Gastric emptying breath tests
11134883 · 2021-10-05
Assignee
Inventors
Cpc classification
A61B5/42
HUMAN NECESSITIES
A61B5/082
HUMAN NECESSITIES
A61K49/0004
HUMAN NECESSITIES
International classification
A61K51/00
HUMAN NECESSITIES
A61B5/00
HUMAN NECESSITIES
A61B5/08
HUMAN NECESSITIES
Abstract
A breath test and a breath test method for diagnosing abnormal gastric emptying using a .sup.13C labeled breath test meal by identifying the time of maximum excretion of .sup.13CO.sub.2 in the breath samples. The test and test method include supplying a subject with the breath test meal, collecting a breath sample of the subject at a plurality of time points after the subject consumes the meal, generating measurements of .sup.13CO.sub.2 excretion from the breath samples, identifying a time T.sub.max at which .sup.13CO.sub.2 excretion is maximal, comparing T.sub.max to a cut-off value. The subject may be diagnosed as having delayed gastric emptying if T.sub.max is greater than the T.sub.max cut-off value.
Claims
1. A breath test method for identifying delayed gastric emptying based on solid phase gastric emptying, comprising: supplying a test subject with a solid breath test meal having a .sup.13C-Spirulina platensis label, wherein the solid breath test meal has a caloric content of between about 200 kCal and about 400 kCal; selecting a plurality of time points at which to collect breath samples from the test subject, the plurality of time points including a first time point and a last time point after the test subject consumes the solid breath test meal, wherein the last time point is four hours after the test subject consumes the solid breath test meal; collecting breath samples of the test subject at the plurality of time points; providing an instrument configured to measure a ratio of .sup.13CO.sub.2/.sup.12CO.sub.2 in the breath samples; measuring the ratio of .sup.13CO.sub.2/.sup.12CO.sub.2 in the breath samples using the instrument; generating .sup.13CO.sub.2 excretion measurements from the breath samples; comparing the .sup.13CO.sub.2 excretion measurements of the breath samples; determining that the .sup.13CO.sub.2 excretion measurement of the breath sample collected at four hours is higher than the .sup.13CO.sub.2 excretion measurements of all the other breath samples collected; determining that the .sup.13CO.sub.2 excretion measurement for each successive breath sample is higher than a .sup.13CO.sub.2 excretion measurement of the respective breath sample collected immediately prior thereto, wherein .sup.13CO.sub.2 excretion for the test subject reaches a maximum after the last time point; determining that the test subject has delayed gastric emptying; and administering a treatment to the test subject after determining that the test subject has delayed gastric emptying, wherein the treatment comprises administering an effective amount of prokinetic agents, antiemetic agents, antihistamine agents or erythromycin lactobionate to the test subject.
2. The breath test method of claim 1 wherein said solid breath test meal has a .sup.13C label in a dose of between about 20 mg to about 80 mg.
3. The breath test method of claim 1 wherein the prokinetic agents comprise dopamine D2-receptor antagonists, metoclopramide, or domperidone.
4. The breath test method of claim 1 wherein the antiemetic agents comprise phenothiazines, prochlorperazine, or thiethylperazine.
5. The breath test method of claim 1 wherein the antihistamine agents include promethazine.
6. The breath test method of claim 1 wherein the instrument is a mass spectrometer or an infrared spectrometer.
7. The breath test method of claim 1 wherein the .sup.13CO.sub.2 excretion measurements are measured in units of kPCD so as to normalize for height, weight, age, and gender of the test subject.
8. The breath test method of claim 1 wherein the time points include 45 minutes, 90 minutes, 120 minutes, 150 minutes, 180 minutes, and 240 minutes after the test subject consumes the solid breath test meal.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
(1) The following drawings are illustrative and therefore do not limit the scope of the invention. The drawings are not to scale (unless so stated) and are intended for use in conjunction with the explanations in the following detailed description.
(2)
(3)
(4)
(5)
(6)
DETAILED DESCRIPTION
(7) For the purpose of promoting an understanding of the principles of the invention, various embodiments will now be described. It will, nevertheless, be understood that no limitation of the scope of the invention is thereby intended; any alterations and further modifications of the described or illustrated embodiments, and any further applications of the principles of the invention as illustrated therein, are contemplated as would normally occur to one skilled in the art to which the invention relates.
(8) Various embodiments include breath test methods that can detect, diagnose and monitor delayed gastric emptying or rapid gastric emptying by obtaining breath sample(s) over the course of the breath test evaluation period and identifying the time at which maximum .sup.13CO.sub.2 excretion occurs. This time point, which may be referred to as the T.sub.max, may be used alone as a single diagnostic value, or may be used in combination with other breath test values and/or scintigraphic values to diagnose delayed or rapid gastric emptying.
(9) The value of T.sub.max for an individual patient may be determined using the results of a gastric emptying breath test. T.sub.max is the time point at which .sup.13CO.sub.2 excretion reaches a maximum during the gastric emptying breath test evaluation period. In some individuals, .sup.13CO.sub.2 excretion increases during the early portion of the test, then reaches a maximum, and then declines during the later portion of the test. In other individuals, particularly individuals with delayed gastric emptying, .sup.13CO.sub.2 excretion increases throughout the duration of the test. In such individuals, .sup.13CO.sub.2 excretion may not truly reach a maximum until a time point that occurs after completion of the test, and therefore it is not possible to identify the time at which maximum .sup.13CO.sub.2 excretion actually occurs. The time point at which the test is completed, such as at four hours, may therefore be used as the time at which T.sub.max occurs in such cases. Though the T.sub.max in such cases does not represent actual time at which .sup.13CO.sub.2 excretion reached a maximum level, the Applicant has discovered that it can nevertheless be used to diagnose gastroparesis.
