Pleural Fluid Cholesterol Level in Differentiating Exudate from Trasudate Pleural Effusion

Correlation of pleural luid cholesterol level with light’s criteria to differentiate exudate from transudate pleural effusion. Classi ication of transudate and exudate clinicallywasdone independently basedon the light’s criteria. Pleural luid cholesterol levels of 100 selectedpatientswereobtained. The cholesterol levelswere comparedwith the earlier obtaineddata to study its speci icity and sensitivity in differentiating exudate from transudate effusion. It was found that pleural luid cholesterol in comparison to protein values in differentiating exudate from transudate showed a sensitivity of 79.55%, speci icity of 91.07%, the positive predictive value of 87.50%, the negative predictive value of 85.00%, with a P-value of <0.001. Comparison of pleural luid cholesterol with LDHvalues showeda sensitivity of 86.36%speci icity of 94.64%, the positive predictive value of 92.68%, the negative predictive value of 89.83%, with a P-value of <0.001. Also, a comparison of pleural luid cholesterol to light’s criteria showed a sensitivity of 100%and86.4% in the transudative group and sensitivity of 100% and 91.1 % in the exudative group, respectively. Routine measurement of pleural luid cholesterol may serve as a valuable diagnostic indicator for differentiating exudate from transudate effusion.


INTRODUCTION
Accumulation of the excess luid in the pleural cavity -pleural effusion, is a luid-illed space around the lung, which can limit the expansion of lung, thus impairing breathing (Jameson et al., 2018). It could develop as a consequence of excessive formation from the related structures -parietal pleurae, interstitial spaces of the lung, or the peritoneal cavity; or maybe a result of reduced luid drainage by the associated lymphatics.
Conditions like congestive cardiac failure, nephrotic syndrome, hypoalbuminemia, liver cirrhosis can cause increased capillary permeability of the lung vasculature, probably mediated by the release of cytokines and in lammatory mediators like VEGF, consequently, producing a transudative effusion (Murray and Nadel, 1987). Whereas, conditions like -post-cardiac surgery when retained blood syndrome ensues, empyema of lung secondary to bacterial pneumonia, parapneumonic effusions, carcinoma lung or metastasis to pleurae -can cause an exudative effusion Diagnosis is usually based on the presentation, examination and a chest x-ray.
Fluid usually should accumulate more than 300ml to be clini-cally detectable. An exudate and transudate can be further differentiated using laboratory methods by analysis of protein and lactate dehydrogenase levels of the pleural luid and serum, where the luid can de ine an exudate to serum protein ration >0.5 and luid to serum LDH ration >0.6 (Wang, 1985).
An exudate can also be characterised by the presence of luid cholesterol >45mg/dL, LDH >0.45 upper limit of that of serum, and a protein >2.9g/dL Cholesterol levels are assumed to have a higher sensitivity and speci icity in comparison to the light's criteria in differentiating exudate from a transudate, based on which the present study was undertaken to compare the reliability of luid cholesterol levels as a diagnostic indicator for differentiation. (Cecil et al., 2012) Aim Correlation of pleural luid cholesterol level with light's criteria to differentiate exudate from transudate pleural effusion

Primary objective
To assess cholesterol level in pleural luid.

Secondary objective
To assess the total cholesterol level in the pleural luid to differentiate exudate and transudate.

Study design
This prospective case study was conducted in Department of General Medicine, Vinayaka Mission Medical College, Karaikal after obtaining approval from the Ethics Review Committee of the Vinayaka Mission Medical College, Karaikal. Written informed consent was taken from all participating patients after explaining the study in their language. Patients were recruited from all units in the inpatient department. We screened 108 patients for pleural effusion, and 100 patients who satis ied the inclusion and exclusion criteria were included in the study. A detailed history regarding the nature and duration of the presenting symptoms was obtained from each patient. List of symptoms regarding breathlessness on exertion, fever, cough and expectoration, pleuritic chest pain, abdominal pain, lower limbs swelling, decreased urine output, loss of appetite, facial puf iness, multiple joint pain, weight loss were enquired. They were also evaluated with the available investigation regarding pleural effusion: Pleural luid-Protein, LDH, cholesterol level, Serum protein, Serum LDH.
Patients were classi ied (according to (Lee and Olak, 1994)) as having a) Exudative effusion if one of the following was met 1. Pleural luid protein level > 2.9 g/dL (conversion to g/L, multiply by 10) 2. Fluid lactate dehydrogenase (LDH) level greater than two-thirds of the upper limit of the standard serum value 3. Fluid cholesterol level > 45 mg/dl. b) Effusions were classi ied as transudative if none of the above criteria was met, or if two of the following conditions were met 1. Fluid LDH level less than or equal to 2/3rd of the upper limit of the standard serum value (222 U/L) 2. Fluid cholesterol level less than or equal to 45 mg/dl.
As the simultaneous serum values of the patients were not consistently available, a criterion suggested by Heffner et al., which does not base on the simultaneous serum values, was chosen. (Figure 1)

