WO2015148387A2 - Procédés de diagnostic et dispositifs permettant de contrôler une glycémie aiguë - Google Patents

Procédés de diagnostic et dispositifs permettant de contrôler une glycémie aiguë Download PDF

Info

Publication number
WO2015148387A2
WO2015148387A2 PCT/US2015/022031 US2015022031W WO2015148387A2 WO 2015148387 A2 WO2015148387 A2 WO 2015148387A2 US 2015022031 W US2015022031 W US 2015022031W WO 2015148387 A2 WO2015148387 A2 WO 2015148387A2
Authority
WO
WIPO (PCT)
Prior art keywords
glycemia
lorenz
blood glucose
acute
quantile
Prior art date
Application number
PCT/US2015/022031
Other languages
English (en)
Other versions
WO2015148387A3 (fr
Inventor
James M. Minor
Original Assignee
Cannon, Alan, W.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Cannon, Alan, W. filed Critical Cannon, Alan, W.
Publication of WO2015148387A2 publication Critical patent/WO2015148387A2/fr
Publication of WO2015148387A3 publication Critical patent/WO2015148387A3/fr

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H15/00—ICT specially adapted for medical reports, e.g. generation or transmission thereof

Definitions

  • This invention relates generally to the field of diagnostic processes, devices, and chaos physics for identifying, monitoring, and controlling acute glycemia.
  • Chronic glycemia commonly denoted AG, is the extreme persistent level of blood glucose (chronic BG) as measured by the AlC-assay and is related to diabetes progression and complications.
  • Acute glycemia is likely when chronic BG plus its acute volatility creates a significant presence at extreme critical levels clinically defined as hypo/hyper-glycemia. Anticipating such acute glycemia is a key factor in both the prevention of sudden serious/dangerous conditions as well as the clinical management of diabetes healthcare as described in established medical reference books (1,2).
  • Acute glycemia (hypo/hyper-BG levels) causes serious consequences to both patient and society in terms of serious injury and healthcare complications due to disease progression and disasters such as transportation/home accidents.
  • attacks of acute glycemia cannot be usefully predicted due to the chaotic nature of BG.
  • a device for monitoring glycemic levels in a patient including: a housing; a processor coupled to memory; an interface for inputting to the processor; a display for displaying results of processing by said processor; and a port configured to receive a blood glucose monitoring strip that is used to take a blood sample from the patient, and from which the processor calculates a blood glucose value; wherein the processor, interface and port are contained in the housing, and wherein the interface is mounted in the housing; and wherein the processor is configured to: calculate a plurality of the blood glucose values from a plurality of blood samples taken from a plurality of the blood glucose monitoring strips; calculate a quantile of the blood glucose values taken over a period of days; create a Lorenz plot for the blood glucose values in the quantile; and display the Lorenz plot.
  • the monitoring is real-time monitoring.
  • the displayed Lorenz plot can be viewed by a user to identify days in which acute glycemia was experienced by the patient.
  • the processor identifies days in which acute glycemia was experienced by the patient, based on values in the Lorenz plot, and displays days in which the acute glycemia was experienced.
  • the processor calculates orbits for the quantile, and applies a quantitative model to the orbits to identify probability trends for acute glycemia,
  • the quantitative model comprises a single metric model.
  • the quantitative model comprises a combined metrics model.
  • the quantitative model comprises logistic regression based on functions of Lorenz factors from the Lorenz plot.
  • the quantitative model comprises calculation of a Lorenz Response Surface (LRS).
  • LRS Lorenz Response Surface
  • a system for monitoring glycemic levels in a patient includes: a processor coupled to memory; and an interface for receiving input to and outputting from the processor; wherein the processor is configured to: receive or calculate a plurality of the blood glucose values from a plurality of blood samples taken from a plurality of the blood glucose monitoring strips; calculate a quantile of the blood glucose values taken over a period of days; create a Lorenz plot for the blood glucose values in the quantile; and output the Lorenz plot.
  • the processor receives the blood glucose values from a blood glucose monitoring device.
  • the processor identifies days in which acute glycemia was experienced by the patient, based on values in the Lorenz plot, and outputs days in which the acute glycemia was experienced.
  • the processor calculates orbits for the quantile, and applies a quantitative model to the orbits to identify probability trends for acute glycemia,
  • the quantitative model comprises logistic regression based on functions of Lorenz factors from the Lorenz plot.
  • the quantitative model comprises calculation of a Lorenz Response Surface (LRS).
  • LRS Lorenz Response Surface
  • a method of monitoring glycemia includes: receiving or calculating inter-day blood glucose values from blood samples taken from a patient on specified days; calculating a quantile from the inter-day blood glucose values; calculating orbits for the quantile; creating a Lorenz plot for the quantile; and identifying at least one of: one or more days during which extreme glycemia was experienced for the quantile, and prediction of when extreme glycemia is expected to be experienced for the quantile.
