Article Text
Abstract
Evidence has shown that mild therapeutic hypothermia (MTH) could improve survival and neurological outcome in patients following cardiac arrest. But this therapy may cause some adverse effects. The authors sought to take a systematic approach to describe the safety aspects and outcome of MTH following cardiac arrest to help clinical practice. MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials, BIOSIS Previews and clinicaltrials.gov were searched up to June 2011. Bibliographies of relevant studies were also reviewed. Comparative studies reporting the mortality or any other studies reporting any kind of adverse events in patients undergoing MTH after cardiac arrest and published in English were included. Of 1742 abstracts, 63 studies were included. Most adverse events potentially associated with therapeutic hypothermia were not significantly different between the hypothermia therapy and the normothermia groups. No significant difference was found in the inhospital mortality, bleeding, pneumonia and bradycardia events between surface and endovascular-cooled groups in this study. Cooling device-related adverse events were generally mild. Serious adverse events potentially attributable to therapeutic hypothermia were seldom reported. MTH was associated with reduced inhospital mortality, mortality at 1 month and at 6 months. Evidence about the safety of MTH in children has been limited. These results suggest that while it may result in some adverse events, MTH is generally safe in patients following cardiac arrest and could improve the short-term and long-term survival of comatose patients after cardiac arrest. But awareness of these adverse events should be kept in mind in clinical practice.
- Mild therapeutic hypothermia
- adverse event
- cardiac arrest
- mortality
- systematic review
- epidemiology
- intensive care
Statistics from Altmetric.com
- Mild therapeutic hypothermia
- adverse event
- cardiac arrest
- mortality
- systematic review
- epidemiology
- intensive care
Introduction
Millions of people around the world die from sudden cardiac arrest every year, often related to coronary heart disease. The global incidence of out-of-hospital cardiac arrest is about 82.9 per 100 000 population within all age groups, and 213.1 per 100 000 population in adult groups.1 Patients who have survived cardiac arrest often have poor neurological outcome.
Two randomised controlled trials (RCTs)2 ,3 and a meta-analysis4 have shown the beneficial effects of a 12–24 h period of mild therapeutic hypothermia (MTH) (32–34°C) in improving survival and neurological outcome of patients following cardiac arrest. The 2010 guidelines from the International Liaison Committee on Resuscitation suggest that therapeutic hypothermia (32–34°C) should be considered for comatose adult patients with spontaneous circulation after cardiac arrest5 and may be beneficial for comatose children following resuscitation after cardiac arrest.6
However, while it could benefit the comatose survivors of cardiac arrest, therapeutic hypothermia also interferes with numerous physiological and pathological processes, and might have unfavourable effects in patients receiving it and potentially put them at risk for adverse events, such as infection, cardiac dysrhythmia and coagulopathy.3 ,7 If not properly managed, these adverse effects could become serious and may negate part or even all of the benefits of therapeutic hypothermia in the patients. Knowledge about the physiological consequences and potential adverse events of therapeutic hypothermia is absolutely necessary for clinicians to carry out high-quality intensive care treatment.
Therefore, we undertook this review of literature to comprehensively investigate the safety aspects and short-term and long-term outcomes of MTH and perform meta-analysis when available.
Methods
Data sources
We searched MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials, BIOSIS Previews and clinicaltrials.gov from the earliest available date to June 2011, using combinations of medical subject headings including induced hypothermia, heart arrest, safety, adverse, complication, death, mortality and other related free-text terms. The electronic search was limited to studies on human subjects and publications in English. We also manually searched the reference lists of relevant review articles and original studies for additional papers not captured by the electronic search and contacted experts in this field.
