One of the missions of the International Liaison Committee on Resuscitation (ILCOR) has been to facilitate the scientific evaluation of the published resuscitation literature [1]. ILCOR was founded in 1992 from organisations from around the world that produced cardiac arrest guidelines. It originally published the universal cardiac arrest algorithms in 1997 [2], then oversaw the publication of a summary of the resuscitation science with international resuscitation guidelines in 2000 

Since then, ILCOR has produced consensus on science statements every 5 years, which included treatment recommendations where international consensus supported this ]. The evidence evaluation process for the resuscitation literature has had to deal with the spectrum of published evidence: from Randomised Controlled Trials (RCTs) and meta-analyses of RCTs to case series and studies using manikins or animal models. The desire to include data from such lower levels of evidence required the resuscitation community to develop their own levels of evidence, which could be used across the variety of studies that had been identified. These levels of evidence were refined further during the 2005, 2010, and 2015review periods. For the 2015 Consensus on Science, ILCOR engaged the skills of an information specialist team to design the search strategies for each of the individual questions asked. The literature identified by these searches was then evaluated according to the methodology developed by the Grading of Recommendations Assessment, Development and Evaluation (GRADE) working group. Using the GRADE process required a paradigm shift in those that were evaluating the literature: evidence was assessed for each of the key outcomes, and a strict process was followed for the evaluation of the quality of the supporting evidence.

The same overarching processes, including the use of the GRADE methodology, will continue to be followed as ILCOR moves forward from the 2015 consensus document. It was recognised that a number of questions that were prioritised to review required similar and overlapping search strategies, which in turn identified a large number of articles that were common to these strategies. For such inter-related questions ILCOR has planned to engage Knowledge Synthesis Units (KSUs) to assist in the evidence evaluation process. These KSUs are collections of individuals who have specific expertise in the identification, retrieval and critical appraisal of the published scientific literature.

In this issue of the journal, the first of these KSU commissioned systematic reviews has been published.

The authors of this review focused on studies that compared strategies for ventilation during the management of cardiac arrests. The inclusion and exclusion criteria, which are key to any systematic review and meta-analysis, are laid out in the article and its online supplement.

As a result of the broad initial search (which identified over 5000 articles), 28 studies with 13 companion reports were evaluated in more detail. As stated above, in keeping with ILCOR’s approach, the methodology that was used to ascertain the quality of evidence (certainty in the estimate of effect) was that developed by the GRADE working group.

In keeping with GRADE, information from RCTs and observational data were listed separately.

The evidence identified was easily collated into specific subsets. In all bar one study identified, the included patients were managed in the out-of-hospital setting. The situation surrounding the delivery of CPR was categorised as bystander CPR, dispatch-assisted CPR, Emergency Medical System (EMS) provider CPR, and in-hospital CPR.

A compression ventilation ratio of 30:2 was the key comparator. It was introduced in 2005 to increase the focus of the resuscitation community on minimising the interruptions to chest compressions, and recognising the deceased ventilation requirements during cardiac arrest. There is still very limited data comparing alternative techniques to the current recommendations (compression:ventilation ratio of 30:2).

We have data from over 10,000 patients in cohort studies to suggest that the use of 30:2 by EMS providers has resulted in improved survival when compared with 15:2.

We have data from a very large cohort study that suggests that the addition of ventilation (using 30:2) is associated with increased overall survival when compared with compression only CPR. This has been also suggested by a cohort study targeting paediatric patients.

Data from a large cluster RCT was unable to show any benefit from the use of a continuous chest compression technique when compared to the currently recommended technique.

Older studies, comparing strategies including previously recommended ratios (eg. 15:2 or a mixture of ratios), are much harder to extrapolate to current practice. There is no data included in this review that helps our understanding about ventilation strategies after the introduction of a definitive airway.

One key lesson learnt was the paucity of comparative data for in-hospital cardiac arrests. The only study included was one that evaluated the use of mechanical CPR, delivered by the “Thumper” in patients brought to the ED.

A number of issues were identified by this literature review. The overwhelming majority of published data is from cardiac arrest registries, and much less from randomised controlled trials. It is however easy to set up search strategies to identify randomised controlled trails that are of interest to the question, so RCTs that addressed the specific questions were unlikely to have been missed. On the other hand, as the authors point out, it is much more difficult to ensure that all data which directly and indirectly addressed the questions posed are identified. As a result, there may be additional valuable data from observational studies that had not been identified.

Throughout any review, frequent engagement of content experts is essential to ensure that the maximum relevant information is retrieved from the identified articles, and potentially misleading combinations of data are avoided. One great example of the pitfalls of systematic reviews was the human albumin saga, where a published meta-analysis caused concern, but a large well performed RCT demonstrated that use of human albumin was “SAFE”.

The next step in this process is for the ILCOR task forces to develop treatment recommendations from this data. Again these will be based on the GRADE process and will include clarifying statements to support their decisions.

We hope to see many more of these well-conducted, collaborative systematic reviews of the resuscitation and first aid science published in the future.

Conflict of interest statement