A model for technology assessment as applied to closed loop infusion systems. Technology Assessment Task Force of the Society of Critical Care Medicine

OBJECTIVES

To test a model for the assessment of critical care technology on closed loop infusion control, a technology that is in its early stages of development and testing on human subjects.

DATA SOURCES

A computer-assisted search of the English language literature and reviews of the gathered data by experts in the field of closed loop infusion control systems.

STUDY SELECTION

Studies relating to closed loop infusion control that addressed one or more of the questions contained in our technology assessment template were analyzed. Study design was not a factor in article selection. However, the lack of well-designed clinical outcome studies was an important factor in determining our conclusions.

DATA EXTRACTION

A focus person summarized the data from the selected studies that related to each of the assessment questions. The preliminary data summary developed by the focus person was further analyzed and refined by the task force. Experts in closed loop systems were then added to the group to review the summary provided by the task force. These experts' comments were considered by the task force and this final consensus report was developed.

DATA SYNTHESIS

Closed loop system control is a technological concept that may be applicable to several aspects of critical care practice. This is a technology in the early stages of evolution and much more research and data are needed before its introduction into usual clinical practice. Furthermore, each specific application and each device for each application (e.g., nitroprusside infusion, ventilator adjustment), although based on the same technological concept, are sufficiently different in terms of hardware and computer algorithms to require independent validation studies.

CONCLUSIONS

Closed loop infusion systems may have a role in critical care practice. However, for most applications, further development is required to move this technology from the innovation phase to the point where it can be evaluated so that its role in critical car practice can be defined. Each application of closed loop infusion systems must be independently validated by appropriately designed research studies. Users should be provided with the clinical parameters driving each closed loop system so that they can ensure that it agrees with their opinion of acceptable medical practice. Clinical researchers and leaders in industry should collaborate to perform the scientifically valid, outcome-based research that is necessary to evaluate the effect of this new technology. The original model we developed for technology assessment required the addition of several more questions to produce a complete analysis of an emerging technology. An emerging technology should be systematically assessed (using a model such as the model developed by the Society of Critical Care Medicine), before its introduction into clinical practice in order to provide a focus for human outcome validation trials and to minimize the possibility of widespread use of an unproven technology.

Outcome of insulin-treated diabetics receiving epinephrine during cardiac arrest

The purpose of this study was to determine the effects of epinephrine in insulin-treated diabetics (DM) compared with nondiabetic (ND) controls during cardiopulmonary resuscitation (CPR). A retrospective analysis from a multicenter study of out-of-hospital cardiac arrest included 62 DM and 1,151 ND. Outcome parameters included return of spontaneous circulation (ROSC), blood pressure (BP), emergency department admissions (EDA), hospital admissions (HA), and hospital discharge (DC). In DM treated with standard-dose epinephrine (SDE), there were trends toward improved ROSC, BP, EDA, and HA compared with ND. DC was not different. With high-dose epinephrine (HDE) these outcomes were not different. Comparison of the outcomes of DM given SDE versus those given HDE were not different. However, none of the DM given HDE (n = 24) were discharged from the hospital. These results suggest that DM may have improved resuscitation rates with SDE and may be adversely affected with HDE. Larger prospective studies are needed to confirm these findings.

Agreement between paramedic-estimated weights and subsequent hospital measurements in adults with out-of-hospital cardiac arrest

OBJECTIVE

To assess the accuracy of paramedic estimates of adult body weights in cardiac arrest cases.

HYPOTHESIS

Paramedics could accurately estimate the weights of out-of-hospital cardiac arrest patients.

DESIGN

Retrospective data analysis of a 15-month, multicenter study involving nontraumatic out-of-hospital cardiac arrest patients. Paramedic estimates of body weights were compared to weights measured in the hospital. Patients were included in the analysis only if both a paramedic weight and a measured in-hospital weight were recorded.

SETTING

Six urban emergency medical services systems.

PARTICIPANTS

The study population included adults with return of spontaneous circulation who subsequently were admitted to the hospital.

MEASUREMENTS

Pearson correlation analysis of paramedic-estimated weights and measured weights.