Study selection
Two reviewers independently screened the identified studies for eligibility, with discrepancies resolved by consensus. Studies meeting the following criteria were included: (1) studies that reported original data about the outcome and safety aspects (such as mortality, complications and any adverse or side effects) of MTH used in comatose patients of cardiac arrest of presumed cardiac origin; (2) studies only reporting the mortality outcomes should be comparative studies (randomised or non-randomised) with one group treated with MTH and the other without; and (3) studies not only reporting the mortality outcomes or just reporting other complications or adverse events potentially related to MTH would be included irrespective of study design. Studies not reporting results of interest were excluded.
Data extraction and quality assessment
Data of included studies were extracted in parallel by two independent reviewers using a standardised form. After this primary extraction, the results from two reviewers were compared and differences were reconciled by a third, independent reviewer. Apart from demographical data of each study, end points extracted consisted of mortality and any kinds of complications and adverse events potentially related to MTH, as defined in each study. All study designs were included in the review in order to find out as many adverse events as possible which might be resulted from the hypothermia therapy. The quality of eligible studies was evaluated in duplicate by two reviewers, using Jadad scale8 for RCTs and Newcastle-Ottawa scale9 for cohort studies. The Newcastle-Ottawa scale was also used to assess the quality of case series.10 Case reports were not evaluated in our study.
Data analysis
Meta-analysis was performed using Review Manager Version 5.1.11 Where possible, dichotomous data were pooled and combined into a summary RR with 95% CI using Mantel–Haenszel fixed-effect model. Heterogeneity among studies was measured using the Cochran's Q statistic and Higgins' I2 statistic; significant heterogeneity was prespecified as p<0.10 or I2>50%.12 Heterogeneity was explored through sensitivity analysis. Studies without control group were not included in the summary RR. We employed a single group analysis to calculate a pooled summary incidence of each end point using the method proposed by Einarson13 when available. The single group analysis, performed by Microsoft Excel 2003 (Microsoft Corp., Redmond, Washington, USA), included all studies involved in the meta-analytical procedure as well as cohorts and case series without control group.
Results
The flow diagram of studies identification through the review is shown in figure 1. Overall, we identified 1742 abstracts from the initial electronic search. Review of abstracts yielded 161 studies requested for detailed evaluation for eligibility. Three studies were retrieved from a review of references of relevant articles. Eventually 63 studies were included in this review.
The main characteristics of included studies with comparator group are described in table 1, and table 2 provides a summary of study characteristics of case series and case reports. Among the included studies, there are seven RCTs, 26 observational comparative studies and 30 non-comparative studies. Only three studies were performed in children. The majority of RCTs and observational comparative studies are of good quality.
Many studies excluded patients with serious infections, major bleeding, trauma, haemodynamic instability or in pregnancy. And the majority of patients involved in the included studies had received medication-induced sedation and paralysis as well as other agents such as vasopressor and insulin as necessary.
Numerous cooling methods were used in these included studies, including surface and invasive cooling. Surface cooling is non-invasive and ranges from simple ice packs and cooling mattress to advanced cooling machines with automatic feedback system. Nasopharyngeal evaporative cooling has been considered as a kind of surface cooling. Invasive cooling methods include the administration of ice-cold fluids intravenously, intravascular cooling catheters, body cavity lavage, extracorporeal circuits and so on. In most studies the target temperature of cooling was 32–34°C, with a duration of 12–24 h.
Complications and adverse events
The pooled results of main complications and adverse events potentially related to MTH reported in the included comparative studies have been presented in figure 2A,B. Most complications and adverse events were not significantly different between the hypothermia and the normothermia groups, except for arrhythmia and hypokalaemia. A significant heterogeneity was found in the analysis of arrhythmia (p=0.007, I2=75%). A sensitivity analysis revealed that the study of Holzer et al 31 demonstrated to be the most responsible for the high heterogeneity, as without this study I2=0% and pooled RR was 1.0 (95% CI, 0.78 to 1.30, p=0.97). As to the hypokalaemia complications, only one study19 involving 53 patients contributed to the pooled result. Some other studies2 ,17 ,29 found no clinically important hypokalaemia in the hypothermia groups. There is a trend of a higher rate of pneumonia in the hypothermia groups than in the normothermia groups (RR, 1.18, 95% CI, 0.99 to 1.40, p=0.06; I2=30%).
