Search result
A total of 362 studies were identified and imported into NoteExpress from Pubmed, Cochrane, and Embase database. And there were not other records when we searched the references manually. As showed in Fig. 1, After glancing over the title and abstracts, a total of 294 studies were removed out. Then we evaluated the remaining 68 original articles by reading the full texts, 1 study was written by other language to describe the contents of article, 18 were not performed surgeries in spine, and 28 were applied with dexamethsone by local irrigation and epidural injection. 10 studies were in accord with the criteria. Two studies were excluded because of the loss of the related data. Finally, 8 studies were selected in our meta-analysis.
Characteristics of included studies
The basic characteristics of all articles were showed in Table 1. The 8 studies published ranging from 2003 to 2019 totally involved 958 cases, 431 cases were in the dexamethasone group and 527 cases in the control group[11, 14-20]. Studies were performed in different countries, including Iran, Thailand, Canada, Germany, Denmark, France, Korea. There were seven studies providing the follow-up period ranging from 2 months to 1 year. Basing on the VAS or NRS scores, all studies evaluated the pain at rest or on movement. Dexamethasone groups received intravenous injection only perioperatively in all studies, and the does of dexamethasone ranged from 8 to 80mg. All studies evaluated the usage of opioids consumption after the spine surgery. And all the studies provided the surgical complications and six studies reported the adverse events, such as sedation, headache, dizziness and sleep disturbance[11, 16, 17, 19, 20]. Several studies evaluated the spine function after the surgery. And three studies evaluated the efficiency of intravenous dexamethasone together with oral pregabalin or ibuprofen[14, 16, 19]. All outcomes of studies were showed in Table 2.
Risk of bias assessment
Cochrane Collaboration tool was used to assess the all RCTs[11, 14-17, 19, 20](Fig 2). All studies included the explicit inclusion and exclusion criteria and the randomization methodology was generated by a computer program. All studies meet the criteria of allocation concealment. And all studies were achieved by double blinding or triple blinding. Six studies provided the complete outcome[11, 14, 15, 17, 19, 20]. A percentage was used to present the risk of each bias item across all included studies, which represented the different levels of risk of bias for each item (Fig 3).
Primary outcomes
VAS scores at 24 hours
Four studies[11, 15, 17, 20] provided the outcomes of VAS scores at 24 hours of movement after the spine surgery. There was no significant heterogeneity of the studies (x2 = 4.56, df = 3 , I2 = 34.2%, P = 0.207); Thus, fixed-effect model was applied. Pooled results suggested that VAS scores at 24 hours of movement in control groups was significantly higher than in dexamethasone groups (SMD = -0.44, 95% CI: -0.67 to -0.21, P < 0.001). And four studies[11, 15, 17, 20] provided the outcomes of VAS scores at 24 hours of rest after the spine surgery. There was no significant heterogeneity of the studies (x2 = 1.21, df = 3, I2 = 0.0%, P = 0.750); Thus, fixed-effect model was applied. Pooled results suggested that there was no significant difference between the two groups (SMD = -0.21, 95% CI: -0.44 to 0.013, P = 0.064) (Fig 4).
VAS scores at 48 hours
Two studies[15, 20] reported the outcomes of VAS scores at 48 hours of movement after the spine surgery. There was no significant heterogeneity of the studies(x2 = 2.3, df = 1, I2 = 56.6%, P = 0.129); Thus, fixed-effect model was applied. Pooled results suggested that there was no significant difference between the two groups (SMD = -0.073, 95% CI: -0.444 to 0.30, P = 0.298). And two studies[15, 20] reported the outcomes of VAS scores at 48 hours of rest after the spine surgery. There was no significant heterogeneity of the studies (x2 = 0.15, df = 1, I2 = 0.0%, P = 0.703); Thus, fixed-effect model was applied. Pooled results suggested that there was no significant difference between the two groups (SMD = 0.112, 95% CI: -0.258 to 0.481, P = 0.553)(Fig 5).
The occurrence of postoperative nausea and vomiting
Six studies[11, 14, 15, 17, 19, 20]provided the occurrence of PONV after the spine surgery. There was no significant heterogeneity of the studies (x2 = 14.73, df = 5, I2 = 66%, P = 0.012); Thus, fixed-effect model was applied. Pooled results suggested that occurrence of PONV in control groups was significantly higher than in dexamethasone groups (RR = 0.52, 95% CI: 0.40 to 0.67, P < 0.001)(Fig 6).
The occurrence of sedation
Five studies[14, 15, 19-21] provided the occurrence of sedation. There was no significant heterogeneity of the studies (x2 = 4.34, df = 4, I2 = 7.9%, P = 0.362); Thus, fixed-effect model was applied. Pooled results suggested thatthere was no significant difference between the two groups (RR = 0.86, 95% CI: 0.58 to 1.26, P = 0.435) (Fig 7).
Opioids consumption within 24 hours and the length of hospital stay
Three studies[11, 15, 20]provided the usage of opioids consumption within 24 hours after the spine surgery. There was no significant heterogeneity of the studies (x2 = 5.43, df = 2, I2 = 63.2%, P = 0.066); Thus, fixed-effect model was applied. Pooled results suggested that the usage of opioids consumption within 24 hours in control groups was significantly higher than in dexamethasone groups (SMD = -0.601, 95% CI: -0.93 to -0.28, P < 0.001). Two studies[17, 21] provided the outcomes of the length of hospital stay. There was no significant heterogeneity of the studies (x2 = 1.05, df = 1, I2 = 5.0%, P = 0.305); Thus, fixed-effect model was applied. Pooled results suggested that the length of hospital stay in control groups was significantly longer than in dexamethasone groups (SMD = -0.207, 95% CI: -0.396 to -0.017, P = 0.003)(Fig 8).