  • the extreme glycemia is acute hypoglycemia or acute hyperglycemia.
  • the method further includes applying quantitative model to the orbits to identify probability trends for extreme glycemia.
  • the quantitative model comprises logistic regression based on functions of Lorenz factors from the Lorenz plot.
  • the quantitative model comprises calculation of a Lorenz Response Surface (LRS).
  • LRS Lorenz Response Surface
  • FIG. 1A-1B show synchronized similarity of CGM and SMBG Quantiles using different Lorenz variables from CGM and SMBG profiles, according to an embodiment of the present invention.
  • Fig. 2 is a Lorenz dynamic flow map for 2% Quantiles of intraday CGM readings, according to an embodiment of the present invention.
  • Fig. 3 shows a diagram of Lorenz geometry invented for a diagnostic that predicts acute glycemia, according to an embodiment of the present invention.
  • Fig. 4 illustrates an example of centroid tracking according to an embodiment of the present invention.
  • Fig. 5 illustrates Lorenz dynamic flow geometry for Q02 according to an embodiment of the present invention.
  • Fig. 6 illustrates Lorenz dynamic flow geometry for Q98 according to an embodiment of the present invention.
  • Fig. 7A-7B illustrate Lorenz dynamic flow geometry for predicting acute glycemia, according to an embodiment of the present invention.
  • Fig. 8 shows exemplary design of experiment values used for calculating
  • Lorenz response surfaces according to embodiments of the present invention.
  • Fig. 9 shows a diagram of tri-section evaluation for acute risk models, according to an embodiment of the present invention.
  • Fig. 10 shows results from using Q02 data to predict next-day acute glycemia risk Q2 , according to an embodiment of the present invention.
  • Fig. 11 shows results from using Q98 to predict next-day acute glycemia risk Q98 .
  • Fig. 12 a fast blood glucose (FB) model for predicting next day Q02 acute glycemia risk, according to an embodiment of the present invention.
  • FB fast blood glucose
  • Fig. 13 a fast blood glucose (FB) model for predicting next day Q02 acute glycemia risk, according to an embodiment of the present invention.
  • FB fast blood glucose
  • Figs. 14A-14B show Lorenz data plots of synchronized chaos between Q02 and FB, according to an embodiment of the present invention.
  • Fig. 15 illustrates an FBQ02 model using combined metrics to predict next day extreme Q02, according to an embodiment of the present invention.
  • Fig. 16 illustrates an FBQ98 model using combined metrics to predict next day extreme Q98, according to an embodiment of the present invention.
  • Fig. 17 shows a diagram of Lorenz dynamic flow geometry invented for low risk DOX to predict extreme BG levels, according to an embodiment of the present invention.
  • Fig. 18 shows low risk Q02 predictive data according to an embodiment of the present invention.
  • Fig. 19 illustrates a low risk Q02 single metric model, according to an embodiment of the present invention.
  • Fig. 20 shows low risk Q98 predictive data according to an embodiment of the present invention.
  • Fig. 21 illustrates a low risk Q98 model according to an embodiment of the present invention.
  • Fig. 22 is a Lorenz two-metric vector diagram according to an embodiment of the present invention.
  • Fig. 23 illustrates a low risk FBQ98 model using combined metrics, according to an embodiment of the present invention.
  • Fig. 24 illustrates a low risk FBQ02 model using combined metrics, according to an embodiment of the present invention.
  • Fig. 25 is a schematic illustration of a blood glucose monitoring device according to an embodiment of the present invention.
  • Fig. 26 is a block diagram illustrating components of a diagnostic system for use on a patient, according to an embodiment of the present invention.
  • Fig. 27 shows events that may be carried out for predicting acute glycemia in a living body according to an embodiment of the present invention.
  • BG blood glucose
  • All blood glucose (BG) units described herein are blood-fluid concentration units, typically mg/dL (milligram/deciliter), unless noted otherwise.
  • the "daily BG mean” refers to a full-day average of BG levels provided by a high rate blood sampling device (continuous glucose monitoring (CGM) device) attached to a patient, typically sampling every 5 to 10 minutes.
  • CGM continuous glucose monitoring
  • the daily mean for each day of serial multiple days in chronological order form a series of inter-day daily means, not necessarily consecutive.
  • Chronic BG is the diabetic persistent level of BG as evaluated by a weighted time average over an extended period of multiple days, typically spanning multiple weeks. This is equivalent to the average of the CGM daily means over the same set of days.
  • Intraday BG levels are device readings of BG at specific times during the day, typically specified by self-monitoring-blood-glucose (SMBG) protocol or CGM sampling rates. These intraday BG levels spanning multiple days in chronological order form a summary series of daily BG events. Patients use the SMBG-protocol when measuring their BG level by skin stick pens of metering devices. The typical seven SMBG-defined events are before and after breakfast, before and after lunch, before and after dinner, and bedtime, and sometimes augmented by before and after snacks.