Seizure was considered as an end point in three comparative studies,3 ,17 ,19 and we did not pool this result for only a few studies reporting this end point and significant heterogeneity among these studies as well as most patients having received medication-induced sedation and paralysis which could influence seizure development. Seizures that occurred in the hypothermia group were not more frequent than that in the normothermia group in these studies.
Three studies2 ,17 ,29 compared the haemodynamic and biochemical values between MTH and normothermia groups, and they found no clinically significant differences between the two groups.
Some studies reported that the body temperature of a few patients receiving MTH reached below 32°C, indicating overcooling.26 ,28 ,36 ,40 ,57 ,59 ,63 ,71
Three studies16 ,27 ,44 reported data about vasopressors used in the hypothermia and normothermia groups, which resulted in a pooled RR of 1.23 (95% CI, 1.09 to 1.40, p=0.001; I2=16%), as shown in figure 3. Another study35 reported significantly higher proportion of patients receiving norepinephrine in the hypothermia group than in the normothermia group.
In the single group analyses, the pooled incidence of adverse events reported in the studies included in this review is summarised in table 3.
Surface versus endovascular cooling
Three studies provided data about adverse events between surface- and endovascular-cooled groups. Pooled results of inhospital mortality, bleeding and arrhythmia (bradycardia) events revealed no significant difference between the two groups, as shown in figure 4. Pneumonia was reported as an end point in two studies28 ,39 and we did not pool this result for significant heterogeneity between studies. Both studies showed no significant difference between the two groups. The study of Tomte and colleagues39 reported that there was a significantly higher rate of sustained hyperglycaemia in the surface-cooled group than in the core-cooled group, whereas hypomagnesaemia occurred significantly more frequently in the core-cooled group.
Device related adverse events
Two studies21 ,50 reported adverse events related to transnasal evaporative cooling devices, including nasal discolouration (13.0%, 23/177), epistaxis (2.8%, 5/177), periorbital emphysema (1.1%, 2/177), perioral bleeding (1.1%, 1/93), coolant in sinus (1.2%, 1/84) and tissue damage (1.2%, 1/84). Most adverse events resolved spontaneously without sequelae.
Five studies40 ,56 ,57 ,71 ,72 reported skin complications related to surface cooling pads, including minimal skin erythema, superficial frostbite, skin tears and ecchymosis. All the skin lesions resolved spontaneously or required only local wound care except for one case whose skin peeled in large pieces after the use of hydrogel pads and who died.
One study68 reported a case of traumatic false aneurysm formation secondary to accidental puncture of a femoral artery with the use of an endovascular cooling system.
Rare complications and adverse events
Brachial plexopathy was reported in a patient treated with MTH (33°C) for 30 h after cardiac arrest with an initial ventricular fibrillation (VF) rhythm.55 Rhabdomyolysis was reported in two male patients treated with MTH (34°C) after cardiac arrest.51 ,52
Mortality
Seventeen comparative studies involving a total of 8201 participants reported inhospital mortality, resulting in a pooled RR of 0.86 (95% CI, 0.83 to 0.89, p<0.00001) for hypothermia versus normothermia groups with significant heterogeneity (p<0.0001, I2=67%), as shown in figure 5. To explore the heterogeneity, a sensitivity analysis of study design using RCTs versus observational comparative studies revealed that all heterogeneity may be due to observational comparative studies, as within RCTs I2=0 and the pooled RR was 0.75 (95% CI, 0.62 to 0.92, p=0.005). Two studies provided data about mortality at 1 month between the hypothermia and the normothermia groups, with a pooled RR of 0.61 (95% CI, 0.45 to 0.81, p=0.0008; I2=0%); and four studies reported mortality at 6 months between these groups, with a pooled RR of 0.73 (95% CI, 0.61 to 0.88, p=0.0009; I2=0%), as shown in figure 5.
