“COMPARATIVE EVALUATION OF CAPILLARY BLOOD GASES AS WITH ARTERIAL BLOOD GASES IN PATIENTS OF COPD WITH ACUTE EXACERBATION PRESENTING TO EMERGENCY DEPARTMENT: A PROSPECTIVE OBSERVATIONAL STUDY”

1. MBBS., DNB EMERGENCY MEDICINE. 2. MBBS, DNB ANAESTHESIA. 3. MRCS (Edinburgh), MRCS (Glasgow), MCEM (U.K.). ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History Received: 20 August 2020 Final Accepted: 24 September 2020 Published: October 2020 Aim: The aim of this study is to investigate the correlation between selected variables of capillary blood gas (CBG) and Arterial blood gas (ABG)values for assessing values of pH, bicarbonate (HCO3), partial pressure of oxygen (pO2), partial pressure of carbon dioxide (pCO2) in patients with acute exacerbation of COPD presenting to the Emergency department of Peerless hospital. Our study on reliability of CBG over ABG is aimed to make blood gas analysis easier, quicker, less painful and equally reliable. Settings And Design: This was a prospective observational study carried out in a tertiary care centre in Kolkata. Materials And Methods: This study was carried out in Peerless hospital and B.K.Roy Research Centre, Kolkata over a period of 1 year. Total of 90 patients who presented to Emergency department with acute exacerbation of COPD patients were included in the study. Informed consent was taken from all the patients recruited in this study. The blood samples were drawn simultaneously from the radial artery by an arterial puncture into a heparinized syringe and the fingertip by a finger prick into a capillary tube of every patient participating in the study. Initially, capillary sample was collected and immediately after that, an arterial sample was taken from the radial artery in order to assess the agreement between the capillary and arterial samples. These samples obtained were analysed immediately by the blood gas analyser (AVL Compact 3, Roche Diagnostics GmbH, Mannheim, Germany) of the Emergency Department for values of acid-base and oxygenation: pH, PO2, PCO2and HCO3 values. Blood gases were obtained only if the patient needed blood gases analysis for clinical decisions. Care was taken to avoid exposing the blood droplet to air, and the arterial sample was continuously turned to avoid clotting. In addition, the measurement of oxygen saturation (SpO2) was also obtained from the finger pulse oximeter in the emergency (Noninvasive pulse oximeter).


557
Cbg Analysis A recent study done by Sergi Vaquer et al in 2014 , evaluation of earlobe arterialized collector in critically ill patients. In this study, 55 patients capillary blood was collected from earlobe using the ear lobe arterialized blood collector after applying vasodilatory cream and earlobe massage. They found that this method was easier, required minimum training and decreased time to yield results and so they can offer good benefits for health care . They also suggested that few more studies should be done focusing on this new equipment (8).

Correlation Between Abg And CBG:
To assess the correlation and agreement between measurements of pH, PO 2 , PCO 2 , HCO 3-, in arterial and capillary blood gas samples and to assess the reliability of capillary measurement, a literature search was carried out to identify the extent of previous work in this area.
Pitkin AP et al, in 1994, took simultaneous ear lobe samples and arterial samples from 40 patients with chronic lung problems. In this study, he found that ear lobe blood gas analysis was accurate with arterial blood gas values of pH, pCO2, pO2 and can be substituted for ABG , but due to lack of knowledge , many centers in UK did not use it(45).

