Introduction

Chronic kidney disease (CKD) is among the challenging health-related issues with growing concern. In 2019, more than 1.4 million deaths were attributable to CKD globally [1]. Pericardial syndrome, including pericarditis, pericardial effusion (PE), cardiac tamponade and the less common chronic constrictive pericarditis, can further complicate CKD and end-stage renal disease (ESRD) cases [2]. Uremic pericarditis occurs before or within the first eight weeks of the dialysis initiation and without timely management, is associated with poor outcomes in patients with CKD [3]. Dialysis-associated pericarditis is a term applied when it manifests more than 8 weeks following the commencement of dialysis. Nevertheless, some evidence suggests that there may not be a clear-cut demarcation between these conditions [4].

The typical pericardial sac holds between 10 to 50 ml of pericardial fluid, functioning as a plasma ultrafiltrate that provides lubrication between the layers of the pericardium. If there is an accumulation of transudative or exudative fluid exceeding 50 mL, it is considered abnormal and referred to as PE [2, 5].

Acute pericarditis and PE might occur simultaneously or alone. Acute pericarditis is characterized by inflammatory condition and is diagnosed based on meeting two or more criteria, including chest pain, pericardial friction rub, electrocardiographic changes, and the presence of PE. However, it tends to be less symptomatic in ESRD patients [2, 6]. Conversely, PE is not always associated with inflammation and, in a significant number of ESRD patients, could be secondary to volume overload [2]. Notably, the majority of cases of PE remain asymptomatic until they reach severe conditions or tamponade develops [7].

The reported prevalence of pericardial disease in patients with CKD varies significantly in literature, possibly due to disparities in reporting time frames, diagnostic and treatment methods, facilities and specialties available at different hospitals. In addition, some studies have lacked clarity in defining uremic and dialysis-associated pericarditis and PE [4]. The prevalence of PE among individuals with CKD, without specifying a time frame relative to dialysis, have been reported to range from 1.9 to 62% [4, 7–14]. Among patients who have undergone pericardial fluid drainage, the etiology of PE was attributed to uremia in 3.78% to 67% of cases; notably, a majority of these studies encompassed patients under maintenance dialysis [15–28].

Due to the paucity of data on diagnosis and severity predictors of pericardial disease among the CKD population in recent years [3], this study aims to address clinical and laboratory features associated with the severity of PE in patients with CKD in two referral hospitals in Tehran, Iran.

Methods

Participants

The study involved a thorough review of clinical records and laboratory data for 1,466 adult patients admitted with International Classification of Diseases 10th Revision (ICD-10) codes indicating CKD, along with 398 patients bearing ICD-10 codes related to PE, irrespective of its underlying cause. Specifically, patients in CKD stages 4 and 5ND with coexisting PE were recruited for the study. The exclusion criteria were as follows: concomitant infectious disease, systemic autoimmune and inflammatory diseases, neoplasms and paraneoplastic syndromes, myocardial infarction within a month, cardiac surgery within a month, aortic dissection and myocarditis, uncontrolled hypothyroidism, drugs such as minoxidil and hydralazine that can result in pericardial involvement. Patient selection is shown in Fig 1, and the reasons of being excluded are demonstrated in S1 and S2 Tables.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Patient selection flowchart based on the ICD-10 codes of admitted patients.

https://doi.org/10.1371/journal.pone.0302200.g001

Results

In this study, out of 1864 patients admitted under ICD-10 codes relevant to CKD/PE, 112 individuals with CKD stages 4 and 5ND, who exhibited PE closely linked to their renal condition, were carefully chosen for inclusion. The average age of these patients was 60.8±16.1 years, with an age range of 25 to 93 years, and 42% were male. Notably, 64.2% of the patients had diabetes. Based on GFR, 41.1% and 58.5% of patients were at stage 4 and 5ND, respectively. Echocardiographic findings indicated that among the patients, 72 (64.3%) presented with mild PE, while 40 (35.7%) exhibited moderate to severe PE. The comprehensive clinical and laboratory characteristics of the study participants are detailed in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. The demographics, clinical characteristics, and paraclinical features of patients among all patients and comparison between groups with mild and moderate-severe pericardial effusion.

