Comparison of analgesic interventions for traumatic rib fractures: a systematic review and meta-analysis

Purpose Many studies report on outcomes of analgesic therapy for (suspected) traumatic rib fractures. However, the literature is inconclusive and diverse regarding the management of pain and its effect on pain relief and associated complications. This systematic review and meta-analysis summarizes and compares reduction of pain for the different treatment modalities and as secondary outcome mortality during hospitalization, length of mechanical ventilation, length of hospital stay, length of intensive care unit stay (ICU) and complications such as respiratory, cardiovascular, and/or analgesia-related complications, for four different types of analgesic therapy: epidural analgesia, intravenous analgesia, paravertebral blocks and intercostal blocks. Methods PubMed, EMBASE and CENTRAL databases were searched to identify comparative studies investigating epidural, intravenous, paravertebral and intercostal interventions for traumatic rib fractures, without restriction for study type. The search strategy included keywords and MeSH or Emtree terms relating blunt chest trauma (including rib fractures), analgesic interventions, pain management and complications. Results A total of 19 papers met our inclusion criteria and were finally included in this systematic review. Significant differences were found in favor of epidural analgesia for the reduction of pain. No significant differences were observed between epidural analgesia, intravenous analgesia, paravertebral blocks and intercostal blocks, for the secondary outcomes. Conclusions Results of this study show that epidural analgesia provides better pain relief than the other modalities. No differences were observed for secondary endpoints like length of ICU stay, length of mechanical ventilation or pulmonary complications. However, the quality of the available evidence is low, and therefore, preclude strong recommendations.


Introduction
Traumatic rib fractures are a common injury among the trauma population and can cause severe pain in both isolated rib fractures and fractures which are a part of more extensive chest injuries [1,2]. Rib fractures are clinically important. Even isolated fractures are associated with significant consequences, such as prolonged pain and disabilities [3]. Rib fractures sustained following blunt chest trauma are a surrogate for significant trauma, particularly in more vulnerable patients [1,4,5]. The number of rib fractures is indicative of the trauma severity. More than 90% of the patients with multiple rib fractures have associated injuries, most commonly involving head, abdomen and/or extremities [1]. An increased number of fractures, older age, and polytrauma patients with rib fractures are associated with increased rates of morbidity and mortality [1,4,5].
The thoracic pain caused by rib fractures or chest contusion limits patients to cough and breathe deeply, which can result in atelectasis and pneumonia. Besides most of these, patients also suffer from a pulmonary contusion, due to their injury. This can lead to an acute respiratory distress syndrome and/or respiratory failure and the need for mechanical ventilation has been reported [6,7].
A combination of adequate pain control, respiratory assistance, and physiotherapy are considered to be the key in the management of patients with fractured ribs [4,8]. In the current practice, different analgesic modalities including epidural catheters, intravenous (patient controlled) narcotics, intercostal, paravertebral or interpleural blocks, oral opioids, or a combination of the aforementioned interventions, are used as therapy [9,10].
The literature on the use of the different analgesic interventions is inconclusive. A clinical guideline supported by the Eastern Association for the Surgery of Trauma recommends epidural analgesia or a multimodal approach over opioids alone in patients with blunt chest trauma [9]. On the other hand, two recently performed systematic reviews and meta-analyses of Duch et al. [10] and Carrier et al. [11] stated that the evidence for the use of epidural analgesia as preferred modality is insufficient, and that there is no firm evidence for benefit or harm of the epidural modality compared to the other interventions.
However, to date, no comprehensive study compared the single modalities independently with each other, including both observational studies and randomized controlled trials. Therefore, the aim of this systematic review and metaanalysis is to compare epidural, intravenous, paravertebral and intercostal analgesia for the primary outcome of pain reduction and the secondary outcomes of mortality during hospitalization, length of mechanical ventilation, length of hospital stay, length of intensive care unit stay (ICU) and complications, in patients with traumatic rib fractures.

Methods
A published protocol for this review does not exist. No ethical committee approval was necessary for this literature review.