Ross Murphy et al, in 2005
in U.K. did a study, which concludes that the values of pCo2, pH and HCo3 had good agreement between capillary blood gas and arterial blood gas samples. This showed that CBG can be used to assess the ventilation of the patient but not the oxygenation. It also states that when we combine with continuous pulse oximetry monitoring, CBG can be used as an alternate to ABG in acute exacerbations of COPD patients (25).
A comparative study of capillary and arterial acid-base parameters in patients with acute respiratory distress was performed by Raymond begin et al in 1975. Small differences were found between the samples for pH, pCO2, pO2 and bicarbonate values, the correlation was >0.97. So they concluded that capillary blood gas can be a valid substitute for arterial blood sample for management of patients with acute respiratory distress (50).
A meta-analysis was conducted in 2006 by Gerald S Zavorsy et all, comparing Po2, Pco2 and pH values between capillary blood gas samples (ear lobe and fingertip) and arterial blood gas sample. They found that earlobe capillary samples were accurately good in predicting arterial pO2 values in hypoxia because the arteriovenous difference in pO2 values will be minimum and also the oxyhemoglobin dissociation curve became more linear in lower pO2 values. For pCO2 values, capillary blood samples accurately reflected the arterial sample values, the ear lobe being the preferred site for sampling. For pH values, blood sampling from any of the capillary site gave accurate values to arterial sample (20). Another study done by Hughes JM et all, also concluded that earlobe arterialized pO2 values and arterial pO2 values were more accurate when the arterial pO2 is lower (41).
Torjussen.W et al in 1967, compared pH and pCO2 of 21 patients drawing arterial samples from femoral artery and capillary samples from ear lobe. No significant differences were found in the values between the two, but there was a tendency of increased pH and pCo2 values in the capillary samples (53).
A study was conducted comparing ABG and CBG in an ED by Kamran Heidari et al in 2013, in which 187 patient's samples were taken and values of pO2, pCO2, HCO3, pH and BE were compared. They concluded that hemoglobin oxygen saturation, pCO2, HCO3, pH, BE and also pO2 showed good statistical correlation between ABG and CBG, and so fingertip CBG samples can be used instead of ABG for the assessment of the above mentioned values (10).
A study investigating the possibility of obtaining the arterial pO2 values in a stored capillary blood gas specimen and measured in an oxygen microelectrode was conducted in 1968 by J Macintyre et al. They concluded that the finger pulp was not a suitable site for obtaining capillary samples even if the finger was massaged with a vasodilatory paste, but if ear lobe was massaged well with thurfyl nicotinate, they gave a po2 value that was in close association with arterial pO2 values in normal, hypoxic as well as hyperoxic patients (52).

558
A study done by Harrison et al on paediatric intensive care unit patients showed CBG values reflected ABG values of pH and pCO2 in most patients and can be used to assess the metabolic status of the patients(39).
Few studies showed good correlation between the values of pH, pCO2 and HCO3, but did not show good correlation between pO2 values(1) (42)(33)(37). They concluded that CBG cannot replace ABG for assessing oxygenation.
Heidi K.Kongstad et all in 2011, conducted a study on 62 patients aged from 34-89 years to compare agreement for values from non-arterialized fingertip samples and arterial blood gas samples. In this study they found that , agreement for values of pH, pCO2 and HCO3 was good between the two sample types, but the agreement for pO2 values were poor. A source of error was noted in the study population as that consisted of outpatient COPD patients (1).
A prospective study on 150 adult patients were conducted by A Sauty et al in 1996, comparing pO2 and pCO2 values from arterialized ear lobe samples and radial artery samples. In that he found, earlobes values of pO2 were lower than arterial samples and also more difference in limits of agreement. Whereas, pCO2 had smaller limits of agreement and so capillary blood gas can be used as an alternate to arterial pCO2 (42).

D Yildizdas et al in 2004
, did a study comparing simultaneously obtained arterial, venous and capillary samples. A total of 116 samples were taken from PICU patients over a period of 2 years. In that he concluded, pH, pCO2 and HCO3 had good correlation among the three samples, whereas they did not recommend CBG and VBG for pO2 estimation (33).
A study was conducted to assess the accuracy of capillary sample for acid base values Kasim Docrat et al in 1965.This study was conducted on 50 patients during their routine abdominal surgery after giving anesthesia. Capillary samples from ear lobe and arterial samples from brachial or femoral artery was taken.
They were assessed for values of pH, pCO2 and bicarbonate values, there was difference in pH and pCO2 values between 2 samples, while bicarbonate had narrow limits (54).
A study comparing arterialized ear lobe sample and direct arterial puncture sample values of pO2 and pCO2 during exercise were compared by Fajac et al , and he concluded that arterialized earlobe oxygen tension is not a good substitute for arterial oxygen tension and should not be used to assess oxygenation during exercise (37).
A descriptive study on 82 patients by Rath A et al in 2018 , for comparing and correlating between arterial and capillary blood gas values in mechanically ventilated patients. He found that Bland Altman analysis for pCO2 and bicarbonate were scattered beyond 2 standard deviations. They concluded that CBG can be used to assess pH and paO2, but may be not for assessing pCO2 and bicarbonate (4

Study Duration
The study was conducted over a period of one year (January 2019 -December 2019).

Study Population
For this study, data was collected after obtaining consent, from all patients who fulfilled the inclusion criteria, presenting to the Emergency Department. To calculate the sample size for this study, we has used the raosoft sample size calculator . (http://www.raosoft.com/samplesize.html).

559
The following data were administered: Population size: 150 Confidential level: 95% Margin of error: 5% Response Distribution: 50% The sample size required for this survey was calculated as 110. 15 patients were eliminated due to sampling error (repeated squeezing of finger) and 5 patients were eliminated due to lack of consent from patient and family member. Totally, 90 samples were evaluated for this study.

Exclusion Criteria
Patients were excluded if: 1. age < 18 years, 2. Pregnant 3. Dyspnoea of causes other than COPD. 4. Those patients whose diagnoses were changed during the course of admission.