https://doi.org/10.1371/journal.pone.0302200.t001

Clinical and laboratory variables on admission and the date of echocardiography

The gender ratio was not statistically significant between groups. Clinically, chest pain and dyspnea were notably more prevalent in patients with moderate to severe PE at admission (p-value: 0.03 and 0.013 respectively), as detailed in Table 1. Laboratory findings revealed a higher NLR (4.6 vs. 3.5, p-value: 0.043) and lower serum albumin (3.1 vs. 3.6 g/dl, p-value: 0.012) in patients with moderate to severe PE compared to those with mild PE.

The average interval between admission and echocardiography was 3 days, ranging from 0 to 24 days. Analysis at the time of initial echocardiography showed significantly higher white blood cell count (WBC) (8.2 vs. 6.8 ×1000/μl, p-value: 0.008), neutrophil percentage (75.7% vs. 68.7%, p-value: 0.020), and count (6000 vs. 4408, p-value: 0.012), and NLR (4.6 vs 3, p-value: 0.012), along with a notably lower lymphocyte percentage (16.2% vs 23%, p-value: 0.004) and serum albumin (3 vs 3.5 g/dl, p-value: 0.023) in patients with moderate and severe PE, as shown in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Laboratory variables at the date of echocardiography.

https://doi.org/10.1371/journal.pone.0302200.t002

Dialysis initiation and pericardial effusion drainage

Among the 112 patients studied, 53 underwent hemodialysis during their hospital stay. There was a significant association between the severity of PE and the initiation of dialysis for any clinical reason (odds ratio (OR): 2.22, 95% Confidence Interval (CI): 1.011–4.9, p-value: 0.045) as shown in Table 3. Pericardial fluid drainage was performed in 21 patients, with 16 experiencing severe and 5 experiencing moderate PE. The composition of the drained pericardial fluid varied, being bloody in 47.1%, serous in 35.3%, and serosanguineous in 17.6% of cases, detailed in Table 4. The need for pericardial fluid drainage was significantly associated with the severity of PE (p-value < 0.001), as indicated in Table 5.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Patients that underwent dialysis during the time of admission due to any indication (before or after the diagnosis of pericardial effusion).

https://doi.org/10.1371/journal.pone.0302200.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Pericardial fluid characteristics of the patients who have undergone pericardial drainage (n = 21).

https://doi.org/10.1371/journal.pone.0302200.t004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. The association between pericardial effusion severity and pericardial fluid drainage.

https://doi.org/10.1371/journal.pone.0302200.t005

Predictors of pericardial effusion severity in CKD patients

Multivariable binary logistic regression analysis, based on admission data, indicated patients with hypoalbuminemia (serum albumin < 3.5 g/dl) (OR: 4.03, p-value 0.014) and NLR level greater than 5.5 (OR: 4.22, p-value 0.015) are more likely to have moderate and severe PE (Table 6). Furthermore, the analysis at the time of echocardiography revealed significant associations between serum albumin and neutrophil count with PE severity; patients with hypoalbuminemia (OR: 5.38, p-value 0.004) and neutrophilia (OR: 7.94, p-value: 0.005) were more likely to experience moderate to severe PE (Table 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Comparison of laboratory indexes at admission and echocardiography time between pericardial effusion groups: Multivariable logistic regression analysis.

https://doi.org/10.1371/journal.pone.0302200.t006

Discussion

Our study highlights dyspnea, chest pain, serum albumin, and an NLR level exceeding 5.5 upon admission, along with hypoalbuminemia and neutrophilia at the time of echocardiography, as significant predictors of moderate to severe PE in CKD patients.

Pericardial involvement in late-stage CKD, often manifesting as pericarditis or PE, is a notable complication. While initially uremic pericarditis was thought to occur without substantial effusion, later research indicates a high frequency of asymptomatic PE in these patients, with 70–100% of those with uremic and dialysis-associated pericarditis experiencing PE [4, 30].