  • SMBG self-monitoring-blood-glucose
  • An "inter-day diabetes state" space is defined by the day-to-day ( inter-day) dynamics of BG daily means, specific intraday events, and other daily metrics.
  • Phase plots or "phase portraits” are the time patterns of data points created by the serial values of inter-day state-space variables as coordinates.
  • time patterns form an ordered structured (deterministic) flow known as an "orbit" according to the geometry of an attractor basin.
  • the "attractor basin” is defined by all possible orbits whose initial phase point is in the attractor basin.
  • the attractor basin for inter-day daily mean is distant from the attractor basin of another inter-day metric by a "vector constant", specific to each patient, which slowly changes over a year.
  • Inter-day metrics are derived from intraday readings. References by the popular term “daily” are ambiguous.
  • the "International Diabetes Center” located at Park Nicollet International Diabetes Center, 3800 Park Nicollet Blvd., St. Louis Park, MN, and founded by Donnell D. Etzwiler, MD, in 1967, provides world-class diabetes care, education, publications and research that meet the needs of people with diabetes and their families.
  • "Acute glycemia” is defined as hypoglycemia for recurrent extreme low BG levels typically ⁇ 71 mg/dL and/or hyperglycemia for recurrent extreme high levels typically BG>249 mg/dL.
  • ROC Receiveiver operating characteristic
  • AUC Average-under-the -curve
  • Synchronization function is a simple combination of specific intraday BG readings that monitors AlC-related chronic BG with high accuracy.
  • An acute-glycemia day is defined as "recurrent hypoglycemia" or
  • N is the count of monitored days in the k-day span.
  • the kinetics and lifespan of glycated red cells imply k equals -90 days, as used in the present invention.
  • X from day d-k+1 up through day d is defined as:
  • N n 0 , (2) where N is the sum of weights of monitored days (md) in the k-day span, and where where Wd-n >0 weights the historical impact of aging glycated red cells. Hence, recent glycations have more impact than older HbAlC cells. W d is designed in accordance with the relative importance of glycation history. Consequently, BG levels from the most recent 30 days have been shown to contribute approximately 50% to HbAlc, whereas those from the period 30-90 days and 90-120 days earlier contribute approximately40% and 10%, respectively.
  • k approximates -90 to 120 days, but the last 30 days have little impact while the first 90 has essentially major and uniform impact as expressed in the definition for "approximate glycemic average/mean”.
  • a "Lorenz" variable pair is a dynamic variable and its embedded (time-delayed) version and is useful for analysis and analytics of a dynamic state space.
  • a "profiler" graphs the impact of known factors on the probability of defined events/classes.
  • FIGs. 1 to 26 illustrate the state-space attractor orbit properties, statistical analyses, and information systems supporting the New Diagnostic that predicts acute glycemia. All statistical analyses and contingency tables are standard methods performed and described by JMP® 8.0.2 provided by SAS Institute Inc., located at SAS Campus Drive, Cary, NC, USA 27513.
  • BG blood glucose
  • BG blood glucose
  • the intraday distribution of BG can be divided into population quantiles Q%.
  • Q10% is the BG level capturing the first (lower) 10% of intraday BG readings.
  • the intraday recurrence time of BG levels at or below Q10% is approximately 0.10 x 24 hours.
  • Extreme population levels are captured by extreme quantiles.
  • Q02% to monitor potential hypoglycemic episodes and Q98% for potential hyperglycemic episodes. Both of these hypoglycemic episodes and hyperglycemic episodes quantiles are classified as acute glycemia.
  • CGM devices Intraday populations are collected by CGM devices or SMBG profiles.
  • the sampling rates of CGM devices are obviously superior to SMBG events, but currently the SMBG protocol is much more prevalent among patients.
  • CGM data provide quantiles directly.
  • step 1 Approximation of quantile (Q) values from SMBG profiles requires two steps: (1) a normalizing transformation of BG readings; and (2) use the normal approximation to calculate quantiles.
  • step 1 the Ln transformation, log b (BG), where the base b can be any value, typically "e”.
  • step 2 a mean value "m” and standard deviation value "s" of the Ln values are calculated.
  • the Q02 and Q98 quantiles are estimated as m- 2.08s and m+2.08s respectively.
  • Fig. 1 shows that both data sources provide similar though not identical quantile information, with the SMBG data values 10 for the Q02 quantile being shown in Fig. 1A and the CGM data values 12 for the Q02 quantile being shown in Fig. IB.
  • Fig. 2 is a Lorenz phase plot 34 of 2% quantile BG readings (Q02%), future versus prior, that shows the inter-day orbits and geometry that support the present invention.
  • Typical inter-day phase-portrait patterns for 2% quantiles (Q02) of intraday CGM distributions as shown are synchronized with other inter-day BG metrics such as the intraday fasting BG levels (FB) that were described in detail in co-pending application no. 13/895,054, filed 05/15/2013, which is hereby incorporated herein, in its entirety, by reference thereto.