(10) To better illustrate the concepts, an example of breath tests results for two different groups of individuals are shown in
(11)
(12) With regard to
(13) Because .sup.13CO.sub.2 excretion curves (plots of kPCD vs. time) rise in normal individuals until reaching a maximum (at approximately 150 minutes on average) and then start to decline again (between 150 and 240 minutes), and the .sup.13CO.sub.2 excretion curves of individuals with delayed gastric emptying continuously rise and often cross through the normal reference range at later time points, there is a decrease in diagnostic sensitivity (for diagnosing gastroparesis) by kPCD metric alone when measured at 180 and 240 minutes. This decrease in diagnostic sensitivity can be observed by referring to the excretion curve of the individual with moderately delayed gastric emptying (line 20) in
(14) The calculation of T.sub.max generally includes obtaining breath sample(s) at multiple time points after a subject ingests a test meal. One exemplary method of using a test meal with unique physiologic .sup.13CO.sub.2 excretion attributes in combination with identification of T.sub.max will now be described. When a clinician orders a breath test including identification of T.sub.max, instructions are provided to the subject. These instructions may preferably include fasting, which may be done by the subject using any acceptable fasting protocol for a period of time before the test is administered. Before the breath test method is performed, a test administrator collects baseline breath samples from the subject. The subject then ingests a test meal that includes a .sup.13C label incorporated into the meal. In some cases, the breath test meal has a caloric content of between about 200 and about 400 kCal, such as preferably between about 200 and about 400 kCal in some embodiments. The caloric content and other features of the test meal may be selected such that T.sub.max can be used to differentiate normal from gastroparetic patients using a test of a desired duration, such as four hours. The test administrator collects a fasting, pre-meal (baseline) breath sample. The subject consumes the meal in a single setting, for example within a specific time frame. The test administrator then collects additional breath samples from the subject at specific time points after the subject ingests the meal. The test administrator generates a gastric emptying breath test measurement such as kPCD from each of the breath samples and uses the measurements to provide a T.sub.max test result. The gastric emptying measurements, expressed in kPCD metric, may also optionally be used to generate other results such as BT_t.sub.1/2.
(15) The test meal includes a .sup.13C label that is incorporated into the meal. In certain cases, a .sup.13C label is incorporated into biomass, which is then incorporated into the test meal. For example, in specific cases, the test meal includes a .sup.13C label incorporated into Spirulina platensis. Spirulina platensis labeled with .sup.13C can be obtained by growing the algal cells in a .sup.13C-enriched environment as is disclosed in commonly assigned U.S. Pat. No. 6,872,516, the disclosure of which is incorporated herein by reference in its entirety.
(16) The test meal includes a .sup.13C label at a dose that provides an excellent signal for .sup.13CO.sub.2 breath testing. In some cases, the dose is between about 20 mg and about 80 mg of .sup.13C label or perhaps between about 40 mg and about 50 mg, such as 43 mg of .sup.13C. In certain cases, the subject ingests .sup.13C-Spirulina platensis at a dose of approximately 100 mg, which contains approximately 43 mg of .sup.13C.
(17) The .sup.13C-Spirulina platensis may be incorporated into an edible food component that forms part of the test meal. The test meal may include any number of edible food components that can be ingested at a single setting. A single setting can be a designated time period, perhaps a period of less than 30 minutes, 20 minutes, or even 10 minutes. In many cases, the test meal can include a main food component as well as any side components and/or liquid components. In certain cases, the test meal includes food components derived from standardized, freeze-dried or lyophilized food components, such as those described in U.S. patent application Ser. No. 10/435,092, the entire contents of which are incorporated herein by reference.
(18) In some particular embodiments, the test meal includes reconstituted lyophilized whole eggs and .sup.13C-Spirulina platensis as a main food component. The .sup.13C-Spirulina platensis in this embodiment is present in an amount of 100 mg (which contains approximately 43 mg of .sup.13C). The test meal can also include bread or crackers as a side component and/or milk or water as a liquid component.
(19) The test administrator directs the subject to deposit a breath sample into a container, such as by blowing through a straw into the bottom of a container such as a glass tube to displace contained air and capture a clean breath sample. The test administrator caps the container such as the tube. Alternatively, the subject may blow into a breath bag container, for example constructed of mylar-type material and having a valve mechanism that operates in a manner that retains the sample in the bag after depositing the breath. The .sup.13CO.sub.2 content measurement of the breath in the container is then measured, such as by an operator inserting the container into an instrument for analysis by Gas Isotope Ratio Mass Spectrometry (GIRMS), such as a mass spectrometer, infrared spectrometer or any other known instrument for measuring of .sup.13CO.sub.2/.sup.12CO.sub.2 ratios in breath, and directing the instrument to perform the measurement.
(20) The T.sub.max value may be used alone or in combination with other breath test measurements or calculations to diagnose rapid or delayed gastric emptying. For example, T.sub.max may be used in combination with breath test results which may be expressed as kPCD and CumPCD. The PCD metric is the Percent Dose (abbreviated PCD) of .sup.13C excreted at time t (in this case, at the single time point) after consumption of the test meal. To provide a more convenient scale, PCD is multiplied by 1000 to produce kPCD at time t. The kPCD value can be calculated as follows:
(21)
(22) where:
(23) DOB=The measured difference in the .sup.13CO.sub.2/.sup.12CO.sub.2 ratio between the single time point measurement and the baseline measurement.
(24) CO.sub.2PR=CO.sub.2 production rate (mmol CO.sub.2/min) calculated using the Schofield equations. The Schofield equations are a set of equations which are fitted to people depending on their age, gender, height, and weight to estimate the basal metabolic rate (BMR), which is intimately related to .sup.13CO.sub.2 production rate. The equations are known in the art and allow the calculation of a specific CO.sub.2 production rate for the individual being tested.
(25) R.sub.s=0.0112372, the ratio [.sup.13CO.sub.2/.sup.12CO.sub.2] in an international reference standard (Pee Dee Belemnite).
(26) 13=The atomic weight of .sup.13C.
(27) 10=A constant factor for converting units.
(28) Dose=The weight (mg) of .sup.13C in the dose of .sup.13C-Spirulina platensis administered to the subject in the test meal. For example, since .sup.13C-Spirulina platensis is approximately 43% by weight .sup.13C, a dose of 100 mg corresponds to approximately 43 mg of 130.
