Detection of COVID-19 by quantitative analysis of carbonyl compounds in exhaled breath | Scientific Reports – Nature.com

Study participants

The research protocol of this study was approved by University of Louisville Institutional Review Board (IRB Number 20.1154). All research was performed in accordance with the Declaration of Helsinki and the relevant guidelines/regulations of the IRB. Informed consent was obtained from all participants. Participants were enrolled from the Travel Clinic of the Division of Infectious Diseases at the University of Louisville and the University of Louisville Health Hospitals in Louisville, Kentucky. The Travel Clinic offered COVID-19 PCR testing required prior to international travel, testing for employees of local businesses that required a negative test result prior to returning to their workplace, and for patients requiring a negative PCR test prior to an out-patient surgical procedure. The majority of recruited participants from the Travel Clinic did not exhibit any symptoms of COVID-19 infection and most of the participants were COVID-19 negative from PCR test. Subjects recruited at the hospitals were patients most with mild COVID-19 symptoms and also subjects with trauma and an incidental SARS-CoV-2 positive test. Written informed consent was obtained from each participant. All participants were tested for SARS-CoV-2 using RT-PCR from nasopharyngeal swab samples. Adult patients aged 18 or over were recruited for the study. Both symptomatic and asymptomatic subjects were included. COVID-19 negative subjects were recruited from the Travel Clinic.

A novel silicon microreactor (Fig. S1) was used to capture carbonyl compounds in breath and then the captured compounds were analyzed by ultra high-performance liquid chromatography-mass spectrometry (UHPLC-MS). Exhaled breath samples were collected in 1L Tedlar bags (Sigma-Aldrich, St. Louis, MO) based on our previous study28,29. The silicon microreactor was fabricated using microelectromechanical systems (MEMS) technology and the device has been characterized for analyzing carbonyl compounds in exhaled breath28. Subjects were instructed to breathe directly into a Tedlar bag through the mouthpiece connected to the bag. A 1L breath sample of a mixture of tidal and alveolar breath was collected. After collection, the mouthpiece was disconnected, disinfected, and then disposed. The Tedlar bag was sealed with the attached valve and placed in a biohazard bag inside a cooler at 4C before transporting to a BioSafety Level 2 Laboratory (BSL-2) for processing and analysis. A nasopharyngeal swab sample for RT-PCR was also collected to test the SARS-CoV-2.

Between March and December 2021, a cohort of subjects with an age range of 1882years were recruited for the study. In Louisville, Kentucky, the Alpha variant of SARS-CoV-2 was dominant reported by the City Health Office during the study period between March and June 2021, so subjects recruited during that period of COVID-19 were attributed to the Alpha wave. The Delta variant was dominant between July and December 202130. Thus, subjects recruited during that period of COVID-19 were attributed to the Delta wave.

All breath samples were transferred to the BSL-2 laboratory in the Division of Infectious Diseases Laboratory at the University of Louisville within 2h of collection for processing. Breath samples were left at ambient temperature for 5min and then evacuated through the silicon microreactors at a flow rate of 7mL/min to achieve above 90% capture efficiencies of carbonyl compounds. The silicon microreactor has thousands of triangular micropillars as shown in Fig. S1 (Supporting Information). The fabrication of silicon microreactors is described in a recent publication28. The surfaces of the channels and micropillars in the microreactors are functionalized with 2-(aminooxy)ethyl-N,N,N-trimethylammonium triflate (ATM) for capture of aldehydes and ketones via oximation reactions. Tedlar bags were connected to the silicon microreactors through deactivated silica tubes. Breath samples were evacuated from the Tedlar bag through the microreactors, then through HEPA filter, and finally through a 75% alcohol in water impinge before entering into air in a BSL-2 hood to avoid contaminations. Detailed characterization of the silicon microreactors and processing of breath samples were reported elsewhere28.

