Fernando Beltramo

Cardiac output and systemic vascular resistance: Clinical assessment compared with a noninvasive objective measurement in children with shock

Purpose: To evaluate physician assessment of cardiac output and systemic vascular resistance in patients with shock compared with an ultrasonic cardiac output monitor (USCOM). To explore potential changes in therapy decisions if USCOM data were available using physician intervention answers. Study design: Double‐blinded, prospective, observational study in a tertiary hospital pediatric intensive care unit. Forty children (< 18 years) admitted with shock, requiring ongoing volume resuscitation or inotropic support. Two to 3 physicians clinically assessed cardiac output and systemic vascular resistance, categorizing them as high, normal, or low. An investigator simultaneously measured cardiac index (CI) and systemic vascular resistance index (SVRI) with USCOM categorized as high, normal, or low. Results: Overall agreement between physician and USCOM for CI (48.5% [&kgr; = 0.18]) and SVRI (45.9% [&kgr; = 0.16]) was poor. Interobserver agreement was also poor for CI (58.7% [&kgr; = 0.33]) and SVRI (52.3% [&kgr; = 0.28]). Comparing theoretical physician interventions to “acceptable” or “unacceptable” clinical interventions, based on USCOM measurement, 56 (21%) physician interventions were found to be “unacceptable.” Conclusions: There is poor agreement between physician‐assessed CI and SVRI and USCOM, with significant interobserver variability among physicians. Objective measurement of CI and SVRI may reduce variability and improve diagnostic accuracy. HighlightsAssessing hemodynamics in patients with septic shock is challengingPhysicians are not accurate in assessing cardiac output and systemic vascular resistanceThere is poor agreement between an objective hemodynamic monitor and physician assessment

Letter to Editor: Reliability of the Ultrasound Cardiac Output Monitor for Pediatric Patients

To the Editors, We appreciate the readers’ interest in our paper and we are pleased to respond to the letter by Lemson and Tibby [1] point by point. First, we acknowledge that the population in our study is older than a general PICU population. Of course, the ideal cohort of study subjects would include a greater variety of ages, sizes, and disease states with both elevated and depressed cardiac outputs, and the samples would be measured at a variety of time points during treatment. However, our study design was only intended to be a first step toward exploring the usefulness of this device in a pediatric setting, as a starting point for further investigation, and we could not meet this lofty standard. Indeed, it would be beneficial to have a greater amount of data from younger patients, but it is difficult to identify a younger subject group undergoing Swan-Ganz catheterization. We believe the study by Knirsch et al. [2] showed the USCOM to be inaccurate because of their use of subjects with intracardiac shunting and valvular malformations which can confound measurements by thermodilution measurement and by ultrasound, respectively. We would also add that we cannot identify a physiologic reason that the USCOM would become inaccurate in a smaller patient population. Second, we agree with the readers that there would be great importance in validating intrasubject variation and reproducibility of cardiac output against a previously accepted, if not standard, measurement. Unfortunately, as fewer pulmonary artery catheters are placed in a PICU setting, this type of study could be very difficult to organize. Third, the readers identified an error in our calculations. We thank them for their thoroughness, and we wish to apologize for this oversight. We calculated a simple percentage error (PE) rather than the method of Critchley et al. [3], and this was not identified during the many reviews of the manuscript by us and the Journal’s reviewers. The readers’ calculations are correct and using the entire dataset the PE is 33 %. However, we noted one outlier believed due to poor acoustic windows as well as change in level of sedation, as we indicated in the original manuscript—and when this sole outlier is removed the PE calculated using (2 9 SD of the bias)/mean value is 17 %. At this point, we respectfully acknowledge the limitations of our study and agree that the USCOM remains incompletely validated, but we disagree with the reader that the technology should not be recommended for use in children. We believe that this technology can be recommended for measurement of cardiac output in children, but that the information yielded by this measurement needs to be used in context and relied upon only to the extent that it agrees with the clinical scenario at hand The USCOM is not yet, and may never be, a replacement for invasive cardiac output measurement including thermodilution measurement, Fick method cardiac output measurement (with its own inherent assumptions and inaccuracies), or simple clinical acumen. But it is a useful tool worthy of adding to our clinical armamentarium. We do hope that there will be future validation studies in smaller children as well as for trends of cardiac output. In the meantime, given its noninvasive nature, ease of use at the bedside, and simple learning & Patrick A. Ross paross@chla.usc.edu

297: EPIDEMIOLOGY AND OUTCOME OF HOSPITALIZED INFANTS WITH PERTUSSIS

Learning Objectives: Pertussis during infancy is associated with significant morbidity and mortality. The objective of this study was to review the epidemiology and outcome of infants discharged with a diagnosis of pertussis and compare them with propensity score matched controls using the Kids Inpatient Database (KID). Methods: The KID is one of the databases developed for the Healthcare Cost and Utilization Project. We used the 2009 and 2012 KIDs for this study. We identified infants aged 1 to 12 mo with a diagnosis of pertussis (ICD 9 code 033.0), and compared their demographic and outcome data to infants without pertussis. In a separate analysis, infants with pertussis were matched 1:1 with correlative propensity score using gender, hospital region, income quartiles, race, ventilation status, and APRDRG severity score and compared their outcomes with controls. Results: A total of 2,900 pertussis cases were identified out of 624,569 discharges (prevalence:4.6/1000). The mean length of stay was longer (6.0 vs 4.5 days; p<0.01) in pertussis patients. Females (49% vs 43%), Hispanics (35% vs 25%), and lower income groups (39% vs 36%) were represented more frequently in the pertussis group (p<0.001). Mechanical ventilation (6% vs 4.7%; p=0.001), pulmonary hypertension [PHTN] (36% vs 1%; p<0.001), and ECMO use (0.65% vs 0.08%; p<0.001) were more frequent in the pertussis group. In intubated pertussis patients 50% had PHTN, 3% required a chest tube, and 10% required ECMO. The mortality rate was significantly higher in patients with pertussis who were ventilated (10.9% vs 0.04%; p<0.001), had PHTN (1.4% vs 0.3%; P<0.001) or required ECMO (60% vs 0.3%; P<0.01). Compared to matched controls, the mortality rate (0.72% vs 0.34%; p=0.048), PHTN (36.4% vs 0.82%; p<0.01) and ECMO usage (0.65% vs 0%; p<0.01) were higher in the pertussis group. The mortality rate in ventilated patients in our study was 11%. Conclusions: Pertussis disproportionately affects females, Hispanics, and lower income groups. Our study presents the national prevalence and outcome of pertussis in hospitalized infants.