(29) The T.sub.max value may be expressed in units of time, such as hours, minutes, and/or seconds. The T.sub.max result may be calculated from the breath test measurement in a variety of ways. In some embodiments, the breath test results may be applied to a mathematical program which calculates the excretion profile by interpolating values between the breath test measurements, such as by using a 5.sup.th order polynomial equation. The time at which T.sub.max occurs may then be identified by the program based on the generated excretion profile. The generated excretion profile may be provided to a clinician, such as in a graphical form like those of
(30) Alternatively, in some embodiments, an individual's T.sub.max value may be determined by simply plotting the individual's kPCD test values versus time and identifying the breath sample time point at which the highest kPCD values occurs on the plot. Alternatively, the breath sample time point at which the highest kPCD value occurs may be determined from the numerical value of the results without using a graphical plot. If the breath test measurements (kPCD values) rise throughout the test period (no peak in the excretion curve is observed), then the last measurement time in the evaluation period, (240 minutes, for example), may be used as the time at which T.sub.max occurs even though maximum .sup.13CO.sub.2 excretion may actually occur much later in very delayed patients. In this way, T.sub.max can be determined and used for diagnosis without the need to mathematically identify T.sub.max with curve fitting models. As will be demonstrated, the graphical approach (or alternatively the numerical approach) is as diagnostically accurate as the curve fitting approach for establishing diagnosis by T.sub.max. Such methods are particularly useful in embodiments using a test meal in which the normal reference range for T.sub.max falls before the time at which the last breath sample is collected during the test evaluation period, and the T.sub.max for individuals with delayed emptying occurs after the time of the last breath sample. An example of such an embodiment is a four hour gastric emptying breath test using a test meal including 43 mg .sup.13C incorporated into Spirulina platensis.
(31) In some embodiments, a T.sub.max may be considered normal if it falls within a normal reference range. For example, for a 4 hour breath test, the T.sub.max may be normal if it is less than 4 hours (that is, if it occurs before the end of the test.)
(32) In certain cases, the T.sub.max can be considered normal if it is less than an upper cut-off point value and greater than a lower cut-off point value. A T.sub.max greater than the upper cut-off value is abnormal and indicates the presence of delayed emptying, and a T.sub.max less than the lower cut-off value is also abnormal and represents rapid emptying.
(33) The cut-off values for T.sub.max for a particular gastric emptying test can be determined statistically using gastric emptying data from a group of normal subjects. For instance, a cohort of normal, healthy subjects can be administered a specific standardized test meal containing a .sup.13C label for breath testing. Each healthy subject in the cohort is tested by collecting and conducting breath test measurements at multiple time points. The T.sub.max may be determined for each individual, such as by using a mathematical program as described above. The T.sub.max results can then be arrayed. From these arrays, an upper and/or lower cut-off point can be selected. For example, a simple way to identify a cut-off point is to array the T.sub.max results for all subjects from high to low. The values residing at the upper and/or lower 95.sup.th percentile of the array of values can then be used as cut-off points. Or, for example, the mean of the test values plus or minus 1.96 standard deviations can serve as cut-off points.
(34) In some embodiments, the upper and lower cut-off points can be identified as the 0.025 fractile (2.5%) bound of the central 0.95 (95%) reference interval of T.sub.max values obtained from healthy subjects. The central 95% reference interval may be calculated using commercially available validated statistical programs such as RefVal and MedCalc. Factors such as the within-subject natural biologic variation of gastric emptying can also be measured and considered when determining the cut-off point between normal and delayed gastric emptying or between normal and rapid emptying. In some embodiments, the 95% level of diagnostic efficiency may be used for determining the cut-off points, while in other embodiments, the diagnostic efficiency level chosen could be set lower at the 90%, 85% or 80% level, for example. In some embodiments, the T.sub.max value can be used to identify a test result as normal, delayed, or borderline delayed or alternatively as normal, accelerated, or borderline accelerated. Methods for identifying the cut off points for such zones are described in U.S. Pat. Pub. No. 2008-0281166, the disclosure of which is hereby incorporated by reference.
(35) In some embodiments, a patient with a T.sub.max occurring prior to the T.sub.max cut off point that demarcates normal from delayed rates of emptying, can be identified and be correctly diagnosed as actually having delayed gastric emptying. The applicant has discovered that in certain unusual circumstances, this can occur in individuals with severely delayed gastric emptying and therefore the diagnosis of delayed gastric emptying may be more accurate if the test properly handles such unusual individuals.
(36) An example of such an unusual patient is demonstrated with reference to
(37) Because T.sub.max occurs late, after the normal reference range, in nearly all individuals with delayed gastric emptying, and a normal T.sub.max only occurs in a very small subset of individuals having severely delayed emptying, a diagnostic rule to capture these individuals is not difficult to devise. For example, a simple kPCD cut-off point at a single time point may be used which would identify all such individuals with severely delayed gastric emptying. For example, a diagnosis of delayed gastric emptying may be made if (1) the T.sub.max occurs on or after a specified cut-off time or (2) T.sub.max occurs before the specified cut-off point and the .sup.13CO.sub.2 excretion rate (kPCD) at a time t is less than a value x.
