The Veterinary Journal 202 (2014) 640–642
Contents lists available at ScienceDirect
The Veterinary Journal j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / t v j l
Short Communication
Diagnostic accuracy of a urine reagent strip to identify bacterial peritonitis in dogs with ascites E.J. Thomovsky a,*, P.A. Johnson a, G.E. Moore b a b
Department of Veterinary Clinical Sciences, College of Veterinary Medicine, Purdue University, 625 Harrison St., West Lafayette, IN 47907, USA Department of Comparative Pathobiology, College of Veterinary Medicine, Purdue University, 625 Harrison St., West Lafayette, IN 47907, USA
A R T I C L E
I N F O
Article history: Accepted 22 September 2014 Keywords: Bacterial peritonitis Diagnosis Urine dipstick Ascites Purulent exudate
A B S T R A C T
The leukocyte esterase test pad on a urine dipstick has been used as a preliminary test for bacterial peritonitis in humans but has not previously been evaluated in dogs. Here, free abdominal fluid from 60 dogs was tested on the leukocyte esterase test pad and results were compared with culture and microscopic analysis. Depending on the ‘gold standard’ comparator, the dipstick had sensitivity of ~60–75%, specificity of ~91–92%, positive predictive value of ~69%, and negative predictive value of ~87–94%. Based on these data, it appears that the leukocyte esterase test pad is most useful for tentative identification of cases in which bacterial infection is unlikely. Therefore a negative test may aid in re-directing clinician attention to alternative diagnoses in dogs with free abdominal effusion, whereas a positive result implies the necessity for further diagnostic tests. © 2014 Elsevier Ltd. All rights reserved.
Bacterial peritonitis is a commonly diagnosed and life-threatening condition in small animal patients with a reported prevalence of 39% in dogs presenting with peritoneal effusion to a veterinary teaching hospital (Bonczynski et al., 2003) and mortality rates reported from 30% to 68% in dogs (Hosgood and Salisbury, 1988; Culp et al., 2009). Associated clinical signs are variable and may not overtly indicate abdominal disease, making the diagnosis of bacterial peritonitis a challenge for clinicians (Culp et al., 2009). In veterinary medicine, the diagnosis of bacterial peritonitis is made if intracellular bacteria are seen on a smear made from abdominal effusion and/or there is a positive bacterial culture derived from the abdominal effusion (Bonczynski et al., 2003). Results of fluid analysis and cytological examination typically are available after 12– 24 h and culture results take 5–7 days. Therefore, availability of a rapid, inexpensive screening test for bacterial peritonitis would be helpful for practitioners. In human medicine, there are multiple publications describing use of the Multistix 8 SG Reagent Strip (Bayer Diagnostics) to predict spontaneous bacterial peritonitis (Butani et al., 2004; Campillo et al., 2006; Nousbaum et al., 2007; Ribeiro et al., 2007; Araujo et al., 2008; Rerknimitr et al., 2010). These reagent strips were reported to have a high degree of specificity (up to 99.2%) (Nousbaum et al., 2007) and high negative predictive values (as high as 99.4%) (Campillo et al., 2006); therefore a negative test result can almost rule out peritonitis in the patient.
* Corresponding author. Tel.: +1 765 4741107. E-mail address: [email protected] (E.J. Thomovsky). http://dx.doi.org/10.1016/j.tvjl.2014.09.020 1090-0233/© 2014 Elsevier Ltd. All rights reserved.
The aim of this study was to determine the sensitivity and specificity of the Multistix 8 SG reagent strips in diagnosing bacterial peritonitis in dogs presenting with free abdominal fluid. All dogs presenting to the Purdue University Small Animal Veterinary Teaching Hospital with ascites and undergoing abdominocentesis as part of their normal diagnostic workup between June 2012 and August 2013 were eligible for inclusion in the study. Samples were obtained using ultrasound guidance, or blindly via a four quadrant technique while animals were restrained in lateral or dorsal recumbency. At the attending clinician’s discretion, some dogs were sedated. Samples were excluded if the fluid amount obtained by abdominocentesis was too little for cytological examination and culture or if colorimetric testing was altered (e.g. PCV >5%). Prior antibiotic therapy was recorded for each patient but was not an exclusion criterion because antibiotic therapy would be expected to alter all test results equally. Immediately following abdominocentesis, a drop of fluid was applied to the leukocyte esterase test pad. After 2 min, the test pad color was compared to the product color guide provided. Individuals performing the test were either the authors (ET, PJ) or one of the clinicians working in the Small Animal Emergency service. Samples were submitted to the clinical pathology laboratory,1 where cytological examinations were conducted by a board-certified clinical pathologist and a cell count and total protein were determined via analyzer (Abbott Cell Dyne, Abbott Hematology). Cultures were obtained by inoculating cotton swabs with two to three drops
1
Clinical Pathology Laboratory, Purdue University, W. Lafayette, IN, USA.