After the breath sample in the Tedlar bag had been completely evacuated through the microreactors, the ATM reacted adducts were eluted from the microreactor using 200 L methanol. ATM-acetone-d6 adduct (5109mol) was added as an internal reference (IR) to the eluted samples. Then, the sample was diluted with water by a factor of 10 for analysis. After processing, all materials including tubes and Tedlar bags were decontaminated according to the laboratory standard procedure for biohazardous waste disposal. The samples were analyzed using a Thermo Scientific UHPLC-MS system equipped with an automatic sampler, a Vanquish UHPLC and a Q Exactive Focus Orbitrap Mass Spectrometer (MS). The UHPLC had an ACQUITY BEH phenyl column (2.1mm100mm, 1.7m, Waters, MA, USA) for the separation of ATM-carbonyl adducts. The liquid flow rate through the column was set to 0.2mL/min. The column temperature was stabilized at 30C. The autosampler tray temperature was set at 8C. 5L of sample volume was injected into the column. The mobile phase A was 0.1% formic acid in water, and mobile phase B was acetonitrile. The mass spectrometer was operated in positive electron spray ionization (ESI) mode with a spray voltage of 3.5kV. Nitrogen was used as sheath, auxiliary, and sweep gas at flow rates of 49, 12, and 2 (arbitrary units), respectively. Full MS mode with the mass range (m/z) from 50 to 500 with a resolution of 70,000 was used to process the breath samples. For MS/MS analyses, a parallel reaction monitoring (PRM) method was used by MS. Chromatographic separation conditions were set via a gradient elution program28. The total chromatographic runtime was 11min. A total of 34 carbonyl compounds were detected for all breath samples and compound concentrations were calculated by comparison of each compound peak area with that of the IR in each breath sample UHPLC-MS chromatogram, including saturated ketones and aldehydes, hydroxy-aldehydes, unsaturated 2-alkenals, and 4-hydroxy-2-alkenals28. A total of 56 features including the 34 carbonyl compound concentrations and 22 derived features of compound ratios and summations including the sum of formaldehyde, acetaldehyde and acetone, the sum of all other carbonyl compounds (OT) and ratios of acetone to butanone were used for statistical analysis (Table S1). Data acquisition and processing were carried out using Thermo Scientific Xcaliber version 4.4. For chemical structure identification of the majority of detected carbonyl compounds, ATM adduct standards were synthesized in-house and used for comparison of retention times and MS/MS spectra28.

There are many classification methods, which include generalized partial least squares, support vector machines, random forests, and logistic regression model to classify the patients into disease and control groups based on breath analysis data31. Prediction (classification) methods involve structured categorical outcome and multiple structured or unstructured covariates32,33. There are no models suited for every condition. Therefore, it is important to identify a good model which takes into account sequential structured covariates for the prediction. Furthermore, the proper identification of key carbonyl compounds through statistical and machine learning techniques requires further advances.

In a typical breath sample analysis, the molecular concentration data on several hundred(s) of endogenous and exogenous VOCs are usually obtained. For the detected VOCs, it may not be required to use all VOCs for the patient classification or the predictive model building process (i.e., training the machine learning models and later use them for class label predictions). Therefore, it is pertinent to select/identify a few metabolic VOCs related to COVID-19 as key features for COVID-19 detection. The selection of key features (here metabolic VOCs) out of many VOCs is called feature selection in machine learning34. Further, it is essential to determine the number of significant VOCs (e.g., feature size or dimension of VOC data), which can be used in the training of the classification model to predict the class type of COVID-19 patients. The selection of significant VOCs saves time for all VOCs present in the breath samples. Thus, the researchers can focus on a few VOCs instead of generating data on all the VOCs present in breath samples of the patients.

The data was first normalized using logarithm (log2) method and then a t-test was used for continuous variables and chi-square test was used for categorical variables35. A p-value less than 0.05 defines statistically significant difference at a 95% confidence interval. All calculations were performed with SAS statistical software36. A logistic regression model was employed for both univariable and multivariable regressions. After the logarithm and quantile methods to normalize the data, it is no longer non-linear. The multivariable logistic prediction model is the most robust one especially when there are less covariates32. The model performance was evaluated by the receiver operator characteristic (ROC) curve with area under the ROC curve (AUC), accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Boxplots were used to visualize the differences between COVID-19 positive and negative groups. A random section of about 67% of samples was used for the training dataset and 33% of samples for testing dataset for all logistic regression models.

Go here to read the rest:

Detection of COVID-19 by quantitative analysis of carbonyl compounds in exhaled breath | Scientific Reports - Nature.com