Patent Application
20230133069
The present invention is related to an apparatus for measuring ammonia concentration in the exhaled breath for the determination and monitoring of chronic kidney disease.
Chronic kidney disease (CKD), a condition in which a patient's kidneys fail to effectively filter their blood, is estimated to affect more than one in seven American adults. It places one of the largest burdens on the healthcare system costing Medicare over $120 billion in 2017 alone. Commonly called the “silent disease,” CKD presents with no symptoms in its early stages; therefore, most patients are diagnosed at an advanced stage leading to poor prognosis. Hemodialysis or kidney transplantation are the best treatments for the most advanced stage of kidney disease known as end-stage renal failure (ESRF). However, due to the high costs and inconvenience of hemodialysis and a lack of available kidneys for transplant, mortality rates remain high for those with ESRF. Therefore, better management of chronic kidney disease in its earlier stages is the key to overcoming the disease.
Blood urea nitrogen (BUN) levels and glomerular filtration rates (GFR) are used as the criteria for diagnosing and staging chronic kidney diseases. Currently, blood and urine tests are necessary to obtain the measurements needed to monitor and manage kidney disease state progression. These tests require patient compliancy as they are performed in clinics where experienced personnel will collect and analyze the samples. Moreover, these sampling procedures are invasive, embarrassing, and can be inaccurate as they have increased chances of preanalytical errors and require a longer time before diagnosis. To resolve the faults of these techniques, new research has shifted to exploiting biomarkers other than blood and urine as a diagnostic aid.
Breath has emerged as an ideal solution to this problem due to its non-invasive nature and ability to take repeat measurements with minimal discomfort to the patient. While a plethora of new breath-analysis research has developed in the past twenty years, examining exhaled breath is an ancient practice. Even in the time of Hippocrates, doctors were “smelling” the breath of patients to diagnose diseases by linking a sweet odor with diabetes and a fishy odor with kidney-related diseases. Since 1971 when Linus Pauling discovered that exhaled breath is a complex gas containing at least 200 volatile organic compounds (VOCs), much work has been done to identify all components of exhaled breath, which is now known to contain more than 3,500 VOCs. The levels of VOCs present in exhaled breath change when a healthy individual enters a pathological state, and this change may be detected and utilized for diagnosis and monitoring of a disease. Therefore, breath analysis has great potential to provide a cost-effective, real-time, quantifiable, diagnostic point-of-care detection method.
In patients suffering from kidney disease, the elevated levels of the VOC ammonia are responsible for the fish-like odor in exhaled breath. Ammonia production results from metabolic processes in the body, primarily protein metabolism by gut bacteria. In a healthy individual, the liver converts this ammonia to urea, and the kidneys filter out the waste through the urine. However, when the kidneys fail to function properly, waste will accumulate in the bloodstream. Ammonia and urea convert reversibly, so urea waste accumulation results in higher systemic levels of ammonium ions in the bloodstream. The excess urea will also increase urea concentration in saliva where it is degraded into gaseous ammonia, and the excess ammonium ions in the blood stream also travel to the lung alveoli where they undergo gas exchange. Both fates result in the increased levels of gaseous ammonia detected in the patient's exhaled breath. Previous studies have documented the association between these increased ammonia concentrations and decreased kidney function. While a healthy individual will typically have an exhaled breath ammonia level in a range from 50 to 1,500 parts-per-billion (ppb), patients suffering from end-stage renal failure can have ammonia levels as high as 15,000 ppb. Therefore, ammonia levels in exhaled breath have the potential to be exploited to track disease-state progression of those suffering from chronic kidney disease.
Currently, there exist techniques such as gas chromatography mass spectrometry, selected-ion-flow-tube mass spectrometry, and laser-acoustic-based spectrometry, which can be utilized to determine VOC levels in exhaled breath. However, diagnostic devices that employ these techniques are bulky, costly, require experienced personnel, and are not suitable for at-home use. And while novel technologies comprised of nanotechnology or smart polymers have recently been explored, they currently do not have the ability to detect trace VOCs at the sensitivity needed for clinical application.
In spite of the advances, those of skill in the art still desire a simple, quick, and low-cost method for measuring ammonia concentration in exhaled breath, thereby leading to improved care and decreased morbidity rates.
It is an object of the invention to provide a simple, quick, and low-cost method, and device, for measuring ammonia concentration in exhaled breath.
