Methodologic dependence of electroretinogram oscillatory potential amplitudes.

The International Society for Clinical Electrophysiology of Vision (ISCEV) protocol for eliciting oscillatory potentials uses a considerably lower flash intensity and a different preconditioning stimulus than the only oscillatory potential protocol used to predict progression of diabetic retinopathy. To determine if the ISCEV protocol will be useful in predicting progression of diabetic retinopathy, summed oscillatory potential amplitudes were measured by both protocols in a population of diabetics. Summed oscillatory potential amplitudes measured by the ISCEV protocol, although smaller, are highly correlated with the summed oscillatory potential amplitudes measured with the higher-intensity flash. Thus, summed oscillatory potential amplitudes measured with the ISCEV protocol should be useful in predicting outcome in diabetic retinopathy. Different signal processing filters used to extract oscillatory potentials from the electroretinogram waveform have a small, but significant, effect on summed oscillatory potential amplitude. Use of the caliper-square method or the summed peak-to-trough method for measuring oscillatory potential heights had an insignificant effect on measured oscillatory potential amplitude.

Predicting outcome in central retinal vein occlusion using the flicker electroretinogram.

OBJECTIVE: To evaluate the usefulness of the 30-Hz flicker electroretinogram (ERG) in predicting outcome in patients with central retinal vein occlusion using an automated system previously developed to measure parameters of the flicker ERG and using information about the a priori probability of developing neovascularization of the iris in central retinal vein occlusion. METHODS: Amplitude and timing were extracted from 30-Hz flicker ERGs of 76 patients with acute central retinal vein occlusion from a previous study. The ability of the ERG to predict the development of neovascularization of the iris based on derived parameters was measured using the area under the receiver operating characteristic curve. RESULTS: The predictive ability of the 30-Hz ERG amplitude, as judged by a two-alternative forced-choice procedure, was 92%. The corresponding value for implicit time was 89%. Using Bayes' theorem, we computed the probability of developing neovascularization of the iris based on a single 30-Hz flicker ERG measurement at initial patient presentation. CONCLUSIONS: The 30-Hz flicker ERG provides useful information about the patients with central retinal vein occlusion.

The care and fitting of Naka-Rushton functions to electroretinographic intensity-response data.

We developed an automated system to estimate parameters of the Naka-Rushton function based on a heuristic model of the electroretinogram intensity-response series. Data from a population of patients with central retinal vein occlusion were used to examine the ability of the derived parameters to predict the development of neovascularization of the iris. The predictive performance of this automated system in central retinal vein occlusion is comparable to that of a human expert.

The variability of the b-wave of the electroretinogram with stimulus luminance.

We measured the variability of the b-wave of the electroretinogram as a function of stimulus luminance in two young normal individuals. We also estimated b-wave variability by examining residuals from Naka-Rushton curves fit to intensity-response data. The change of variability with amplitude was similar with both techniques. The standard deviation of b-wave amplitude rose with b-wave amplitude at low stimulus intensities. At higher intensities, the standard deviation of b-wave amplitude became constant. The point at which the standard deviation became constant was Log K for the eye, as determined by fitting the data with the Naka-Rushton equation. These changes suggest that the mechanisms underlying the growth of the b-wave with luminance change near Log K.

Pattern discrimination perimetry in patients with glaucoma and ocular hypertension.

We compared the results of the pattern discrimination perimeter to the program 30-2 on the Humphrey Field Analyzer (Humphrey, Inc., San Leandro, California) in 93 consecutive patients with ocular hypertension and glaucoma and 30 control patients. In 20 patients with ocular hypertension, a significantly greater number of glaucomatous defects were noted on pattern discrimination perimetry (ten patients) than on the program 30-2 (two patients) (P less than .05, Wilcoxon signed rank test). The diversity in diagnoses found on pattern discrimination testing was not explained by age, intraocular pressure, refraction, number of glaucoma medicines, race, presence of vascular disease, optic disk status, or pupil size. In contrast, in 73 patients with glaucoma no statistical difference in the severity of diagnoses was noted between perimeters (P greater than .05, Wilcoxon signed rank test). These results suggest the potential value of pattern discrimination perimetry as a visual function test in patients with glaucoma and in defining subsets of patients with ocular hypertension not found with conventional automated perimetry.

Automated estimation of implicit time and amplitude from the flicker electroretinogram.

The electroretinogram (ERG) has been shown to yield sensitive and specific information about the development of neovascularization in ischemic disorders such as diabetic retinopathy and central retinal vein occlusion. However, even though the test is quick, easy to perform, and carries few risks, the ERG is poorly utilized in clinical situations because of the extended training period necessary for test interpretation. We have developed an algorithm that estimates the phase and amplitude of a 30-Hz flicker ERG and that is relatively insensitive to many forms of interference in ERG recordings. These estimates can be directly compared to established criteria for the risk of proliferative retinal disease.

Computerized threshold determination for automated ABO/Rh tests.

Automated microplate blood-grouping systems rely on a pair of thresholds to determine whether reactions are positive or negative. The determination of these thresholds is a critical step in the quality control process for automated microplate blood grouping. A simple method has been developed for automatically setting these thresholds. A more sophisticated method, which required substantial computational complexity, was also tried. The more complex method did not perform better than the simple method described here.

A one-step Du test in microplates.

A one-step Du test, developed for use in automated microplate systems, uses anti-D with 0.6 percent dextran to potentiate the reaction. Because the washing and reagent-adding steps of the antiglobulin test are not required, the Du test can be performed in the same microplate as the ABO/Rh test. A set of reactions prepared with this technique was visually interpreted and also classified by an automated microplate ABO/Rh system. Visual interpretation of reactions resulted in a sensitivity and specificity close to those of the antiglobulin test, although the sensitivity of the test was reagent-dependent. When the automated microplate blood grouping system was used, the test was not as sensitive or as specific as the antiglobulin test, although it may be sufficient for many applications.