  • the ordered flow patterns (orbits) described in the definitions are characteristic of the deterministic phase dimensions discovered by the prior application no. 13/895,054.
  • phase portraits are contained within the circular/ellipsoidal attractor-basin geometry of the diabetes inter-day state space discovered in the prior invention.
  • the diameter of these orbits reflects the degree of inter-day metric fluctuations (scatter).
  • ANOVA analysis of variance
  • the orbits 35 are the consequence of the endocrine system attempting control of BG facing the challenge of diabetes, by analogy similar to the planetary orbits created by gravity.
  • Orbit-point labels e.g., see 12, 13, 14, 15, 17, 18, 19 , 22-24, and 29 in Fig. 2 indicate monitored days (mdays) in chronological order as distributed over a fixed interval of time.
  • the Lorenz pair now-day quantile is located by the horizontal X axis 36 and next-day quantile is on a vertical Y-axis 37 grid. Note in Fig. 2 the orbits 35 flow about the attractor centroid 38, approximated as (90, 90) mg/dL.
  • a complete 360-degree non-closed orbit 35 cycle requires typically 6 to 8 days.
  • Some days may be missing (for example, empty circle 16 in Fig. 2) in the sequence but the orbit structure is evident.
  • the orbits flow clockwise with the occasional interruptions (noise blips).
  • low next-day values can imply acute glycemia.
  • point 17 would predict a hypoglycemic next day, as verified by day 18. Therefore, one can graphically monitor BG trends for acute alerts using CGM or SMBG intraday data.
  • a user when viewing the plot 34 of Fig. 2 can readily identify that the person being monitored had acute hypoglycemia (blood glucose (BG) value of less than 70mg/dl) for at least 28 minutes on day 18, Monday (2% of 24 hours is 28 minutes).
  • acute hypoglycemia blood glucose (BG) value of less than 70mg/dl
  • BG blood glucose
  • QOl, Q04, or any other preselected quantile it can likewise be determined when a patient has acute hypo or hypoglycemia for at least the amount of time represented by each respective quantile.
  • the Q98 quantile would show days that a patient experienced acute hyperglycemia for at least 28 minutes on any particular day.
  • the labels also contain lifestyle information in terms of work day versus free days (weekend). Consequentially, Fig. 2 indicates for this subject that the weekend lifestyle increases his/her hypoglycemic risks.
  • Fig. 3 diagrams the basin geometry of the attractor indicating its centroid 38 (center of attraction) and radius 40.
  • the horizontal X-axis 42 is the current-day axis while vertical Y-axis 44 is the next-day axis.
  • the large circle represents the attractor region 46 as centered on its trending centroid 38.
  • the solid horizontal lines 48 represent acute glycemia thresholds.
  • the two dotted next-day vectors 50 indicate the hypoglycemic/hyperglycemic directions.
  • the small solid dot 52 is a typical orbital point indicating the next point in the flow direction of the small arrow 54. Negative angular flow is clockwise rotation.
  • Fig. 4 illustrates tracking of the shifting of centroid location over time (centroid tracking-small squares) using the algorithms defined above with regard to approximate glycemic average/mean, exact glycemic weighted average/mean and modification of glycemic average defined above, with k set to 6 to cover 7 days typical of complete cycles.
  • TheQ02 values are shown as the large circles 39 and the centroid values are shown as the small squares 38.
  • the centroid 38 is updated over complete weekly cycles.
  • the centered Lorenz variables in Fig. 5, CGMQ02 are data vectors minus their trending centroids. Hence, they are plotted relative to stationary (0,0) coordinates, avoiding the complexity of trending centroids, thereby de-trending the data. Hence, they plot all combinations of Q02 relative to their respective centroids.
  • Fig 6 indicates hyperglycemic Lorenz pairs as big dots 41.
  • a CGM device or comprehensive SMBG strategy can be initiated to evaluate the BG quantiles, such as Q02 near the right horizontal axis or Q98 near the left horizontal axis.
  • ⁇ the metric angle
  • the Q-centroid denoted C Q
  • C Q (sin «9)e BOW +cos «9)e Bex( ) / (cos «9) + sin «9)) ) (4)
  • Q can be either Q98 or Q02 .
  • the centroid tracking formula (4) can be used to track the centroid of any metric used.
  • the angles in the orbital positions of any metric plotted about their centroids tend to show increasing risk of extreme acute glycemia (including acute hypoglycemia and hyperglycemia) with increases counterclockwise rotation of the metric value from the horizontal axis (Fig. 5).
  • BG metric and model may be used.
  • a simple, parsimonious linear model is used for next-day probabilities of quantile levels using functions of factors that locate the attractor centroid and vector position within the attractor basin of active dynamic variables in terms of Lorenz coordinates (as defined above), e.g., functions of trending centroids 38, radial distance 40, and hypo/hyper angular positions respectively.
  • Lorenz coordinates as defined above
  • theta is the Lorenz angle relative to defined references in Fig 3 and delta shifts the max/min locations of the cosine function away from 0 and pi. Delta is optimized by nonlinear fit to the data.
  • b and a can contain linear functions of factors independent of ⁇ .
  • the a/b function version is optimized by linear regression.
