Clinica Chimica Acta 430 (2014) 164–170

Contents lists available at ScienceDirect

Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim

Invited critical review

Laboratory diagnostics of spontaneous bacterial peritonitis Giuseppe Lippi a,⁎, Elisa Danese b, Gianfranco Cervellin c, Martina Montagnana b a b c

Laboratory of Clinical Chemistry and Hematology, Academic Hospital of Parma, Parma, Italy Laboratory of Clinical Chemistry and Hematology, University of Verona, Verona, Italy Emergency Department, Academic Hospital of Parma, Parma, Italy

a r t i c l e

i n f o

Article history: Received 1 December 2013 Received in revised form 9 January 2014 Accepted 11 January 2014 Available online 6 February 2014 Keywords: Laboratory diagnostics Peritonitis Spontaneous bacterial peritonitis Peritoneal fluid Procalcitonin

a b s t r a c t The term peritonitis indicates an inflammatory process involving the peritoneum that is most frequently infectious in nature. Primary or spontaneous bacterial peritonitis (SBP) typically occurs when a bacterial infection spreads to the peritoneum across the gut wall or mesenteric lymphatics or, less frequently, from hematogenous transmission in combination with impaired immune system and in absence of an identified intra-abdominal source of infection or malignancy. The clinical presentation of SBP is variable. The condition may manifest as a relatively insidious colonization, without signs and symptoms, or may suddenly occur as a septic syndrome. Laboratory diagnostics play a pivotal role for timely and appropriate management of patients with bacterial peritonitis. It is now clearly established that polymorphonuclear leukocyte (PMN) in peritoneal fluid is the mainstay for the diagnosis, whereas the role of additional biochemical tests is rather controversial. Recent evidence also suggests that automatic cell counting in peritoneal fluid may be a reliable approach for early screening of patients. According to available clinical and laboratory data, we have developed a tentative algorithm for efficient diagnosis of SBP, which is based on a reasonable integration between optimization of human/economical resources and gradually increasing use of invasive and expensive testing. The proposed strategy entails, in sequential steps, serum procalcitonin testing, automated cell count in peritoneal fluid, manual cell count in peritoneal fluid, peritoneal fluid culture and bacterial DNA testing in peritoneal fluid. © 2014 Elsevier B.V. All rights reserved.

Contents 1.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. Primary peritonitis or spontaneous bacterial peritonitis . . . . . . . . 2. Clinical signs and symptoms . . . . . . . . . . . . . . . . . . . . . . . . 3. Complications and prognosis . . . . . . . . . . . . . . . . . . . . . . . 4. Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. State-of-the-art of laboratory diagnostics of spontaneous bacterial peritonitis 5.1. Macroscopic and microscopic examination of peritoneal fluid . . . . . 5.2. Biochemical analysis of peritoneal fluid . . . . . . . . . . . . . . . 5.3. Microbiological analysis of peritoneal fluid . . . . . . . . . . . . . . 6. Future perspectives in laboratory diagnostics of peritoneal fluid . . . . . . . . 7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction The peritoneum is the serous membrane that forms the lining of abdominal cavity, covers and supports most of intra-abdominal organs, ⁎ Corresponding author at: U.O. Diagnostica Ematochimica, Azienda OspedalieroUniversitaria di Parma, Via Gramsci, 14, 43126 Parma, Italy. Tel.: +39 0521 703050, +39 0521 703791. E-mail addresses: [email protected], [email protected] (G. Lippi). 0009-8981/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cca.2014.01.023

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

164 165 165 165 166 166 166 167 167 168 168 169

and also serves as a conduit for their blood and lymph vessels and nerves. From a biological perspective, the peritoneal membrane is a sterile, semi-permeable membrane with multiple pores, which allows a flux of solutes and water from the vascular system to the peritoneal cavity and vice versa, mainly through a diffusion mechanism [1]. The term peritonitis designates an inflammatory process involving the peritoneum. Although peritonitis may be occasionally sterile (e.g., due to chloridric acid or to bile salts), the most frequent cause is represented by infections. Bacterial peritonitis (BP) is hence

