Journal of Emerging Trends in Engineering and Applied Sciences (JETEAS) 4(6): 859-862

© Scholarlink Research Institute Journals, 2013 (ISSN: 2141-7016)

jeteas.scholarlinkresearch.org
Journal of Emerging Trends in Engineering and Applied Sciences (JETEAS) 4(6):859-862 (ISSN: 2141-7016)

Diagnosis of Kidney Failure by Analysis of the Concentration of

Ammonia in Exhaled Human Breath
Ima O. Essiet

Department of Electrical Engineering,

Bayero University, PMB 3011 Kano, Nigeria.
__________________________________________________________________________________________
Abstract
The human mouth contains many kinds of substances both in liquid and gaseous form. The individual
concentrations of each of these substances could provide useful insight to the health condition of the entire
body. Ammonia is one of such substances whose concentration in the mouth has revealed the presence or
absence of diseases in the body. Kidney failure is one disease which is identified by extremely high ammonia
content in human breath. This disease is as a result of the kidneys’ inability to process the body’s liquid waste
thereby resulting in high blood urea nitrogen (BUN) level. The result of this is the release of urea throughout the
body which is dissipated in the form of ammonia through oral breath. This paper proposes an affordable
ammonia breathalyzer for the diagnosis of kidney failure in humans with 85% success rate. The purpose of this
research is to provide an affordable and reliable means of detecting kidney failure in hospitals and even in
homes. This will help patients of the disease seek medical help quickly (especially those in rural areas). The
potential impact of this research is the availability of this means of BUN detection to less privileged people.
__________________________________________________________________________________________
Keywords: ammonia, BUN, kidney failure, metal oxide semiconductor.

INTRODUCTION diseases. Breath testing devices first appeared in 1784

The ammonia odour in the mouth of kidney failure when Lavoisier detected CO in exhaled breath of
patients is associated with high levels of blood urea guinea pigs (Gouma et al, 2009). Since then,
nitrogen (BUN). When the kidneys are not colorimetric assays and gas chromatography columns
functioning well, blood urea nitrogen will accumulate have been used to analyze VOCs (volatile organic
in the body. The excess urea will be decomposed into compounds) in human breath in quantities varying
ammonia under the effect of urea enzymes in the from millimolar (10-3M) to picomolar (10-12M)
gastrointestinal tract. Ammonia will be discharged concentrations (Gouma, 2009).
through the oral cavity(mouth) to cause ammonia
smells in the mouth. Also, high BUN in saliva can This article involves the development of an ammonia
also be decomposed into ammonia to make the breath analyzer which was tested on candidates with
ammonia odour in the mouth worse. and without the occurence of kidney failure in order
to compare the ammonia concentrations in their
Kidney failure reaches advanced stages in many breath. It was established that those with kidney
patients because it is not quickly diagnosed. This failure had an unusually high concentration of
problem has led to many unnecessary deaths from the ammonia in their breath compared to those who did
disease especially in rural areas where there is limited not exhibit the disease. The breathalyzer is
access to standard medical facilities. implemented using a metal oxide semiconductor
sensor which has been trained by a neural network to
Haemodialysis is a treatment that involves the use of identify ammonia in gaseous form (e.g. in human
specially designed machines to reduce the BUN breath). The sensor’s sensitivity to ammonia is in the
levels in patients who suffer from kidney failure. This parts per million (ppm) range.
due to the fact that diseased kidneys can no longer
filter the blood’s urea-nitrogen. As a result, there is a MATERIALS AND METHODS
build up of these compounds in the body to the point Metal oxide semiconductor odour sensors have been
that it becomes toxic. At this point, patients are said used extensively over the years to identify volatile
to suffer from kidney failure. Although analysis of organic compounds (VOCs) in varying
body fluids (blood, sputum,urine) for disease concentrations for different purposes. They have been
diagnosis and monitoring is routine clinical practice, used in agriculture (Kizil & Lindley, 2009), food
human breath analysis methodologies that exploit the industries (Balasubramanian, 2009 & Berna, 2010),
non-invasive nature of such an approach are still as well as for security purposes (Khawaja, 2009,
under-developed. Since the time of the ancient Schwarzbock, 2012, Pawar, 2012 & Kodogiannis,
physician Hippocrates, exhaled breath was 2008) to mention a few. Metal oxide sensors operate
recognized as a non-invasive tool to diagnose on the basis of the Sauerbrey equation:
859
Journal of Emerging Trends in Engineering and Applied Sciences (JETEAS) 4(6):859-862 (ISSN: 2141-7016)

