Background: Sepsis is a critical condition resulting from a dysregulated immune response to infection, leading to severe organ dysfunction and increased healthcare costs. Early identification of sepsis severity is crucial for effective management. This study investigates the correlation between hypoalbuminemia, arterial blood gas (ABG) pH levels, and the severity of sepsis, as assessed by the Sequential Organ Failure Assessment (SOFA) score, in critically ill patients. Methods: A retrospective analysis was conducted on 100 adult patients admitted to the ICU of a tertiary care hospital in South India from June 2023 to December 2023. Data collected included demographic characteristics, comorbidities, SOFA scores, ABG results, and microbiological cultures. Statistical analyses were performed using descriptive statistics, Chi-square tests, ANOVA, and Pearson’s correlation coefficient, with a significance threshold set at p < 0.05. Results: The majority of patients exhibited acidemia, with a significant inverse correlation between pH levels and SOFA scores (p < 0.001), indicating that lower pH is associated with greater sepsis severity. However, no significant relationship was observed between hypoalbuminemia and SOFA scores (p = 0.566). The most frequently identified pathogens included Klebsiella and E. coli. Conclusion: The study demonstrates that ABG pH levels serve as a valuable early indicator of sepsis severity in critically ill patients. While hypoalbuminemia was prevalent, its correlation with sepsis severity was not statistically significant. These findings underscore the importance of prompt ABG analysis in managing sepsis and suggest the need for further research to elucidate the role of hypoalbuminemia in sepsis outcomes.
Sepsis is a life-threatening condition resulting from altered host immune response to ongoing infection which can cause severe organ dysfunction1. Uncontrolled sepsis is one of the main reasons for rapid deterioration in critically ill patients. Sepsis is also a leading reason for the increase in overall hospital stay and medical health care costs. Early identification of the severity of sepsis on presentation can modify the outcome in patients by adequate fluid management and help in choosing appropriate antibiotics. Various parameters like Hypoalbuminemia and early PH change can predict the severity of sepsis.[1]
An acid-base condition that can be difficult to manage is metabolic acidosis, particularly in the sick. It may arise from either a decrease in bicarbonate (HCO3-) ions or increased hydrogen (H+) ions concentration. Metabolic acidosis is linked to a higher death rate in severely ill individuals, regardless of the cause.[2]
The European Society for Clinical Nutrition and Metabolism (ESPEN) guidelines state that serum protein indicators do not accurately indicate nutritional status in the intensive care unit (ICU) context, but rather reflect an acute phase response. There is a reduction in protein synthesis in acutely ill and dilution by the systemic inflammatory response, these nutritional indicators can be low during the acute stage of sepsis. Serum albumin is primarily involved in the physiological regulation of capillary membrane permeability, plasma oncotic pressure, ligand binding, and transport. In addition to participating in free radical scavenging and serving as a store and carrier for numerous endogenous and foreign substances, albumin also possesses antioxidant and circulatory maintenance qualities. Tumor necrosis factor alpha can cause phosphorylation of C/EBPβ, which is one of the transcription factors that curtail the active transcription of the gene responsible for albumin synthesis.[3]
On taking this into account we took blood albumin level and PH at the time of admission and correlated with SOFA scoring to assess the severity of sepsis.
Aim
The study aims to correlate hypoalbuminemia, PH, and severity of sepsis. We categorized the sepsis severity using Sequential Organ Failure Assessment (SOFA) scoring, which was correlated with blood hypoalbuminemia and PH taken from the ABG measurement at the initial presentation.
A retrospective analysis of adult patients above 18 years of age admitted to ICU in a tertiary care hospital and teaching institute between June 2023 and December 2023 was taken and analyzed.
We collected the following data during the first 24 hours of admission at the ICU: demographic characteristics, comorbid medical conditions, factors predisposing to infection, Sequential assessment of Organ dysfunction (SOFA) assessment, severe sepsis, septic shock, biochemical and hematologic alterations. We also included the data of microbiological isolate from blood, sputum, endotracheal aspirate, and urine.
Data Analysis
Analysis was done by descriptive statistics.
Analysis was done using the Chi-square test, ANOVA, and students' unpaired t-test.