G. Lippi et al. / Clinica Chimica Acta 430 (2014) 164–170

defined as an inflammation of the serous membrane that covers the viscera and the peritoneal cavity due to bacterial contamination. It is conventionally classified as (i) primary, or spontaneous BP (SBP), when a bacterial infection spreads to the peritoneum across the gut wall or mesenteric lymphatics or – less frequently – from hematogenous transmission in combination with an impaired immune system and in absence of an identified intra-abdominal source of infection or malignancy; (ii) secondary BP, when the infection is a consequence of a gastro-intestinal perforation; and (iii) tertiary or recurrent BP, defined as persistence or recurrence of intraabdominal infection in the presence of apparently appropriate therapy [1,2]. The most important infective agents involved in different types of peritonitis are reported in Table 1. Due to the remarkable differences existing in clinical significance and management of the aforementioned forms of BP (i.e., the secondary and tertiary forms of PB are mainly of surgical competence), this article is focused on clinical and laboratory diagnostics of SBP. 1.1. Primary peritonitis or spontaneous bacterial peritonitis SBP, firstly described in 1971 by Conn and Fessel [3], is the infection of a previously sterile ascitic fluid that frequently represents a complication of liver cirrhosis. It affects about one third of cirrhotic patients [4,5], in absence of visceral perforations or other intra-abdominal infections, as abscess, acute pancreatitis or cholecystitis [6,7]. SBP is characterized by poor outcome and high mortality, ranging from 10 to 50% at the first in-hospital episode [8–10]. Bacterascites, which is instead defined as the presence of a positive culture of ascitic fluid without an increased peritoneal leukocyte count, has a much lower prevalence, ranging from 2 to 3% in outpatients, but reaching 11% in hospitalized patients [5]. Portal hypertension, changes in intestinal flora and impaired immunity, typically involving cirrhotic patients, are the main causes of bacterial overgrowth and translocation from the intestinal lumen to mesenteric lymph nodes or other extraintestinal organs and sites [6,11–13]. Changes in intestinal flora and bacterial overgrowth are represented mainly by the increased growth of Gram-negative aerobic bacilli from Enterobacteriaceae family (such as Escherichia coli and Klebsiella spp.) [14–16], due to failure of intestinal clearance [17], and in association with impaired small-bowel motility and decreased intraluminal concentration of bile salts [18,19]. Only a limited number of intestinal bacteria can efficiently translocate from the lumen of the gut into mesenteric lymph nodes, and these include E. coli, Klebsiella pneumoniae and other Enterobacteriaceae [20–22]. Gram-positive bacteria are typically involved in 23–40% of SBP and comprise Streptococci and (less often) Staphylococci, whereas anaerobic

165

bacteria (e.g., Bacteroides, Clostridia, Lactobacillus) are more frequently isolated from multiple organisms SBP [23]. Listeria monocytogenes has been only occasionally identified in cases of SBP [24,25]. The third factor predisposing to SBP in cirrhotic patients is the impaired immunity, which is characterized by reticuloendothelial system depression, leukocyte dysfunction, and altered ascitic fluid defenses [8,16]. 2. Clinical signs and symptoms Since infections in the peritoneum can be generalized or localized, the signs and symptoms of peritonitis are then highly variable. In a classical secondary peritonitis, these include swelling, bloating and tenderness in the abdomen. The pain may vary from dull aches to severe, localized or diffuse, sharp pain, and is frequently accompanied by fever (many of patients have a temperature that exceeds 38 °C, although patients with severe sepsis may become hypothermic) and chills, loss of appetite, thirst, nausea and vomiting [26–28]. Abdominal pain is more intense with motion or touch and is often lessened when patients take a fetal position [29]. Obstruction to gas or stool, oliguria, low blood pressure and tachycardia may occur in the most severe forms of generalized peritonitis. The clinical presentation of SBP is highly variable and this condition may manifest as a relatively insidious colonization – without signs and symptoms – or it can rapidly develop as a septic syndrome [6,30]. Since suggestive symptoms and signs are frequently absent in patients with SBP, the available guidelines suggest a diagnostic paracentesis in all ascitic patients admitted to the hospital [5,31–33]. Very rarely, hepatic encephalopathy may be the only manifestation of SBP. Also in tertiary peritonitis the presenting symptoms are nonspecific and insidious in onset (e.g., low-grade fever, anorexia, weight loss) [26]. 3. Complications and prognosis Despite remarkable developments in earlier detection, medical and surgical therapy, the average mortality rate of SBP remains elevated, approaching 30% [34,35], and ranging from b 5% in low-risk patients to approximately 90% in those at higher risk. Most information on the predictive factors associated with poor outcome comes from studies carried out in cirrhotic patients with SBP. In this setting well recognized indicators of mortality include advanced age [36], child score N2, the presence of bacteremia [37], lack of infection resolution, modification of antibiotic treatment and culture positivity [9,38], nosocomial origin [39], and the presence of CARD15/NOD2 (nucleotide-binding oligomerization domain-containing protein 2/caspase recruitment domain-containing protein 15) gene variants [40], along with increased concentrations of

Table 1 The most important infective agents involved in different types of peritonitis.