E.J. Thomovsky et al./The Veterinary Journal 202 (2014) 640–642
Table 1 Disease diagnoses for dogs included in the study. Presenting disease Neoplasia Ascites (secondary to right sided heart failure or right atrial mass) Peritonitis (secondary to GI perforation) Ascites (secondary to PLE/PLN) Pancreatitis Nephrotic syndrome Renal failure Duodenal or gastric ulcers without perforation Mesenteric torsion Hepatic fibrosis GDV Bile peritonitis Septic uroabdomen Prostatic abscess Open
Number of dogs enrolled 12 13 7 6 3 2 2 1 1 1 1 1 1 1 8
GI, gastrointestinal; PLE, protein-losing enteropathy; PLN, protein-losing nephropathy; GDV, gastric-dilatation-volvulus.
of abdominal fluid and placing them in culture medium prior to submission to the diagnostic laboratory.2 Upon arrival in the diagnostic laboratory, samples were inoculated onto blood agar, MacConkey agar, anaerobic blood agar and then placed into thioglycolate enrichment broth. Cultures were handled according to standard microbiology procedures. Patient signalment and final diagnosis plus abdominal fluid test results were recorded. Leukocyte esterase test results were considered positive if the reading was ‘small,’ ‘moderate,’ or ‘large’ and negative if ‘trace’ or ‘negative’ was identified. Samples were considered positive if intracellular bacteria were observed within neutrophils or if there was positive growth during culture of the fluid. The observers recording dipstick results were blinded to final diagnosis because the laboratory results were not available at the time the dipstick analysis was conducted. The results of the leukocyte esterase dipstick test were compared to the cytology and culture results for calculation of sensitivity, specificity, positive predictive value, and negative predictive value. Separate calculations were conducted using different reference standards for diagnosis of bacterial peritonitis as follows: (1) only the results of cytology; (2) only the results of culture; and (3) either the results of cytology or culture. A total of 60 dogs of a variety of breeds and ranging in age from 8 months to 14 years (mean 8 ± 3.1 years) were enrolled in the study. The dogs presented for a range of diseases (see Table 1) that were divided into the following cytological categories: non-septic exudates (7), neoplastic effusions attributed to lymphoma (6), modified transudates (25), pure transudates (10), and septic exudates (12). In all comparisons, the leukocyte esterase test pad on the reagent strip had a specificity of 91.1% (CI: 78.8–97.5%) or greater depending on the comparison being conducted (see Table 2 for more details). The highest sensitivity for the reagent strips (75%; CI: 42.8–94.5%) was found when comparing reagent strip results with the positive cytology results (see Table 2). Table 3 presents more detail on the relationship between the dipstick results and the type of effusion determined through cytological examination. A total of 18 cases were culture-positive. Nine of these cases were also positive on the reagent strip: seven were diagnosed as
641
septic effusions based on cytology and the remaining two were neoplastic effusions. The specificity of the leukocyte esterase test pad on the Multistix 8 SG Reagent Strips ranged between 91.1% and 95.3% depending on the standard used for diagnosis of bacterial peritonitis. These specificities were similar to those found in the human literature when the reagent strip was used to test patients with abdominal effusion for spontaneous peritonitis. Specificities reported in the human literature were 98% (Campillo et al., 2006), 96% (Ribeiro et al., 2007), 99.2% (95% CI: 98.8– 99.6) (Nousbaum et al., 2007), 98.5% (95% CI: 94–99.7%) (Araujo et al., 2008), 94.5% (Rerknimitr et al., 2010), and 97.7% (95% CI: 93–98%) (Tellez-Avila et al., 2012). In the human literature, peritonitis was defined as polymorphonuclear (PMN) cell counts ≥250 cells/μL, which is consistent with using cytology results in which an exudate would have a total nucleated cell count ≥250 cells/μL. The specificity from our study comparing the reagent strip results to cytology results was 91.7% (CI: 80.0–97.7%). The sensitivity for the leukocyte esterase test pad in this study when compared with cytology results was 75% (CI: 42.8– 94.5%). This is very similar to or in some cases greater than the sensitivities found in the human literature that were reported as: 45.7% (Campillo et al., 2006), 86% (Ribeiro et al., 2007), 