  • another linear representation is a quadratic polynomial of ⁇ , which has a maximum zone and a minimum zone with discontinuity at ⁇ 2 ⁇ radians. This polynomial can also interact with other factors and is optimized also by linear logistic regression
  • Figs. 8 and 9 introduce both the Lorenz Response Surface (LRS-polynomial) and Design of Experiment (DOX) methods.
  • LRS-polynomial Lorenz Response Surface
  • DOX Design of Experiment
  • a logistic Lorenz Response Surface can be defined as:
  • Prob of extreme Q-levels exp(linf) (5), where linf is a linear combination of any of the following 11 polynomial Lorenz factors :
  • metric_LC becomes FB_LC.
  • MO is a constant
  • Angle is its clockwise angle
  • LC is the Ln Centroid
  • Lr is the Ln radial distance r.
  • the radial distance r may also be used directly.
  • C is denoted mn7 in the specific models; and ":" indicates a logistic regression variable. Appending [: Subject] to a factor indicates subject differences of factor impact.
  • the clockwise angle is negative from either the left (hypo) or right (hype) horizontal axis 60, depending on model target for acute risk.
  • the left horizontal axis 60A is the zero-angle reference for hypoglycemia and the right horizontal axis 60B is the reference for hyperglycemia, as illustrated in Fig. 7.
  • the earliest effective test zone is expected to be in general at - ⁇ radians. However, for subjects already in a severe risk condition this angle will be nearer to the vertical axis 62 located at - ⁇ /2 radians.
  • the polynomial logistic model can be transformed to the Cos(angle + ⁇ ) representation as follows:
  • Fig. 8 a design of experiment (DOX) evaluation 70 is described according to an embodiment of the present invention.
  • the data in Fig. 8 show that this linear function (LRS function that linearly combines up to 11 terms as described above are optimized by regression fit of the data) is effectively trained using at least 2 orbit cycles. Extreme Q levels are inherently in these cycles. Hence, using fitting concepts such as generalized least squares, the function can be tuned to predict the desired probabilities. JMP*SAS approximates this fitting process using logistic regression to model acute risk in terms of Lorenz factors.
  • Angular functions are either trigonometric functions or polynomials. Using the SAS profiler, these regression functions identify important angles for next-day predictions. Cross terms of these functions with other factors identify how these important angles may shift with variation in these other factors.
  • the model is evaluated by predicting next-day extreme quantile BG levels in terms of truth/contingency tables and by the impact of Lorenz factors.
  • Nominal logistic regression measures the impact of Lorenz factors by statistical significance and profiler analyses.
  • the significance (“Prob>ChiSq" «0.05) of the likelihood ration (L-R) evaluates the importance of each term in the model.
  • the G-efficiency score 72 of 79.90542 reflects that the precision of the model is very good, as the G-efficiency score is set to a scale of 0 to 100, where a 100 score is perfectly precise.
  • the prediction date must be within 2 or 3 days of the prior (predictor) date. 780 monitored days distributed over several months satisfied this restriction. For specific models missing parameter values can reduce this number to between 780 and 770. For special models, the number of days were reduced by design.
  • the truth table, section (1) in fig 9 reveals the total of days used by each model. CGM provides readings every 5 or 10 minutes per day (typically 300 to 600 values per day). SMBG events typically are before and after meals and snacks as well as bedtime (typically 6 to 7 per day). For example fasting BG is before breakfast and is the most typical event among diabetes subjects.
  • Special Fig. 9 diagrams the tri-section analysis for all models showing: (1) truth table 80, (2) significant factors 82, and (3) profiler graph 84.
  • the truth table 80 counts correct and incorrect model predictions.
  • the significant factors 82 show the importance of model terms where prob>chisq is less than 0.05.
  • the profiler graph 84 uses model profiles relating acute risk to Lorenz factors, typically the angle theta ( ⁇ ).
  • Figs. 10 to 13 report tri-section evaluations using quantiles or fasting BG (FB) to predict next-day extreme acute quantiles.
  • BG fasting BG
  • All of the examples of Figs. 10-13 show excellent model performance since, truth tables 80 indicate 90% correct predictions; and significant factors chart 82 shows that most of the terms are significant (important).
  • the profiler charts 84 have a profile alert angle generally at about -3 radians for acute glycemia.
  • the present invention relies on the property of synchronized chaos (which states that all metrics tend to be at the same orbital angles over time) to combine FB angle monitoring with strategic quantile sampling.
  • FB readings provide Lorenz angles.
  • patient should take distribution info, i.e., by CGM device or full SMBG profiles.
  • the angle factors are calculated from FB readings. All other factors are calculated from distribution Q levels.
  • Figs. 14A-14B show Lorenz data plots of synchronized chaos between Q02 1410 and FB 1420, according to an embodiment of the present invention. Although not identical, it can be observed that the small dots 1412, 1422 show similar pattern placement. Likewise the large dots 1414, 1424 show similarities in location on the plots.
