efficiency and safety of ondansetron in preventing post-anesthesia shivering during cesarean section. The literature were searched from their inception to October 2020 without restriction of language. All randomized controlled trials investigating the efficacy of ondansetron versus placebo in preventing shivering during cesarean section under neuraxial anesthesia were included. The meta-analysis was conducted using Stata software. Eleven randomized controlled studies with a total of 748 individuals were finally included in our meta-analysis. Our results manifested that intravenous ondansetron compared with intravenous placebo significantly reduced the incidence of post-anesthesia shivering (PAS) (RR 0.53, 95% CI 0.14–0.68). Subgroup analysis according to doses of ondansetron indicated that the efficacy of 4 mg doses of ondansetron (RR 0.37, 95% CI 0.21–0.64) is equivalent to that of 8 mg doses of ondansetron (RR 0.61, 95% CI 0.47–0.81) in preventing PAS. In addition, the intravenous ondansetron led to a lower incidence of hypotension than intravenous placebo (OR 0.47, 95% CI 0.32–0.70). We could not demonstrate differences in the incidence of bradycardia between intravenous ondansetron and intravenous placebo. Our results found that intravenous ondansetron was effective in preventing shivering during cesarean section under neuraxial anesthesia, and had an advantage in reducing the incidence of hypotension compared with intravenous placebo.


Introduction
In the past few decades, the cesarean section (CS) has been chosen by more and more pregnant women. The global rate of cesarean birth has doubled to 21% in the past 15 years, and is increasing annually by 4% [1,2]. This might be due to the improvement of operation and anesthetic techniques which reduces the risk of postoperative complications [3]. Another reason is that the safety of cesarean section has been recognized by obstetrician-gynecologists and patients. Meanwhile, the cesarean section compared to vaginal delivery is thought to prevent incontinence, prolapse, and sexual dissatisfaction, increasing its attractiveness [4]. Finally, the psychology of cesarean sections in some countries was already widespread among affluent women and their doctors [5]. But some obstetrician-gynecologist are reluctant to choose a cesarean section because of a significantly increased risk of surgical complications, including maternal death, cardiac arrest, surgical site infections, or anesthesia-related complications [3,6,7].
Neuraxial anesthesia (NA), composing spinal anesthesia, epidural anesthesia, or spinal-epidural anesthesia, was recommended for cesarean sections in 2007 by American Society of Anesthesiologists [8,9]. However, NA is associated with PAS, which may lead to extreme discomfort, anxiety, and fear. More importantly, PAS affects maternal metabolic activity and leads to increased oxygen consumption, which may increase the risk of cardiovascular 1 3 and cerebrovascular diseases caused by hypoxia. Serious adverse events may occur if the patient has cardiopulmonary insufficiency [10].
The perioperative clinical probability of PAS was as high as 80% [11], so preventing perioperative PAS will become particularly significant. Currently, medicines include: α2-agonists, opiates, tramadol, magnesium sulfate, corticosteroids, nefopam, and serotonin (5-hydroxytryptamine) 5-HT3 antagonists still have remained the most popular way to prevent PAS [12]. Unfortunately, the overall quality of the pharmacological treatment is low. Ondansetron, a 5-HT3 antagonist, is generally used as an antiemetic and is widely used clinically. At the same time, ondansetron was also considered to prevent hypotension, bradycardia, and PAS [13]. But its efficiency in the prevention of PAS during cesarean section under neuraxial anesthesia remains controversial [13,14]. Therefore, we conducted a meta-analysis to explore the effect of ondansetron in preventing PAS during cesarean section under neuraxial anesthesia.

Data sources and searches
We searched The Cochrane Library, PubMed, Web of Science, Embase, China National Knowledge Infrastructure for all randomized controlled trials that investigated the efficacy of ondansetron versus placebo to prevent shivering during cesarean section under neuraxial anesthesia published up to August 2020. In our literature search, we applied no language restrictions, and the following keywords were included: ondansetron, cesarean section, and shivering.