Pearson's correlation coefficient analyzed the correlation.
The statistical package SPSS version .25.0 was used to do the analysis. P value less than 0.05 is considered significant.
Data of patients above the age of 18 who were admitted to the ICU with sepsis were retrospectively analyzed from June 2023 to December 2023 over 6 months.
Patients who were below the age of 18 were excluded from the study.
The sample size for this study was determined using the formula for estimating a population proportion with a specified margin of error. The formula used is:
n = \frac{4pq}{d^2}
Were,
n is the sample size
p is the estimated proportion of the population with the event of interest
q = 1-p
and
d is desired the margin of error
Based on existing literature, analyzing Sepsis in patients within Intensive Care Units (Unit) – (add ref), the value of p was chosen as 0.5 (i.e., q=1-p=1-0.5=0.5) and d as 0.1.
Hence it was decided that a total of 100 participants was required for the successful conduct of this study.[4]
100 patient's medical records were analyzed for the study between June 2023 and December 2023. Out of these patients, males were the majority (n=68) and females were less in number (n=32).
Gender |
Count |
Male |
68 |
Female |
32 |
Table 1: Gender Distribution |
The mean age of the patients recruited for the study was 61.8 with a standard deviation of 14.6. The minimum age of the patient recruited for the study was 28 years and the maximum age was 92 years old. The age of the group studied showed a range of 64 and a median of 61.8.
|
Age |
Number |
100 |
Median |
64.8 |
Mean |
61.8 |
Standard Deviation |
14.63 |
Range |
64 |
Minimum Age |
28 |
Maximum Age |
92 |
Table 2: Age Distribution |
Age group (Years) |
Frequency |
Percentage |
28-45 |
17 |
17 |
46-64 |
33 |
33 |
>=65 |
50 |
50 |
Table 3: Age-wise Distribution |
Age-wise Distribution of Patients Recruited
The majority of the patients studied belong to ages more than 65 years (n=50), while the age group 28-45 years had the least number of people (n=17). This might be due to the increased propensity of elderly patients to develop sepsis.
|
Creatinine |
Mean Arterial Blood Pressure |
Total Bilirubin |
Glasgow Coma Scale |
Platelet |
SOFA |
Number |
100 |
100 |
100 |
100 |
100 |
100 |
Mean |
2.337 |
90.040 |
1.684 |
11.530 |
248497.000 |
5.210 |
Median |
1.725 |
92.000 |
.950 |
12.000 |
239000.000 |
5.000 |
S.D |
1.817 |
22.330 |
2.348 |
2.096 |
126337.123 |
2.5556 |
Range |
10.410 |
113.000 |
17.000 |
9.000 |
540000.000 |
11.0 |
Minimum |
.510 |
37.000 |
.200 |
6.000 |
10000.000 |
1.0 |
Maximum |
10.920 |
150.000 |
17.200 |
15.000 |
550000.000 |
12.0 |
Table 4: Statistical Analysis of SOFA Scoring |
The mean value for the sofa score is 5.2 with an SD of 2.5. The maximum SOFA score is 12 and the minimum is 1 in the patients recruited for the study. The range of SOFA scoring is 11.
SOFA Score |
Age Group |
Total |
||||
28- 45 |
45 - 64 |
.>=65 |
||||
|
0 -1 |
Count |
1 |
4 |
1 |
6 |
% |
5.9% |
12.1% |
2.0% |
6.0% |
||
2- 7 |
Count |
10 |
23 |
42 |
75 |
|
% |
58.8% |
69.7% |
84.0% |
75.0% |
||
8 - 11 |
Count |
5 |
6 |
7 |
18 |
|
% |
29.4% |
18.2% |
14.0% |
18.0% |
||
>11 |
Count |
1 |
0 |
0 |
1 |
|
% |
5.9% |
0.0% |
0.0% |
1.0% |
||
Total |
Count |
17 |
33 |
50 |
100 |
|
% |
100.0% |
100.0% |
100.0% |
100.0% |
||
a. X2=11.209 p=0.082 ns |
||||||
Table 5: SOFA Score vs Age Group |
The majority of the age group falls into the category of SOFA score between (2-7) around 75% of total people studied. The statistical analysis of the study is not significant as the p-value corresponds to 0.082.