(38) The values for t and x will depend on the particular test method (e.g., size of the meal, amount of .sup.13C included in test meal) and the limits of the established reference range. However, the time t may be selected to fall at a time or within a time window at which the maximum difference in .sup.13CO.sub.2 excretion occurs between normal and delayed individuals. In a 43 g .sup.13C test meal, this time window is between 90 minutes and 150 minutes (as demonstrated in
(39) If the T.sub.max result is greater than the higher end cut-off point of the normal reference range, or if the .sup.13CO.sub.2 excretion (kPCD) is less than x at time t, a clinician diagnoses the subject as having delayed gastric emptying. Alternatively, if T.sub.max is less than a lower end cut-off point for the normal reference range, a clinician diagnoses the subject as having accelerated gastric emptying. The clinician can then investigate the cause of the disorder and select an intervention program, for example for gastroparetic patients, that may include actions such as dietary changes, weight loss and/or pharmacologic intervention. For example, the clinician may prescribe for the patient a treatment plan which the patient may then implement. Examples of dietary changes include consuming meals with lower fat content and lower residue, and smaller and more frequent meals. Examples of pharmaceutical agents which may be prescribed for the treatment of gastroparesis include prokinetic agents to improve gastric emptying such as dopamine D2-receptor antagonists such as metoclopramide, type II dopamine antagonists such as domperidone, or other agents such as erythomycin lactobionate. Some subjects with severely delayed gastric emptying may suffer nutritional deficiency and therefore the clinician may prescribe intravenous feeding intervention which may be administered to the patient. Symptomatic treatment may also be prescribed and may be implemented by the patient, including antiemetic agents such as phenothiazines such as prochlorperazine and thiethylperazine or antihistamine agents such as promethazine.
(40) After intervention, a separate follow-up breath test may be conducted and the pre and post intervention T.sub.max values and/or .sup.13CO.sub.2 excretion values at time t can be obtained. A clinician can then compare the post-treatment result with the original result, to determine whether the disorder is improving. In such cases, the T.sub.max value and kPCD value at time t is first used as a diagnostic health test and then is later used to monitor treatment results.
(41) When using a unique test meal containing 27 grams of powdered egg and 100 mg .sup.13C-Spirulina (43% by weight .sup.13C) the maximum .sup.13CO.sub.2 excretion rate in normal subjects rarely exceeds 209 minutes (<2.5% chance) AND gastroparetic subjects rarely reach a maximum excretion rate before 240 minutes. Therefore T.sub.max clearly discriminates between normal and delayed subjects. The excellent utility of T.sub.max as a breath test diagnostic metric can be confirmed by evaluating its diagnostic precision. For example, the diagnostic precision of T.sub.max (used individually or in combination with a kPCD cutoff value x at time t) can be confirmed by comparing diagnosis by T.sub.max to diagnosis obtained by the gold standard scintigraphy method. (Diagnostic precision is then expressed using standard equations for calculating sensitivity, specificity, concordance, NPV and PPV). The diagnostic precision of T.sub.max can be determined against multiple scintigraphic metrics used for diagnosis, such as fraction of meal emptied at time t, or scintigraphic t.sub.1/2. T.sub.max and BT_t.sub.1/2 (a metric calculated from kPCD results using Multiple Linear Regression (MLR) equations that generate a value equivalent to a Scintigraphic t.sub.1/2) values may be co-reported with kPCD test results obtained at 90, 120 and 150 minutes. All test results across the 3 metrics are diagnostically useful and may be used individually or in combination by a clinician. A breath test report may also include a graphical display of the subject's breath test results (.sup.13CO.sub.2 excretion curve) across all measured time points out to 4 hours, for example. The subject's .sup.13CO.sub.2 excretion curve (kPCD metric) may be displayed relative to the normal reference range established for the measured time points, such as the 90, 120 and 150 minute time points where discrimination between normal and delayed subjects is best. The breath test results and the graphical display of .sup.13CO.sub.2 excretion curve may provide clinicians with a comprehensive gastric emptying profile from which a diagnosis may be made. Diagnosis of gastroparesis may be made based on T.sub.max alone or in combination with a cutoff value x at time t, and/or in combination with one or more other breath test metrics, such as BT_t.sub.1/2 and kPCD test results, optionally including a graphical display.
(42) Examples of comprehensive patient profiles which may be provided to clinicians and used by a clinician for patient diagnosis and monitoring based on the results of a gastric emptying breath test are shown in the reports depicted in
(43) In each .sup.13CO.sub.2 excretion graph 106 of
(44) Each report further includes a summary section 108 indicating whether the gastric emptying breath test results are consistent with normal gastric emptying or delayed gastric emptying, for example. In these examples, the reports further include explanations for the clinician, such as an explanation of the T.sub.max diagnostic algorithm 109 that was applied in the report.
(45) Consumption of the entire test meal of a specific caloric value and dose of carbon-13 has been important for breath testing results as the normal reference ranges and cut-off points are established from healthy subjects on this basis. Furthermore, linear regression equations used to estimate t.sub.1/2 by breath test are derived from the relationship of breath test values to scintigraphic values in patients who have consumed the entire meal. Hence, if a test subject was unable to completely consume the test meal, practitioners of breath tests generally reported inconclusive test results for measurements of gastric emptying and/or suggested that the test subject be re-scheduled for the breath test procedure. However, another particularly useful attribute of using T.sub.max for diagnosis is that it is not always necessary for the subject to consume the full meal in order for a diagnosis to be made. For example, even if a test subject could not completely consume the test meal, the patient may be classified as having delayed gastric emptying if T.sub.max in such a subject occurs after the T.sub.max cut-off value for the test, such as if T.sub.max occurs at 240 minutes, or if the subject has a kPCD value that is less than a cut-off value at time t, such as less than 33 at 120 minutes for a four hour breath test. In this way, the patient may be diagnosed as having delayed gastric emptying using T.sub.max without needing to be re-scheduled for a repeat test procedure. This is particularly helpful in those patients presenting with significant symptoms of nausea, vomiting, bloating and early satiety.
(46) The applicants performed both statistical analyses and diagnostic precision evaluations on test subjects using the breath test T.sub.max metric as compared to diagnosis by scintigraphy. In addition, to demonstrate the exceptional utility and diagnostic performance of T.sub.max compared to other breath test metrics, diagnosis by all 3 useful breath test metrics (kPCD, BT_t.sub.1/2 and T.sub.max) were compared to diagnosis by scintigraphy. These analyses and evaluations are discussed in the experimental section below.