Aerobic

Spontaneous bacterial peritonitis

Secondary peritonitis

Tertiary peritonitis

Gram-negative Escherichia coli Klebsiella

Gram-negative Escherichia coli Enterobacter Klebsiella Proteus Fusobacterium sp. Pseudomonas aeruginosa Chlamydia trachomatis Gram-positive Streptococci Enterococci Staphylococci Listeria monocytogenes Bacteroides (B. fragilis) Eubacteria Clostridia Peptostreptococci Peptococci

Gram-negative Pseudomonas aeruginosa Enterobacter

Gram-positive Streptococci Staphylococci Listeria monocytogenes Anaerobic

Fungi

Bacteroides Clostridia Lactobacilli

Gram-positive Enterococci Staphylococcus

Candida

166

G. Lippi et al. / Clinica Chimica Acta 430 (2014) 164–170

serum bilirubin and creatinine. In these patients, approximately half of all deaths occur after resolution of infection and are consequent to development of complications such as upper gastrointestinal bleeding, renal dysfunction, hepatic encephalopathy and paralytic ileus. Among these complications, renal impairment is probably the strongest independent predictor of mortality and occurs as a result of a decreased arterial blood volume, mediated by vasoactive cytokines, with a resultant increased renin–angiotensin–aldosterone system activation [41]. In a study by Follo et al. [42], the mortality rate in 252 consecutive episodes of SBP was 100% when associated with progressive renal impairment, 31% when associated with steady renal impairment, and only 7% in patients with preserved renal function. The stronger predictors of poor outcome both in SBP and secondary BP patients include the concurrent development of sepsis and subsequent multiple organ failure (MOF). Sepsis is a complex, multifactorial, evolutive syndrome, which may progress to conditions of varying severity. Several studies have described an increased risk of death along with transition from sepsis to severe sepsis and septic shock [43]. In the context of BP, severe sepsis represents the virtual threshold separating stable from critical conditions. When improperly treated, sepsis may cause functional impairment of one or more organs or systems and finally lead to MOF [44]. Strong correlations between mortality rates and number of failing organs have been described in the literature [45]. It is also noteworthy that several scoring systems have been developed to assess the clinical prognosis of patients with BP in the past decades. The most widely used include the APACHE-II score [46], the simplified Acute Physiology Score (SAPS)-II [47], the Mannheim Peritonitis Index (MPI) [48], the Multiple Organ Dysfunction Score (MODS) [49], and the Sepsis-related Organ Failure Assessment (SOFA) score [50]. Most of these scores are based on host criteria, systemic signs of sepsis, and complications related to organ failure. Although valuable for comparing patient cohorts and institutions, these scores have limited significance in the specific, day-to-day clinical decisionmaking process for each individual patient. 4. Treatment The current approach to BP is multidisciplinary, includes medical and/or surgical interventions, and mainly consists of timely hemodynamic resuscitation, empirical antimicrobial therapy and source control measures [51]. Empirical antibiotic therapy should be initiated immediately after the diagnosis of BP, or once infection is considered likely, but should never be delayed for obtaining results of radiographic or microbiological examinations. Systemic antibiotics are administered on the knowledge of the probable composition of the infecting flora. When the source of contamination is unknown, coverage is usually directed against aerobic Gram-negative organisms and anaerobes [52]. The optimal duration of antibiotic therapy must be individualized and is guided by the underlying pathology, severity of infection, speed and effectiveness of source control, and patient response to therapy. An appropriate source control procedure is then compulsory for nearly all patients with intra-abdominal infection(s) and entails physical measures undertaken to drain infected foci and modify factors in the infectious milieu that promote microbial growth or impair host antimicrobial defenses [53]. 5. State-of-the-art of laboratory diagnostics of spontaneous bacterial peritonitis Due to the adverse prognosis when SBP is left untreated, the diagnosis of disease and the identification of the underlying cause should be performed as soon as possible. As mentioned, the therapy should not be delayed until the final diagnosis is available. Since clinical characteristics and physician assessment are usually insufficient for establishing a certain diagnosis or for excluding SBP [54], diagnostic paracentesis with appropriate ascitic fluid analysis is virtually unavoidable for a timely and accurate

