Agreement between rectal and tympanic membrane temperatures in marathon runners

STUDY OBJECTIVE: To determine the agreement between rectal temperature and infrared tympanic membrane temperatures in marathon runners presenting to a field hospital at the finish line. METHODS: The subjects of this prospective, blinded, controlled study were runners 18 years or older who were triaged to the acute care medical area at the finish line for suspected hypothermia, hyperthermia, dehydration, or altered mental status. Rectal and tympanic temperatures were measured simultaneously in all subjects for whom rectal temperature measurement had been deemed necessary and recorded on separate data cards. RESULTS: Of the 239 runners treated in the acute care medical area, 37 required rectal temperature measurement and were enrolled in the study. The mean rectal temperature was 38.45 degrees +/- 1.20 degrees C (range, 35.9 degrees to 41.5 degrees C). The mean tympanic membrane temperature was 37.81 degrees +/- 95 degrees C (range, 36.3 degrees to 40.4 degrees C). Pearson's correlation coefficient revealed a moderate correlation (r = .6902, P = .00023). The mean temperature difference between the two thermometers, mean rectal minus mean tympanic membrane, was .64 degrees C (95% confidence interval, .35 degrees to .93 degrees C). Sixty-Two percent of the tympanic membrane readings were within 1 degree C of their rectal counterparts. Agreement ranged from 1.16 degrees (+2 SD) to -2.95 degrees (-2 SD). The 95% confidence interval was 1.67 degrees to -2.95 degrees C. CONCLUSION: We were able to demonstrate only a moderate correlation between the two thermometer readings, with a wide spread between the limits of agreement. This spread could be clinically significant and therefore limits the usefulness of tympanic temperature in the marathon race setting. Because of the potentially large and clinically significant differences in rectal and tympanic temperatures and the limitations inherent in our study, we cannot endorse the use of tympanic temperature in the setting of a marathon event.     

Unreliability of the infrared tympanic thermometer in clinical practice: a comparative study with oral mercury and oral electronic thermometers

BACKGROUND: This study was designed to determine the magnitude and frequency of measurement errors with infrared tympanic thermometers in the clinical setting. METHODS: In a convenience sample of 137 adult inpatients, we compared body temperatures measured by a Diatek 9000 Infrared Aural Thermometer and an IVAC 2090 CoreCheck Tympanic Thermometer between themselves, in right versus left ears, and against concurrently measured oral temperatures using both an electronic thermoprobe and conventional glass mercury thermometer. RESULTS: There was a significant between-brand difference of 0.6 degrees C (IVAC 相似文献   

Infrared temperature measurement in the ear canal with the DIATEK 9000 Instatemp and the DIATEK 9000 Thermoguide. Comparison with methods of temperature measurement in other body parts