Eight comparative studies reported inhospital mortality in patients with initial VF rhythm, with a pooled RR of 0.79 (95% CI, 0.68 to 0.91, p=0.001; I2=2%), and five comparative studies in patients with initial non-VF rhythm, with a pooled RR of 0.89 (95% CI, 0.81 to 0.97, p=0.009; I2=31%), as shown in figure 6. One study42 reported no statistical significance of mortality outcome in patients with initial pulseless electrical activity or asystole (PEA/A) rhythm between the hypothermia and normothermia groups without a detailed outcome.
MTH in child population
There are two observational comparative studies in child population. One study22 reported that no statistically significant differences were found between the hypothermia and normothermia groups in hypothermia-related adverse events except for blood pressure being significantly lower in the hypothermia therapy group in patients 1–12 months of age (p=0.002) and patients 1–18 years of age (p=0.017). The 30-day mortality was higher in the hypothermia group than in the normothermia group but without statistical significance (58.6% vs 36.0%, p=0.054), and the 6-month mortality was significantly higher in the hypothermia group than in the normothermia group (69.0% vs 38.0%, p=0.009), whereas there is no significant difference after an adjusted analysis. Another study25 found that the hypothermia and normothermia groups had similar rates of haemorrhage, receipt of red blood cell transfusions, intermittent arrhythmias, infection and seizures, whereas the hypothermia group received more electrolytes supplementation and insulin infusions than the normothermia group, and mortality was similar between the two groups (55.0% vs 55.3%, p=1.0). One prospective case series70 that included 12 paediatric patients found no serious adverse events attributable to MTH, and documented complications included hyperglycaemia, hypoglycaemia, hypokalaemia, hypomagnesaemia, hypocalcaemia, hypophosphatemia, bradycardia, hypotension, leucopenia and thrombocytopaenia.
Discussion
There is substantial evidence that MTH could benefit the postcardiac arrest patients in survival as well as neurological outcome, especially those out-of-hospital cardiac arrests with an initial cardiac rhythm of VF. And MTH has been recommended as the standard of care for comatose patients after VF cardiac arrest by the International Liaison Committee on Resuscitation and the American Heart Association. It is so far the only therapeutic method that improves the neurological outcome after cardiac arrest in RCTs.
The main protective effects of MTH may result from reduction of brain metabolism and free radical production, inhibition of excitatory amino acid release, attenuation of the immune response during reperfusion, and inhibition of apoptosis.75 However, as it can disturb many physiological processes, MTH may result in unfavourable effects in the patients undergoing this therapy.
This review sought to collect data from all the available literature concerning the safety aspects of MTH used in comatose patients following cardiac arrest and to describe the safety profile and outcome of MTH from several aspects.
Overall, as presented in our review, most complications potentially related to MTH have no significant difference between the hypothermia and the normothermia groups, in agreement with a previous meta-analysis.4 Serious adverse events have not been observed to a significant extent in most studies so far. This study showed no significant difference in the inhospital mortality, bleeding, pneumonia and bradycardia events between the surface- and endovascular-cooled groups. Surface and endovascular (invasive) cooling methods are two major methods used to induce and maintain hypothermia. Compared with endovascular cooling, simple surface cooling seems to be associated with greater temperature fluctuations and more frequent overcooling, which may result in serious complications,32 although seldom reported. However, endovascular cooling and surface cooling with automatic feedback system could provide more stable temperature maintenance.36 ,71 Several cooling methods can be used in clinical practice. And further research is still required to identify the most effective and safe methods used to induce hypothermia.
Device related adverse events are not common, and most adverse events are trivial and self-limited. Transnasal cooling uses a relatively small area for heat exchange,50 which may be the cause of adverse events relating to this method described above. Skin lesions related to other surface cooling devices were also reported and serious cases were seldom reported. Adverse events related to the endovascular cooling method may include bleeding or haematoma, and as long as the cooling procedure is carried out properly, this kind of adverse event can seldom happen.