RESULTS

Among the 133 study patients, the correlation coefficient (R) for paramedic estimates and the actual measured weight was 0.93. Paramedic estimates of weight were within 10% of the measured weights in 74% of the patients, and within 20% of measured weights in 93% of the patients.

CONCLUSION

Paramedic weight estimates correlated well with measured weights.

Initial countershock in the treatment of asystole

BACKGROUND

Routine provision of defibrillatory countershock (CS) in the initial management of asystolic cardiac arrest has been advocated because certain cases of ventricular fibrillation (VF) may present as asystole (AS).

OBJECTIVE

To determine the value of initial CS versus endotracheal intubation and pharmacologic therapy alone in the treatment of asystolic cardiac arrest.

DESIGN/PARTICIPANTS

A retrospective analysis of data collected prospectively during a multicenter study of out-of-hospital cardiac arrest. The study subjects were all patients whose initial cardiac arrest rhythm was AS and were treated with standard advanced cardiac life support (ACLS).

SETTING

Six urban emergency medical services (EMS) systems.

INTERVENTION

Patients in AS were treated initially with CS followed by ACLS therapy (CS Group), and were compared to those patients receiving endotracheal intubation and pharmacologic therapy alone (No CS Group).

OUTCOME MEASURES

Those receiving initial CS were compared to those not receiving CS using both Chi-square and logistic regression analysis. Outcome parameters included: rates of return of spontaneous circulation (ROSC), emergency department admission, hospital admission and hospital discharge.

RESULTS

Of the 194 patients presenting with AS, 77 received CS as their initial therapy. Of these, 13 (16.9%) had ROSC compared to 27 of the 117 (23.1%) from the No CS Group (P = 0.30). Emergency department and hospital admission rates were not significantly different; 13.0% versus 18.0% (P = 0.36), and 13.0% versus 11.1% (P = 0.69) for CS versus No CS, respectively. None of the patients in the CS Group were discharged alive versus two (1.7%) from No CS (P = 0.52). Of 42 patients with bystander-witnessed cardiac arrests, 13.3% in the CS Group had ROSC compared to 40.7% in the No CS Group (P = 0.07). Emergency department admission rates were 6.7% for the CS Group and 33.3% for the No CS Group (P = 0.07); while hospital admission rates were 6.7% and 22.2%, respectively (P = 0.39). When these comparisons were adjusted for bystander-initiated CPR, CPR interval, and paramedic response interval, the P-values became 0.10, 0.05 and 0.17, respectively.

CONCLUSIONS

Although, statistically, the results for both groups were not distinguishable, outcomes for asystolic patients had a tendency to be better when the initial therapy did not involve CS. Larger study populations are recommended to confirm these preliminary observations.

Effect of first-responder automated defibrillation on time to therapeutic interventions during out-of-hospital cardiac arrest. The Multicenter High Dose Epinephrine Study Group

STUDY OBJECTIVES

The effect of automated defibrillation provided by basic emergency medical technician (EMT) first-responder units on the time intervals to other critical interventions in the management of out-of-hospital cardiac arrests is unknown. The purpose of this study was to define and compare elapsed time intervals to basic CPR, paramedic arrival, initial countershock, endotracheal intubation, IV access, and initial adrenergic drug therapy in first-responder automated defibrillation/paramedic versus basic EMT/paramedic emergency medical services systems.

DESIGN

Prospectively collected data from a 15-month multicenter study of out-of-hospital, nontraumatic cardiac arrests were analyzed. The mean time intervals to critical therapeutic interventions between first-responder automated defibrillation/paramedic and basic EMT/paramedic groups were compared using the Student's t-test with Bonferroni correction.

SETTING

Three first-responder automated defibrillation/paramedic and three basic EMT/paramedic urban emergency medical services systems.

PARTICIPANTS

1,578 patients with out-of-hospital cardiac arrest.

INTERVENTIONS

The first-responder automated defibrillation/paramedic group received initial ECG analysis and/or automated countershock by first-responder/EMTs; the basic EMT/paramedic group received initial ECG analysis and/or manual countershock by paramedics.