RESULTS
Out of 100 patients, 83 patients were males, and 17 patients were females in this study, as depicted in Figure 2. In this study group, the age group of 51-60 years constituted the majority of patients (Figure 3). Diabetes is considered the most common comorbidity in our study (Table 1). The above table (Table 2)shows a comparison of Cholesterol group with Protein group. In Cholesterol group when comparing with Protein group transudates were 44 and exudates were 56. The above table (Table 3) depicts the Cholesterol test was comparatively better than          The area is 0.863 with standard error 0.040. The asymptomatic 95% Con idence Interval lower bound was 0.783 and upper bound 0.942 (Table 4, Figure 4). The above table (Table 5) shows a comparison of Cholesterol group with LDH group. In Cholesterol group when comparing with LDH group transudates were 44 and exudates were 56. The above table (Table 6) depicts the cholesterol test was comparatively better than the LDH group, where the P-value is <0.0001, which is very highly signi icant. The area is 0.913 with standard error 0.033. The asymptomatic 95% Con idence Interval lower bound was 0.848 and upper bound 0.977 (Table 7, Figure 5). Table 8 shows the mean serum LDH levels to be 231.02 and mean serum protein as 6.10mg.dl. This igure shows Pleural Fluid Cholesterol Sensitivity is 100% in both transudative and exudative type, whereas 86.4% in transudative and 91.1% in exudative type according to light's criteria. (Figure 6)