Literature search and eligibility criteria
This systematic review and meta-analysis was written in accordance to the PRISMA guidelines for reporting systematic reviews and meta-analyses [12]. Two reviewers (JP, DS) independently performed a structured literature search, on September 16th 2017, to identify comparative studies investigating epidural, intravenous, paravertebral and intercostal interventions for blunt chest trauma with traumatic rib fractures. Three different electronic databases (PubMed, EMBASE and CENTRAL) were used to perform a systematic search. The search strategy included keywords and MeSH or Emtree terms relating to traumatic rib fractures, analgesic interventions, pain management and complications. The full search syntax is provided in Appendix Table 2. The search was not restricted by date or any other limits.
After screening of all titles and abstracts of the identified studied, full texts were obtained of the remaining relevant studies. Two reviewers (JP, DS) read the full-text articles, removed duplicates and made a final selection of relevant studies. Reference lists of retrieved articles were checked and citation tracking was performed using Web of Science, to identify articles not found in the original search. Figure 1 shows a flowchart of the search strategy.
Manuscripts were eligible for inclusion if published in English, French or Dutch language and available in full-text. Studies describing mixed cohorts of patients with blunt chest trauma, including traumatic rib fractures, were also eligible for inclusion. Animal studies, abstracts for conferences, studies including patients below 16 years of age, case reports and studies with less than five patients were excluded. There were no further restrictions for inclusion.
Authors were approached if additional information was needed or if full-text was not available.

Quality assessment
The methodological quality of the articles was independently assessed by two reviewers (JP, DS) using the validated methodological index for non-randomized studies (MINORS) score [13]. Additional criteria, described in Appendix Table 3, were defined to make further distinction in quality between the included studies. The quality was determined by means of the total MINORS score. Studies were not excluded based on the quality assessment. Disagreement was resolved by discussion with a third independent reviewer (MJ), followed by consensus.

Data extraction
Data were retrieved by two independent reviewers (JP, DS). Data extracted included first author, year of publication, country, study design, setting and treatment groups. For each treatment group, age, sex, type of analgesia and injury severity score (ISS) were extracted. The extracted data were shown as mentioned in the original studies. If exact pain scores were not given, an estimation of the scores was made on the basis of the figures. Outcomes were retrieved including confidence intervals (CI's) and/or p values.

Outcome measures
The predefined primary outcome was the reduction of pain, preferably expressed in a Numeric Rating Scale (NRS). Secondary outcomes were mortality during hospitalization, length of mechanical ventilation, length of hospital stay, length of intensive care unit stay (ICU) and complications.

Data analysis
Data were pooled according to the analgesic modalities that were compared. Meta-analyses were performed if the endpoints were reported by two or more studies. If the extracted data were initially noted as median with an interquartile range, the mean and standard deviation (SD) were estimated as follows: the reported median value was used as mean value, and the standard deviation was estimated by dividing the interquartile range with 1.35. Statistical heterogeneity was assessed by visual inspection of the forest plots and estimated by means of the I 2 , Tau 2 and Cochran's Q (Chi-square test). A random-effects model was used if high heterogeneity was present (where I 2 > 75% reflects a high heterogeneity). Odds ratios and 95% confidence intervals (95% CI) were calculated for dichotomous variables. Studies that reported zero events in one or both arms were included by adding a continuity correction of 1.0 to all cells in the 2 × 2 table of that study [14]. p values < 0.05 were considered statistically significant.
After the primary statistical analyses, sensitivity and subgroup analyses were conducted. In the sensitivity analyses on study design, only RCTs were included. In the sensitivity analyses on time, only studies published after the year 2000 were included. In the sensitivity analyses on quality, arbitrarily all studies with more than 16 points were included [15]. A sensitivity analyses on outlier studies was conducted. For the subgroup analyses on etiology, only studies describing cohorts with solely traumatic rib fractures were included. Studies describing mixed cohorts of patients with blunt chest trauma were excluded.
All statistical analyses were performed using Review Manager (RevMan, Version 5.3.5 Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014).

Quality assessment
The total MINORS score of the included articles are listed in Appendix Table 3. On average the included articles scored 15.7 ± 2.9 points, with a range of 11-23 points.

Epidural analgesia versus intravenous analgesia
The results of the studies comparing epidural with intravenous analgesia are summarized in Appendix Table 5. Metaanalyses are shown in Fig. 2. Of the 11 included studies, 4 studies [16,20,21,28] examined pain scores on different intervals after treatment with epidural or intravenous analgesia. One study [16] described lower pain scores at all intervals of the study period in the group that received epidural analgesia (p < 0.05). Significant lower pain scores on coughing were found in the first 24 h in the epidural group (p < 0.05). One study [20] found significantly lower pain scores at all intervals (p < 0.05), except on the baseline interval (p = 0.82), in the group that received epidural analgesia. One [28] study found significant differences (p < 0.05) in pain relief on day 1 and on day 3 in favor of the patients that received epidural analgesia, no differences were found on day two. One study [21] reported that the improvement in pain was more pronounced in the group that received epidural analgesia, but no significant difference was found between the two groups (p = 0.08). The results on pain relief are shown in Table 1.