EXPERIMENTAL
(47) For these Examples, Applicant analyzed data from a study conducted by Applicant as described in Szarka L, Camilleri M, Vella A, et al. A Stable Isotope Breath Test with a Standard Meal for Abnormal Gastric Emptying of Solids in the Clinic and in Research. Clin Gastroenterol H. 2008; 8: 635-643. In the Szarka study, the reported cohort size was 129 subjects who were symptomatic for and suspected of having gastroparesis. However, Applicant excluded 14 subjects due to minor clinical protocol violations, such that the remaining cohort included 115 total subjects. Prior to conducting the Szarka study, the normal reference ranges and the applicable cut-off points for demarcating normal from delayed emptying were established for both breath test (GEBT) and scintigraphy by testing a cohort of 60 healthy subjects with the exact same test meal as that utilized in the Szarka study, a dual labeled test meal that included both .sup.13C-labeled Spirulina platensis with 43 mg .sup.13C and 0.5 m Ci .sup.99mTc sulphur colloid. The normal reference range for each method is the 95% reference interval of the data sets collected from the healthy subjects.
(48) The dual label procedure is a superior method of evaluating agreement between two testing methods. The normal, day-to-day biologic variability of gastric emptying is large and well documented in the literature. For example, when given the same test meal on consecutive days, a test subject will display significant variability in their rate of gastric emptying (on average, the percent coefficient of variation is approximately 12%). Hence, conducting both test methods on the same patient concurrently in the Szarka study and when determining the normal reference ranges eliminated the confounding effects of conducting the breath test on one day and scintigraphy on another day.
Example 1
(49) Applicant analyzed the breath test results of the 60 normal subjects to identify the T.sub.max for each subject using a fifth order polynomial curve fitted to the kPCD values at each time point for each subject. The mean time of peak excretion, T.sub.max, was 161 minutes. Applicant identified a normal T.sub.max 95% reference range from the T.sub.max values of the 60 normal subjects using Medcalc (version 14.1, 2013). The 95% reference range was 104.6-209.4 minutes.
(50) The excretion profiles (kPCD vs. time) of the 60 normal subjects were averaged and the results are shown in
Example 2
(51) In Example 2, Applicant analyzed the breath test data from the 115 test subjects participating in Applicant's clinical study as described in the Szarka study. As with the 60 normal subjects, Applicant analyzed the breath test results of the test subjects to identify T.sub.max for each subject by using a fifth order polynomial curve that was fitted to the kPCD values at each time point. While the 95% T.sub.max reference range limit (the “cut-off′ point”) identified in Example 1 was approximately 209 minutes, Applicants discovered that most of the gastroparetic test subjects (as diagnosed by scintigraphy) had T.sub.max values that well exceeded 209 minutes. Most delayed subjects' T.sub.max occurred at the end of the evaluation period (at 240 minutes). Applicants further discovered that a small subset of gastroparetic test subjects (as diagnosed by scintigraphy) had a T.sub.max that occurred before 209 or 240 minutes, but also had very low excretion profiles (kPCD values vs. time). That is, they were frankly delayed. An example of such a subject is shown in line 40 of
(52) Based on these results, Applicant decided to use the following criteria for diagnosing delayed gastric emptying in the following examples. Curve fitting was omitted and the kPCD value obtained at each time point was used for the evaluation. If maximum excretion rate (kPCD value) occurred at 240 minutes, the patient had delayed gastric emptying. Additionally, if T.sub.max occurred prior to 240 minutes, but kPCD at 120 minutes was less than 33 min.sup.−1, then the subject was likewise classified as having delayed gastric emptying. The value of 33 min.sup.−1 kPCD was selected as being 25% below the lower reference range limit at 120 minutes.
Example 3
(53) Applicant next used the breath test data and the scintigraphic data obtained from the 115 test subjects to establish a diagnosis for each subject by each test method. Diagnosis (positive or negative for delayed emptying) was established by comparing each test subject's breath test result and, likewise, their scintigraphic test result, to pre-determined cut off points for each test method as described above. The breath test's (GEBT) diagnostic performance vs. scintigraphy was evaluated using kPCD, BT_t.sub.1/2 and T.sub.max metrics. The applicable cutoff points (“COP's”) for the kPCD metric and the breath test t.sub.1/2 metric (BT_t.sub.1/2) are shown in Table 1 below. The COPs for these breath test metrics are the applicable 0.025 fractile bounds (2.5% limit) of the 95% reference interval (normal range) obtained from the 60 healthy subjects. For diagnosis using T.sub.max, the subject was diagnosed as gastroparetic if T.sub.max occurred at 240 minutes, or if T.sub.max occurred before 240 minutes but kPCD at 120 minutes was less than 33 min.sup.−1.
(54) TABLE-US-00001 TABLE 1 COP's kPCD COP (min.sup.−1) Time (min) BT_t.sub.1/2 COP 45 90 120 150 180 240 (min) GEBT 12.9 33.1 44.2 49.3 46.5 39.2 92.3 Scint_FE COP Time (min) Scint_t.sub.1/2 COP 45 90 120 150 180 240 (min) Scintigraphy 0.181 0.494 0.678 0.774 0.875 0.957 91.0
(55) When using the breath test kPCD metric and scintigraphic fraction emptied (FE) metric, test results lower than the COP indicate delayed emptying. Using the 90-minute measurement time as an example, a test subject having a kPCD value<33.1 min.sup.−1 and a fraction of test meal emptied (FE) by scintigraphy<0.494 would be classified as delayed by both test methods.
(56) Conversely, t.sub.1/2 represents the time it takes to empty ½ of the test meal from the stomach. When test subjects are delayed, they take longer to empty the stomach. Hence, t.sub.1/2 values increase as function of increasing gastric delay. By way of example, a test subject having a BT_t.sub.1/2 value>92.3 minutes and a scintigraphic t.sub.1/2 value>91.0 minutes would be classified as delayed by both methods.