Many CKD-related PE cases initially show no symptoms, even with significant effusion, posing a challenge for early diagnosis [31, 32]. Our findings corroborate this, showing that a majority of PE patients did not report chest or epigastric discomfort. Specifically, 81% of patients with PE and 60.8% with moderate to severe PE were asymptomatic in these regards. Despite not being a definitive symptom, dyspnea was noted in 54% of PE patients, increasing to 71.4% in moderate to severe cases. Furthermore, 70.7% showed peripheral edema, indicative of volume overload [33].

This finding is consistent with the observations made by Frommer et al., who reported clinical and radiological signs of volume overload in a group of 50 non-dialysis ESRD patients with PE. In their study, 36% of the patients had asymptomatic PE, and 6% displayed clinical features of pericarditis without PE [6]. Building upon our earlier observation that PE, even large ones, often remains asymptomatic in CKD patients, it’s crucial to consider the implications of undiagnosed PE, particularly in those requiring dialysis. In such cases, if PE goes unrecognized, the typical high ultrafiltration rate in dialysis can result in decreased central venous pressure and venous return. This scenario can lead to hypotension, a serious complication [28, 34]. Therefore, heightened clinical vigilance for CKD-related PE and knowledge of its predictors are essential in preventing these severe outcomes [7].

In evaluating our findings in relation to existing literature, it is important to consider the variances in methodologies and study populations. Prior research in this field displays a wide range of participant demographics, including individuals on dialysis, those not on dialysis, and mixed groups. Furthermore, the comparative analyses in these studies vary greatly, from comparing patients with and without PE to examining PE groups of different severity levels. The variation may also arises from differences in reporting timeframes, definitions of PE and its severity grades, and disparities in diagnostic and treatment methodologies. This diversity necessitates a nuanced approach when comparing results across different studies.

In Leehay’s study of 47 ESRD patients (11 patients had not undergone dialysis), significant differences were noted in dyspnea, tachypnea, low voltage ECG, pleural effusion presence, and left shift in WBCs between patients with PE exceeding 250 cc and those with less than 250 cc [38]. Frommer et al., found a higher incidence of PE in diabetics (58.3%) compared to non-diabetics (29%), with lower serum albumin in diabetics. However, the small sample size was a limitation [9]. In the case-control study by Ravi et al., involving 84 stage 4 and 5 CKD patients with PE, 44% were receiving dialysis. Notably, 47% of these patients exhibited moderate to large PE. The study identified higher heart rate (OR: 1.290 per 10bpm), elevated potassium level (OR: 1.949 per 1‐mEq/L), and lower corrected calcium (OR: 1.33 per 1‐mg/dL) as independent predictors for the presence of any PE. A specific finding was that a corrected calcium level below 8 mg/dL showed a high specificity for moderate and large PE compared to those with no effusion. Furthermore, in patients not on dialysis, corrected calcium was the sole significant predictor for PE. Notably their study, did not find an association between PE and a worse short-term prognosis [31].

In a study involving 2820 outpatient ESRD patients, 79.5% were undergoing dialysis at the time of echocardiography. Among these, 54 exhibited moderate to large PE which 75.9% had moderate PE, 24.1% had large PE, and 13% experienced tamponade and 70.4% of these 54 patients received intermittent dialysis. The study evaluated clinical and laboratory factors, including serum albumin and calcium, comparing PE patients with those without PE in a 1:2 ratio. The findings suggested that the duration of hemodialysis was a key protective factor against developing moderate to large PE. Patients in this study were followed for an average of 39 months, revealing a 10-year survival rate of 87%, indicating a favorable prognosis [35].

In the study by Yoshida et al., echocardiography was performed on 150 patients with chronic uremia, both before and during hemodialysis treatment. The study revealed significant difference in systolic blood pressure, dilation of the left atrial chamber, anemia, and hypoproteinemia between patients with and without PE. However, no notable differences were observed in creatinine, uric acid, calcium, or body weight changes across patients with varying degrees of PE [10].