RESULTS

Elapsed time intervals in minutes +/- SD for first-responder automated defibrillation/paramedic versus basic EMT/paramedic groups, respectively, were as follows: Collapse to CPR, 4.3 +/- 3.9 versus 5.4 +/- 5.2 (P = .017); collapse to countershock, 10.7 +/- 5.9 versus 13.0 +/- 6.0 (P = .017); collapse to paramedic arrival, 13.0 +/- 5.4 versus 10.3 +/- 6.1 (P = .0001); paramedic arrival to IV access, 5.1 +/- 3.9 versus 7.0 +/- 5.0 (P = .0001); paramedic arrival to endotracheal intubation, 4.8 +/- 4.0 versus 6.8 +/- 5.8 (P = .0001); paramedic arrival to initial adrenergic drug therapy, 7.4 +/- 4.5 versus 8.2 +/- 4.7 (P = .015); collapse to IV access, 17.7 +/- 6.1 versus 16.6 +/- 7.4 (P = .10); collapse to endotracheal intubation, 17.3 +/- 6.4 versus 16.6 +/- 7.8 (P = .32); collapse to initial adrenergic drug therapy, 20.4 +/- 6.7 versus 18.1 +/- 7.2 (P = .010). The time intervals from paramedic arrival to IV access, endotracheal intubation, and initial adrenergic drug therapy remained shorter in the first-responder automated defibrillation/paramedic systems despite stratification by presenting cardiac rhythm.

CONCLUSION

First-responder automated defibrillation/paramedic systems provide not only shorter times to initial countershock, as compared with basic EMT/paramedic systems, but by having delegated initial countershock to first-responders, they also allow for significantly shorter times from paramedic arrival to IV access, endotracheal intubation, and initial adrenergic drug therapy interventions.

Effectiveness of a human monoclonal anti-endotoxin antibody (HA-1A) in gram-negative sepsis: relationship to endotoxin and cytokine levels

Gram-negative sepsis is caused by endotoxin-induced release of tumor necrosis factor (TNF) and other cytokines. HA-1A is a human monoclonal antibody that binds specifically to endotoxin. HA-1A should prevent death in endotoxemic patients and reduce serum levels of TNF and interleukin-6 (IL-6). This hypothesis was tested in 82 septic patients who were randomly allocated to receive a single intravenous 100-mg dose of HA-1A or placebo. Pretreatment endotoxemia was detected in 27 patients (33%). Death occurred within 28 days of treatment in 8 (73%) of 11 placebo recipients and in 5 (31%) of 16 HA-1A recipients (P = .02). The median decrease in serum TNF level 24 h after treatment was 12 ng/L in patients given HA-1A and 0 ng/L in placebo recipients (n = 65; P = .04). For IL-6, this was 204 ng/L in patients given HA-1A and 44 ng/L in placebo recipients (n = 67; P = .4). Thus, HA-1A reduces mortality in septic patients with endotoxemia and lowers serum TNF levels.

A comparison of standard-dose and high-dose epinephrine in cardiac arrest outside the hospital. The Multicenter High-Dose Epinephrine Study Group

BACKGROUND

Experimental and uncontrolled clinical evidence suggests that intravenous epinephrine in doses higher than currently recommended may improve outcome after cardiac arrest. We conducted a prospective, multicenter study comparing standard-dose epinephrine with high-dose epinephrine in the management of cardiac arrest outside the hospital.

METHODS

Adult patients were enrolled in the study if they remained in ventricular fibrillation, or if they had asystole or electromechanical dissociation, at the time the first drug was to be administered to treat the cardiac arrest. Patients were randomly assigned to receive either 0.02 mg of epinephrine per kilogram of body weight (standard-dose group, 632 patients) or 0.2 mg per kilogram (high-dose group, 648 patients), both given intravenously.

RESULTS

In the standard-dose group 190 patients (30 percent) had a return of spontaneous circulation, as compared with 217 patients (33 percent) in the high-dose group; 136 patients (22 percent) in the standard-dose group and 145 patients (22 percent) in the high-dose group survived to be admitted to the hospital. Twenty-six patients (4 percent) in the standard-dose group and 31 (5 percent) in the high-dose group survived to discharge from the hospital. Ninety-two percent of the patients discharged in the standard-dose group and 94 percent in the high-dose group were conscious at the time of hospital discharge. None of the differences in outcome between the groups were statistically significant.