A particular feature of the invention is the ability to provide a device which can be used without medical personnel thereby allowing a patient to monitor and track a disease, particularly CKD, associated with increased ammonia in exhaled breath.
A particular advantage is the ability to detect diseases associated with high ammonia levels in exhaled breath early thereby allowing for early medical intervention.
These and other advantages, as will be realized, are provided in an apparatus for measuring breath ammonia concentration in exhaled breath. The apparatus comprises a means to monitor breath flow rate, a sample collection portion, capable of capturing exhaled breath, and an analysis portion capable of responding to the concentration of ammonia in said exhaled breath from said breath sample. The analysis portion comprises a colorimetric sensor comprising an active component on a substrate. The colorimetric sensor changes colors proportional to breath ammonia concentration in said exhaled breath.
Yet another embodiment is provided in a method for determining breath ammonia in exhaled breath comprising:
providing an apparatus comprising:
a sample collection portion capable of capturing exhaled breath; and an analysis portion capable of responding to the concentration of ammonia in the exhaled breath wherein the analysis portion comprises a colorimetric sensor comprising an active component on a substrate wherein the colorimetric sensor changes color proportional to breath ammonia concentration in the exhaled breath;
calibrating the apparatus by passing a series of control gases simulating exhaled breath with known ammonia concentrations through the apparatus to determine a relationship between a control color change and the known ammonia concentrations;
having a patient exhale breath into the apparatus;
measuring a clinical color change; and determining a breath ammonia concentration in said exhaled breath based on said relationship.
The present invention is related to a simple, at-home method for patients to routinely track the breath ammonia levels in their exhaled breath. This is accomplished by a system capable of detecting gaseous ammonia in exhaled breath. More specifically, the present invention is related to an apparatus which a patient can exhale breath into wherein the exhaled breath is captured and the concentration of breath ammonia in the exhaled breath can be determined by colorimetric methods.
Provided herein is a non-invasive, exhaled breath-based test that provides the patient an indication of their health in real-time. The method is fast, user-friendly, and low-cost allowing patients to regularly track their breath ammonia levels from the comfort of their home thereby allowing a patient to assess their health more often and at a lower cost providing clinicians frequent assessment of their patients' levels to assist in making therapeutic decisions to help slow the progression of CKD or other diseases which cause a high breath ammonia concentration in exhaled breath.
Of particular importance is the ability for at-home monitoring for patients with chronic kidney disease (CKD). At-home monitoring of ammonia is also important for other conditions such as for individuals with the rare metabolic disorder known such as a urea cycle disorder (UCD), which can lead to critical life-threatening increases in ammonia concentration in the blood stream. Patients suffering from hepatic encephalopathy (HE) have also been shown to have increased levels of ammonia in their blood and these patients may benefit from the invention. Another particular feature is the ability to track progression of these diseases allowing patients to better manage their ammonia levels and reduce the risk of a life-threatening ammonia spike.
The invention will be described with reference to the figures which are integral, but non-limiting, part of the specification provided for clarity of the invention. Throughout the various figures similar elements will be numbered according.
An embodiment of the invention is illustrated schematically in
The outlet of the internal cylinder is preferably placed approximately 5 mm (0.2″) from the sensor such that the stream of exhaled breath impinges in a perpendicular direction onto the central region of the sensor by being press fit into the bottom of the inlet tube. This arrangement directs the gas sample to impede onto the exposed area of the sensor to achieve functional contact. Changing the distance from the end of the internal cylinder to the sensor can adjust the sensitivity of the apparatus.
The colorimetric sensor comprises an active component capable of changing color proportional to the amount of breath ammonia impinging on the colorimetric sensor wherein the active component is on a substrate. The active component of the colorimetric sensor is preferably selected from the group consisting of Bromocresol Green (BCG) preferred, Bromophenol Blue (BPB), Bromocresol Purple (BCP), Titanium(IV) Oxide, Berthelot's reagent, and Hydroquinone. Bromocresol Green (BCG) is preferred. Bromocresol Green (BCG) (ACS Regent, #114359-5G), Bromophenol Blue (BPB) (ACS Reagent, #114391-5G), Bromocresol Purple (BCP) (ACS Reagent, #B5880-5G), Titanium(IV) Oxide (TiO2) (Nanopowder, 21 nm, #718467-100G), and Hydroquinone (Reagent plus, #H9003-100G) are all commercially available Sigma Aldrich. Bromophenol Blue is yellow in acid and turns blue in a pH range of 3.0-4.6. Bromocresol Purple is yellow in acid and turns purple in a pH range of 5.2-6.8. Bromocresol green is yellow in acid and turns green in a pH range of 3.8-5.4.