diagnosis of disease, for the differential diagnosis with other conditions that may cause ascites, and for the subsequent patient management. 5.1. Macroscopic and microscopic examination of peritoneal fluid The analysis of ascitic fluid encompasses both macroscopic and microscopic examination. The macroscopic examination is aimed to define color (which varies from white, yellow, green, red, brown and black) and clarity (the peritoneal fluid typically ranges from clear, cloudy, or opalescent), although the presence of abnormal findings (i.e., haziness, cloudiness, or bloody appearance) has a very poor sensitivity for both diagnosing and ruling out SBP [55]. It is now universally agreed that laboratory diagnostics of SBP should be essentially based on leukocyte count and leukocyte differential in the ascitic fluid [5]. More specifically, the polymorphonuclear leukocyte (PMN) count is the mainstay for differentiating SBP from other causes of ascites. Some diagnostic thresholds of PMN count have been proposed, which are characterized by different values of diagnostic sensitivity and specificity. Indeed, the cut-off of 250 PMN/μL is associated with optimal sensitivity whereas, and rather understandably, higher thresholds (e.g., 500 PMN/μL) provide a much better specificity that is however counterbalanced by a lower sensitivity. General consensus has now been reached on the choice of the 250 PMN/μL cut-off, since this threshold would reduce the number of false negative cases [5]. It is noteworthy, however, that even this lower cut-off may be occasionally associated with suboptimal diagnostic sensitivity. Campillo et al. performed leukocyte and PMN counts in ascitic fluids from patients with SBP [56] and found different values according to the type of infecting bacteria. In particular, the mean PMN count in patients with peritonitis due to Staphylococcus was below the 250 PMN/mm3 threshold (i.e., 87 ± 200 PMN/μL) and was also substantially lower than that of patients with SBP due to other etiologies such as Staphylococcus (650 ± 1359 PMN/mm 3), Enterococcus (771 ± 1686 PMN/μL) or Enterobacteriaceae (3275 ± 8342 PMN/μL). González-Navajas et al. also found that the percentage of PMN in ascites was substantially lower in patients with SBP due to Gram-positive bacteria than in those with SBP due to Gram-negative bacteria (5.3 ± 9.2 versus 46.0 ± 28.7%; p = 0.01) [57]. For an appropriate estimation of PMNs, it has also been suggested that subtraction of one leukocyte per 250 red blood cells should be made to adjust for potential presence of contaminating blood in patients with hemorrhagic ascites (i.e., with a red blood cell count N 10 000/μL). According to the current Clinical and Laboratory Standards Institute (CLSI) recommendations, nucleated cell count and differential should be performed in EDTA-anticoagulated ascitic fluid by means of manual microscopy, using the same hemocytometer chamber, and after preferential staining with May-Grunwald–Giemsa [58,59]. Cytocentrifugation, with approximately 20-fold concentration of cells, is advisable because this minimizes cell distortion and produces a uniform monolayer of cellular elements [59]. When the specimen is excessively bloody or the nucleated cell count is markedly increased, the fluid should be diluted using isotonic saline or other appropriate fluids. Although result reporting varies widely across different laboratories, it is now accepted that the number of nucleated cell elements may be expressed in SI (i.e., 109/L for nucleated cell elements and 1012/L for erythrocytes) or conventional units (i.e., cells/μL) [58]. Despite optical microscopy has represented the cornerstone of ascitic fluid analysis for decades, novel opportunities have recently emerged. Most clinical laboratories are now equipped with several types of hemocytometers and urine cytofluorimeters which hold premises for the analysis of other sample matrices such as cerebrospinal fluid (CSF) and pleural and peritoneal fluids [58]. The leading drawbacks of automated flow cytometry for ascitic fluid assessment are represented by inappropriate classification of nucleated elements that may be present in biological fluids different from blood and urine (i.e., macrophages, mesothelial cells, malignant cells and other less common elements

