In this meta-analysis, 12 trials involving 2,991 patients analyzed the relationship between PRF and the neurological outcomes and/or mortality of patients. Results of this meta-analysis suggest that the patients who experienced PRF after TTM had similar neurological outcomes compared with the normothermia patients regardless of the definition body temperature of PRF. Meanwhile, results suggest that PRF with a broader definition (body temperature > 37.8°C or greater) after TTM have no significant association with mortality. However, the patients who experienced PRF defined as a higher body temperature (> 38.5℃) possessed a significantly higher mortality than the normothermia.
A concise meta-analysis conducted by Makker et al. [33] had found that PRF was associated with a significantly worse neurological outcome, which findings were not consistent with the present authors’ study. Possible reasons for the above discrepancy results may include the following: (1) all the studies included in the previous meta-analysis were outdated. Some studies have emerged in recent years suggesting that PRF may not be associated with unfavorable neurological outcomes. Several studies even suggested that PRF was associated with favorable neurological outcomes.[17, 21] The inclusion of new studies influenced the results of previous meta-analysis; (2) according to the meta-analysis operation requirements, the previous meta-analysis was a less rigorous meta-analysis, which only extracted the OR value of the studies but did not concern the specific details and data of studies.[31]
The present meta-analysis reveals that PRF with a broader definition (body temperature > 37.5°C, > 37.8℃ or greater) is not significantly associated with neurological outcome. (Fig. 2, Fig. 3) It is notable that in the analysis, when PRF was defined as a higher body temperature (> 38.5℃), four included studies showed the same result. (Fig. 4) However, sensitivity analysis suggest that the study conducted by Nobile et al. [20] was worthy of attention. The study obviously influenced the result and heterogeneity of the analysis. (Supplementary Table 3). The possible reason is that study conducted by Nobile et al. [20] involved CA patients who stayed in ICU, while the other three studies [18, 24, 25] focused on OHCA patients.
In this meta-analysis, when PRF was defined as body temperature > 37.8℃ or > 38℃, PRF was not significantly associated with mortality. (Fig. 5) Notably, the PRF was associated with higher mortality when PRF was defined as > 38.5℃. (Fig. 6) Therefore, after the TTM, the results suggest that active temperature control is necessary when the body temperature was > 38.5℃. However, one study indicated that the implementation of controlled normothermia to prevent PRF was not associated with favorable neurological outcome.[34] And in an in vitro study, the prevention of PRF obviously aggravated apoptosis of cells and release of inflammatory factors. [35] This study suggested that PRF was related to activation of inflammatory response and programmed cell death following the ischemia-reperfusion injury caused by CA. Since most works are focused on the clinical application of TTM, including optimal cooling temperature[5, 36, 37], practical methods of cooling for temperature control[6, 38, 39] and rate of rewarming following TTM[40, 41], it is necessary to investigate the PRF with different body temperature and body temperature control in the future.
The following are the advantages of this study: (1) this meta-analysis evaluated PRF after TTM in CA patients and assisted physicians to recognize the clinical effects of body temperature management and patient prognosis following ROSC. (2) the PRF group was divided into the > 38°C and > 38.5°C subgroups in order to provide a better understanding for physicians to make respective clinical strategies about different body temperature.
Nonetheless, the following are the limitations of this meta-analysis: (1) meta-analysis was a secondary analysis of original studies. High heterogeneity existed between different studies and was caused by many reasons, such as sample size, type of CA, TTM treatment strategies, etc. The difference between the original studies could have influenced statistical analysis. (2) all of the studies selected into this meta-analysis were observational studies. The data of observational studies were analyzed for certain purpose, which was considered a methodological flaw. Therefore, more RCTs were needed in the future to investigate the affect of PRF.