Temperature of the tympanic membrane is recommended as a "gold standard" of core-temperature recording. However, use of temperature probes in the auditory canal may lead to damage of tympanic membrane. Temperature measurement in the auditory canal with infrared thermometry does not pose this risk. Furthermore it is easy to perform and not very time-consuming. For this reason infrared thermometry of the auditory canal is becoming increasingly popular in clinical practice. We evaluated two infrared thermometers-the Diatek 9000 Thermoguide and the Diatek 9000 Instatemp-regarding factors influencing agreement with conventional tympanic temperature measurement and other core-temperature recording sites. In addition, we systematically evaluated user dependent factors that influence the agreement with the tympanic temperature. MATERIALS AND METHODS: In 20 volunteers we evaluated the influence of three factors: duration of the devices in the auditory canal before taking temperature (0 or 5 s), interval between two following recordings (30, 60, 90, 120, 180 s) and positioning of the grip relative to the auditory-canal axis (0, 60, 180 and 270 degrees). Agreement with tympanic contact probes (Mon-a-therm tympanic) in the contralateral ear was investigated in 100 postoperative patients. Comparative readings with rectal (YSI series 400) and esophageal (Mon-a-therm esophageal stethoscope with temperature sensor) probes were done in 100 patients in the ICU. The method of Bland and Altman was taken for comparison. RESULTS: Shortening of the interval between two consecutive readings led to increasing differences between the two measurements with the second reading decreasing. A similar effect was seen when positioning the infrared thermometers in the auditory canal before taking temperatures: after 5 s the recorded temperatures were significantly lower than temperature recordings taken immediately. Rotation of the devices out of the telephone handle position led to increasing lack of agreement between infrared thermometry and contact probes. Mean differences between infrared thermometry (Instatemp and Thermoguide, CAL-Mode) and tympanic probes were -0.41 +/- 0.67 degree C (2 SD) and -0.43 +/- 0.70 degree C, respectively. Mean differences between the Thermoquide (Rectal-Mode) and rectal probe were -0.19 +/- 0.72 degree C, and between the Thermoguide (Core Mode) and esophageal probe -0.13 +/- 0.74 degree C. DISCUSSION: Although easy to use, infrared thermometry requires careful handling. To obtain optimal recordings, the time between two consecutive readings should not be less than two min. Recordings should be taken immediately after positioning the devices in the auditory canal. Best results are obtained in the 60 degrees position with the grip of the devices following the ramus mandibulae (telephone handle position). The lower readings of infrared thermometry compared with tympanic contact probes indicate that the readings obtained represent the temperature of the auditory canal rather than of the tympanic membrane itself. To compensate for underestimation of core temperature by infrared thermometry, the results obtained are corrected and transferred into core-equivalent temperatures. This data correction reduces mean differences between infrared recordings and traditional core-temperature monitoring, but leaves limits of agreement between the two methods uninfluenced.     

Are electronic thermometry techniques suitable alternatives to traditional mercury in glass thermometry techniques in the paediatric setting?

Three thermometers, mercury in glass, Becton-Dickinson digital and IVAC tympanic membrane thermometer, were compared. The study was designed to test the null hypothesis that there is no difference between the thermometer recordings, meaning that electronic thermometry is a suitable alternative to traditional mercury in glass thermometry. Children aged between 0 and 15 years old were entered into the study, n = 114. Standard procedures were used throughout the study to minimize the risk of errors in the data collection and technical errors were recorded. The data were analysed using graphical techniques described by Bland and Altman. The differences between the temperature recordings were plotted against their mean (the estimated true value), with the bias and limits of agreement (2 standard deviations from the mean) calculated for the mercury in glass/digital, mercury in glass/tympanic and digital/tympanic temperature recordings. The range of temperatures were 35.1 degrees C-38.5 degrees C, with a mean of 36.85 degrees C for the mercury in glass thermometer, 36.8 degrees C for the digital thermometer and 36.65 degrees C for the tympanic thermometer. The bias for each comparison was 0.36 degrees C, 0.21 degrees C and 0.17 degrees C and the limits of agreement were wide, -0.516 degrees C to 1.234 degrees C, -0.84 degrees C to 1.252 degrees C and -0.940C to 1.244 degrees C for the mercury in glass/digital, mercury in glass/tympanic and digital/ tympanic recordings respectively. A difference of 0.2 degrees C is usually accepted for clinical practice, therefore the null hypothesis was rejected.     

Supralingual temperatures compared to tympanic and rectal temperatures

OBJECTIVE: To evaluate the validity, reliability, sensitivity, and specificity of the PaciTemp supralingual digital pacifier thermometer as compared to the Thermoscan Instant tympanic and glass-mercury rectal thermometers. METHOD: Eighty-one children under the age of 2 years had temperatures taken sequentially at three body sites: supralingual, tympanic, and rectal. Corrections were calculated between the readings of the three types of thermometers. Percentage of agreement was done to examine sensitivity and specificity. RESULTS: Using the glass-mercury measurement as the standard, both the supralingual and tympanic measurements showed an overall specificity of 62.8% and sensitivity of 63.3%. Correlation between rectal and supralingual was 0.62, and correlation between rectal and tympanic was 0.71. CONCLUSIONS: The Paci-Temp provides temperature readings that are similar to the tympanic method as compared to the rectal method. Further research on at-home thermometers is needed.     