Inclusion and exclusion criteria
Studies meeting the following criteria were eligible for inclusion: (I) adult pregnant women; (II) Randomized Controlled Trials investigating the efficacy of ondansetron versus placebo in preventing shivering during cesarean section under neuraxial anesthesia; (III) Outcomes (primary outcome: incidence of shivering, secondary outcomes: incidence of hypotension and bradycardia). Studies meeting the following criteria were excluded: (I) we excluded commentaries and documents without sufficient data; (II) overlapping with previous published data or articles; (III) We excluded studies that used different agents alone or in combination with ondansetron.

Quality assessment
According to the Cochrane Collaboration's tool (version 5.1.0) for assessing the risk of bias [15], two authors (JYZ, and JPL) independently assessed the risk of bias in all included studies, and any differences were resolved through discussion. Any disagreement in abstracted data was arbitrated by a third author if necessary. Risk of bias tables for each study included the following domains: random sequence generation, allocation concealment, blinding of participants/personnel, blinding of outcomes assessment, incomplete outcome data, and selective outcome reporting (see Table 1).

Data extraction
Two researchers (GHZ, and JPL) were required to independently abstract all data from each included study: first author, publication date, country, age, American society of anesthesiologists grading, the definition of PAS, injection anesthesia mode and location, the primary outcome and secondary outcomes, and the number of participants with the ondansetron group and the placebo group. Table 2.

Measure of effect and statistical analysis
We assessed the efficacy of ondansetron versus placebo for preventing shivering during cesarean section under neuraxial anesthesia on three outcomes: the incidence of PAS, the incidence of hypotension, and the incidence of bradycardia. The three outcomes were binary count data, were presented as relative risk (RR) or odds ratio (OR) with a 95% confidence interval (CI), and a fixed-effects model was employed for the meta-analysis. All of the statistical analyses were conducted with the Review Manager Version 5.3 (The Cochrane Collaboration, Software Update, Oxford, UK). We estimated the percentage of variability contribution to heterogeneity with the I 2 statistic. I 2 values of ≥ 50% indicated substantial heterogeneity. All p values were calculated from two-tailed tests of statistical significance with a type I error rate of 5%. The funnel plot test was used to assess the publication bias.

Trial sequential analysis
The shortcomings of classical meta-analysis can be solved by trial sequential analysis (TSA). TSA can determine whether the current sample size is enough and to prevent duplicated updates from increasing the risk of random errors in the meta-analysis. The trial sequential monitoring boundaries are that TSA forms monitoring boundaries by correcting random errors. The conventional threshold Z (Z = 1.96), is the traditional significant horizontal line (α = 0.05). Required information size (RIS) means the number of cases required for a meta-analysis to achieve a statistically significant difference. Consequently, for the primary outcomes, we conducted trial sequential analyses (TSAs) using TSA software (version 0.9 Beta; Copenhagen Trial Unit, Copenhagen, Denmark). The TSA was performed with a 5% risk of type I error, a 20% risk of type II error, and a power of 80%. The cumulative Z curve crossed the trial sequential monitoring boundary or the required information size, suggesting that the statistical evidence is firm for this meta-analysis. When analysing dichotomous outcomes using a fixed-effects model, we practically anticipated an intervention effect of 47% risk ratio reduction.

Literature search and trial characteristics
We searched the Cochrane Library, PubMed, Embase, China National Knowledge Infrastructure, Web of science, a total of 749 records were identified. We excluded 496 articles by analyzing the titles and abstracts. After excluding duplicate studies, 15 articles remained. After evaluating the full texts of these 15 articles. 2 studies without placebo control group were excluded, 2 studies without sufficient data also were excluded. Eleven randomized controlled studies with a total of 748 individuals were finally included in our meta-analysis (see Fig. 1). All included trials were published between 2013 and 2020. The 11 trial participants were all over 18 years of age. Of these 11 studies, 5 were conducted in Africa (4 from Egypt [16][17][18][19] and 1 from Nigeria[13]), 5 were conducted in Asia (3 from China [11,20,21], 1 each from India [22] and Iran [23]), 1 was conducted in Oceania (1 from Australia [14]). The trial drugs were given before anesthesia in most studies, but one was given after delivery of the baby [17]. The dosage of ondansetron was either 4 mg or 8 mg in included studies. In 11 trials, 9 trials used spinal anesthesia, and 2 trials used spinal-epidural anesthesia. Besides, most of the injection location is in L3-4. The main characteristics of the eleven studies are listed in Table 2.
The result of the TSA showed that the cumulative Z curve crossed the conventional boundary, the trial sequential monitoring boundary and the required information size. The cumulative Z curve entered the benefits range, indicating that not only can we draw a positive conclusion, but also that it is unnecessary to launch more trials, which established sufficient and conclusive evidence.