Arterial Blood Gas Analysis
The patients were divided into acidotic (ph <7.35, normal (7.36-7.45), and alkalotic groups (ph >7.45).
|
Frequency |
Percent |
|
|
Acidotic |
60 |
60.0 |
Normal |
25 |
25.0 |
|
Alkalotic |
15 |
15.0 |
|
Total |
100 |
100.0 |
|
Table 6: Arterial Blood Gas Analysis |
The table shows majority of the critically ill patients had acidosis (n=60) while 25 patients had normal pH and 15 patients had alkalotic pH.
N |
Mean |
Std. Deviation |
Minimum |
Maximum |
||
Acidosis |
60 |
5.750 |
2.6078 |
1.0 |
12.0 |
|
Normal |
25 |
4.440 |
2.3819 |
1.0 |
10.0 |
|
Alkalosis |
15 |
4.333 |
2.1602 |
1.0 |
9.0 |
|
R |
-0.380 |
|||||
P |
0.000 |
|||||
N |
100 |
|||||
Table 7: SOFA Score vs pH |
||||||
The mean sofa score of patients with acidosis was 5.75 where it is slightly lesser for patients with normal pH (mean=4.44) and alkalotic pH (mean=4.3).
The analysis was done with ANOVA and the r showed -0.380 which indicates that the SOFA score increases as the PH decreases.
|
Ph<7.35 |
7.35-7.45 |
Ph>7.45 |
Total |
||
SOFA Score |
0 -1 |
Count |
2 |
2 |
2 |
6 |
% |
3.3% |
8.0% |
13.3% |
6.0% |
||
2 - 7 |
Count |
43 |
20 |
12 |
75 |
|
% |
71.7% |
80.0% |
80.0% |
75.0% |
||
8 - 11 |
Count |
14 |
3 |
1 |
18 |
|
% |
23.3% |
12.0% |
6.7% |
18.0% |
||
>11 |
Count |
1 |
0 |
0 |
1 |
|
% |
1.7% |
0.0% |
0.0% |
1.0% |
||
Total |
Count |
60 |
25 |
15 |
100 |
|
% |
100.0% |
100.0% |
100.0% |
100.0% |
||
a. x2=5.63 P=0.466 NS |
||||||
Table 8: SOFA Score and ABG |
The table shows that the group with higher sofa scores had more acidotic patients, whereas patients with a lower sofa score had pH normal or alkalotic.
The analysis was done with ANOVA and the r showed -0.380 which indicates as the SOFA score increases the PH decreases.
The study's p-value is near zero, showing a significant correlation. The study shows an inverse correlation of PH in sepsis patients studied.
The graph shows a negative correlation of pH with SOFA scoring as an indicator of sepsis. The graph implies that as SOFA score increases the pH of the patient decreases.
The analysis shows patients with hypoalbuminemia have a slightly greater SOFA score (mean=5.369) compared to patients with normal albumin levels (mean=4.9)
Albumin group |
N |
Mean |
Std. Deviation |
t |
|
Hypoalbuminemia |
65 |
5.369 |
2.7075 |
.8480 |
|
Normal albumin levels |
35 |
4.914 |
2.2540 |
p=0.399 ns |
|
|
|
Hypoalbuminemia |
Normal |
Total |
|
0-1 |
Count |
6 |
0 |
6 |
|
% |
9.2% |
0.0% |
6.0% |
||
2 - 7 |
Count |
45 |
30 |
75 |
|
% |
69.2% |
85.7% |
75.0% |
||
8 - 11 |
Count |
13 |
5 |
18 |
|
% |
20.0% |
14.3% |
18.0% |
||
>11 |
Count |
1 |
0 |
1 |
|
% |
1.5% |
0.0% |
1.0% |
||
Total |
Count |
65 |
35 |
100 |
|
% |
100.0% |
100.0% |
100.0% |
||
a. x2=5.006 P=0.171 NS |
|||||
SOFA Score |
Albumin |
||||
R |
-0.58 |
||||
P |
0.566 |
||||
N |
100 |
||||
Table 9: SOFA Score vs Albumin Group |
|||||
Hypoalbuminemia was present in 65 people and normal albumin level in 35 people diagnosed with Sepsis. While applying a one-sample chi-square test we found this percentage difference is significant (p<0.001)
The patient group studied showed a p-value of 0.566 which implies a non-significant relation between hypoalbuminemia and SOFA score.