The measurement of the single-breath transfer factor for carbon monoxide and its components using the Morgan Transflow system

In contrast to the standard single-breath transfer factor for carbon monoxide (TLCO), there are no specific guidelines or recommendations for the measurement of its components, the pulmonary capillary blood volume (VC) and membrane component (DM), by the Roughton and Forster method. Ten randomly selected heart transplant patients (three life-long non-smokers, seven ex-smokers > 1 yr, age range 24-55 years) were assessed on two occasions using either the standard or high-oxygen mixture as the first inspired gas in random order. Ten normal subjects (all non-smokers, age range 23-54 years) were assessed on two occasions using either a long protocol (30 min waiting time between repeat measurements in an individual set) or a short protocol (5 min waiting time). Two technically acceptable results of TLCO were used to derive a mean value for DM and VC for each set of measurements (Transflow, P. K. Morgan, Kent, U.K.). The different sequences of gas mixtures produced no significant differences between the values obtained in ten heart transplant patients for mean TLCO (mmol min-1 kPa-1) (standard first 5.13 +/- 1.15, high-oxygen first 5.14 +/- 1.12; limits of agreement -0.57 to 0.59 for DM or for VC. The long or short protocol produced no significant differences between the means of TLCO (mmol min-1 kPa-1) (long 8.0 +/- 1.9, short 8.0 +/- 1.9; limits of agreement -0.5 to 0.5), DM or VC. This allows the development of a standard test protocol of short duration (about 40 min) making it practical for clinical use without compromising the precision or reproducibility of the results obtained.     

Validity and reliability of measurements taken by the Peak 5 motion analysis system

Many new measurement tools are becoming available for physiotherapists; however, there is often a lack of data to support their validity and reliability. This study attempts to address this issue with respect to the Peak 5 video analysis system. Five experiments were conducted. Experiment A investigated the validity of static angular measurements by filming a universal goniometer. A high level of agreement was recorded with a mean difference between the two instruments of 0.2 degree. Experiment B investigated the validity of angular joint velocity by filming a Biodex isokinetic dynamometer. A high level of agreement was recorded with a mean difference between the two instruments of 0.96 degree s-1. Experiment C investigated the reliability of reflective marker placements on the hip, knee and ankle on 17 healthy volunteers, in weight bearing on three separate days. Reliability was high with the 95% confidence interval (CI) for marker placement being 1.5 cm. Experiment D investigated the normal intrasubject variation of knee joint 'critical angle' and angular velocity measures over time, by filming 17 healthy volunteers performing an eccentric step test on three separate days. For 'critical angle' measures the standard deviation was 5.72 degrees and the 95% CI was +/- 11.44 degrees. For angular velocity measures the standard deviation was 8.51 degrees s-1 and the 95% CI was +/- 17.02 degrees s-1. Experiment E repeated analysis of the reflective marker placement (mean standard deviation 0.6 cm), the 'critical angle' (mean standard deviation 1.32 degrees) and the angular velocity (mean standard deviation 1.29 degrees s-1 showed high consistency, when analysing three pieces of film three times. Data generated in this study suggest that the Peak 5 provides valid angular and angular velocity data when compared to other measuring devices. Reliable data are produced when filming healthy volunteers performing an eccentric step test. These data will be used in the next stage of the project investigating Anterior Knee Pain Patients.     

Comparison of three rewarming methods in a postanesthesia care unit

Postoperative hypothermia is problematic because patients in postanesthesia care units (PACUs) often feel very cold, and unrecognized or prolonged postoperative hypothermia can aggravate patients' underlying cardiovascular disorders. The researchers compared three methods of rewarming PACU patients who had undergone laparotomy procedures. Patients were assigned randomly to three groups. Each patient in group one received the standard PACU rewarming intervention (ie, two warmed thermal blankets and a hospital bedspread). Each patient in group two received the standard PACU rewarming intervention plus a reflective blanket. Each patient in group three received the standard PACU rewarming intervention plus a reflective blanket and a reflective head covering. Nurses measured patients' vital signs on admission to the PACU and every 15 minutes thereafter until patients' sublingual temperatures reached 36 degrees C (96.8 degrees F). No significant temperature differences occurred among patients in the three groups, but an inverse relationship existed between patients' PACU admission temperatures and the time they required to reach normothermia.     