m remained high after the same time interval. A 1kΩ

f  2.3  10 6  f 02  (1) potentiometer was used to adjust the reference
A voltage so that the circuit can function properly in
Where f is change in frequency, f0 is the resonant environments where the air quality varies. A toggle
frequency of the sensor, m is change in mass switch is used for powering on the circuit. The
caused by deposition of the volatile compound on the constructed prototype circuit is housed in a plastic
sensor and A is the area of the sensor. The above casing and the circuit diagram is shown in Figure 1.
equation is responsible for biasing the sensor’s
78L05 Switch
sensing element based on the fact that an increase in
IN OUT
the concentration of the odorant (ammonia) results in
COM
a change in the element’s fundamental frequency. 27 Ohm
Thus, the sensor is biased based on the detection of Red
LED
the odorant’s intensity which is one of the 1k 50% 1k
1k 1k
components of an odour (Pawar, 2012).
LM324
390 Ohm 33 Ohm
Due to the fact that the gas sensor’s desirable +
1.47k
performance is vital to ensuring that the breathalyzer 12V 1k
Green
gives accurate results, there is a need to simulate the Sensor
LED
1k 50%
chosen sensor’s response to molecules of ammonia
220 Ohm
gas. This was achieved by testing the sensor’s
response to ammonia gas using a neural network. The
results of the neural simulation revealed that the
sensor identified ammonia in various concentrations Figure 1. Electric circuit diagram of the breath
with a 93% success rate (Essiet & Dan-Isa, 2013). analyzer
The sensor being used in the design of the
breathalyzer has an ammonia sensitivity of CIRCUIT TESTING AND RESULTS
0.033V/1ppm. The circuit is powered on by means of the toggle
switch. In order for the sensor’s internal heating
However, in spite of this wide application one of the element to give accurate results, the circuit was left
greatest challenges facing breath analyzer designs is on for 1 – 2 minutes without any breath sample being
that of cost. Most forms of the breath analyzer that introduced. A rubber tube was attached to the sensor
have been successfully implemented are too through which each candidate exhaled. The circuit
expensive for use by the general public. As a result, gave an output voltage of between 0.6V and 1.1V
their widespread use has been severely limited. The (DC) for good breath condition (between 18ppm and
breath analyzer proposed in this article has been 33ppm ammonia), which activated the green Light
designed to address this limitation. The components Emitting Diode (LED). This voltage range is the
in the sensor’s circuitry are affordable compared to reference for the comparator’s inverting input.
most of the existing implementations.
The breath of ten candidates with kidney failure
DESIGN OF THE AMMONIA BREATH symptoms was tested using the proposed breathalyzer
ANALYZER SENSOR CIRCUIT as well as ten candidates without the ailment. Each
After extensive neural network simulation tests of candidate was made to exhale through a rubber tube
various ammonia sensors, the TGS 2602 sensor was attached to the ammonia sensor for about 10 seconds
selected. Based on the electrical characteristics of the while the circuit’s DC output voltage was monitored
chosen gas sensor, the circuit was designed with a to obtain the variation in ammonia concentration over
DC power supply of 12V but was regulated to 5V in the time interval. The candidates without kidney
order to supply the sensor circuit and LM324 failure were each observed to have breath ammonia
comparator Integrated Circuit (IC). A red LED was concentrations which did not exceed 50ppm(1.65V)
used to indicate proper circuit operation. The output within the first 5 seconds of exhalation. This
voltage of the sensor circuit along with the sensor’s ammonia concentration caused the green LED on the
sensitivity value were used to estimate the breath analyzer to go off before the exhalation period
concentration of ammonia in the exhaled breath expired indicating that breath ammonia
samples. A green LED was illuminated when the concentrations were normal. Candidates with kidney
comparator’s threshold value was exceeded (typically failure, however, had concentrations of up to
at the start of exhalation). For candidates without 100ppm(3.3V) with the worst cases being up to
kidney disease, the green LED went off after about 5 120ppm(3.96V). Results of breath analyzer tests for
seconds of exhalation. This indicated that their breath both sets of candidates are depicted in graphical form
ammonia levels began to drop after this interval. in figures 2 and 3.
However, the opposite was the case for kidney
disease candidates because their ammonia levels
860
Journal of Emerging Trends in Engineering and Applied Sciences (JETEAS) 4(6):859-862 (ISSN: 2141-7016)