A particularly preferred colorimetric sensor comprises Bromocresol Green (BCG) Powder (ACS Regent, #114359-5G), methanol solution (Fisher Chemical, CAS 67-56-1), on a Whatman grade 3 MM cellulose chromatography paper substrate from a 25 mm diameter (Whatman, product #1030-025) are particularly suitable for demonstration of the invention. A housing unit comprising polycarbonate, polypropylene, and Teflon® components, a 1-L sampling bag, a pressure gauge, and a mouthpiece is also suitable for demonstration of the invention.
In an embodiment the internal cylinder can be made from a relatively hydrophobic polymer such as polypropylene with the donut, or annular, shaped holder made from a relatively hydrophobic polymer such as polytetrafluoroethylene. All other components can be made from polycarbonate. Sample volume can be controlled by the addition of a sample bag to collect the exhaled breath. A Tedlar sample bag which is at last 0.1 L to no more than 5 L is preferable with a 1-L Tedlar sample bag (Cole-Parmer, item #EW-86561-00) being particularly suitable for demonstration of the invention.
A pressure gauge is preferably used to demonstrate the invention thereby providing feedback regarding back pressure to the user during sampling to control the flowrate of the sample. A pressure gauge from Grainger, item #491064, is particularly suitable for demonstration of the invention.
A mouthpiece is preferably attached to the inlet tube of the housing unit. A particularly preferred mouthpiece comprises a one-way valve to trap the exhaled breath sample inside the unit and eliminates the chance of rebreathing the exhaled breath. A Carolina Diagnostic Solutions, item #SKU: 20980 mouthpiece is particularly suitable for demonstration of the invention.
In an exemplary embodiment colorimetric disc sensors can be formed from 50 mg of bromocresol green powder dissolved in 50 mL methanol solution referred to herein as BCG solution. To demonstrate the invention a 150 μL of BCG solution would be pipetted onto the middle of a 25 mm diameter Whatman 3 MM chromatography paper preferably laid flat on a plate made from a hydrophobic polymer such as polyethylene, polypropylene or polytetrafluoroethylene. This would be left to dry for about 30 minutes and then stored in a sealed package, such as a relatively thick plastic bag, to prevent exposure to ambient air prior to use.
The substrate is preferably selected from chromatography paper and particularly Whatman grade 3 MM cellulose chromatography paper.
The components are assembled to create the apparatus which a user can breathe into. The components can be assembled by press fitting the internal cylinder preferably about 0.2″ (5 mm) into the bottom of the inlet tube. The column would then be placed around the internal cylinder and preferably screwed onto mating threads located on the bottom of the inlet tube. The side arm can be glued to the side of the column with aligned orifices such as the 0.375″ (9.525 mm) holes. A colorimetric disc sensor is placed in the bottom cap and covered with a hydrophobic polymer, preferably polypropylene or polytetrafluoroethylene, donut or annular shaped holder. The bottom cap can then be screwed onto the bottom of the column such as with mating threads. This geometry leaves approximately 0.2″ (5 mm) between the outlet of the internal cylinder and the colorimetric disc sensor, with the flow of exhaled breath directed to impinge on the central region of the sensor disc in a perpendicular direction. The pressure gauge can be screwed into the top of the inlet tube, and the mouthpiece can be press fit into the side of the inlet tube. The sampling bag can be press fit into the side arm.