  • Figs. 15 and 16 report tri-section (80, 82, 84) evaluations using synchronized chaos to combine FB angle monitoring with strategic quantile sampling. Using fig 9, Figs. 15 and 16 verify the success of this approach to predict next-day extreme quantiles, as all sections 1, 2, and 3 report very good results. Section 80 reports greater than 92% prediction success for 760 to 773 monitored days. Section 82 shows most Lorenz terms to highly significant, and section 84 supports the critical angle for predicting next-day glycemia to near minus three radians as expected.
  • Fig. 17 shows a strategy using extreme such as mild or moderate but not necessarily acute quantiles to train the model or graph-evaluator to be an alert system for acute glycemia.
  • extreme non-acute Q levels would be Q02>60 and Q98 ⁇ 250 mg/dL.
  • the diagram 100 applies to Q monitoring with focus on data near the axes.
  • Figs. 18 to 21 evaluate this strategy and show that the training/alert combination steps work very well for Q02 and Q98 using models/graphs based on a single metric.
  • the additional bar graphs show how well this special strategy extrapolates to predict serious acute events as an alert system. Note this result verifies the ability of the model to predict acute risk beyond the coverage of its (DOX) training data.
  • Figs. 18 and 19 use Q02 distributions only near -3 radians (84) to predict next-day extreme low Q02 based on 418 days of Q02 data.
  • the model requires only three significant terms to fit this restricted data (82)
  • the truth tables (80) show perfect prediction for extreme lows and >90 correct predictions and alert extrapolations.
  • the alert bar chart 86 shows the correct alerts and the three missed alerts (dark shades 87).
  • Figs 20 and 21 show the same results for Q98 alerts using predictor data near -3 radians (84) within 546 days of Q98 data.
  • Chart (82) indicates 4 significant model Lorenz terms.
  • Truth table (80) indicates >98 correct predictions on predictions and alert extrapolations.
  • the alert bar chart shows one missed alert 87 (dark shaded).
  • Fig. 22 is a vector diagram 120 relating two metrics and their synchronized property to the Q-centroid 138 (C Q ) instead of using the tracking definitions described above.
  • C Q Q-centroid 138
  • Q is the known distributional quantile
  • M is the metric centroid for Q
  • FB is the popular fasting metric.
  • Theta estimated by FB and FB tracking is the proper counterclockwise geometric angle relative to the horizontal axis.
  • Subtracting 2pi converts it to the clockwise rotations used in the Lorenz diagrams.
  • the vector relation doesn't care about angle versions.
  • the small circle captures FB orbits that provide the angle.
  • the large circle captures Q orbits; however, the Q centroid M is calculated by CQ (as in equation (6) above.
  • Figs. 23 and 24 show results from using such combined metric factors and sub-acute (non-acute defined above) extreme Q levels to develop a model system for prediction of impending episodes acute glycemia.
  • This relates to the strategy depicted by Figs. 17 to 21.
  • Fig. 23 is like Fig. 21 and Fig. 24 is like Fig. 18 except FB data indicates Lorenz angle near -3 for evaluating Q98 properties using a COM device; thereby significantly reducing patient data burden.
  • Fig. 9 guidance and the alert-based histograms, this limited-Qdata method also works extremely well.
  • FIG. 25 is a schematic illustration of a blood glucose monitoring device 2500 according to an embodiment of the present invention.
  • Device 2500 includes a main body 2516 that houses one or more processors 502 (also referred to as central processing units, or CPUs) that are coupled to memory, including at least one of primary storage 2506 (typically a random access memory, or RAM) and primary storage 2504 (typically a read only memory, or ROM).
  • processors 502 also referred to as central processing units, or CPUs
  • primary storage 2506 typically a random access memory, or RAM
  • primary storage 2504 typically a read only memory, or ROM
  • primary storage 2504 acts to transfer data and instructions uni-directionally to the CPU and primary storage 2506 is used typically to transfer data and instructions in a bi-directional manner Both of these primary storage devices may include any suitable computer-readable media such as those containing instructions for carry out the algorithms and procedures described above for predicting acute glycemia.
  • Storage 2504 and/or 2506 may be used to store programs for processing blood glucose data and storage 2506 may be used to store data such as blood glucose values from blood glucose readings/samplings, and results from processing the blood glucose values from the blood glucose readings.
  • An interface 2510 at least part of which may be operated by a user to carry out various operations, such as changing modes of values displayed on the display 2514 is provided. Also interfacing with the processor 2502 is a port that receive a blood glucose monitoring strip that is used to take a blood sample, and from which the processor 250 calculates a blood glucose reading in a manner known in the art.
  • the device 2500 may be connectable to a network, such as the Internet, and/or to a private network, such as by WiFi, Bluetooth, or the similar types of connectivity. This connectivity can be used to export blood glucose readings data and/or data used to predict acute glycemia achieved by processing according to techniques described above.
