We found that in successfully resuscitated patients, the risk of cardiac arrest re-occurring during the flight was high during both primary and secondary missions. Therefore, MCDs are beneficial for CPR during airlift to the hospital. Employing MCDs was feasible and safe in the HEMS; standing orders and crew training were carefully implemented, as previously described (16). There were only two MCD malfunctions (one on the scene, one in flight).
Unnecessary prolonged CPR before terminating efforts is reported in several studies and should be avoided, as it blocks the HEMS crew and therefore wastes EMS and hospital resources (17, 18). In our study, the time until ROSC or termination of CPR was usually <30min, indicating determined decision-making on the scene in routine cases. However, in patients being airlifted with ongoing CPR, we found that there is a relevant proportion of cases without recommended indication regarding reversible causes of cardiac arrest, suggesting an individual interpretation of the situation. The poor prognosis in this group of patients, with only 3% survival at discharge, is in accordance with recent studies (19, 20). However, precise criteria for patient selection and timing are still lacking. On the other hand, there is an increasing amount of literature showing that carefully selected patients may benefit from early transport with ongoing CPR for further treatment (19-21).
A relevant finding of our study was the fact that in 20 of 271 patients with ROSC (7%), a second cardiac arrest occurred. A pre-installed MCD was a safe and fast way to initiate chest compressions in flight. However, in many HEMS worldwide, manpower and space in the cabin are severely limited and it is impossible to install a MCD or to perform sufficient and efficient manual chest compressions in flight. Thus our findings support the need for a preinstalled device in this patient group to prevent a relevant delay in CPR in case of re-arrest.
A shockable cardiac arrhythmia was associated with a five times higher chance for survival in the multivariate logistic regression analysis (OR 0.176, 95% CI 0.084 to 0.372, p<0.001). These findings are in a line with previous data which demonstrated that survival was best in the group with shockable rhythms in OHCA (22). These findings could help create further SOPs on the indication, use and duration of MCDs.
Despite the evidence against general use of MCDs, case reports and the recent ERC and AHA guidelines report beneficial effects during prolonged CPR in special circumstances, such as accidental hypothermia, intoxication, or pulmonary embolism (23-25). However, providing high-quality chest compressions with minimal hands-off time is challenging during evacuation, e.g., in alpine environments and during transport in a helicopter. Accidental hypothermia from a fall in a crevice, submersion in ice water, or burying in an avalanche is more common in HEMS accidents than in accidents treated by ground-based EMS (25, 26). Therefore, according to the current CPR recommendations, a robust and feasible way to perform CPR is crucial in this patient group.
Few studies have looked at the feasibility of the different MCDs in mountain HEMS. In general, there are different underlying mechanisms and theories driving MCDs, the cardiac pump theory (LUCAS™) and the thoracic pump theory (AutopulseÒ) (27). Requirements for HEMS with winch rescue are different from requirements for ground-based EMS. In a simulation study and additionally in a case series, Pietsch et al. demonstrated the feasibility of using MCDs, even in remote areas and adverse environments requiring winch rescues with ongoing chest compression (4, 28). In our study, one of the 69 trauma patients with known data survived to hospital discharge. This may indicate that the trauma patient sample was too small, or that the wrong patients were transported to the hospital.
All of the existing MCD studies, and even the recent Cochrane review, excluded studies explicitly including patients with cardiac arrest caused by trauma, drowning, hypothermia and toxic substances. These conditions are mostly excluded from cardiac arrest intervention studies because they have a different underlying pathophysiology, therapy and thus prognosis (29). To our knowledge there is no study reporting data on outcomes of traumatic CA and MCDs so far. Thus, this study would be the first reporting data in this field of non-cardiac-caused CAs and MCDs. In contrast, hospital discharge rates ranged between 13% in women and 23% in men suffering non-traumatic cardiac arrest, indicating that employing MCDs resulted in very good survival rates in this group. Our clear findings against the use of MCDs in traumatic CA could potentially influence resuscitation guidelines in the future.
MCDs offer additional benefits for the HEMS crew. During a simulated helicopter transport, Rehatschek et al. showed that MCDs improved the efficacy of chest compressions, reduced physical stress, and led to enhanced cognitive performance of the EMS crew as compared to manual CPR (30). This may enable better monitoring of clinical developments, and leaves hands free for additional interventions. However, a theoretical basis and evidence regarding those positive effects on human error, as well as data on non-technical skills, are still limited.
An obvious reason to implement MCDs in HEMS is to ensure conformity with flight transport regulations, such as the requirement for the entire crew to have latched seatbelts during ongoing chest compressions in flight. Further, with a growing number of contagious diseases (e.g., COVID-19), MCDs can enhance safety while still allowing provision of high-quality CPR with maximal distance to the patient on the scene and during transport.
The strength of our study is the large patient cohort we examined. The HEMS missions in our study are very diverse, consisting of a mixture of urban, rural, mountainous, and alpine areas, thus the results may be extrapolated to other EMS settings. Due to the retrospective study design we could investigate application of MCDs in daily practice of HEMS without the artificial limitations of a prospective trial protocol. In addition, the fact that the involved HEMS were physician-staffed meant that the course of CPR efforts in our study was not heavily influenced by legal considerations, since physicians could pronounce a patient dead on the scene. In contrast, paramedic-staffed services – depending on local legislation and protocols – might require extensive CPR efforts before termination, thus altering the role of MCDs.
The retrospective observational nature of our study has several limitations as well. First, there was no systematically gathered information regarding MCDs from the HEMS missions except for the question of whether the device was used or not. We extracted all relevant information from the medical HEMS reports, but quality varied greatly. Second, there may have been missions where a MCD was used but not documented in the HEMS database. Since we were unable to check all missions manually, we could not include them, and therefore we might have underestimated the true proportion of HEMS missions involving MCDs, a potential source of selection bias. Third, a multivariable logistic regression was performed adjusting the analysis for confounding factors, however, residual confounding due to unmeasured factors is possible. And last, we could not compare our findings with a control group since identifying a set of comparable HEMS missions in the years preceding the implementation of MCDs in our organization was not possible.