Epidural analgesia versus intercostal block
The results of the studies comparing epidural analgesia with intercostal blocks are summarized in Appendix Table 6. Meta-analyses are shown in Appendix Fig. 3. As a consequence of insufficient data and variability of outcome measurement, meta-analyses were only possible for the length of hospital and ICU stay.
Two studies [22,26] reported on pain scores. One study [26] described solely pain scores of the group that received intercostal blocks. Placement of the intercostal catheter resulted in significant improvement in pain severity (p < 0.05). No comparison was made with the historical control group that received epidural analgesia. According to one study [22], epidural analgesia provides better control of pain than the intercostal modality. The mean VAS scores that were observed during hospitalization were 2.2 ± 0.74 at rest and 3.05 ± 0.88 with cough in the epidural group, respectively 3.3 ± 1.01 and 4.95 ± 0.99 in the intercostal group.
Three studies [22,25,26] reported on the length of hospital stay. The average number of days of hospitalization was 7.1 ± 2.3 with epidural analgesia and 6.0 ± 2.7 with intercostal blocks. One study [26] was not included for pooled analysis because the standard deviations were not reported. Pooled analysis of the two remaining studies showed no significant differences [95% CI, MD − 0.13 Two studies [22,25] reported on the length of ICU stay, pooled analysis showed no significant differences [95% CI, MD − 0.37 (− 0.93, 0.19), I 2 = 0%, p = 0.20].

Epidural analgesia versus paravertebral block
The results of the studies comparing epidural analgesia with paravertebral blocks are summarized in Appendix Table 7. Meta-analyses are shown in Appendix Fig. 4. Two studies reported on pain scores. One study [6] found no significant intergroup difference in mean pain scores either at rest (p = 0.426) or on coughing (p = 0.721) on different intervals, and one study [30] described that there was no difference between both groups in the mean change of pain during hospital admission (Table 1).
Three studies [6,30,31] reported on the length of hospital and ICU stay. The average number of days of hospitalization was 8.3 ± 1.7 with epidural analgesia and 8.6 ± 2.6 with paravertebral blocks, respectively, 4.5 ± 2.1 and 4.6 ± 1.9 for the length of ICU stay. Pooled analysis showed no significant differences for the length of hospital stay [95% CI, MD 0.

Intercostal block versus intravenous analgesia
One study [23] compared intravenous analgesia with intercostal blocks. The average number of hospital days and the VAS pain scores were reported, and are summarized in Appendix Table 8, respectively, Table 1. Significant differences in pain relief were described on different intervals, in favor of the intercostal blocks.

Paravertebral block versus intravenous analgesia
One study [32] compared paravertebral blocks with intravenous analgesia. The mortality and the VAS pain scores were reported, and are summarized in Appendix Table 9, respectively Table 1. Significant differences in pain relief were described on different intervals, in favor of the paravertebral blocks.

Sensitivity and subgroup analyses
The sensitivity and subgroup analyses are shown in Appendix Table 10. The results remained non-significant for all secondary outcomes in the group comparing epidural analgesia with intravenous analgesia and in the group comparing epidural analgesia with paravertebral blocks.