The graph shows a mild inverse relationship between hypoalbuminemia with severity, however, there is no statistical significance found in our study.
The cultured organism in blood, sputum, urine, and endotracheal aspirate has been recorded. The majority of patients grew Klebsiella (n=6) and E coli (n=5) followed by Acinetobacter (n=3), candida (n=3), coagulase-negative staphylococcus aureus was grown in 2 patients, acid-fast bacilli in 1 patient.
In our study we found majority of patients had Hypertension (n=18), diabetes mellitus (n=13), chronic liver disease (n=3), ischemic heart disease (n=4), Chronic obstructive lung disease (n=5), acute kidney injury, cardio vascular accident (n=2), asthma (n=1), prostatomegaly (n=1), and anxiety (n=1).
The ABG is an important test that measures blood pH, which falls between 7.35- 7.45 normally. It also reports the partial pressure of oxygen (PaO2) and carbon dioxide (normal values PaO2: 75–100 mmHg, PaCO2: 35–45 mm Hg), bicarbonate (HCO3), where normal is between 22–26 mEq/L, lactate (normal value: 0.5–2 mEq/L), base excess (normal values −2 to +2 mEq/L).
Acid-base conditions like metabolic acidosis are more common among critically ill patients. It may be brought on by a rise in the concentration of H+ or a fall in the concentration of bicarbonate (HCO3-) ions. Sepsis, cardiogenic shock, severe hypoxemia, hepatic failure, and inebriation are among the several causes of metabolic acidosis. Anion gap (AG) acidosis can be categorized as excessive or normal in metabolic acidosis.[5] One typical cause of high AG metabolic acidosis is lactic acidosis. It is further divided into Type-A and Type-B (aerobic and anaerobic). Despite having complete mitochondrial activity, type-A lactic acidosis is linked to the decreased perfusion of the tissues and insufficient oxygen supply. Acidosis owing to aberrant carbohydrate metabolism, such as in failure of the liver to optimally function, occurs in lactic acidosis (type B) when there is no tissue hypoxia and a sufficient supply of oxygen to the cells.[6] Lactic acidosis develops in response to the degree of elevated lactate, the body's ability to function as a buffer, and the presence of other diseases that cause tachypnoea and alkalosis (such as sepsis or liver illness). Consequently, acidemia, alkalemia, or normal pH may be linked to hyperlactatemia.[7]
Numerous frequent aetiologies, such as lactic acidosis, hyperchloremic acidosis, renal failure, and ketoacidosis, can cause metabolic acidosis. Although the underlying illness process linked to each of these subtypes has discrete clinical consequences, acidosis may either accelerate or ameliorate the detrimental effects of these disorders. Patients with sepsis and other severe illnesses frequently experience metabolic acidosis, which is linked to a poor prognosis. Increases in arterial pCO2 can cause respiratory acidosis, whereas a variety of inorganic or organic fixed acids can cause metabolic acidosis.
One of the main reasons for the decrease in blood pH is the onset of metabolic acidosis due to decreased organ perfusion and increased lactic acid.
The release of inflammatory mediators such as TNF alfa, I.L-6, and NOS increases inflammation and decreases blood pressure.
Our study has found a correlation in the severity of sepsis with acidosis in blood, which is almost similar to the results obtained by Samantha et.al published in the Indian Journal of Critical Care Medicine in October 2018.[2] In our study, the major cause of a decrease in blood pH is due to metabolic acidosis as evidenced by a decrease in bicarbonate mean value (18meq/L) whereas the mean partial pressure of CO2 (Pco2) remains almost normal.
Noritomi et al did a longitudinal study in 2009 that showed an increase in mortality in critically ill patients who did not correct the acidosis during the study compared to the patients who corrected their acidosis and had a mortality benefit.[8] Our study found a decrease in pH (acidosis) as the severity of sepsis increases, with SOFA score being the objective parameter.
