Page 23 - Journal of Special Operations Medicine - Spring 2016

interrogation of our hypothesis, we present several case The patient was admitted to the pediatric ICU (PICU),
studies to demonstrate the potential clinical and opera- where monitoring began (Figure 1). CRI after arrival in
tional utility of the CRI in a variety of emergency medi- the PICU was initially 0.4 but quickly declined to <0.3
cal and human performance settings that are relevant to within 1 hour (i.e., loss of compensatory reserve). As the
military medicine. day progressed, the patient complained of increasing ab-
dominal pain. Three hours after admission to the PICU,
the patient received 1L of lactated Ringer’s solution,
Materials and Methods
which was successful in restoring his compensatory re-
For clinical cases 1–6, consent or waiver of consent was serve to 0.8, but after 2 hours, the CRI drifted to <0.3.
obtained from each patient after study approval was ob- Subsequent infusions of packed red blood cells (PRBCs)
tained from the Colorado Multiple Institution Review and saline restored some reserve, but the CRI eventually
Board (IRB) or the Brooke Army Medical Center IRB. fell to <0.2. The patient’s HR and BP remained relatively
CRI data were obtained from small pulse oximetry- stable throughout this period; 14 hours after admission
based DataOx or CipherOx devices (Flashback Technol- to the PICU, the patient reached a nadir CRI of 0.12,
ogies; http://www.flashbacktechnologies.com) and vital with an HR of 148 bpm, and BP of 115/48mmHg.
sign data were recorded on a BedMasterEx system (Excel Shortly thereafter, he had a large emesis and rapidly de-
Medical; http://excel-medical.com). These devices were compensated, becoming unresponsive, hypotensive, and
applied in the emergency department (ED) and used to bradycardic. This prompted an emergent exploratory
record data as a patient traveled from the ED to radiol- laparotomy, where two jejunal perforations were dis-
ogy and the operating room or intensive care unit (ICU). covered. Significant reduction of this patient’s compen-
Deidentified PPG waveform data from select patients of satory reserve within the initial hours in the PICU was
interest were analyzed independently without any refer- contrasted by his stable BP and adequate urine output.
ence to clinical scenarios or therapeutic interventions. Successful surgery was reflected by postoperative resto-
Clinical information from the patient medical records ration maintenance of his CRI to >0.8.
was then correlated with CRI results.
Figure 1 The compensatory reserve measured over 29 hours
For experimental cases 7 and 9, written consent to use in a pediatric trauma patient admitted to the PICU with acute
collected data was obtained from each subject before their blood loss due to trauma followed by development of sepsis.
participation in a laboratory demonstration. For case 8,
consent was obtained after receiving study approval from
the University of Texas Southwestern Medical Center
IRB. For study case 8, CRI values were calculated from
PPG waveforms recorded from a Finometer BP monitor
(Finapres Medical Systems; http://www.finapres.com)
during exposure to progressive lower-body negative
pressure (LBNP) with and without exposure to a heat
stress (protocol details are presented under Results). For
study cases 7 and 9, CRI values were recorded directly
and continuously on a CipherOx device during graded
exercise (protocol details presented under Results).

Results Each bar represents 1 hour. Bar colors: green, Compensatory Reserve
Index (CRI) >0.6; yellow, CRI ≤0.6 and >0.3; red, CRI ≤0.3. LR, lac-
Case 1: Patient With Acute Blood Loss tated Ringer’s; NS, normal saline; OR, operating room; PICU, pediat-
ric intensive care unit; pRBC, packed red blood cells.
Due to Trauma Followed by Sepsis
A 15-year-old, unhelmeted, male bicyclist was struck by
an automobile and dragged 250 feet. Initial evaluation Case 2: Patient With Acute Appendicitis
noted tachycardia and hypotension, followed by 4L of A 10-year-old boy presented to the ED following 12
intravenous (IV) crystalloid resuscitation. He arrived at hours of abdominal pain. He was febrile and normoten-
a Level I trauma center ED with an HR of 146 beats sive (systolic BP [SBP], 118mmHg) but tachycardic with
per min (bpm) and BP of 99/61mmHg after a 1-hour an HR of 108 bpm, and a white blood cell count of 18.6
transport. A focused abdominal sonography for trauma × 10 L. CT scan of the abdomen showed evidence of
9
(FAST) examination was negative for free intra-abdom- acute appendicitis. His last oral intake was 10 hours be-
inal fluid, and trace intraperitoneal free fluid was noted fore the initiation of monitoring, and he received antibi-
on an abdominal computed tomography (CT) scan. otics and maintenance fluids. The patient’s preoperative

Machine Learning and Hemodynamic Instability 7