In CKD patients at earlier stages, PE tends to arise from factors unrelated to renal disease. Interestingly, in late-stage CKD patients, we did not observe any significant association between serum creatinine, GFR, and BUN with the severity of PE. Previous studies have similarly reported a lack of correlation between these parameters and the presence or severity of PE [8, 10, 31]. However, Matsumoto et al.’s study in 131 uremic patients, observed that among 65 patients who did not undergo dialysis, 25 of them displayed PE and serum creatinine levels exceeding 5 mg/100 ml. The authors noted an increased likelihood of uremic PE in correlation with elevated creatinine levels within this subgroup [36]. Additionally, Frommer et al. noted in pre-dialysis ESRD patients that those with PE tended to have lower BUN levels [9].

It’s important to highlight that the NLR and platelet-to-lymphocyte ratio (PLR) have recently been recognized as innovative inflammatory markers associated with adverse outcomes in a range of medical conditions [37]. However, the exploration of their relationship with PE in CKD patients has not been extensively pursued in existing research. This gap suggests a potential area for future studies to investigate the role of these markers in the context of PE among CKD patients PE, particularly in moderate to severe cases, carries the risk of progressing to tamponade. This risk is significant, with about one-third of severe PE cases, regardless of cause, evolving into tamponade. It’s crucial to note that occurrence of tamponade isn’t solely contingent on the volume of PE; the rate of fluid accumulation within the pericardial space also plays a pivotal role [2]. Our current study established that the severity of PE can serve as a predictor for the necessity of surgical intervention. None of the patients with mild PE required drainage, whereas all with severe PE and about one-fifth of those with moderate effusion underwent this procedure, aligning with previous research that associates the extent of PE with an increased likelihood of surgical intervention [38].

In Bataille et al.’s study of 44 CKD patients with PE, 43% were undergoing hemodialysis, 7% were on peritoneal dialysis, 11% had undergone transplantation, and 39% were at CKD stage 4 or 5. They reported that severe effusion often necessitated surgical drainage, as did 35% of mild to moderate effusion cases with serum albumin below 3.1 g/dl. Only 7% of patients with mild to moderate effusion and albumin levels above 3.1 g/l underwent drainage. The study concluded that large effusions in CKD patients should be promptly drained. For mild to moderate effusions, serum albumin levels can guide the decision for drainage. However, the study didn’t specify if there were changes in effusion volume between initial diagnosis and drainage, despite potential delays up to 136 days [39]. Although the risk of drainage in PE is influenced by its severity, directly comparing our results with their study is challenging due to differences in population groups, methodologies, and data reporting. Additionally, some studies report no significant relationship between serum albumin and PE in mixed populations of dialysis and non-dialysis CKD patients [31, 35], further illustrating the variability in findings across different research contexts.

Our study did not find a significant relationship between the severity of PE and in-hospital mortality, aligning with previous studies that suggest a favorable short-term prognosis for CKD patients with PE who receive prompt management [31, 35]. Further details on the review of related studies can be found in S3 Table.

Conclusions

Despite advancements in diagnosing and treating CKD, PE remains a significant concern for CKD patients. We have identified hypoalbuminemia, neutrophilia, and NLR as predictive factors for moderate and severe PE in CKD patients with PE. These findings can aid physicians in recognizing and addressing moderate to large PE in this population. However, prospective studies with larger sample size are necessary to validate our results.

Supporting information

Acknowledgments

We extend our sincere thanks to Dr. Fatemeh Pourmotahari (Dezful University of Medical Sciences) and Dr. Hadi Pashapour (Shahid Beheshti University of Medical Sciences) for their invaluable statistical guidance. Additionally, we are grateful to Dr. Shohra Qaderi (Harvard Medical University) and Dr. Noosha Sameeifar (Shahid Beheshti University of Medical Sciences) for their assistance in copyediting this paper. We would also like to thank Dr. Mariah Farrokhi Moghaddam (Shahid Beheshti University of Medical Sciences) for her valuable advice throughout this research. We also express gratitude to the Clinical Research Development Center of Shahid Modarres Educational Hospital, affiliated with Shahid Beheshti University of Medical Sciences in Tehran, Iran, for their invaluable support, cooperation, and assistance throughout the study period.