DISCUSSION
The present study indicates that Pleural luid Cholesterol may serve as a valuable diagnostic indicator as compared to light's criteria to differentiate exudates from transudates. In this study, the total number of patients with pleural effusion were 100, in which 83 were males, 17 were females. In this study, the age group between 51-60 years were 33%, between 41-50 years were 22%, between 31-40 years were 20%, above 60 years were 19% and between 20-30 years were 6%. In this study, out of 100 patients, 63 were diabetic, and 37 were non-diabetic, 45 patients were hypertensive, and 55 patients were non-hypertensive. Tubercular pleural effusion was of exudative type. Out of 33 patients suffering from tuberculosis, Pleural Fluid Cholesterol levels classi ied everyone to be having an exudative type of pleural effusion, whereas, light's criteria gave confusing results in 6 patients.
As per the study done by (Patel and Choudhury, 2013), the sensitivity and speci icity in classifying a transudate based on the cholesterol levels, with a cut -off point of greater than 60mg/dL, was found to be 98% and 100% respectively, where one of the patients with a tuberculous effusion of the assessed 49 patients was classi ied wrongly as transudate.
Although, there was found to be no misdiagnosis of 11 transudate patients.
In the same way, Chronic kidney disease patients with pleural effusion have transudative type of effusion which were classi ied correctly in all study subjects by pleural luid cholesterol measurement compared to light's criteria which classi ied 5 of 23 study subjects as the exudative type of pleural effusion.
A study by (Mitra, 2012), Symptomatic pleural effusion was present in 6.74% patients of CKD (stages 3 to 5) and 5.88% of post-transplant patients. (Mitra, 2012) in their study showed an incidence of 6.7% for symptomatic pleural effusion among the chronic kidney disease patients -stage 3 to 5, and 5.88% among the post-transplant group. (Kumar et al., 2015), in their study, showed that most of the unilateral, blood-tinged, lymphocytepredominant, exudative effusions among the chronic kidney disease patients were secondary to tuberculosis; with TB being a leading cause of pleural effusions among the CKD patients only secondary to cardiac failure with a reported incidence of 28% and 31% respectively. Of the 35 patients they have assessed, unilateral effusion was found in 57%, of which majority of effusions were moderate (60%) followed by minimal (25%) and massive (15%). The other signi icant causes they have described are uremic effusion (14%), parapneumonic effusions(11%), malignancy (9%) and connective tissue disorders (2%).
Moreover, 8 of 18 patients who were suffering from alcoholic liver disease and had pleural effusion were classi ied as exudative type as per light's criteria whereas, all were classi ied correctly into the group of transudative effusion by Pleural luid Cholesterol measurement.
Pleural effusion in alcoholics can be related to a study done by Jalal Assouad et al., which suggested that Passage of ascites through diaphragmatic defects appears to be the leading cause of PE complicating cirrhosis.
According to Jose (Alonso, 2010),~10% of cirrhotic patients with pleural effusion due to the hepatic cause have protein concentrations in the exudate range as a result of diuresis. (Paramothayan and Barron, 2002), in their study had similar results as that of our study with pleural luid cholesterol diagnosing six cirrhotic cases correctly as transudates whereas 1 of 6 was a false negative according to light's criteria.
And also, malignancy and related pleural effusion were of an exudative type which was classi ied correctly in all seven patients by Pleural luid Cholesterol. Still, light's criteria classi ied one of the patients as transudative effusion. These indicate the ef icacy of Pleural luid Cholesterol over Light's criteria for the classi ication of pleural effusion. (Hamm et al., 1987), showed when exudates differed from transudates with a cut-off level of 60mg/dL, 5% were incorrectly segregated. They also showed that in the malignant effusions, the mean cholesterol level was 94mg/dL, and the elevated cholesterol levels in the exudates were not dependant on the serum cholesterol levels. (Guleria et al., 2003), suggested criteria that exudative can be best recognised with pleural luid cholesterol levels and triglyceride levels more than or equal to 60mg/dL and 40mg/dL respectively, and, a luid to serum ratio of cholesterol and triglyceride greater than or equal to 0.4 and 0.3 respectively. They also demonstrated a sensitivity and speci icity of 88% and 100% respectively for the pleural luid cholesterol, and, 98% and 84% respectively for luid to serum cholesterol ratio, with an accuracy of 92% for luid cholesterol among the exudates group which were on a comparison basis superior to the criteria proposed by the light et al. (Light et al., 1972) suggested criteria with luid serum protein ratio greater than 0.5, luid LDH more than or equal to 2/3 rd upper limit of that of serum and a ratio of luid to serum LDH more than 0.6, suggesting a sensitivity and speci icity of 99% and 98% respectively in classifying an exudate from a transudate, but, which when evaluated by other researches failed to show the same result; with reproduced values from other studies showing a speci icity ranging in between 70-86%. Also, it was highlighted that of the patients with transudate variant, 25% were classi ied as having exudate by light's criteria.
Mis-diagnosis of exudate from a transudate can be a leading factor for increased morbidity, misuse of, and increased cost of health care resources; mainly by the alteration of treatment leading to inappropriate management. This suggests a necessity of tests with higher diagnostic accuracy.
In our study, we found a signi icantly low misdiagnosis rate with the use of luid cholesterol levels for classi ication, in comparison to light's criteria.
All these on a conclusive basis, with synergistic support, form our personal experience about this study showing a higher sensitivity and speci icity of 100% draw a suggestion that the pleural luid cholesterol might be a new landmark in the pleural exudate diagnosis, and it could be a promising test for avoiding misdiagnosis.

Limitations of the study
For one thing, the detailed mechanism of cholesterol in the pathogenic process has not been systematically investigated; second, the standard cut-off value of pleural cholesterol has not been founded, there were several cut-off values ranged from 38 to 65 mg/dl. Further work should aim to identify the cutoff value of pleural cholesterol that provides optimal diagnostic accuracy.

CONCLUSIONS
In conclusion, pleural luid cholesterol showed higher accuracy, speci icity and sensitivity, statistically signi icant P value of <0.0001 when compared to the light's criteria. It also facilitates in the reduction of healthcare-associated costs by cutting down investigations like plasma protein and LDH, serum protein and LDH. Hence, in short, assessment of pleural luid cholesterol is way more ef icient, costeffective and an easy method for classifying exudate from transudate and must be applied in routine clinical practice. As the study involved a small sample size, we suggest that further studies on large samples are warranted.