(57) Applicant then compared diagnosis obtained by breath test on each subject to diagnosis obtained by scintigraphy on each patient. Breath test diagnoses were established for each patient by each breath test metric (kPCD, BT_t.sub.1/2 and T.sub.max). For purposes of this analysis, Applicant considered the scintigraphy results as the true results, or the gold standard. Applicant classified each subject's breath test diagnosis, with reference to the scintigraphic diagnosis, as a true positive (TP), true negative (TN), false positive (FP) or false negative (FN). Using these classifications by each method, Applicant then calculated the breath test's diagnostic performance statistics versus scintigraphy using standard calculations for sensitivity, specificity, concordance, negative predictive value (NPV) and positive predictive value (PPV). Calculations for PPV and NPV incorporated prevalence. The 95% confidence intervals (CI) for the PPV and NPV values were calculated by a statistical bootstrap procedure with 1,000 iterations.
(58) Breath test (GEBT) performance using the three GEBT metrics was compared to three scintigraphic metrics commonly used for diagnosis. The GEBT metrics were: kPCD, BT_t.sub.1/2, and T.sub.max. The scintigraphic metrics were: scintigraphic fraction emptied (FE) at corresponding time points, scintigraphic t.sub.1/2, and scintigraphic fraction of test meal retained at 4 hours.
(59) Table 2 shows a comparison of the results for diagnosis by breath test using kPCD metric vs. diagnosis by scintigraphy using the scintigraphic fraction of test meal emptied metric (Scint_FE). As previously described, when kPCD is used as the sole metric for diagnosis, there is a propensity for the kPCD values of subjects with delayed gastric emptying to rise into the pre-established reference range at later time points. This occurs because at these later time points normal subjects' .sup.13CO.sub.2 excretion rates are declining while the .sup.13CO.sub.2 rates of delayed subjects are still rising. Hence, the breath test metric kPCD has an inherent propensity for false negative diagnoses at 180 and 240 minutes when compared to scintigraphic diagnoses. This is reflected in the poor breath test (kPCD) performance characteristics tabulated in Table 2 at 180 and 240 minutes. Thus, diagnostic utility for kPCD metric is best at 90, 120 and 150 minutes.
(60) TABLE-US-00002 TABLE 2 Breath Test Performance: kPCD vs. Scint_FE Diagnosis by kPCD vs. Scint_FE at time t 45 90 120 150 180 240 Classification (N = 115) TP* 16 41 48 40 29 15 TN.sup.† 84 54 45 49 56 62 FP.sup.‡ 6 11 13 13 8 7 FN.sup.§ 9 9 9 13 22 31 Performance Statistics (%) Sensitivity 64.0 82.0 84.2 77.4 56.9 32.6 Specificity 93.3 83.1 77.6 79.0 87.5 89.9 Concordance 87.0 82.6 80.9 78.3 73.9 67.0 PPV.sup.¶ 72.7 78.8 78.7 75.9 78.4 68.2 (95% CI**) (56.7-89.6) (69.8-88.7) (70.5-86.8) (66.6-84.3) (66.8-89.6) (49.2-85.8) NPV.sup.†† 90.3 85.7 83.3 80.3 71.8 66.7 (95% CI) (85.6-95.2) (78.4-92.9) (74.9-91.7) (71.2-87.2) (65.5-79.5) (61.8-71.6)
(61) Table 3 below shows a comparison of the results for diagnosis by T.sub.max to diagnosis by Scint_FE for each time t. Table 3 (compared to Table 2) shows that T.sub.max is much more diagnostically sensitive than diagnosis by kPCD and very useful at later time points. Because T.sub.max occurs significantly earlier in normal subjects than in delayed subjects, normal and delayed subjects can be reliably differentiated with this metric. As shown in Table 3, diagnosis by T.sub.max has excellent agreement with diagnosis by Scint_FE at each time point.
(62) TABLE-US-00003 TABLE 3 Diagnostic Performance of T.sub.max vs. Scintigraphy (FE) Diagnosis by T.sub.max vs. Scint_FE at time t, measured in minutes from the end of test meal consumption 45 90 120 150 180 240 Classification TP 22 44 47 45 43 40 TN 64 61 57 59 59 61 FP 26 4 1 3 5 8 FN 3 6 10 8 8 6 Performance Statistics (%) Sensitivity 88.0 88.0 82.5 84.9 84.3 87.0 Specificity 71.1 93.8 98.3 95.2 92.2 88.4 Concordance 74.8 91.3 90.4 90.4 88.7 87.8 PPV.sup.¶ 45.8 91.7 97.9 93.8 89.6 83.3 (95% CI) (37.6-56.0) (84.1-97.9) (93.4-100.0) (87.0-100.0) (81.5-97.4) (74.2-92.6) NPV 95.5 91.0 85.1 88.1 88.1 91.0 (95% CI) (90.8-100.0) (85.7-96.9) (77.8-92.4) (81.6-94.6) (81.4-94.9) (84.4-96.9)
(63) Table 4 shows the results for diagnosis by kPCD at 90, 120, and 150 minutes, as well as for diagnosis by T.sub.max and BT_t.sub.1/2, as compared to diagnosis established by Scint_t.sub.1/2. This comparison shows that all three breath test metrics agree well with diagnosis by Scint_t.sub.1/2. Diagnosis by T.sub.max and/or BT_t.sub.1/2 improves diagnostic confidence and is complementary to diagnosis by kPCD at 90, 120, and 150 minutes. Furthermore, it can be seen that T.sub.max and BT_t.sub.1/2 metrics have similar performance characteristics as compared to scintigraphy, albeit T.sub.max is best.