Discussion
This systematic review and meta-analysis of both RCTs and cohort series focused on the analgesic therapy for patients with traumatic rib fractures. Results of this study show that overall epidural analgesia provides better pain relief than the other modalities. In three studies [16,20,28] significant differences (p < 0.05) were found in the improvement of pain in favor of epidural analgesia when compared with intravenous analgesia. In one study [21], the reduction of pain appeared to be more definite in the group that received epidural analgesia. With respect to the secondary outcomes, our systematic review and meta-analysis failed to show significant differences between the analgesic modalities. Most of these outcome parameters are multifactorial and heterogeneously determined. Therefore, the relationship between the intervention and the secondary outcome parameters is influenced by multiple underlying factors, other than the type of analgesia. To alleviate the influence of these factors, heterogeneity corrections and sensitivity analyses were conducted. As a result, the trends that were initially observed in the group comparing epidural analgesia with intravenous analgesia for length of ICU stay (p = 0.11) and length of mechanical ventilation (p = 0.14), were not consistent after excluding outlier studies [24].
A recent systematic review and meta-analysis on this subject by Duch et al. [10], found a significant increased intervention effect for the reduction of pain, in favor of epidural analgesia, when compared with the paravertebral or intercostal modality. Because these results were based on only two studies and no significant differences were found on the other outcomes, they concluded that there was no firm evidence to assume that epidural analgesia has advantages over the other modalities. Likewise, a systematic review of 2008 from Carrier et al. [11], reported that there was no improvement in mortality, length of hospital and ICU stay, or duration of mechanical ventilation, if epidural analgesia was compared with other analgesic interventions. Our results differ from theirs in several aspects. Most importantly, our study showed that there is evidence that epidural analgesia results in better pain relief than the other modalities. The results of our secondary outcomes are in accordance with the aforementioned reviews, and seem to rely on a multifactorial basis. In contrast to the studies of Duch et al. [10] and Carrier et al. [11], we included observational studies. Therefore, we were able to include several (new) studies [16-20, 23, 25-27, 29-32] resulting in a larger patient database.
The current guideline of the Eastern Association for the Surgery of Trauma (EAST) recommend epidural analgesia or a multimodal approach over opioids alone, for pain relief in patients with blunt chest trauma [9]. In comparison with this guideline of the EAST, our study differs in certain respects. First, a major distinction is that in our study, the results of the single modalities were separately compared with each other. In the guideline of the EAST, the single modalities were compared with the merged results of larger groups. The epidural, paravertebral and intercostal modalities were in particular compared with the results of patients receiving "non regional'' analgesia, and the interpleural modality was compared with "other regional modalities''. Analysis to demonstrate the differences between the single modalities were not implemented. Second, four studies [4,47,49,50] using mixed cohorts of patients, in which the analgesic interventions used in the control group were not extractable, were also excluded in our study. Third, we were able to include six new studies [16,17,27,[30][31][32].
A potential advantage of our method is that by comparing the single analgesic interventions, subtle differences might be more accurately ascertainable. Besides, because the studies were compared separately, our method and results might approach closer to reality.
Another strength of this systematic review is that a considerable amount of extra studies was included due to inclusion of observational studies. In addition, as stated in recently published systematic reviews [15,51,52], the inclusion of both RCTs and observational studies might lead to more study power. If observational studies are of sufficient quality, the results will correspond with those of an RCT [15,51,52]. Furthermore, it appears to give a better reflection of common clinical practice, which might improve the generalizability and applicability of the outcomes of a systematic review [51,52].
On the other hand, the included studies were of low methodological quality, as assessed using the MINORS score. Therefore, the overall quality and applicability of the available evidence is low, and there is potentially a high risk of bias. Besides, merely a small amount of studies investigated the management of pain. Of the studies reporting on pain, patient samples were overall small, outcome measurements varied and exact pain scores were often not or poorly reported. Pooled analyses for pain in patients with traumatic rib fractures were not feasible due to inadequate reported data. Conversion of pain scores to one comprehensive score was not performed due to increase of bias. Furthermore, the studies were overall difficult to compare because of the heterogeneity in the study method and investigated endpoints. Analgesia-related complications such as nausea, vomiting, catheter inflammation, hypotension, respiratory depression, itching and rash, were also not frequently reported. However, pulmonary complications, which are considered to be important complications in patients with traumatic rib fractures, where in general adequately reported and could be properly investigated. As described in the results, there were no significant differences in the occurrence of pulmonary complications between the three analgesic therapies.
Pooled analyses between epidural and paravertebral was for a greater part determined by the large sample size of Malekpour et al. [31]. As we could only include three studies in these analyses, this might have influenced the outcome.
The value of the different analgesic modalities in critical care patients is insufficiently described. Only one of our included studies compared epidural analgesia with parenteral analgesia in mechanically ventilated ICU patients with flail chest [17]. This RCT described a significant difference in the length of ICU stay, the duration of mechanical ventilation and the change in tidal volume in the first 24 h of ICU admission, in favor of epidural analgesia.
The type of medication is not reflected in our analysis. The different modalities were compared, as described in the baseline characteristics (Appendix Table 4). However, it could be relevant if only opioids were administered, or if local anesthetics were also applied. Furthermore, there was insufficient information about any additional pain medication and whether escape medication was prescribed.
Although there seemed to be significant differences between the different analgesic therapies, further research on the analgesic therapy for traumatic rib fractures is desirable to extend our knowledge of the reduction of pain. Many different pain assessment tools are used in the current practice. The NRS pain score at breathing/coughing seems to be the most reliable outcome parameter, since it reflects the influence of pain on function of the ribcage. To compare the results of pain reduction more homogeneously, future studies should use a universal pain assessment tool. Second, besides pain measurement, there should also be data available on the use of other multimodal treatments started, the daily total opioid consumption and efficacy of the interventional analgesic therapy. On account of the increasing contraindications and the high probability of failure of the epidurals, research into safe and effective pain management by other analgesic methods must be continued.
Another future perspective is to determine the contribution of surgical rib fixation for the primary and secondary outcomes as described in this systematic review.

Conclusion
Results of this study show that epidural analgesia provides better pain relief than the other modalities. No differences were observed for secondary endpoints like length of ICU stay, length of mechanical ventilation or pulmonary complications. However, the quality of the available evidence is low, and therefore, preclude strong recommendations.
Funding The authors received no financial support for the research, authorship, and/or publication of this article.

Conflict of interest
The authors declared no conflicts of interest with respect to the research, authorship, and/or publication of this article.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.