A difference of 5 degrees C between ear and rectal temperatures in a febrile patient

A 4-year-old boy with a history of seizures triggered by fever presented at an emergency department (ED) with tachycardia, skin vasoconstriction, and a rectal temperature of 42.2 degrees C. However, his ear temperature (as repeatedly measured in two ears, by two experienced nurses, and with two infrared thermometers) was between 36.4 degrees C and 37.6 degrees C. Antipyretic therapy resulted in skin vasodilation, a rapid decrease of rectal temperature, restoration of heart rate, and disappearance of the difference between the two temperatures. Seizures did not occur. This case shows that infrared ear thermometry cannot be recommended in EDs as the procedure of choice for detecting fever in small children, especially when they are vasoconstricted.     

Reliability of infra-red tympanic thermometry (ITT)

BACKGROUND: The ability to take a patient's temperature quickly, easily, and accurately is desirable in a variety of situations. However, in order to have confidence in the values obtained, it is important to quantify the reliability of the measurement. PURPOSE: The purpose of this study, therefore, was to determine the reliability of temperature measurements obtained by infra-red tympanic membrane thermometers. METHOD: This study examined intra-tester, inter-tester and inter-instrument reliability using two commercially available tympanic membrane thermometers. Forty-four college students (mean age 20.2 +/- 3.6 yr) had their tympanic membrane temperature taken by two investigators, each using two different instruments (ThermoScan Pro-1 and FirstTemp Genius). RESULTS: The results indicated a statistically significant difference between testers for each instrument. Similarly, there was a significant statistical difference between instruments within each tester. Correlations were moderately high (r = 0.66-0.88) between testers for each instrument, whereas correlations within instruments varied considerably between testers (r = 0.35-0.78). The intra-tester reliability was good for both testers using both instruments with all coefficients of variation (CV) less than 2%. Statistically significant differences were found between testers for each instrument and between instruments for each tester. A 2 x 2 ANOVA revealed a significant main effect for Tester and a significant Tester x Instrument interaction. CONCLUSION: We conclude that individual investigators are able to obtain reliable temperature measures when using either the FirstTemp Genius or ThermoScan Pro-1 instruments under resting laboratory conditions. However, inter-tester and inter-instrument reliability should be considered when collecting or comparing data.     

Standard and multifrequency tympanometry in normal and otosclerotic ears

OBJECTIVES: The primary goal of this study was to evaluate alternative tympanometric parameters for distinguishing normal middle ears from ears with otosclerosis. A secondary goal was to provide guidelines and normative data for interpreting multifrequency tympanometry obtained using the Virtual 310 immittance system. DESIGN: Nine tympanometric measures were examined in 68 normal ears and 14 ears with surgically confirmed otosclerosis. No subjects in either group had a history of head trauma or otoscopic evidence of eardrum abnormalities. Two parameters, static admittance and tympanometric width, were derived from standard low-frequency tympanometry and two parameters, resonant frequency and frequency corresponding to admittance phase angle of 45 degrees (F45 degrees), were derived from multifrequency tympanometry. RESULTS: Differences between normal and otosclerotic ears were statistically significant only for resonant frequency and F45 degrees. Group differences in resonant frequency were larger when estimated using positive tail, rather than negative tail, compensation. Group differences in both resonant frequency and F45 degrees were larger when estimated from sweep frequency (SF), rather than sweep pressure, tympanograms. Test performance analysis and patterns of individual test performance point to two independent signs of otosclerosis in the patient group; 1) an increase in the stiffness of the middle ear, best indexed by F45 degrees derived from SF recordings, and 2) a change in the dynamic response of the tympanic membrane/middle ear system to changes in ear canal pressure, best indexed by tympanometric width. Most patients were correctly identified by only one of these two signs. Thus, optimal test performance was achieved by combining F45 degrees derived from SF recordings and tympanometric width. CONCLUSIONS: The findings confirm the advantage of multifrequency tympanometry over standard low-frequency tympanometry in differentiating otosclerotic and normal ears. Recommendations for interpreting resonant frequency and F45 degrees measures obtained using the Virtual Immittance system are also provided. In addition, the relationship among different tympanometric measures suggests a general strategy for combining tympanometric measures to improve the identification of otosclerosis.     