It is preferable to calibrate the apparatus by providing a controlled amount of a control gas comprising a known amount of ammonia gas in a known volume of humidified carbon dioxide and air to mimic exhaled breath with NH3 concentration from 0 to 12 ppm. Particularly preferred is a 5% CO2-in-air gas cylinder which flows through a water bath, such as at 44° C. to humidify the CO2-air stream to about 94% relative humidity, followed by the addition of an ammonia-in-air gas stream to the flowing CO2-in-air into the apparatus. Mass flow controllers would be utilized as would be apparent to one of skill in the art. For the purposes of demonstration of the invention a 5% CO2-in-air cylinder is suitable for use which can be purchased from Airgas (part #X02Al95C2000117). The ammonia source can be purchased as 20-ppm or 30-ppm ammonia-in-air stock cylinders from Airgas (part #X02Al99CP580081 and X02Al99CP5837Q1, respectively). A regulator suitable for use with the ammonia cylinders is available commercially from Airgas (part #Y111226CC10-AL). Mass Flow Controllers from Cole-Parmer, referred to as Masterflex Proportional Flowmeter Controller, Mass, 10 L/min Gas (item #UX-32907-7)1 and 5 L/min Gas (item #HV-32907-69) are suitable for demonstration of the invention.
As would be realized, the sensor disc changes color in proportion to the ammonia concentration. The color change of a sensor disc after a test run can be analyzed by a quantifying color scale, such as RGB (red-green-blue) color with either control or accounting for ambient lighting conditions. For example, a sensor disc can be placed on a 2″×2″ white plastic plate under an LED-lit enclosure, to control the lighting, and the disc can be photographed such as by a smartphone camera followed by RGB color analysis to quantify the color change. Alternatively, the color change could be simply determined by comparing it by eye against a set of standard color patches that are calibrated to known ammonia concentrations.
To prepare the gas flow system, 100 mL of DI water can be added to a 500 mL side-arm flask and placed in a water bath set to 44° C. The mass flow controllers can be connected to the gas cylinders with dry PVC tubing. The 5% CO2-in-air gas line is preferably humidified by its introduction into the headspace of a sealed side arm flask filled with DI water placed in the water bath. The humidified air could exit out the side arm of the flask and combined with the ammonia-air stream to create the final desired concentration of gaseous ammonia. Dry PVC tubing is preferably used to connect the combined airflow system to the housing unit.
The system can be first flushed with dry 5% CO2 in air for 15 seconds for calibration. A colorimetric disc sensor would be placed in the bottom cap of the housing unit, covered with the, preferably Teflon, holder, and the apparatus would be sealed. The exhaled breath collection bag, such as a 1-L Tedlar bag, would be attached to the side arm and its valve fitting would be opened.
For calibration purposes the mass flow controllers would be set as needed to obtain the desired NH3 concentration such as at a combined total flowrate of 4 liters per minute (LPM) with an exemplary range of flow rates suitable for calibration provided in Table 1.
After exposure of the colorimetric sensor to the simulated exhaled breath stream, for the purposes of calibration, or exhaled breath, in actual use, the colorimetric sensor would be immediately removed from the apparatus, placed face up under the LED-lit enclosure when controlled lighting is desired, and the color determined such as by photography, such as with a smartphone camera, followed by determination of the color of the colorimetric sample using a quantitative color scale.
For the purposes of demonstration of the invention, the images could be saved and imported into an RGB color-analysis system. It is preferable to average the RGB values for a group of N samples. A representative graph of the average change in Red, Green and Blue is illustrated in
A Plot and accompanying lines of best fit of the total color change response of the calibration sensors is illustrated graphically in
The calibration results were based on 90 colorimetric disc sensors. Repeat samples were completed at increasing levels of gaseous ammonia to create a plot of total color change versus ammonia concentration. One-L gas samples provided adequate sensitivity and 1 L of exhaled breath was an appropriate amount to ask of a patient based on preliminary studies. Typical users took approximately 15-20 seconds to exhale 1-L of exhaled breath through the apparatus. This noted time for testing is the basis for the use of 4 LPM as the calibration flowrate. Furthermore, a user can easily maintain a pressure reading of 4-5 water column inches (in. wc) on the pressure gauge during use, and a 4 LPM flowrate corresponded to a 4 in. wc pressure on the pressure gauge. These constants were selected to replicate end-user considerations. Due to observations seen in the preliminary work, two gas cylinders were used for calibrations (one cylinder of 5% CO2-in-air, and the other with either 20 ppm or 30 ppm NH3-in-air). Sensors were exposed to 1-L of gas sample and immediately imaged. The images were processed and analyzed using a MATLAB code.