Prompt identification of severe acidosis can help in prompt resuscitation, early identification, and correction of anion gap if present, need for mechanical ventilation, and proper choice of antibiotics.
Usually, the severity of sepsis is identified by various scoring systems like SOFA (Sequential Assessment of Organ Failure) and APACHE 2(Acute Physiology and Chronic Health Evaluation) which consist of blood parameters like creatinine and total bilirubin, platelet which have significantly higher laboratory turnaround time whereas Arterial blood gas analysis will be available in early hours itself. So early ABG analysis helps to assess the severity of sepsis and early aggressive intervention if required.[9]
The second parameter we included was albumin levels in blood. The majority of the patients had hypoalbuminemia and were diagnosed and treated as sepsis. However, the statistical correlation was not significant for hypoalbuminemia and severity of sepsis (assessed by SOFA scoring) in our study.
The distribution of albumin between the extravascular and intravascular compartments is affected by critical disease. The rates at which the protein is synthesized and degraded also alter. Early on in a serious illness, the serum albumin concentration will frequently decline sharply. It will not rise once again till the sickness has recovered. The kinetics of intravenous albumin administration will vary significantly between critically sick patients and healthy individuals.
Critically ill patients may experience a significant decrease in the rate of albumin production. Increases in the transcription rates of genes for positive acute-phase proteins, such as C-reactive protein, and decreases in the production of blood albumin due to the ineffective transcription of messenger ribonucleic acid are usually seen in the immediate response to inflammation or sepsis. TNF-a and IL-6 both work to decrease the transcription of genes, reduced hepatic synthesis of albumin also reduces during the time of inflammation. Albumin has several anti-inflammatory and antioxidant effects, so a decrease in albumin levels can be detrimental for patients with sepsis. Albumin helps to bind the antibiotics and increase their half-life in the blood. Serum albumin primarily modulates plasma oncotic pressure and capillary membrane permeability, as well as ligand binding and transport. Albumin stores and transports various chemicals scavenges free radicals, and offers antioxidant and circulatory protection. Serum albumin is commonly utilized to treat many illnesses such as hypovolemia, shock, burns, surgical blood loss, and trauma.[10] Inflammation suppresses serum albumin levels and inhibits its synthesis. Total protein (TP) includes albumin and γ-globulin, which aid in humoral immunity.
The study conducted by Isabel et al in 2016 showed a significant correlation between hypoalbuminemia and sepsis. Our study finds a correlation between hypoalbuminemia and sepsis, but the statistical significance of hypoalbuminemia with the severity of sepsis using SOFA score was not significant. The expected correlation could be attained by a larger sample size.
The majority of the patients in our study diagnosed with sepsis fall into the elderly category (age >=65 years) attributed to their reduced immunity and comorbidities.
The major co-morbidity in our study was hypertension followed by diabetes which can also potentially affect pH in the arterial blood gas analysis.[11]
Hypoalbuminemia and arterial Ph can predict the severity of sepsis. Albumin being a positive phase reactant reduces with severity of sepsis. Albumin is an anti-inflammatory as well as an antioxidant. It also binds with antibiotics administered and increases their half-life in blood in sepsis patients. Albumin helps to maintain the hydrostatic pressure in the blood and can help in maintaining blood pressure in sepsis and septic shock patients. Evaluating and treating hypoalbuminemia is of paramount importance in patients with sepsis. Arterial blood gas analysis of patients with sepsis helps in measuring the intensity of sepsis, underlying electrolyte imbalance, know the cause of acidosis if present.
In our study, we found a negative correlation between arterial blood pH and with severity of sepsis. Several causes can be attributed to acidosis in blood, hypoperfusion, and lactic acidosis being major factors. So, prompt evaluation of arterial blood gas can help in fluid management, correction of electrolytes, identifying the need for mechanical ventilation, and optimal choice of antibiotics depending on severity.
Limitations
Our study did not take into account the general nutritional status of the patients. Several comorbidities can act as confounding factors. Since it is a retrospective study follow-up treatment could not be assessed.