A comparison of radiographic and electrogoniometric angles in adolescent idiopathic scoliosis

STUDY DESIGN: This was a cross-sectional study of a consecutive group of adolescent patients presenting to a scoliosis clinic for routine assessment or monitoring of their scoliosis, excluding postsurgical patients. SUMMARY OF BACKGROUND DATA: In vitro studies suggested electrogoniometry could be useful in the evaluation of scoliosis. No prior in vitro study had been performed. OBJECTIVES: To determine the reliability and validity of an electrogoniometric instrument, the Metrecom Skeletal Analysis System, in assessing adolescent idiopathic scoliosis. METHODS: Thirty-one patients were examined, radiographed, and scanned with the Metrecom Skeletal Analysis System twice by two different examiners. The magnitudes of the curves derived from the Metrecom Skeletal Analysis System scans were compared with each other and with the Cobb angles measured from standing radiographs. RESULTS: The intraclass correlation coefficient (a measure of agreement, ranging from 0 to 1, where 1 represents complete agreement) for the intraexaminer reliability of the Metrecom Skeletal Analysis System ranged from 0.71 to 0.83. The interexaminer reliability intraclass correlation coefficient of the Metrecom Skeletal Analysis System was 0.58, with a mean difference between examiners of 5.5 degrees (SD = 5 degrees), and limits of agreement (mean difference +/-2 SD) ranging from -4.5 degrees to 15.6 degrees. The Metrecom Skeletal Analysis System and the radiographically derived Cobb angle correlation was 0.64, but the mean difference between the methods was 3.7 degrees (SD = 11.1), with limits of agreement from 18.4 degrees to 25.9 degrees. CONCLUSION: The Metrecom Skeletal Analysis System does not provide sufficient clinical precision to substitute for the Cobb angle measured from spinal radiographic measurements in the management of adolescents with scoliosis.     

A vaccine consisting of recombinant Borrelia burgdorferi outer-surface protein A to prevent Lyme disease. Recombinant Outer-Surface Protein A Lyme Disease Vaccine Study Consortium

The purpose of this study was to determine whether infrared tympanic membrane thermometry can replace mercury-in-glass temperatures as an assessment tool for detecting fevers earlier and more reliably in a pediatric oncology outpatient setting. A total of 313 patient visits had infrared tympanic temperatures (obtained by using the LighTouch LTX Pedi-Q thermometer (Exergen; Watertown, MA) and axillary temperatures taken simultaneously (obtained by using mercury-in-glass thermometers). Those patients with a normal axillary temperature and an elevated tympanic measurement of 38 degrees C or higher had a follow-up axillary temperature conducted that evening to determine whether an elevated tympanic temperature predicted on coming fever or infection. The mean tympanic temperature was 37.5 degrees C and the mean axillary temperature was 36.8 degrees C, a difference of 0.7 degree C (P = .0001). Sixty-two (20%) patients had discrepant temperatures with a febrile tympanic and normal axillary measurement. Three (5%) of these patients were febrile at their follow-up axillary reading. Tympanic thermometry resulted in a significantly higher temperature reading than the axillary method. Elevated tympanic temperatures were not predictive of oncoming fever or infection. Fevers were not missed when using the tympanic method. To prevent unnecessary medical intervention, it is recommended that mercury-in-glass thermometers verify elevated tympanic temperatures.     