For the purposes of demonstration of the invention a representative MATLAB code is:
clear
cic
close all
% create blank arrays to store sample values
differenceinR_array=[ ];
differenceinG_array=[ ];
differenceinB_array=[ ];
colorchangeinRGB_array=[ ];
for i=1:3%
read images based on their stored file name; the number after Patient, this will change for each patient tested
filename=strcat(‘Patient-6-’,num2str(i),‘.png’);
sample=imread(filename);
sample_lab=rgb2lab(sample);
ab=sample_lab(:,:,2:3);
ab=im2single(ab);
% use k-means to recognize two color clusters in the image
[L, centers]=imsegkmeans(ab,2,‘NumAttempts’,3);
% get the size of the 2D array to iterate through the R,G,B layers
dim=size(sample);
len=dim(1)*dim(2);
% choose which labels from k-means to get indices of; label 1 is the cluster with more pixels
morepixel=find(L==1);
lesspixel=find(L==2);
% compute averages for the RGB values of the two clusters
r1=sum(sample(morepixel))/length(morepixel);
g1=sum(sample(morepixel+len))/length(morepixel);
b1=sum(sample(morepixel+(2*len)))/length(morepixel);
r2=sum(sample(lesspixel))/length(lesspixel);
g2=sum(sample(lesspixel+len))/length(lesspixel);
b2=sum(sample(lesspixel+(2*len)))/length(lesspixel);
if r2>r1
% a larger r2 value indicates r2, b2, g2 are the yellow part of the sensor
mask1=L==2;
% sample_lab(:,:,x) is just one of the L, a, or b values
% this applies the mask over the image, only showing the parts in the mask
cluster1=double(sample_lab(:,:,1)).*double(sample_lab(:,:,3)).* mask1;
% computes the difference in values by substracting yellow part of sensor from blue part of sensor
differenceinR=(r1-r2);
differenceinG=(g1-g2);
differenceinB=(b1-b2);
% outputs the RGB values and change in color for the sensor
fprintf(‘Sample: %.1f\n’,i)
fprintf(‘Average RGB Values for Yellow Area of Sensor R: %.2f G: %.2f B: %.2f\n’,r2,g2,b2)
fprintf(‘Average RGB Values for Blue Area of Sensor R: %.2f G: %.2f B: %.2f\n’,r1,g1,b1)
fprintf(‘Change in Values R: %.2f G: %.2f B: %.2f\n’,differenceinR,differenceinG,differenceinB)
else
mask1=L==1;
cluster1=double(sample_lab(:,:,1)).*double(sample_lab(:,:,3)).* mask1;
% computes the difference in values by substracting yellow part of sensor from blue part of sensor
differenceinR=(r2-r1);
differenceinG=(g2-g1);
differenceinB=(b2-b1);
% outputs the RGB values and change in color for the sensor
fprintf(‘Sample: %.1f\n’,i)
fprintf(‘Average RGB Values for Yellow Area of Sensor R: %.2f G: %.2f B: %.2f\n’,r1,g1,b1)
fprintf(‘Average RGB Values for Blue Area of Sensor R: %.2f G: %.2f B: %.2f\n’,r2,g2,b2)
fprintf(‘Change in Values R: %.2f G: %.2f B: %.2f\n’,differenceinR,differenceinG,differenceinB)
end
% create figures to show the original images
figure(1)
subplot(3, 5, i)
imshow(sample)
title(filename)
% create figures to show the clustered images
figure(2)
subplot(3, 5, i)
imshow(cluster1)
title(filename)
% computes total change in color
totalcolorchange=sqrt((differenceinR)A2+(differenceinG)A2+(differenceinB)A2); % outputs color change to the command window
As the concentration of NH3 increases, the colorimetric disc sensor becomes increasingly more green. The sensor's original color is yellow and has average RGB values of 155, 120, and 15, respectively. Exposure to basic, gaseous ammonia changes the sensor's color to green. These value changes are detected by a MATLAB code that clusters the image into two groups with the inherent assumption that there will be two groups within the image. This explains why MATLAB outputs a slight color change for the 0 ppm samples even though there is no discernable color change of the 0 ppm sensor. It is further recognized that there is no color change of the sensor as the MATLAB code does not detect the circular area in the 0 ppm sample that was exposed to the analyte. The 1 ppm and 2 ppm samples clearly indicate a color change.