G. Lippi et al. / Clinica Chimica Acta 430 (2014) 164–170

such as LE cells, Reed–Sternberg cells, mast cells and megakaryocytes), along with the poor sensitivity for detecting pathological samples with low cell counts, typically between 1 and 100 elements/μL. On the other hand, the use of automated instrumentation has a number of advantages compared to optical microscopy, which include fully automated sample preparation, shortened turnaround time, less intra- and interobserver variability, major degree of standardization, and no need of trained personnel for microscopic analysis. It is also understandable that the use of hemocytometers and urine cytofluorimeters requires less stringent analytical criteria for the analysis of peritoneal fluid than for CSF, since the diagnostic thresholds are much higher in the former case (e.g., 250 PMN/μL in peritoneal fluid versus 5 PMN/μL in CSF) [58]. It is hence not surprising that a number of studies have now confirmed that several automated hemocytometers display high correlation with nucleated cell count by manual microscopy (e.g., correlation coefficients comprised between 0.98 and 1.00) and very modest bias, along with optimal agreement at SBP cut-off (from 96% to 99%) [60–62]. Similarly, recent evidence suggests that automated urine cell analyzers also exhibit optimal performance at the conventional diagnostic thresholds of SBP and may hence be considered a reliable perspective for initial screening of patients [63,64]. Despite continuous evolution and improvement of instrument software, hemocytometers and urine cytofluorimeters have been specifically designed for identifying (and distinguishing) nucleated cells in blood and urine, so that their analytical performance for detecting other cell types that may be present in pathological ascites is still unsatisfactory. Particular concern has been expressed about the risk of false positive cases, wherein some nucleated cells may be counted as leukocytes. Moreover, not all blood and urine analyzers fulfill the quality criteria necessary for ascitic fluid analysis, so manufacturers should provide a specific statement of intended use that clearly defines which body fluids have been cleared by a regulatory agency for testing [58]. Finally, the laboratory should identify the lower limits for nucleated cell counting, below which the use of flow cytometry for the analysis of biological fluid may be unreliable. According to these limitations, automated flow cytometry is not intended to completely replace microscopic cell counting and classification so far, but it represents a valuable and suitable approach for initial screening of peritoneal fluid in patients with suspected SBP. 5.2. Biochemical analysis of peritoneal fluid Despite the value of cell count and differentiation in ascitic fluid is now undisputed, the role of biochemical testing is by far less certain [65]. The biochemical tests that are conventionally performed in ascitic fluid include pH, glucose, lactate dehydrogenase (LDH) and lactate (and corresponding arterial-ascitic gradients), although none of them has been demonstrated to be sufficiently sensitive or specific for identifying SBP, due to the considerable overlap of values across the various forms of peritonitis [66]. A specific algorithm however has been proposed by Akriviadis and Runyon for differentiating secondary from SBP in the setting of neutrocytic ascites, which include ascitic fluid total protein N10 g/L, ascitic glucose b 2.8 mmol/L (i.e., b50 mg/dL) or LDH value greater than the upper limit of the reference range for serum [67]. As regards ascitic glucose, although a decreased concentration has been reported in bacterial or tubercular peritonitis and carcinomatosis, its value does not significantly differ in the initial phase of SBP from that of sterile peritoneal fluid, so this measurement has an overall unsatisfactory sensitivity [68]. Although promising data has been published about the assessment of leukocyte esterase in peritoneal fluid, a systematic review where the evidence of prospective clinical studies has been systematically analyzed showed that the diagnostic performance of leukocyte esterase reagent strips varies widely, with a sensitivity comprised between 0.45 and 1.00 and a specificity comprised between 0.81 and 1.00, respectively [69]. Another more recent review of the literature concluded