(64) TABLE-US-00004 TABLE 4 Breath Test (GEBT) Performance vs. Diagnosis by Scintigraphic t½ Diagnosis by kPCD at 90, 120 and 150 Diagnosis by T.sub.max and minutes vs. Scint_t.sub.1/2 BT_t.sub.1/2 vs. Scint_t.sub.1/2 90 Min. 120 Min. 150 Min. T.sub.max BT_t.sub.1/2 Classification TP 41 44 40 44 44 TN 54 48 52 61 55 FP 11 17 13 4 10 FN 9 6 10 6 6 Performance Statistics (%) Sensitivity 82.0 88.0 80.0 88.0 88.0 Specificity 83.1 73.8 80.0 93.8 84.6 Concordance 82.6 80.0 80.0 91.3 86.1 PPV.sup.¶ 78.8 72.1 75.5 91.7 81.5 (95% CI) (70.5-88.4) (63.9-81.8) (66.0-85.7) (84.7-97.9) (72.6-90.7) NPV 85.7 88.9 83.9 91.0 90.2 (95% CI) (78.2-92.9) (81.2-96.3) (75.9-91.5) (84.8-97.2) (83.1-96.5)
Example 4
(65) As shown in Table 5, diagnosis by T.sub.max using the upper limit of the statistically generated T.sub.max 95% Reference Range (104-209 minutes) produces performance statistics akin to simply using 240 minutes as the T.sub.max COP for demarcating normal from delayed emptying.
(66) TABLE-US-00005 TABLE 5 T.sub.max vs. Scintigraphy: T.sub.max Curve Fit COP (209 minutes) vs. 240 Minute COP T.sub.max > 209 min T.sub.max > 209 min Gastroparetic if (no caveat) kPCD.sub.120 < 33 T.sub.240 = maximum or vs. vs. kPCD.sub.120 < 33 Scint_t.sub.1/2 > Scint_t.sub.1/2 > vs. 91.0 min 91.0 min Scint_t.sub.1/2 > 91.0 min Classification N = 115 TP* 43 46 44 TN.sup.† 59 58 61 FP.sup.‡ 6 7 4 FN.sup.§ 7 4 6 Performance Sensitivity 86.0 92.0 88.0 Specificity 90.8 90.8 93.8 Concordance 88.7 91.3 91.3 PPV.sup.¶ 87.8 88.5 91.7 NPV** 89.4 93.7 91.0
(67) Diagnosis by breath test was compared to diagnosis by Scint_FE at 240 minutes. Relevant scintigraphic literature indicates that 4 hours (240 minutes) may be the best single measurement time for diagnosis of gastroparesis. This is because patients symptomatic for gastroparesis are not always frankly delayed or frankly normal throughout a gastric emptying evaluation. However, with regards to evaluating patients symptomatic for gastroparesis, later time points may be generally more sensitive in detecting delayed emptying than earlier time points. Therefore, breath test diagnostic performance, when utilizing kPCD.sub.(t), BT_t.sub.1/2 and T.sub.max metrics, was compared to Scint_FE diagnosis obtained at 240 minutes. The results are shown in Table 6 below.
(68) TABLE-US-00006 TABLE 6 Breath Test (GEBT) Performance vs. Scintigraphic Classification at 4 Hours Diagnosis by GEBT: kPCD.sub.(t), T.sub.max and BT_t.sub.1/2 vs. Diagnosis by Scint_FE at 240 minutes kPCD at kPCD at kPCD at 90 min 120 min 150 min T.sub.max BT_t.sub.1/2 Classification (N = 115) TP 36 38 36 40 37 TN 53 46 52 61 52 FP 16 23 17 8 17 FN 10 8 10 6 9 Performance Statistics (%) Sensitivity 78.3 82.6 78.3 87.0 80.4 Specificity 76.8 66.7 75.4 88.4 75.4 Concordance 77.4 73.0 76.5 87.8 77.4 PPV.sup.¶ 69.2 62.3 67.9 83.3 68.5 (95% CI**) (59.9-79.5) (54.6-71.3) (58.0-78.0) (74.5-92.4) (59.5-79.0) NPV 84.1 85.2 83.9 91.0 85.2 95% CI) (76.9-91.9) (77.2-93.4) (75.9-91.8) (84.8-97.0) (77.6-93.5)
(69) As shown in Table 6 above, diagnosis by T.sub.max shows excellent agreement with diagnosis by Scint_FE at 240 minutes and is superior to either diagnosis by kPCD or BT_t.sub.1/2.
Example 5
(70) The positive predictive value (PPV) and negative predictive value (NPV) of the breath test (GEBT) are important indicators of how confident one can be that the breath test result (diagnosis), in the absence of a comparative scintigraphic test result, is correct. A high PPV value means that when the breath test yields a positive result (for example “delayed”), it is most likely correct in its assessment. Likewise, a high NPV value means that a negative diagnosis by breath test is most likely correct in its assessment compared to scintigraphy. The PPV and NPV values associated with diagnosis by each breath test metric vs. scintigraphy are reported in prior Tables 2-5. As can be seen, the PPV and NPV values associated with diagnosis by T.sub.max are superb, very good by BT_t.sub.1/2 and good by kPCD. However, PPV and NPV values for a disorder are influenced by the prevalence of the disorder in the population being tested. Breath testing may be utilized in a referred patient population that is symptomatic for accelerated or delayed emptying (gastroparesis). Hence, the need to assess breath test PPV and NPV values across varying prevalence fractions in the intended use population is important and of particular interest to regulatory authorities if the test is intended for commercial use.
(71) Based on an extensive scintigraphic literature review, the prevalence of gastroparesis in referral centers that conduct gastric emptying evaluations may range from 27% to 57%. To see what effect changing the prevalence would have on the diagnostic performance of T.sub.max, the PPV and NPV values were calculated using a prevalence of 27% and of 57%. The PPV and NPV values and their two-sided 95% confidence intervals were calculated using a statistical bootstrap procedure in which the set of 115 test subjects was sampled 1000 times with replacement to best estimate the PPV and NPV values. Standard equations that incorporate prevalence were used to compute the PPV and NPV values.
(72) The results are shown in Tables 7 and 8 below. A mean PPV or NPV value of 1.0 represents 100% confidence that the breath test result would be consistent with a true scintigraphic test result.