MATLAB calculates the change in RGB values between the two segments of the sensor by taking the RGB values of the green region and subtracting the RGB values of the yellow region. The averages and standard deviations of these changes are shown in
The change in red values had the lowest standard deviations and therefore the change in red value is preferably used for the calibration plots. However, as seen in Table 3, after the concentration of ammonia reaches greater than about 4,000 ppb (4 ppm), the groups of 6,000-8,000, 8,000-10000, and 10000-12000 are not significantly different, while the color change between these three groupings is statistically significant. A similar trend was identified for the green and blue values. The blue values also showed similarity at lower levels of ammonia.
The use of the white donut cover over the sensor protects the outer yellow region of the sensor with the central region of the sensor exposed to the gas stream.
The sensor's change in color is due to the basic ammonia reacting with the bromocresol green indicator embedded in the substrate. This is a surface reaction which is limited by the available BCG reactant on the exposed area of the sensor. Therefore, the technology will reach a saturation limit.
To enhance sensitivity of the design, the MATLAB code computes and sums the changes in red, green, and blue values to obtain the total color change by the following equation:
ΔRGB=√{square root over ((ΔR)2+(ΔG)2+(ΔB)2)}
Averages and standard deviations of the total color change were received from the MATLAB code. Tables 8 through 10 provide the statistics and ANOVA one-way comparisons of the total color change by NH3 concentration. As seen in Tables 9 and 10 statistically significant differences (p<0.001) were detected at 1,000 ppb intervals of humidified, gaseous ammonia at clinically relevant concentrations (between 0 and 6,000 ppb). Table 10 provides the grouping information, where the saturation effect is seen to begin after 4,000 ppb NH3 concentration. While there is still a clear increase in value, the samples of groupings of 6000-8000, 8000-10000, and 10000-12000 are not significantly different, while the color change between these three groupings is statistically significant.
Total color change (on the x-axis) is plotted against NH3 concentration [ppb] (on the y-axis) in the final calibration plot, provided in
These two lines of best fit and the inflection point of 4,000 ppb were chosen for multiple reasons. First, it is clear from the graph that there is a linear to nonlinear shift that occurs around the 4,000 ppb data point. Additionally, 4,000 ppb was selected based on findings in the literature and minimization of residual error. Based on published literature on ammonia levels in the exhaled breath, the critical range of ammonia concentrations was taken to be from 0-4,000 ppb and more preferably 0-12,000 ppb.
The y-intercept of the line of best fit was adjusted to be zero as a 0 ppb sample should theoretically provide zero change in color. Visually, there is no color change in the 0 ppb sensor samples. However, due to the inherent assumption embedded in the MATLAB code that detects two groups in the imported images, the program will find two groupings in any image and compute the difference between them. When extracting data from the 0 ppb images, it segments two groups that are extremely similar in RGB values which is why the total color change is so minimal. To accommodate for this issue, the y-intercept of the linear calibration equation is adjusted to zero. This linear equation also provided a correlation value of 0.997.
A rational function was used to characterize the higher region of the data since the data appears to be reaching a vertical asymptote around 115. After about 4,000 ppb, even as concentrations of gaseous ammonia increased, the color change was trending toward that vertical asymptote, or saturation limit. This phenomenon is due to the nature of the technology; the reaction between BCG and gaseous ammonia is a surface reaction which depends on the amount of BCG reactant available. Therefore, as the amount of unreacted BCG available decreases to zero, the total change in color will reach a maximum saturation limit. The rational function was determined in Excel using data solver to minimize the sum of residual squares. Further, this equation provided a correlation value of 0.9775.
The least squares method was used to confirm 4,000 ppb was an appropriate inflection point with the aforementioned lines of best fit. 3,000, 4,000, and 5,000 ppb were investigated as points of inflection, and 4,000 ppb was confirmed as the appropriate choice as it minimized the sum of the residuals.
The calibration plot is provided in
The apparatus was evaluated for patients with chronic kidney disease (CKD) at stages 1, 3, and 5, with patients at stage 1 disease representing the control group that should have minimal elevation of breath ammonia concentration. The components of the apparatus were sterilized by autoclaving before use, and individual, disposable mouthpieces and sample bags were used for each test. The device was assembled in-clinic.