167

that leukocyte esterase reagent strips are usually characterized by a high negative predictive value (N0.95 in most studies), thus supporting their use as a preliminary screening tool for diagnosis of SBP. Interestingly, Téllez-Ávila et al. recently performed a prospective study in 223 consecutive patients with ascites (49 of whom were finally diagnosed with SBP) attending to an emergency department [70] and found that sensitivity, specificity, negative predictive value and positive predictive value for identifying SBP of two different leukocyte esterase reagent strips were 0.80–0.78, 0.98 (for both), 0.90–0.91 and 0.94 (for both), respectively. It was hence concluded that the use of leukocyte esterase reagent strips may be a useful tool for the screening of SBP in emergency settings. The serum-ascites albumin gradient (SAAG), which is conventionally defined as the difference between serum and peritoneal albumin concentration, has been proven as a reliable marker of portal hypertension, using a diagnostic threshold of ≥11 g/L. However, its role in differentiating SBP from secondary forms of disease is much more controversial [65]. According to the recent recommendations of the European Association for the Study of the Liver (EASL) [71], the ascitic total protein concentration should always be assessed because patients with protein concentration b 15 g/L in ascitic fluid have an increased risk of developing SBP (Level A1 recommendation), thus taking benefit from preventive antibiotic treatment (Level A1 recommendation). This is supported by evidence that ascitic fluid proteins do not increase during episodes of SBP, whereas patients with the lowest protein concentration were found to be the most likely to develop peritoneal infection(s) [65]. Additional laboratory investigations, including amylase, C reactive protein (CRP) and cytology should only be performed in cases when the diagnosis is uncertain or in the suspicion of pancreatic disease, cancer or tuberculosis [71]. 5.3. Microbiological analysis of peritoneal fluid In general, the large majority of ascitic fluid infections are spontaneous in nature, monomicrobial and characterized by low-colony-count [72]. According to current recommendations, Gram staining is rarely helpful for diagnosing SBP and for the accurate identification of pathogens, due to the low number of bacteria that are typically found in the infected fluid (i.e., usually b1 bacterium/mL). In particular, Chinnock et al. retrospectively reviewed all peritoneal fluid analyses performed in an urban 3-hospital system [73] and reported that Gram stain had a sensitivity of 0.10 and a specificity of 0.97 for detecting SBP. Similarly, the classical culture techniques are not effective to demonstrated bacterial growth in up to two-third cases of SBP [5]. Conversely, it is currently recommended that inoculation of ascitic fluid (10 mL) in blood culture bottles for both aerobic and anaerobic culture should be performed at bedside in all patients with suspected SBP (i.e., bacterial growth can be detected in up to 90% of cases when inoculation is performed at the bedside as compared with only 40% to 60% with conventional culture) [5,71]. The use of non-radiometric systems (e.g., colorimetric BacTec) has also remarkably improved the time to diagnosis, since these techniques are much faster than using conventional blood culture bottles [5]. Even inoculation of ascitic fluid into blood culture bottles is, however, not foolproof. Culture-negative neutrocytic ascites (i.e., negative results of ascitic fluid culture associated with a PMN count of ≥ 250 cells/μL) may be encountered in up to 50% of patients with SBP [74], and this may be due to a variety of reasons. First, inappropriate preanalytical procedures (e.g. contamination or transportation delays) as well as poor culturing techniques may impair bacterial growth, thus generating false negative results. As mentioned, empirical antibiotic therapy must be initiated immediately after the diagnosis of SBP even without results of ascitic fluid culture according to current practice [5,71], so that the possibility of a false negative bacterial culture should always be considered in patients receiving antibiotics at the time of paracentesis. False negative results of ascitic fluid culture may also be observed in patients with late-stage resolving infections.