CONCLUSIONS

In this study, we were unable to demonstrate any difference in the overall rate of return of spontaneous circulation, survival to hospital admission, survival to hospital discharge, or neurologic outcome between patients treated with a standard dose of epinephrine and those treated with a high dose.

Glucocorticoid treatment does not improve neurological recovery following cardiac arrest. Brain Resuscitation Clinical Trial I Study Group

Glucocorticoids are commonly given to patients with global brain ischemia, although their efficacy has not been proved. The database of the Brain Resuscitation Clinical Trial I, a multi-institutional study designed to evaluate the effect of thiopental sodium therapy on neurological outcome following brain ischemia, was used for a retrospective review of the effects of glucocorticoid treatment on neurological outcome after global brain ischemia. This study included 262 initially comatose cardiac arrest survivors who made no purposeful response to pain after restoration of spontaneous circulation. The standard treatment protocol left glucocorticoid therapy to the discretion of the hospital investigators. This resulted in four patient groups that received either no, low, medium, or high doses of glucocorticoids in the first 8 hours after arrest. Neurological outcome was scored using a modification of the Glasgow Cerebral Performance Category Scale. None of the steroid regimens statistically improved mean group survival rate or neurological recovery rate over that observed in the group that did not receive steroids. The routine clinical practice of administrating glucocorticoids after global brain ischemia may be associated with serious complications and is not justified.

Charcoal hemoperfusion for the treatment of a combined diltiazem and metoprolol overdose

We describe the management of a 49-year-old woman who ingested 1,200 mg diltiazem and 500 mg metoprolol. The patient was bradycardic and hypertensive, requiring large doses of inotropic agents and temporary transvenous pacing. Charcoal hemoperfusion was used as an adjunct technique to hasten drug elimination; it rapidly lowered plasma drug levels, with a parallel improvement in the patient's clinical condition.

Continuous intravenous infusion of sodium thiopental for managing drug withdrawal syndromes

This report describes the successful use of sodium thiopental by continuous infusion in three patients with drug withdrawal syndromes and the concomitant need for mechanical support of ventilation. This approach provided the sedation necessary so the patients could be effectively ventilated while simultaneously controlling the manifestations of drug withdrawal.

High-frequency jet ventilation versus conventional ventilation after surfactant displacement in dogs

High-frequency jet ventilation (HFJV) was compared with conventional ventilation (CV) after surfactant displacement with diocytyl sodium sulfosuccinate (OT) via ultrasonic nebulization. After aerosol delivery, dogs were separated into 3 groups and followed for 2 h. Ventilator settings were not changed in group I (CV) and group II (HFJV) after OT delivery. In group III, drive pressure was increased to 40 psi. Adequacy of oxygenation varied directly with peak inspiratory pressure (PIP) rather than airway pressure (Paw) in both HFJV and CV. In all groups, immediately after OT administration PaO2 decreased and there was a slight increase in Paw; PIP was significantly elevated only in groups I and III. Two hours later, PaO2 had returned to baseline in groups I and III, but had not improved significantly in group II. Paw remained the same in all groups. These data demonstrate that in noncompliant lungs, oxygenation is not improved unless a high PIP is used to establish the critical opening pressures needed to rerecruit alveoli. In this noncompliant lung model, HFJV was not effective at low Paw values, and thus offered no apparent advantage over CV.

Problems in brain death determination

During the last decade there has been philosophical acceptance of the concept that the state of brain death is equivalent to total patient death. The application of this concept to clinical medicine has been associated with major problems in both the diagnosis of brain death and the medical management of the brain dead patient. In our experience with 176 consecutive cases of suspected brain death over a seven-year period, we have found that a standardized protocol applied by experienced clinicians will minimize these problems.

Prepubescent gonococcal vulvovaginitis

Investigation of a five-year-old girl with gonor rheal vulvovaginitis revealed four other cases in the household. This illustrates the importance of investigation of all family members in such cases.