After assembly, patients were handed the device and asked to place the mouthpiece in their mouth, inhale through their nose and exhale through the mouthpiece into the device while maintaining between 4 and 5 inches of water column (wc) static pressure reading on the pressure gauge. This breathing technique was continued until the 1-L Tedlar gas-sampling bag at the outlet arm of the device was full, thus providing control of sample volume (1-L total volume). The sample bag was sealed and removed, and the bottom cap containing the colorimetric disc sensor was removed. The sensor was removed, placed under the LED-lit enclosure to control for ambient lighting, and the resulting color change was photographed using a smartphone camera. The image was analyzed using the MATLAB program to extract the total color change of the sensor. The total color change was used to calculate the breath ammonia levels in the exhaled breath via the calibration curves as previously described. These results were documented in the secure database after de-identification.
Patients completed this study during their routine clinic visit, so blood tests were also performed around the time of testing. The metrics determined from the blood sample which included blood urea nitrogen (BUN), creatinine levels, and estimated glomerular filtration rate (eGFR) were also entered into the secure database after de-identification. These metrics were compared to the results obtained from the exhaled breath test to evaluate the level of correlation.
For the clinical evaluation, a set of at least three patients each with stage 1, 3, and 5 CKD were tested. Three independent tests for each patient provided statistical data on the performance of the exhaled breath ammonia sensors. The breath ammonia concentrations were averaged together for each patient to calculate their mean value, thus providing a group sampling of at least N=3 for each CKD stage. The values were statistically analyzed by one-way analysis of variance (ANOVA) with post-hoc Tukey multiple comparison to determine if the measured breath ammonia levels were significantly different (p<0.05) between the stage 1, 3, and 5 patients. Values for each group were plotted versus BUN, blood creatinine values, and eGFR, which are normally determined at each clinical visit, to assess the level of correlation between these metrics. Significantly elevated breath ammonia levels between CKD patient populations were predicted, and it was anticipated that there will be a correlation between exhaled breath ammonia and BUN, eGFR, and creatinine levels. The compiled patient data from pilot clinical trials is provided in Table 11. The blood urea nitrogen (BUN) levels plotted against detected breath ammonia (NH3) levels, N=3, per patient are provided in
The clinical trial was completed with pediatric patients suffering from stage 1, 3, and 5 chronic kidney disease. Patients were anywhere from 8 to 18 years of age, and clinic appointments were completed at all times of the day. Further, since patients were recruited during their visit and not before, they were not required to adhere to any dietary, fasting, or oral hygiene conditions prior to testing.
The results demonstrate the potential for providing evidence for the deployment of the developed technology as a simple, low-cost way for patients to personally monitor their disease progression from the comfort of their home. Images of the colorimetric sensors after three independent exhaled breath tests were completed by a patient at stage 1, 3 and 5 CKD respectively. There was a distinct increase in color change observed between the stages. The repeatability of the system was also observed. While the standard deviations for the three tests is minimal for those of lower CKD stages, this variation is increased for stage 5 patients.
Table 11 provides the compiled patient data from the ten patients who participated in the pilot clinical trial. The average detected breath ammonia level, N=3, and standard deviations are provided for each patient as well as their corresponding blood urea nitrogen (BUN) level, creatinine level, estimated glomerular filtration rate (eGFR), and stage of disease.
Table 12 provides the statistics on detected levels of NH3 in exhaled breath by stage of chronic kidney disease. Table 13 shows the results of ANOVA with 95% confidence comparing the levels of breath ammonia by stage of disease (F=39.62 and p<0.001). Further, Table 14 shows the results of Tukey simultaneous tests for differences of means with 95% confidence for the levels of breath ammonia by stage of disease. Stage 5 patients were statistically significantly different than stage 1 and 3 respectively (p<0.001), but there was not a statistically significant difference (p=0.565) between stage 1 and 3. However, applying the Tukey Method to the corresponding BUN levels of patients grouped by stage of disease did not give a statistically significant difference (p=0.055) between stage 1 and 3 either. Therefore, it can be concluded that the prototype successfully detected levels of breath ammonia in the exhaled breath, or breath ammonia, of patients staged 1, 3, and 5, and the detected NH3 concentrations strongly correlated with BUN levels.
Variability in results seen from patients of the same stage may be explained by differences in diet, dental hygiene, and breathing pattern.