168

G. Lippi et al. / Clinica Chimica Acta 430 (2014) 164–170

As regards nucleic acid amplification tests, a recent prospective study including 102 consecutive patients with ascites reported a rather poor sensitivity of this analysis for detecting SBP (i.e., bacterial DNA could not be detected in approximately half of patients with culturenegative SBP) [75]. Controversial results were provided in separate investigations [76–78], which are mainly attributable to differences in analytical sensitivity of various DNA extraction methods and in the sequence of primers. Therefore, nucleic acid amplification can only be considered as a complement, and not a substitute, of conventional culture methods. 6. Future perspectives in laboratory diagnostics of peritoneal fluid Interesting evidence is emerging from the measurement of procalcitonin (PCT) in serum of patients with SBP. Su et al. performed a literature search to identify original studies that reported the diagnostic performance of PCT alone or in combination with other biomarkers for the diagnosis of SBP [79]. The following meta-analysis showed that serum PCT (cut-offs comprised between 0.58 and 0.75 ng/mL) displayed a high accuracy for diagnosis of SBP, with a pooled area under the curve (AUC) value of 0.95 (95% confidence interval [95% CI], 0.82–0.99), a sensitivity of 0.86 (95% CI, 0.73–0.94), and a specificity of 0.80 (95% CI, 0.72–0.87). In two out of three studies included in the meta-analysis, the sensitivity of PCT was 0.95, whereas it was 0.50 in the remaining investigation that used a higher cut-off for supporting a high diagnostic specificity. Even more interestingly, the positive likelihood ratio of 7.73 (95% CI, 0.91–65.64) was considered to be sufficiently high for using PCT as a diagnostic test, whereas the low negative likelihood ratio (0.14; 95% CI, 0.01–1.89) was deemed suitable to suggest discontinuation of antibiotics therapy in combination with a negative culture. The concentration of PCT in serum or plasma was also found to be a better marker than CRP or Interelukin-6 for distinguishing SBP from other causes of ascites. These results were then confirmed in two subsequent investigations. Cekin et al. measured serum PCT and CRP levels in 101 patients with ascites (20% of whom with infectious peritonitis) [80]. Using receiver characteristic curve (ROC) analysis, procalcitonin (cut-off b 0.61 ng/mL) displayed an AUC of 0.98 for diagnosing SBP. In a recent study on 84 patients with chronic severe hepatitis B, 42 of whom with SBP, Yuan et al. reported a good diagnostic performance for both PCT (cut-off, 0.48 ng/mL; AUC, 0.89; 95% CI, 0.81–0.96) and CRP (cut-off, 16.15 mg/L; AUC, 0.86; 95% CI, 0.78–0.94) for the identification of SBP [81]. The sensitivity and specificity of serum PCT were 0.95 and 0.79, respectively. Neutrophil gelatinase-associated lipocalin (NGAL) belongs to the family of lipocalins and mainly acts as an endogenous bacteriostatic agent that interferes with siderophore-mediated iron acquisition. The leading endogenous sources of this protein include activated neutrophils, tubular cell of the kidney, cardiomyocytes, and epithelia of the prostate, uterus, salivary glands, lung, liver, trachea, stomach, bowel and colon [82]. NGAL mainly exists as a monomeric form (which is prevalently synthesized by tubular cells), along with a homodimeric form (which is mainly released by activated neutrophils) and a heterodimeric form, bound to matrix metalloproteinase 9 (MMP-9), which is also prevalently present in the kidney. Owing to the fact that the current commercial immunoassays for measuring NGAL do not distinguish one molecular species from the others [83], and that the protein is actively released by PMN, some recent studies have assessed the potential usefulness of NGAL for diagnosing SBP. Axelsson et al. first described a 10-fold increase of NGAL plasma levels in patients with acute peritonitis [84]. Leung et al. measured NGAL concentration in peritoneal dialysate effluent in patients with continuous ambulatory peritoneal dialysisrelated peritonitis and found that the concentration of this biomarker was higher in patients with Gram-positive or Gram-negative peritonitis than in those with culture-negative peritonitis [85]. In a following investigation, Martino et al. observed a remarkable increased concentration of NGAL in the peritoneal fluid of dialysis patients with SBP and found

that NGAL assessment in peritoneal fluid had an AUC of 0.99 [86]. Lippi et al. also measured NGAL and LDH in 111 peritoneal fluids, 23% of which from patients with SBP [87], and found an AUC of 0.88 for LDH, 0.89 for NGAL and 0.94 for their combination (both tests positive) for identifying bacterial infections. The sensitivity was 0.81 for LDH, 0.96 for NGAL (cut-off of 120 ng/mL) and 0.80 for their combination, whereas the specificity was 0.87 for LDH, 0.75 for NGAL and 0.95 for their combination. Interestingly, the diagnostic performance of total proteins (AUC, 0.80) and glucose (AUC, 0.71) in peritoneal fluid was consistently worse than that of LDH, NGAL, or their combination. More recently, Lacquaniti et al. studied 30 patients with peritonitis and 30 patients undergoing continuous ambulatory peritoneal dialysis (CAPD) [88] and reported that NGAL levels in peritoneal fluid were higher compared with baseline values at the onset of peritonitis and, even more importantly, that the assessment of this biomarker in peritoneal fluid showed a good diagnostic performance for identifying treatment failure. Interestingly, it has also recently been reported that the concentration of NGAL in CSFs displayed an AUC of 0.94 (95% CI, 0.89 to 0.99) for identifying acute bacterial meningitis, with a sensitivity of 1.00 and a specificity of 0.74 at a diagnostic threshold of 13 ng/mL [89]. Taken together, this clinical evidence suggests that NGAL in peritoneal fluid may be regarded as a putative biomarker for rapid screening of patients with suspected SBP and, possibly, for monitoring effectiveness of treatment. Further and larger studies are needed, however, to confirm these preliminary findings. Additional and appareling perspectives emerge from two recent proteomic studies. Tyan et al. performed 2-dimensional gel electrophoresis (2DE) coupled with reverse phase nano-high performance liquid chromatography electrospray ionization tandem mass spectrometry (RP-nano-HPLC–ESI-MS/MS) followed by peptide fragmentation pattern in the peritoneal dialysate of 12 patients before and after peritonitis [90] and detected as many as 350 significant spots. After excluding proteins that were ultrafiltered from circulation, ten putative proteins were identified as being differentially expressed (i.e., N2-fold or 50%) before and after peritonitis. More specifically, downregulation was found for apolipoprotein A-I, heat shock 70 kDa protein 1A/1B, interalphatrypsin inhibitor heavy chain H4, fibrinogen gamma and beta chains, ceruloplasmin, zinc-α-2-glycoprotein, and α-1-antitrypsin, whereas up-regulation was observed for haptoglobin and antithrombin. An altered plasma proteome has also been found by Thongboonkerd et al. in plasma of pigs before and 12 h after peritonitis-induced sepsis [91]. After resolution by 2DE and staining with SYPRO Ruby fluorescence dye, 36 spots were found to be significantly modified in plasma. Subsequent analysis with quadrupole-time-of-flight (Q-TOF) MS and MS/MS allowed the identification of 22 proteins which were up-regulated in sepsis and five proteins which were instead down-regulated. 7. Conclusions Despite the vast array of potential analyses (Table 2), the current laboratory diagnostics of SBP entail a limited number of conventional investigations, which basically include PMN count in peritoneal fluid and peritoneal fluid culture. However, some emerging tests may provide a significant contribution to the diagnosis and therapeutic management of this disorder. These basically include serum procalcitonin, along with assessment of NGAL and bacterial DNA in peritoneal fluid. The use of leukocyte esterase reagent strips is another appealing opportunity for those healthcare settings where timely diagnosis is pivotal and conventional laboratory resources are not easily available (e.g., the emergency department). According to clinical and analytical evidence available so far, it seems hence reasonable to suggest a tentative algorithm for rapid and efficient diagnosis of SBP, which is based on a reasonable integration between optimization of human/economical resources and the gradually increasing use of invasive and expensive testing (Fig. 1). Owing to the high sensitivity, which has been reported to be approximately 0.95 using a diagnostic threshold of 0.5 ng/mL or higher, serum PCT may help to