Study Methodology:-
All patients with COPD exacerbation, attending Peerless hospital Emergency Department, who were thought to require ABG analysis by the treating physician, were identified and enrolled by Emergency Department doctor during the 12-month study period. A written informed consent was obtained from each patient or his or her family member before enrolment into this study and the patient was included only if the consent was obtained for the study.
The blood samples were drawn simultaneously from the radial artery by an arterial puncture into a heparinized syringe and the fingertip by a finger prick into a capillary tube of every patient participating in the study. Initially, capillary sample was collected and immediately after that, an arterial sample was taken from the radial artery in order to assess the agreement between the capillary and arterial samples.
These samples that were obtained was analysed immediately by the blood gas analyser (AVL Compact 3, Roche Diagnostics GmbH, Mannheim, Germany) of our Emergency Department for values of acid-base and oxygenation: pH, PO 2 , PCO 2 , HCO 3 and they were recorded.
Blood gases were only obtained if the patient needed blood gases analysis for clinical decisions and further management. In addition, the measurement of oxygen saturation (SpO 2 ) was also obtained from a finger pulse oximeter (Noninvasive pulse oximeter).

Obtaining Arterial sample:
Under strict aseptic precautions, arterial punctures were carried out and 1ml of arterial blood was collected with a heparinized syringe (BD ABG syringe) and was transferred as soon as possible for assessment by the blood gas analyser.

Obtaining Capillary sample:
After aseptic cleaning, the fingertip was punctured with a no 18 needle and blood gas samples were obtained by "contact" with the capillary tube's tip. The first drop of the blood was discarded. A short manual massage was necessary in some instances. After collection of blood in the capillary tube, two sides of the tube was kept closed by fingers to avoid air bubbles and was sent for analysis by the analyser.

Data analysis:
The data that were obtained were entered in an excel format and was sent for statistical analysis by a statistician. All the documented data were analyzed using Pearson's Chi Square test correlation coefficient test to assess the strength of the relationship between the arterial and capillary gas values.

Statistical Methods
Categorical variables are expressed as Number of patients and percentage of patients and compared across the groups using Pearson's Chi Square test for Independence of Attributes/ Fisher's Exact Test as appropriate.
Continuous variables are expressed as Mean and Standard Deviation and compared using paired t test.

Association between continuous variables captured using Spearman's Rank Correlation Coefficient
The statistical software SPSS version 20 has been used for the analysis. An alpha level of 5% has been taken, i.e. if any p value is less than 0.05 it has been considered as significant.  The above pie diagram shows majority of patients (17.8%) belonged to age group 71-80 years, followed by equal percentage (15.6%) between age groups 31-40years and 51-60 years.  The above pie chart shows that among the total patients, 60% were males and 40% were females.   The above bar diagram and the tables show that the maximum pCO2 of 48 in CBG is seen in the age group above 81 years and the maximum pCO2 of 45.56 in ABG is seen in the age group 71-80 years. The minimum pCO2 values of 40 in both ABG and CBG were seen in the age group 31-40 years. The p values in all age groups are insignificant (>0.05) The above bar diagram shows that maximum normal level of pCO2 of 92.8% in CBG was seen in the age group of 31-40 years and minimum normal value of 20% was seen in the age group of 81-90 years.  The above bar diagram shows maximum normal level of pCO2 of 92.8% in ABG in age group of 51-60 years and minimum normal value of 50% was seen in the age group of 71-80 years.       The above bar diagram shows overall average pCO2 value in ABG was 43.26 and overall average pCO2 value in CBG was 43.76.    In 85% cases, both the methods measure abnormal and normal PCO2 in similar direction. In 14% cases, both the methods measure abnormal and normal PCO2 is opposite direction.

Gender
There is a fair association between ABG and CBG in terms of predicting "normal vs abnormal "PCO2 level.

Results And Observation:-
for HCO3 values in ABG and CBG:-Age     The above bar diagram shows maximum normal level of HCO3 of 92.3% seen in CBG in the age group of 21-30 years and minimum normal level of 37.5% seen in the age group of 71-80 years. The above bar diagram shows maximum normal level of HCO3 of 71.4% in ABG in the age group of 31-40 years and minimum normal level of 25% in the age group of 71-80 years.         Total 90 100.0

Figure 20: Pie chart depicting comparison of agreement and disagreement of HCO3 values in ABG and CBG
In 58.9% cases, both the methods measure abnormal and normal HCO3 in similar direction. In 41.1% cases, both the methods measure abnormal and normal HCO3 in opposite direction. Hence, within normal range or within abnormal range HCO3 comparison is not making any sense.   The above bar diagram and the tables show that the maximum pH of 7.42 in CBG and ABG were seen in the age group 31-40 years. The minimum HCO3 values of 7.32 in CBG is seen in age groups 71-80 years and 81-90 years and 7.32 in ABG was seen in the age group 71-80 years. The p values were insignificant in all age groups.