(73) The PPV and NPV values associated with diagnosis by T.sub.max were compared to diagnosis by scintigraphic fraction emptied (Scint_FE) at each measurement time throughout the 4-hour evaluation period and shown in Table 7. The upper and lower limits of the 95% Confidence Intervals around the mean values of the PPV and NPV values are co-reported. Calculations were performed at 3 levels of prevalence: 27%; the actual prevalence of gastroparesis found at each measurement time in the cohort of 115 test subjects (Szarka study); and 57%. The prevalence at each time point for the cohort of 115 test subjects was established with reference to the pre-established Scint_FE cut-off point at each measurement time. As noted, a statistical bootstrap procedure using 1000 samplings of the data from the 115 test subjects with replacement was used to obtain the best estimate of the true PPV and NPV values.
(74) TABLE-US-00007 TABLE 6 Bootstrap-Derived T.sub.max PPV and NPV Values Across Changing Prevalence of Gastroparesis (Diagnosis by T.sub.max vs. Scint_(FE) Time Points (minutes) Prevalence (%) Statistic 95% CI 45 90 120 150 180 240 27 PPV Mean PPV 0.5332 0.8459 0.9481 0.8703 0.8017 0.7414 SD 0.0454 0.0676 0.0495 0.0652 0.0719 0.0675 Upper 0.6261 0.9582 1.0000 1.0000 0.9490 0.8828 Lower 0.4490 0.7148 0.8336 0.7400 0.6610 0.6162 NPV Mean NPV 0.9423 0.9556 0.9383 0.9443 0.9406 0.9474 SD 0.0303 0.0160 0.0177 0.0167 0.0181 0.0189 Upper 1.0000 0.9848 0.9691 0.9761 0.9749 0.9811 Lower 0.8749 0.9226 0.9013 0.9112 0.9044 0.9095 Study PPV Mean PPV 0.4595 0.9197 0.9798 0.9391 0.8979 0.8317 Prevalence SD 0.0469 0.0369 0.0205 0.0324 0.0397 0.0469 45 min = 21.7 Upper 0.5604 0.9794 1.0000 1.0000 0.9740 0.9255 90 min = 43.4 Lower 0.3762 0.8405 0.9336 0.8701 0.8153 0.7416 120 min = 49.6 NPV Mean NPV 0.9554 0.9126 0.8514 0.8808 0.8827 0.9103 150 min = 46.0 SD 0.0235 0.0299 0.0362 0.0341 0.0350 0.0313 180 min = 44.4 Upper 1.0000 0.9686 0.9242 0.9462 0.9487 0.9693 240 min = 40.0 Lower 0.9081 0.8568 0.7778 0.8163 0.8137 0.8441 57 PPV Mean PPV 0.8023 0.9503 0.9844 0.9589 0.9338 0.9096 SD 0.0287 0.0242 0.0154 0.0225 0.0276 0.0283 Upper 0.8572 0.9880 1.0000 1.0000 0.9852 0.9643 Lower 0.7450 0.8998 0.9472 0.9107 0.8748 0.8519 NPV Mean NPV 0.8255 0.8589 0.8112 0.8272 0.8173 0.8363 SD 0.0813 0.0462 0.0469 0.0457 0.0486 0.0524 Upper 1.0000 0.9477 0.8975 0.9194 0.9155 0.9355 Lower 0.6611 0.7689 0.7181 0.7412 0.7253 0.7372
(75) Other than the 45 minute time point, the PPV and NPV values associated with diagnosis by T.sub.max were superb and consistent at each time point and across changing prevalence. The 45 minute time point was not useful for diagnosis of gastroparesis when using either breath test or scintigraphy as this is the time point at which the least number of positive subjects were identified by either test method throughout the evaluation period.
(76) To further explore confidence in the T.sub.max metric, PPV and NPV values across changing prevalence fractions were calculated when diagnosis by T.sub.max was compared to diagnosis by scintigraphic t.sub.1/2 across changing prevalence fractions. BT_t.sub.1/2 was co-explored in the same analysis.
(77) The PPV and NPV values associated with diagnosis by T.sub.max and BT_t.sub.1/2 were compared to diagnosis by scintigraphic t.sub.1/2. Calculations were performed at 3 levels of prevalence: 27%; 43.5%; and 57%. The prevalence number of 43.5% is the prevalence of gastroparesis found in the 115 subjects where classification (normal or delayed) was established relative to the scintigraphic t.sub.1/2 cut-off point of 91.0 minutes. As noted, a statistical bootstrap procedure using 1000 samplings from the 115 test subjects data set with replacement was used to best estimate the true PPV and NPV values.
(78) TABLE-US-00008 TABLE 8 Bootstrap-Derived BT_t.sub.1/2 and T.sub.max PPV and NPV Values Across Changing Prevalence of Gastroparesis (Diagnosis by BT_t.sub.1/2 and T.sub.max vs. Scint_t.sub.1/2 ). Prevalence (%) Statistic 95% CI BT_t.sub.1/2 T.sub.max 27 PPV Mean PPV 0.6846 0.8442 SD 0.0669 0.0673 Upper 0.8187 1.0000 Lower 0.5616 0.7148 NPV Mean NPV 0.9503 0.9547 SD 0.0180 0.0173 Upper 0.9830 0.9844 Lower 0.9124 0.9165 Study PPV Mean PPV 0.8152 0.9178 Prevalence SD 0.0455 0.0368 43.5 Upper 0.9074 0.9793 Lower 0.7258 0.8468 NPV Mean NPV 0.9023 0.9134 SD 0.0349 0.0317 Upper 0.9649 0.9719 Lower 0.8307 0.8478 57 PPV Mean PPV 0.8842 0.9497 SD 0.0313 0.0240 Upper 0.9418 1.0000 Lower 0.8212 0.8998 NPV Mean NPV 0.8443 0.8566 SD 0.0506 0.0492 Upper 0.9416 0.9461 Lower 0.7440 0.7537
(79) Once again, PPV and NPV values associated with T.sub.max were superb across changing prevalence. BT_t.sub.1/2 also exhibited very good PPV and NPV values, but they were not as good as T.sub.max.
(80) The examples show that T.sub.max may be used for diagnosing gastroparesis with a very high level of confidence when prevalence is low or high in the intended use population. T.sub.max is therefore a unique and useful tool for diagnosing delayed gastric emptying.