  • embodiments of the present invention further relate to computer readable media or computer program products that include program instructions and/or data (including data structures) for performing various computer- implemented operations.
  • the media and program instructions may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well known and available to those having skill in the computer software arts.
  • Examples of computer-readable media include, but are not limited to, magnetic media such as hard disks, floppy disks, and magnetic tape; flash drive devices, optical media such as CD-ROM, CDRW, DVD-ROM, or DVD-RW disks; magneto-optical media such as floptical disks; and hardware devices that are specially configured to store and perform program instructions, such as read-only memory devices (ROM) and random access memory (RAM).
  • Examples of program instructions include both machine code, such as produced by a compiler, and files containing higher level code that may be executed by the computer using an interpreter.
  • Fig. 26 is a schematic illustration of a typical AG diagnostic system 2600 that may be used to perform procedures described above.
  • the monitoring device 2500 provides serial BG values to be processed by the techniques described herein for predicting acute glycemia.
  • the processing may be carried out in the processor(s) 2502 of the device, or, alternatively, the BG values can be outputted from the monitoring device 2500 to an external computer 2602 that can perform the processing. Further alternatively, if the device 2500 processes the BG values, the results from processing can be outputted to an external computer 2602 for storage and/or further output such as by printing, or can be outputted directly to a printer for printing.
  • Outputting can be wireless, such as described above, or wired, either by direct connection 2604 to the computer or printer 2602, or over a network 2606, such as the Internet or other network.
  • Time series and/or statistical analysis of the geometric flow patterns in the BG phase plots provide the critical information on the diabetic status and their trends for each patient and hence prediction of acute glycemia.
  • Fig. 27 shows events that may be carried out for predicting acute glycemia in a living body according to an embodiment of the present invention. Processing of events are carried out by one or more processors of an apparatus according to an embodiment of the present invention., one or more or all of which may be incorporated into a handheld personal blood glucose monitoring device 2500, such as by modifying a commercially available blood glucose monitoring device such as the ONE TOUCH ULTRA 2 Blood Glucose Monitor (Lifescan, Inc., Milpitas, California) or any other commercially available blood glucose monitor.
  • ONE TOUCH ULTRA 2 Blood Glucose Monitor Lifescan, Inc., Milpitas, California
  • some or all events can be carried out by one or more processors located externally of a blood glucose monitoring device (e.g., external computer 2602), using blood glucose readings taken from the blood glucose monitoring device 2500 or from multiple glucose monitoring devices or clinical blood monitoring instrumentation. Still further alternatively, these events can be performed by non-portable apparatus including blood monitoring capability, such as apparatus in a hospital, lab or doctor's office.
  • intra- day and inter-day blood glucose measurements can be made at specified times.
  • the system may receive blood glucose values from an external source.
  • intra-day and inter-day blood glucose values are either received or calculated from the blood glucose values.
  • the blood glucose values may be normalized.
  • At event 2706 at least one quantile of the blood glucose values are calculated.
  • orbits are calculated for at least one specified quantile.
  • orbits are calculated for at least two quantiles preselected as being representative of potential acute hypoglycemia and potential acute hyperglycemia, respectively, such as Q02 and Q98 quantiles, or other preselected quantiles.
  • a Lorenz plot is created for at least one specified quantile, each of which may be a predetermined quantile.
  • the Lorenz plot(s) can be displayed on the display of a device, such as on display 2514 or on the display of another system, such as an external computing device monitor.
  • the Lorenz plot(s) can be outputted to a printer where they can be printed out and/or exported for review by other parties, or storage and/or can be stored internally in the device that created the plot(s).
  • this invention is a real-time device.
  • the Lorenz plot(s) can be read/reviewed by a user to visually identify days exhibiting any of extreme glycemia,, acute hypoglycemia and/or acute hyperglycemia.
  • the system can apply one or more of the quantitative models described herein to determine probability trends for acute glycemia.
  • the system identifies probability values greater than a predetermined probability value as values associated with days exhibiting extreme glycemia.
  • the system identifies probability values greater than clinical values for acute glycemia and identifies those days considered to include extreme glycemia for the quantile considered.
  • the system outputs results which may include identification of days of extreme glycemia, day of acute glycemia, and or predictions as to when one or more days including extreme glycemia and/or acute glycemia are expected in the future.