Figure 22: Bar diagram depicting age wise distribution of % of normal values of pH in CBG
The above bar diagram shows 100% normal level of pH in CBG in the age group of 21-30 years and 31-40 years and minimum normal value of 20% was seen in the age group of 81-90 years.     The graph shows 75% of females had normal pH values in CBG and 77.7% of males had normal pH values in CBG.  The graph shows 75% of females had normal pH values in ABG and 79.6% of males had normal pH values in ABG.
579  The above bar diagram shows overall average pH value in ABG was 7.37 and overall average pH value in CBG was 7.38.    In 96% cases, both the methods measure abnormal and normal pH in similar direction. In 3.3% cases, both the methods measure abnormal and normal pH is opposite direction. There is a fair association between ABG and CBG in terms of predicting "normal vs abnormal "pH value.     The above bar diagram shows maximum normal level of pO2 of 40% seen in CBG in the age group of 81-90 years .  The above bar diagram shows maximum normal level of pO2 of 78.5% in ABG in the age group of 51-60 years.

Results and observation for pO2 values in ABG and CBG:- AGE
 GENDER    The graph shows 25% of females had normal pH values in CBG and 14.8% of males had normal pH values in CBG.       In 43% cases, both the methods measure abnormal and normal pO2 in same direction. In 56.7% case, both the methods measure abnormal and normal pO2 is opposite direction. Hence, within normal range or within abnormal range PO2 comparison is not making any sense because % of disagreement is moderately high.

Discussion:-
Morbidity and mortality from COPD are considerable and increasing. By the year 2020, COPD is predicted to become the third leading cause of death worldwide (exceeded only by heart disease and stroke.(40)(43)(29).
In the present study, we took a total of 90 patients with acute exacerbation of COPD, who came to the emergency department of Peerless Hospital and B.K.Roy Research Centre, Kolkata. This was a prospective observational study.
To our knowledge this study was to evaluate and compare the applicability and reliability of fingertip capillary blood gas analysis technique with arterial blood gas analysis. Capillary blood gas measurement from fingertip has been extensively evaluated in various physiological and pathological situations. Nevertheless, contradictory results as to its accuracy have been reported, especially for PO2 measures.
This can be explained by the physiological changes in the body as the oxygen levels in the arteries differ from that of capillaries, as capillaries are the small blood vessels which connect arteries to veins.
In a recent analysis reported that, although agreement was not high, fingertip capillary estimations could be used for clinical management since they followed arterial values (50), (45). The main sources of variability were attributed to different collection techniques and procedures.
As per the study conducted and the calculations done by pearson's chi square test method, the mean of the ABG and CBG values of pH, PCO2 and HCO3 showed that they can be safely considered as consistent predictors whereas Po2 showed significant difference between CBG and ABG values and also wide limits of agreement, which was similar to few studies done by Heidi.K.Kongstad et al (1) By the similar test performed for the pH, the mean of ABG and CBG values showed p value>0.05 and insignificant difference, and they also showed good agreement for predicting normal and abnormal values. Hence we conclude that CBG can be used as a consistent predictor for ABG in estimating pH values.
On the other hand, for PO2 both the methods ABG and CBG, yield significant difference in PO2 values. Further comparison of normal and abnormal values also showed significantly high differences. Hence, there is high disagreement in values of CBG over ABG and CBG cannot be used as an alternate to ABG for assessing oxygenation status of the patient, which was similar to the few previous study results. (1)(42)(25)(33).
In view of bicarbonate comparison between CBG and ABG , the mean values had p value>0.05 and there was insignificant difference, so CBG and ABG were consistent predictors of HCO3 values, but when normal and abnormal values were correlated there was quite high % of disagreement , the comparison is making no sense, which is similar to a study done by Rath et al (4).
By the similar test performed for pCO2, ABG and CBG were consistent predictors for pCO2 values with insignificant difference, and they also showed good agreement for predicting normal and abnormal values. Hence we conclude that CBG can be used as a consistent predictor for ABG in estimating pCO2 values.
Another important finding of the study was high capillary sampling failure rate. Total sample size was 110, but 15 patients were eliminated due to sample failure. High sampling failure ratios were associated with insufficient blood flow delivery to the collecting system. This was attributed to a reduced capillary blood flow in the fingertip , which was noticed mainly in old age patients and patients in shock.