G. Lippi et al. / Clinica Chimica Acta 430 (2014) 164–170 Table 2 Spectrum of laboratory tests for diagnosis and differential diagnosis of spontaneous bacterial peritonitis (SBP). Peritoneal fluid analysis 1) Macroscopic and microscopic examination a) Color and clarity b) Leukocyte count and differential 2) Biochemical analysis a) pH b) Glucose c) Lactic acid d) Lactate dehydrogenase (LDH) e) Leukocyte esterase f) Total protein g) (Serum)-ascites albumin gradient 2) Microbiological analysis a) Gram staining b) Peritoneal fluid culture c) Nucleic acid amplification Serum or plasma analysis 1) Procalcitonin 2) Neutrophil gelatinase-associated lipocalin (NGAL) 3) Amylase 4) C reactive protein

Suspected SBP

Serum Procalcitonin

NO SBP

PMN 5.0 ng/mL

Paracentesis

Automated cell count

PNM >250/µL

NO SBP

PMN 250/µL

NO SBP

Negative

Empiric antibiotic therapy

Peritoneal fluid culture

Positive

Nucleic acids amplification

Specific antibiotic therapy

Negative

Positive Consider non-bacterial etiology

is b250 cells/μL, SBP can be safely ruled out as for current guidelines, due to the optimal agreement with optical microscopy. Owing to the modest but still clinically meaningful number of false positive cases, a peritoneal fluid PMN value N 250 cells/μL should be further confirmed with microscopic cell count. The presence of SBP may hence be excluded when the manual PMN count is b 250 cells/μL, whereas empiric antibiotic therapy, accompanied with peritoneal fluid culture, should be immediately started when the PMN count exceeds this threshold. In the presence of bacterial growth, targeted antibiotic therapy should be established, whereas nucleic acid amplification may be advisable in the case of negative peritoneal fluid culture. Identification of bacterial DNA would then allow diagnosing neutrocytic ascites and starting targeted antibiotic therapy, whereas non-bacterial etiologies should be considered in the case of negative nucleic acid amplification. Indeed, this diagnostic algorithm is aimed to integrate but not replace existing guidelines and represents an innovative approach to be further tested in large prospective studies and cost-effective analyses.

References

timely rule out SBP in those patients with concentration of this biomarker below the diagnostic cut-off. When serum PCT concentration is N0.5 ng/mL, paracentesis should be performed and followed by automated cell count in peritoneal fluid. When the PMN count