Initial Rhythm and Outcomes for Inhospital Cardiac Arrest in Kenya

Document Type



Internal Medicine (East Africa); Emergency Medicine (East Africa)


Background: Health care resource allocation remains challenging in lower middle income countries (LMIC) like Kenya with the co-existence of both communicable and non-communicable diseases in hospitals. In-hospital cardiac arrest and resuscitation remains poorly elucidated.

Methods: This was a prospective study of cardiac arrest in 6 hospitals in Kenya from 2014 to 2016. Existing resuscitation teams were utilized to collect data during resuscitation using a standardized protocol. Patients with do-not-resuscitate (DNR) orders, trauma, post-surgical and pregnancy related complications were excluded. The Modified Early Warning Score (MEWS - SBP, HR, respiration rate, temperature, responsiveness) was determined based on worst parameters at least 4 hours prior to the arrest and divided in to low (<3), intermediate (3-4) or high risk (>4).

Results: 353 patients with cardiac arrest were included over 19 months. The mean age was 58.6 years, 53.5% were male, and admission diagnoses included cardiovascular disease (26%), sepsis 28.9%, and cancer 14.5%. The median MEWS was 5 (IQR 3-7) and low, intermediate and high MEWS was found in 21.2%, 21.5% and 57.2% respectively. The mean time to CPR was 0.83 minutes (IQR 0-26 minutes). The initial rhythm was asystole in 47.6%, PEA in 38.2%, VT/VF in 5.4% and unknown in 8.8%. Return of spontaneous circulation (ROSC) occurred in 29.2% of patients with the median time to ROSC being 28 minutes (IQR 7.5-16.5). ROSC was 17.3% in asystole, 40.7% in PEA, 57.9% in VT/VF and 25.8% in patients with an unknown rhythm. Of all patients, 16 (4.2%) were discharged alive.

Conclusion: Non-shockable rhythms account for the majority of the cardiac arrests in hospitals in a LMIC, and are associated with unfavourable outcomes. Future work should be directed to training healthcare personnel in recognizing early warning signs and implementing appropriate measures in a resource-scare environment.

Publication (Name of Journal)