Psychological health in a non-clinical sample of obese women

PURPOSE: A new semi-continuous thermodilution cardiac output (CCO) system has been developed recently (Opti-Q and Q-vue Abbott critical care system). The aim of this study was to compare the accuracy and reproducibility of this new device with conventional ice-bolus thermodilution cardiac output (BCO). METHODS: Fifteen critically ill patients who needed pulmonary artery catheterization were prospectively investigated. Eighty seven paired data using BCO and CCO methods were compared. Reproducibility was assessed from 90 BCO and 87 CCO determinations by calculation of the mean standard error (SEM) and according to Bland and Altman methodology. RESULTS: The BCO and CCO ranged from 2.46 to 11.20 L.min-1 and from 1.75 to 10.05 L.min-1 respectively. Bias (mean difference between BCO and CCO) was null (0.002 L.min-1, P = 0.98), precision (SD of the bias) was 0.74 L.min-1 and the limits of agreement (mean difference +/- 1.96 SD) ranged from -1.45 to 1.45 L.min-1. The threshold to consider two cardiac outputs as different (3 x SEM) was equivalent for BCO and CCO (0.54 and 0.465 L.min-1 respectively). According to the Bland and Altman method, reproducibility of CCO was greater than that of BCO; bias of repeated measurements of BCO and CCO were 0.15 L.min-1 (P < 0.05) and 0.047 L.min-1 (NS), respectively. CONCLUSION: Compared with BCO, this new device was accurate but cannot be considered as interchangeable regarding the limits of agreement. Reproducibility of CCO was superior to BCO.     

Influence of temperature and activity on the metabolic rate of adult Drosophila melanogaster

We measured metabolic rates of adult male Drosophila melanogaster allowed to evolve in the laboratory at 18 and 25 degrees C and compared these with measurements of metabolic rates of flies collected along a latitudinal gradient in Australia. Metabolic rates of flies that had evolved in the laboratory at low temperature were 5-7% higher than those of flies allowed to evolve at high temperature. Metabolic rates of field collected increased with latitude when measured at 18 degrees C but not at higher temperature (25 degrees C) and were about 9% greater in high latitude (approximately 41'00) flies than low latitude (16'53) flies. Metabolic rate was strongly influenced by measurement temperature; estimated Q10s ranged from 1.79 to 2.5 for measurements made at 18 and 25 degrees C. Metabolic rate scaled isometrically with body mass; the estimated slope of a ln-ln regression of metabolic rate and body mass was 1.03 +/- 0.1. We used our measures of metabolic rate and activity to estimate the minimum cost of transport (MCOT) while walking. The estimates of MCOT have high standard errors (lab, 34.30 +/- 14.2 ml O2/g/km; and field, 38.0 +/- 17.0 ml O2/g/km); however, they differ by only 3-9% from predicted values based on allometric relationships reported in the literature.     

Acoustic streaming: comparison of low-amplitude linear model with streaming velocities measured by 32-MHz Doppler

To further investigate the dependency of fMRI signal changes on echo time TE, we measured T2 and T2* values, obtained from human blood samples at various oxygenation levels and used them in a simple model to calculated signal enhancement in fMRI. In addition, the longitudinal relaxation time T1 of human blood was determined for reference. All measurements were performed at 23 degrees C to reduce blood cell metabolism during the measurement procedure. At 23 degrees C T1 values of 1434 +/- 48 ms for arterial human blood were obtained after correcting for hematocrit content, as hematocrit values ranged fro 28% to 34% only. The T2 relaxation times obtained are 181 +/- 23 ms for venous and 254 +/- 26 ms for arterial human blood, T2* relaxation times corrected for inhomogeneities of the static magnetic field (B0) are 42 +/- 2.8 ms and 254 +/- 32 ms, respectively. Furthermore, absolute and relative signal changes in fMRI experiments are calculated. The results from these model calculations reveal that contrast in fMRI can be optimised by choosing an appropriate echo time.     