Some studies have shown that different phases of a single exhale may have different levels of trace compounds in them. Breath exhalation may be broken into three phases. Phase I, or the start of exhalation, is mostly anatomical dead space, i.e. consisting of the air that was just inhaled but not mixed with alveolar air. Phase II is where exhaled breath has mixed with alveolar air; this phase is characterized by a large increase in carbon dioxide. Phase III of exhalation is where carbon dioxide levels have plateaued at their peak value. This breath phase contains the most water vapor, comes from the deepest reserves of one's lungs, and thus is the most indicative of the concentrations of analytes present in the blood stream. Therefore, it may be assumed that Phase III exhaled breath will contain a larger amount of breath ammonia since it is more representative of gas concentrations at the alveolar interface which come from the blood stream.
During clinical trials, patients were asked to inhale through their nose and exhale through their mouth into the device while maintaining a constant 4-5 in. wc pressure on the pressure gauge. However, it was observed that some patients completed the test with an increased amount of shallower breaths and others completed the test with only one or two large exhales. Nevertheless, both still maintained the constant 4-5 in. wc pressure during exhalation. Patient 5, stage 1 with BUN level 18 mg/dL, completed the tests with two large exhales and provided a breath ammonia concentration of 3,530 ppb. While Patient 8, stage 1 with BUN level 12 mg/dL, completed the test via a more tidal and shallower breathing pattern and provided a breath ammonia concentration of 173 ppb. If it is assumed that the shallow, tidal breathing pattern contains more Phase I breath while large exhalations contain more Phase III breath, then this difference in breathing pattern could account for the difference in breath ammonia levels observed. This is especially true if Phase III exhalation does contain more water vapor, which has been shown to improve the sensitivity of colorimetric technologies for this particular device.
In one embodiment, it is preferable that the first portion of the exhaled breath be excluded from the data such as the first 200 mL of exhaled breath to minimize the influence of oral cavity contamination or dilution of samples.
The results obtained from the pilot clinical trials provide sufficient evidence to support the deployment of this technology as an easy-to-use, quick, at-home, and low-cost way for a patient to assess and manage their kidney health. While there may be variability between users, there was a decreased deviation observed in the independent tests performed by each patient. Further, the results obtained from Patient 3 before and after hemodialysis eliminated the user-to-user variability and controlled the other considerations such as diet and dental habits, as this patient did not partake in any activities other than hemodialysis in between the testing sessions. The response in sensor trended nicely with the corresponding BUN levels showing that the sensor was selectively responding to ammonia in the exhaled breath.
Statistically significant differences were found between stages 1 and 5 and stages 3 and 5, but not for stages 1 and 3 CKD. However, a strong correlation was found between blood urea nitrogen levels and breath ammonia levels providing evidence that this prototype system could be used as a non-invasive way of tracking BUN levels, thus providing the potential to greatly improve the current cycle of care for the CKD patient population.
The color response of the simulated exhaled breath was evaluated using three sizes of the calibrated Tedlar bags (0.5, 1.0, and 3.0 L volume) to control the total volume of exhaled breath passed through the system during a test when using a 4.0 LPM total gas flow rate. When the concentration of NH3 in the simulated exhaled breath was varied over the physiological range of 0 to 8 ppm, the response of the color-indicating disc increased as the volume of exhaled breath was increased per test. The volume of the exhaled-breath collection bag is thus shown to directly increase the color response, and thus the sensitivity of the breath test for NH3. The results are presented in
The color response of the simulated exhaled breath was evaluated using three different flow rates of simulated exhaled breath (2, 4, and 6 LPM), each using the 1-L Tedlar bag for breath-volume control. The response of the color-indicating disc slightly increases as the flow rate of simulated exhaled breath is decreased. Comparison with the prior slide shows that the color response is not as sensitive to gas flow rate as it is the volume of exhaled breath. The results are presented in
Simulated exhaled-breath system with the 1-L Tedlar bag to control breath volume and a 4.0 SLPM total gas flow rate, the color response of the simulated exhaled breath was evaluated with and without the small-diameter internal cylinder in place. With the small-diameter internal cylinder in place to focus the stream of simulated exhaled breath on the center of the color indicating disc, the color response is about 100% greater than the color response that is obtained without the small-diameter internal cylinder in place, thus providing greatly enhanced sensitivity. The results are provided in
The results of patients having various stages of chronic kidney disease was studied. The exhaled-breath test was used by patients exhibiting various stages of chronic kidney disease, thus having different BUN levels, which indicate the degree of urea clearance by the kidneys. Each data point shown in
The invention has been described with reference to preferred embodiments without limit thereto. One of skill in the art would realize additional embodiments which are described and set forth in the claims appended hereto.