Discussion
As a way of the body's response, shivering is an involuntary muscle tremor [24]. Post-anesthesia shivering may be thermoregulatory in most patients with the mechanism of central thermoregulatory central control [25]. The shivering-response thresholds of pregnant women may change because maternal metabolic rates during delivery are different from others [26]. Although there are beneficial thermoregulatory effects, shivering places the body under increased physiological stress, which may be harmful. The response of shivering during neuraxial anesthesia may double oxygen consumption and carbon dioxide production, interfere with blood pressure, electrocardiogram, and pulse oximeter monitoring, which can be adverse factors for the delivery of women with ischemic cardiovascular disease [10,27]. Serotonin (5-hydroxytryptamine ) is a biogenic amine which presents in the brain and spinal cord, and plays an important role in neurotransmission. In animal models, researchers found that direct intraventricular injections of 5-HT influence body temperature and shivering [28]. As a In our meta-analysis, we have reviewed trials reporting the efficacy of ondansetron in the prevention of PAS during cesarean section under neuraxial anesthesia from 11 trials with 748 participants. All trials are placebo-controlled randomized controlled trials, and the combined results using a fixed-effect model showed that ondansetron reduced the incidence of cesarean section shivering under neuraxial anesthesia compared with placebo. Meanwhile it has an advantage in reducing the incidence of hypotension. Neuraxial anesthesia in parturient for cesarean section is frequently associated with hypotension and the incidence varies from 70 to 80% without any prevention. Prolonged episodes of hypotension lead to organ ischemia, loss of consciousness, cardiovascular collapse, uteroplacental hypoperfusion, and fetal acidosis [29]. So it should be recognized that ondansetron as an appropriate agent is effective in reducing the incidence of post-anesthesia hypotension during cesarean section. In a previous meta-analysis in 2018 [30], the researcher found that ondansetron had no significant effect on PAS prevention during cesarean section under neuraxial anesthesia, which was inconsistent with our results. The previous meta-analysis only included three studies published before 2014 that assessed the efficiency and safety of ondansetron during cesarean section. Our analysis included more studies than did the previous meta-analysis, especially included many studies that published after 2014. Therefore, this inconsistent result may be due to the limited number of studies that included in their study. Moreover, the previous meta-analysis was designed to assessed the incidence of nausea/vomiting as the primary outcome and the incidence of PAS as a secondary outcome. Our analysis was more concerned with the efficiency and safety of ondansetron in preventing PAS during cesarean section. Also, the other reason was that previous study had different inclusion criteria. In addition, we provide more conservative estimates by further applying TSA. The results of the TSA suggested that our meta-analysis establishes sufficient and conclusive evidence.
Although our study included relatively high-quality randomized controlled studies and a larger sample size, similar to other systematic reviews and meta-analyses, in our study there are substantial statistical and clinical heterogeneity. All participants in included trials were from three continents (Africa, Asia, and Oceania), so the limitation of regional populations in included trials may cause sources of

Inclusion
Our results found that intravenous ondansetron was effective in preventing shivering during cesarean section under neuraxial anesthesia, and had an advantage in reducing the incidence of hypotension compared with intravenous placebo.
Funding This research was carried out without funding.

Conflict of interest
The authors declare that they have no conflict of interest.
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.