Considerations in the use of laboratory robots: aspects of safety and accuracy.

A general-purpose laboratory robot was configured to prepare blood samples for the HBsAg and anti-HIV tests. Several issues were studied in order to eliminate problems and optimize the system. These included pipetting accuracy, intersample carryover, reagent interference, splashing and aerosol generation and user-related issues. A number of general guidelines were developed to assist others in developing and implementing systems for this and similar applications.

The status of automation and data processing in the United States blood banking community.

A survey of the extent of automation and data processing in the blood banking community was conducted by questionnaire. The likelihood that a facility had some type of automation was related to the volume of blood products produced or transfused. Facilitates that collected blood or collected and transfused blood were more likely to have automation for ABO and Rh testing than those that only transfused blood. Automation for many other blood bank tests is unavailable. Many facilities had data processing capability, which is used primarily for accounting.

Cost analysis of routine laboratory testing in blood centers.

The direct costs of ABO, Rh, Du, syphilis, and antibody screening tests were investigated. Questionnaires from 58 blood centers were analyzed to compare cost-effectiveness among methods of testing and annual collection volume as well as differences in cost of the various methods of testing within the same-size center. We found that "expensive" automated equipment cannot be justified on the basis of direct costs in centers processing less than 100,000 units of blood per year and that there is a wide variation in costs among centers using the same equipment.

Application of robotics in blood banking.

To demonstrate the feasibility of using robots in blood banking applications several prototype systems were developed. Activities associated with sample testing and component preparation were examined. In one project, a general-purpose laboratory robot (Zymate Laboratory Automation System, Zymark, Inc., Hopkinton, Mass.) was configured to prepare samples for microplate-based ABO/Rh testing. In a second project, this same robot was configured to carry out specific steps in evaluating bar-coded labels, as part of a quality control procedure. A fluid-handling robot (Sampler 505, Tecan AG, Hombrechtikon, Switzerland) was used to prepare the dilution of serum samples for the evaluation of an anti-HTLV-III test kit. It was then configured to aspirate, dilute and dispense samples for anti-HTLV-III and HBsAg testing. This robotic system is now in field trial. The use of large industrial robots for automating component production was also considered. The key element in this design was the development of a fixture that would hold the blood bag set during the balancing, centrifugation and expressing steps. However, a fixture which was capable of performing these operations and that was equivalent in size and adaptable to a standard centrifuge bucket could not be fabricated.

A robotic system to prepare samples for HTLV-III testing.

A robotic handling system was adapted to perform the sampling and dilution steps needed in an assay to detect antibodies to the HTLV-III virus, the causative agent of AIDS. The system reduced the labor required to prepare the samples and provided standardization and accuracy in the preparation of the samples.

Estimating medullary chemoreceptor blood flow from ventilatory-CO2 response transients: theory and data from anesthetized dogs.

We examined the tracer wash-in technique for measuring blood flow for a specific experimental model in which medullary chemoreceptor blood flow is estimated using carbon dioxide as the tracer. Using a mathematical model, we calculated the flow values that would be estimated when blood flow is a function of the tracer concentration, CO2. The estimate was compared with three indices of time-varying blood flow: the steady-state blood flow before and after the step change of CO2 and the time-average blood flow. Estimated blood flow was found to be most closely related to the final value of flow. Perfusion estimates from five pentobarbital anesthetized dogs were 102 ml/100 g/min (S.D. 35) at an average of PaCO2 of 60.7 Torr.

Automated determination of ABO/Rh in microplates.

A semi-automated system for determining the ABO group and Rh type of blood samples has been developed using a commercially available automated microplate (ELISA) reader and a microcomputer. Optimization of serologic, measurement and interpretation parameters was accomplished without significantly changing an existing manual procedure. The first pass noninterpretation rate of this system in the laboratory prior to field trials is 7.1%. A commercial system of this type should be cost-effective as a primary instrument for small to medium sized blood centers and transfusion services.

Interaction of chemoreceptor effects and its dependence on the intensity of stimuli.

We have studied the way in which carotid and medullary chemoreceptor effects combine to control respiration in pentobarbital anesthetized dogs. The carotid sinuses were autoperfused from a membrane oxygenator system to produce either normoxic or hypoxic-hypercapnic blood without producing brain stem hypoxia. The dogs breathed CO2 in O2, N2 mixtures regulated to produce isocapnic levels of arterial partial pressure of CO2 (PaCO2) of 36, 54, and 68 Torr at PaO2 = 87. Tidal volume (VT) increased (P less than 0.05) from a control value of 317 ml to 545 at PaCO2 = 54 but did not increase further at PaCO2 = 68. Carotid chemoreceptor stimulation increased VT to 432 ml at PaCO2 = 36 (P less than 0.05) but did not increase VT at the higher levels of PaCO2. Thus, VT has reached a maximum at PaCO2 = 54 Torr and was not increased further by either chemoreceptor drive. Breathing frequency (f) increased (P less than 0.05) from an eupnic value of 20.2 breaths/min to 35.5 and 41.3 at PaCO2 = 54 and 68, respectively. Carotid chemoreceptor stimulation increased (P less than 0.05) f by 7.7, 5.2, and 5.0 breaths/min at each of the levels of PaCO2. Tidal volume exhibited a less-than-additive (P less than 0.05) combination of chemoreceptor effects but f was not significantly different from additive. We conclude that, in pentobarbital-anesthetized dogs, f continues to increase at chemoreceptor drives beyond those at which VT reaches a maximum. This VT nonlinearity may be responsible for the different interactions found in these experiments.