Table 1. Confusion matrix showing accuracy of

results obtained for ammonia breath test samples
Kidney failure Non kidney
Accuracy
(predicted) failure
%
(predicted)
Kidney failure
(actual) 9 1 90.0
Non kidney
failure 2 8 80.0
(actual)
Overall
85.0
Accuracy

From the table above, it can be seen that 9 out 10

kidney failure cases were correctly identified while 8
out of 10 non-kidney failure cases were correctly
identified. The overall accuracy of the sensor is 85%.
The limitation of this study is that the circuit
designed can accurately diagnose kidney failure in a
Figure 2. Oral ammonia concentrations for 10 maximum of 5 patients at a time. Beyond this
candidates with kidney failure number, it is observed that the circuit begins to
malfunction due to overheating of the sensing
element. The circuit then has to be switched off for
about 30 minutes to allow the sensing element to
sufficiently cool down.

CONCLUSION
Oral ammonia concentrations of 60 – 120 ppm were
observed for candidates with kidney failure.
Candidates with less severe cases of kidney failure
had lower breath ammonia concentrations compared
to those with advanced symptoms. Candidates
without kidney failure had half the concentration of
oral ammonia compared to those with the disease.
Also, it was observed that oral ammonia
concentrations gradually decreased over the 10
second exhalation interval for all candidates tested.
The results obtained using the proposed breath
analyser show that breath ammonia analysis is an
effective and accurate method of diagnosing kidney
failure. It is therefore recommended that breath
Figure 3. Oral ammonia concentrations for 10 ammonia analysis be used by medical practitioners
candidates without kidney failure for the diagnosis of kidney failure.

From the figures above, it can be seen that the REFERENCES

candidates with kidney failure have approximately Balasubramanian, S. (2009). Neural Networks-
twice the concentration of oral ammonia content Integrated Metal Oxide-Based Artificial Olfactory
compared to those without kidney failure. Oral System for Meat Spoilage Identification. Journal of
ammonia concentrations of up to 120 ppm were Food Engineering , 91-98.
obtained for the worst of the kidney failure
candidates. Candidates without kidney failure had Berna, A. (2010). Metal Oxide Sensors for Electronic
oral ammonia concentrations of no more than 52 ppm Noses and Their Application to Food Analysis.
at most and 23 ppm at the least. It was also observed Sensors , 3882-3910.
that the candidate with the highest concentration of
oral ammonia (figure 2) also had the worst case of Davies, S. S. (1997). Quantitative Analysis of
kidney failure among the candidates tested. Table 1 Ammonia on Breath of Patients In End-Stage Renal
below shows a confusion matrix for breath analysis Failure. Kidney International , 223-228.
tests carried out on both sets of candidates. This is to
demonstrate the accuracy of the neural network Essiet, I. & Dan-Isa, A. (2013). Practical
simulations carried out to determine the suitability of Discrimination of Good and Bad Cooked Food Using
the TGS 2602 sensor to detect ammonia in human Metal Oxide Semiconductor Odour Sensor. Acta
breath. Periodica Technologica .
861
Journal of Emerging Trends in Engineering and Applied Sciences (JETEAS) 4(6):859-862 (ISSN: 2141-7016)

Gouma, P. (2009). Nanosensor and Breath Analyzer

for Ammonia Detection in Exhaled Human Breath.

Khawaja, J. (2009). ASIC Gas Sensors Based on

Ratiometric Principles. Warwick: University of
Warwick.

Kizil, U. & Lindley, J.A. (2009). Potential Useof Gas

Sensors in Beef Manure Nutrient Content
Estimations. African Journal of Biotechnology ,
2790-2795.

Kodogiannis, V. (2008). Artificial Odour

Discrimination System Using Electronic Nose and
Neural Networks for the Identification of Urinary
Tract Infections. IEEE Transactions on Information
Technology in Biomedicine , 707-708.

Pawar, N. (2012). Nanostructured Fe2O3 Thick Film

As Ethanol Sensor. International Journal on Smart
Sensing and Intelligent Systems , 441-457.

Phillips, M. (2002). Disease Markers in Exhaled

Breath. New York: Marcel Decker.

Prasad, A. (2003). Reactively Sputtered MoO3 Films

for Ammonia Sensing. Thin Solid Films , 46-51.

Risby, T. &. (1999). Clinical Application of Breath

Biomarkers of Oxidative Stress Status. Free Radical
Biol. Med. , 1182-1192.

Schwarzbock, T. (2012). Market Review on

Available Instruments for Odour Measurement.
Berlin: Kompetenzzentrum Wasser.

862