Landscapes

  • Medical Informatics (AREA)
  • Engineering & Computer Science (AREA)
  • Public Health (AREA)
  • Health & Medical Sciences (AREA)
  • Pathology (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Biomedical Technology (AREA)
  • Epidemiology (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Investigating Or Analysing Biological Materials (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

La présente invention concerne des dispositifs, des systèmes et des procédés pour surveiller des niveaux de glycémie chez un patient, en calculant une pluralité de valeurs de glucose dans le sang à partir d'une pluralité de prélèvements sanguins, en calculant un quantile des valeurs de glucose dans le sang relevées sur un certain nombre de jours ; et en créant un tracé de Lorenz pour les valeurs de glucose dans le sang dans le quantile.
PCT/US2015/022031 2014-03-22 2015-03-23 Procédés de diagnostic et dispositifs permettant de contrôler une glycémie aiguë WO2015148387A2 (fr)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US201461969150P 2014-03-22 2014-03-22
US61/969,150 2014-03-22
US201462051943P 2014-09-17 2014-09-17
US62/051,943 2014-09-17

Publications (2)

Publication Number Publication Date
WO2015148387A2 true WO2015148387A2 (fr) 2015-10-01
WO2015148387A3 WO2015148387A3 (fr) 2015-11-19

Family

ID=54196551

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2015/022031 WO2015148387A2 (fr) 2014-03-22 2015-03-23 Procédés de diagnostic et dispositifs permettant de contrôler une glycémie aiguë

Country Status (1)

Country Link
WO (1) WO2015148387A2 (fr)

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7440786B2 (en) * 2002-03-08 2008-10-21 Sensys Medical, Inc. Method and apparatus for presentation of noninvasive glucose concentration information
DK1702559T3 (da) * 2005-03-15 2009-01-19 Hoffmann La Roche Fremgangmsåde og system til at analysere glucosemetabolisme
US8306610B2 (en) * 2006-04-18 2012-11-06 Susan Mirow Method and apparatus for analysis of psychiatric and physical conditions
US9218453B2 (en) * 2009-06-29 2015-12-22 Roche Diabetes Care, Inc. Blood glucose management and interface systems and methods
US20130116526A1 (en) * 2011-11-09 2013-05-09 Telcare, Inc. Handheld Blood Glucose Monitoring Device with Messaging Capability
WO2013173499A2 (fr) * 2012-05-15 2013-11-21 Minor James M Procédés et dispositifs de diagnostic pour surveiller une glycémie chronique

Also Published As

Publication number Publication date
WO2015148387A3 (fr) 2015-11-19

Similar Documents

Publication Publication Date Title
ES2755879T3 (es) Procedimiento y sistema para proporcionar tanto un valor de glucemia medio real estimado como un valor de glucohemoglobina (HbA1C) estimado a partir de mediciones puntuales estructuradas de glucemia
CA2404262C (fr) Procede, systeme et produit informatique ameliorant l'exploitation des donnees de glycemie obtenues par auto-controle
Gerstein et al. Albuminuria and risk of cardiovascular events, death, and heart failure in diabetic and nondiabetic individuals
EP3614394B1 (fr) Systèmes d'optimisation de l'insuline et procédés de test comportant un critère de sortie ajusté prenant en considération le bruit du système associé aux biomarqueurs
Xing et al. Optimal sampling intervals to assess long-term glycemic control using continuous glucose monitoring
KR101766473B1 (ko) 임상 결과를 평가하기 위한 방법 및 시스템
US9336355B2 (en) Apparatus and method for generating a condition indication
Lane et al. Continuous glucose monitors: current status and future developments
EP3361480A1 (fr) Analyse de médiane de glucose, de la variabilité et du risque d'hypoglycémie de guidage de thérapie
EP3779995A1 (fr) Systèmes et procédés permettant d'optimiser le dosage d'insuline
Schaupp et al. Taking a closer look—continuous glucose monitoring in non-critically ill hospitalized patients with type 2 diabetes mellitus under basal-bolus insulin therapy
EP2545185A1 (fr) Procédé et système pour la sureté, l'analyse et la supervision de l'action d'une pompe à insuline et d'autres modes d'administration de l'insuline dans des cas de diabète
US11363991B2 (en) Method and system for analyzing glucose monitoring data indicative of a glucose level and a non-transitory computer readable medium
Peyser et al. Hypoglycemic accuracy and improved low glucose alerts of the latest Dexcom G4 Platinum continuous glucose monitoring system
US20190046112A1 (en) Systems and methods for treating and estimating progression of chronic kidney disease
Hirsch et al. Professional flash continuous glucose monitoring with ambulatory glucose profile reporting to supplement A1C: rationale and practical implementation
US11571170B2 (en) Method for providing alert of potential thyroid abnormality
Shin et al. Costs during the first five years following cancer diagnosis in Korea
US9465910B2 (en) Diagnostic methods and devices for monitoring chronic glycemia
CN106999041B (zh) 用于健康相关问题的管理的智能记录
US11798685B2 (en) Diagnostic methods and devices for controlling acute glycemia
WO2015148387A2 (fr) Procédés de diagnostic et dispositifs permettant de contrôler une glycémie aiguë
Sun et al. Time in range estimation in patients with type 2 diabetes is improved by incorporating fasting and postprandial glucose levels
ES2962508T3 (es) Visualización de valores biológicos
US11793471B2 (en) Method and system for monitoring a diabetes treatment plan

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15770281

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15770281

Country of ref document: EP

Kind code of ref document: A2