Malignant melanoma in Europe: changes in mortality rates (1970-90) in European Community countries

We measured metabolic rates (mL O2 h-1, converted to kcal d-1), deep body temperatures (degree C), and skin temperatures (degree C) and calculated whole-animal thermal conductances (mL O2 g-1 h-1 degree C-1) of five 3-yr-old harbor seals (Phoca vitulina concolor) at air temperatures between -20 degrees and 35 degrees C. The mean thermal neutral zone of these seals extended from a lower critical temperature of -12.9 degrees +/- 1.6 degrees C (SD) to an upper critical temperature of 28.6 degrees +/- 1.7 degrees C. Hyperthermia was observed at an ambient air temperature of 35 degrees C. Mean standard metabolic rate was 1,553.6 +/- 168.2 kcal d-1, about 1.2 times the value expected for adult animals of similar body mass (mean mass = 49.2 +/- 7.5 kg). Mean deep body temperature increased from 37.5 degrees +/- 0.30 degrees C at an ambient temperature of 30 degrees C and reached 39.3 degrees +/- 0.33 degrees C at an ambient temperature of 35 degrees C. Skin temperature decreased with decreasing ambient temperature but remained well above ambient temperature. Mean whole-animal thermal conductance decreased from an ambient temperature of 35 degrees C until it reached a minimum value of 0.007 mL O2 g-1 h-1 degree C-1 at -4.0 degrees C; it then increased with a further decrease in ambient temperature. In comparison to the thermal limits of the same seals during their first year of life, the results indicate a broadening of the thermal neutral zone with age: an 11 degrees C decrease in the lower critical temperature and a 3.5 degrees C increase in the upper critical temperature. These findings suggest that warm ambient air temperatures should not pose any particular thermoregulatory problems for larger and older harbor seals, even beyond the limits of their current annual distribution.     

Clinical trial of an ex vivo arterial blood gas monitor

PURPOSE: The purpose of this study was to test the performance of a patient attached, on demand ex vivo arterial blood gas (ABG) monitor, and to compare the frequency of ABG analysis using the monitor, where the monitor was operated by intensive care unit (ICU) staff on shock trauma and neurosurgical intensive care patients for < or = 6 days, with standard clinical laboratory analysis. MATERIALS AND METHODS: The ABG monitor (SensiCath; Optical Sensors Inc., Minneapolis, MN) incorporates fiber optic pH, PCO2, PO2, and thermistor temperature sensors in a 0.3-mL sensor chamber that attaches in line with the patient's arterial pressure tubing and connects via a fiberoptic cable to a bedside instrument. The monitor and standard clinical laboratory performance were compared following an institutionally approved protocol. Adult ICU patients (n = 30) were studied for whom an arterial cannula was required, the expected ICU stay was > 72 hours, > or = 2 ABG analyses/day were anticipated, and informed consent had been obtained. Paired comparison ABG analyses and quality assurance checks were performed daily. The frequency of ABG analyses in this study, for which monitor values were used for clinical decision making, was compared with the frequency previously reported for the same ICUs, for which the monitor and laboratory results were compared but only the latter were used for clinical decision making. RESULTS: Five hundred ABG analyses, 436 over the first 72 hours, were obtained using the monitor for patient management over 3,248 patient hours (85 +/- 47 hours/patient). Monitor-laboratory comparison ABG analyses (n = 258) indicated stable performance over 6 days: For pH, the range of laboratory measurements was 7.200 to 7.540, accuracy (mean difference between monitor and laboratory measurement) was +0.013, and precision (standard deviation of difference between monitor and laboratory measurements) was +/-0.031. For PCO2, range: 18 to 78.5, accuracy: -0.8, precision: +/-3.4 mm Hg. For PO2, range: 41.0 to 344.0, accuracy: +2.3, precision: +/-12.8 mm Hg. The frequency of ABG analyses obtained using the monitor (ie, 15.0 +/- 11.6 ABGs/patient/72 hours) was significantly greater than that using the clinical laboratory (ie, 8.8 +/- 4.2 ABGs/patient/72 hours) (P = .01). CONCLUSION: The ABG monitor provides performance comparable to standard clinical laboratory analysis for < or = 6 days (< or = 144 hours), consistent with ICU arterial cannula changeout schedules. More frequent ABG analyses are obtained by critical care practitioners using the monitor compared with the clinical laboratory system, suggesting that clinical decision making based on ABG data may be limited by the frequency of ABG analysis.