If appropriate management has been taken, the majority of complications potentially related to MTH could be avoided. As although hypothermia could possibly lead to adverse effects mentioned above, these complications did not differ significantly between the hypothermia and normothermia groups in this study. For example, increased usage of vasopressor could help patients receiving MTH keep haemodynamics stable, which could be influenced by hypothermia.76 It is shown that relevant supportive care is of great importance for MTH.
Our study has shown that MTH could reduce the inhospital mortality of postcardiac arrest patients, in accordance with another study,77 and especially mortality at 1 month and at 6 months in patients after cardiac arrest, suggesting that MTH could improve the short-term and long-term survival of postcardiac arrest patients. Our study showed marked reduction of mortality in patients with initial VF rhythm receiving MTH, comparable with a previous report.78 And MTH also reduced mortality of patients with an initial non-VF rhythm (PEA/A) in our meta-analysis.
Interestingly, our study has revealed that MTH might also be beneficial for comatose survivors of non-VF cardiac arrest. Patients with initial PEA/A rhythm are always associated with a poor prognosis,79 and studies of MTH within this population are controversial at present. This review indicates that MTH might also be used in patients with non-VF rhythm and may improve their outcome. However, most studies involved in the analysis of inhospital mortality of patients with non-VF cardiac arrest were observational studies. Further studies are needed to determine this effect.
The applications of MTH in children are conflicting,80 and there are few studies about MTH in the child population. Available data from our literature review indicated that MTH in children seemed to be not as effective as in adults to improve outcome of cardiac arrests. Meanwhile, no significant difference was found in hypothermia-related adverse effects between the hypothermia and normothermia groups. It appears that the usage of MTH in child population may be safe, but the efficacy of this therapy needs further research in this population.
Our study has several limitations. First, this study possesses all the inherent limitations associated with a systematic review and the methods of meta-analysis, which needs to be taken seriously. Second, as we included retrospective studies in our review, rates of adverse event could be underestimated, for some events may not be fully recorded in these studies, and some non-comparative studies in the analyses may be limited by the relatively small sample size, potential for patient selective bias and confounding. Third, as adverse events were not common in most studies, these studies may be underpowered to detect the differences in the event rates between the hypothermia and normothermia groups. On the other hand, this could also indicate a low risk of adverse events in the hypothermia therapy. Furthermore, the results of our review come from mixed-up analyses that did not separate each hypothermic temperature level between 32 and 34°C, for most studies included stated merely the body temperature level of hypothermia as 32–34°C. Further studies are needed to determine the optimal temperature level of hypothermia therapy. In addition, although serious adverse events have been seldom reported, the possibility of serious adverse events related to MTH cannot be ruled out, and clinicians still need to pay attention to the serious adverse events related to MTH. Finally, some studies included in our review were low-level, observational studies, which may dilute the conclusions of this review. Therefore, although there is a large literature base supporting the results of this review, these conclusions should be interpreted cautiously.
Conclusions
Although there are some complications potentially associated with hypothermia therapy, most of these complications are preventable or counteracted by appropriate supportive care in a timely manner, and so serious complications are rare, and thus MTH is generally a safe therapeutic method for comatose patients following cardiac arrest and could improve the short-term and long-term survival of comatose patients after cardiac arrest. Nevertheless, possible complications related to MTH should be kept in mind and monitored in the clinical practice in order that timely care could be taken to minimise these complications.
Further research is needed to determine the most effective and safe hypothermia therapeutic methods, the effect of MTH in patients with non-VF cardiac arrest, the optimal temperature level of hypothermia therapy and the effect of MTH in paediatric patients.
References
Footnotes
-
Funding This work was partly supported by an academic grant from Program for Changjiang Scholars and Innovative Research Team in University (registration number IRT0935).
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Primary survey