Automated closed-loop resuscitation of multiple hemorrhages: a comparison between
fuzzy logic and decision table controllers in a sheep model
#MMPMID28265453
Marques NR
; Ford BJ
; Khan MN
; Kinsky M
; Deyo DJ
; Mileski WJ
; Ying H
; Kramer GC
Disaster Mil Med
2017[]; 3
(?): 1
PMID28265453
show ga
BACKGROUND: Hemorrhagic shock is the leading cause of trauma-related death in the
military setting. Definitive surgical treatment of a combat casualty can be
delayed and life-saving fluid resuscitation might be necessary in the field.
Therefore, improved resuscitation strategies are critically needed for prolonged
field and en route care. We developed an automated closed-loop control system
capable of titrating fluid infusion to a target endpoint. We used the system to
compare the performance of a decision table algorithm (DT) and a fuzzy logic
controller (FL) to rescue and maintain the mean arterial pressure (MAP) at a
target level during hemorrhages. Fuzzy logic empowered the control algorithm to
emulate human expertise. We hypothesized that the FL controller would be more
effective and more efficient than the DT algorithm by responding in a more rigid,
structured way. METHODS: Ten conscious sheep were submitted to a hemorrhagic
protocol of 25 ml/kg over three separate bleeds. Automated resuscitation with
lactated Ringer's was initiated 30 min after the first hemorrhage started. The
endpoint target was MAP. Group differences were assessed by two-tailed t test and
alpha of 0.05. RESULTS: Both groups maintained MAP at similar levels throughout
the study. However, the DT group required significantly more fluid than the FL
group, 1745 ± 552 ml (42 ± 11 ml/kg) versus 978 ± 397 ml (26 ± 11 ml/kg),
respectively (p = 0.03). CONCLUSION: The FL controller was more efficient than
the DT algorithm and may provide a means to reduce fluid loading. Effectiveness
was not different between the two strategies. Automated closed-loop resuscitation
can restore and maintain blood pressure in a multi-hemorrhage model of shock.
free
free
free