45.3% (36.6–54.3%) (Nousbaum et al., 2007), 80% (95% CI: 58.7–92.4%) (Araujo et al., 2008), 80% (Rerknimitr et al., 2010), and 77.5% (95% CI: 64–88%) (Tellez-Avila et al., 2012). Several hypotheses were advanced by these authors to explain the low sensitivities. Some authors proposed that leukocyte esterase test pads were negative in the presence of low numbers of neutrophils (Nousbaum et al., 2007; Araujo et al., 2008). Others (Campillo et al., 2006) suggested that the sensitivity was limited by the need for leukocyte lysis – which typically occurred with chronicity – to produce a positive leukocyte esterase test pad result. Looking critically at the cases of peritonitis diagnosed by either culture or cytology, there were three cases in which the reagent strip was erroneously negative in cases of bacterial peritonitis. Two dogs had intestinal perforations and the third dog had a prostatic abscess with secondary bacterial peritonitis. No definite reason was determined for the false negative results in these cases but it is possible that the fluid was tested from these dogs so early in the course of disease (i.e. shortly after perforation or abscess rupture) that leukocyte lysis was limited, leading to a false negative result as suggested by Campillo et al. (2006). We suggest that using the reagent strip as an initial diagnostic test to determine if a patient might have bacterial peritonitis is worthwhile because reagent strips are readily available, minimal in cost, require no special training or skills, and testing is easily conducted with a single drop of peritoneal fluid. With its high specificity and negative predictive value, the leukocyte esterase test pad on the reagent strip can be used by practitioners to determine how intensely bacterial peritonitis should be suspected. In general, if the reagent strip is positive, follow up testing is mandatory, including cytology of the fluid, paired glucose and/or lactate testing between the fluid and blood, and/or imaging/exploratory surgery of the abdomen. If the reagent strip is negative, the likelihood that the patient has bacterial peritonitis is low but the authors would recommend re-testing of the reagent strip at a future date and/or follow-up testing if the patient’s condition does not improve.
Conflict of interest statement
2 Indiana Animal Disease and Diagnostic Laboratory, W. Lafayette, IN, USA; accredited by the American Association of Veterinary Laboratory Diagnosticians (AAVLD).
None of the authors of this paper has a financial or personal relationship with other people or organizations that could inappropriately influence or bias the content of the paper.
642
E.J. Thomovsky et al./The Veterinary Journal 202 (2014) 640–642
Table 2 Detailed results of the performance of the leukocyte esterase test pad on the Multistix 8 SG Reagent Strips. Cytological analysis
Positive leukocyte esterase on reagent strip Negative leukocyte esterase on reagent strip Prevalence of bacterial peritonitis Sensitivity of reagent strip Specificity of reagent strip Positive predictive value reagent strip Negative predictive value reagent strip
Culture results
Cytology or culture results
Positive
Negative
Positive
Negative
Either positive
Both negative
9/60 3/60
4/60 44/60
9/60 6/60
4/60 41/60
11/60 6/60
2/60 41/60
12/60 = 20% 75% (CI:42.8–94.5%) 91.7% (CI:80.0–97.7%) 69.2% (CI:38.6–90.9%) 93.6% (CI:82.5–98.7%)
15/60 = 25% 60% (CI:32.3–83.7%) 91.1% (CI:78.8–97.5%) 69.2% (CI:38.6–90.9%) 87.2% (CI:74.3–95.2%)
17/60 = 28% 64.7% (CI:38.3–85.8%) 95.3% (CI:84.2–99.4%) 84.6% (CI:54.6–98.1%) 87.2% (CI:74.3–95.2%)
Positive cytology results indicate the presence of intracellular bacteria diagnosed by a clinical pathologist. Positive culture results indicate bacterial growth under either aerobic or anaerobic conditions. Positive for the combined cytology and culture results mean that bacterial growth was found under aerobic or anaerobic culture and/or intracellular bacteria were seen during cytological examination.
Table 3 Detailed results of the dipstick analysis compared to cytological effusion type. Cytological effusion type
Dipstick results
Septic exudate
3/12 false negatives
Non-septic exudate
3/7 false positives
Neoplastic Pure transudates Modified transudates
1/6 false positives 0/6 false positives 0/25 false positives
Dog 1 2 3 1 2 3 1 – –
References Araujo, A., Lopes, A., Michalczuk, M.T., Stifft, J., Nardelli, E., Escobar, G., Rossi, G., Alvares-da-Silva, M.R., 2008. Is there yet a place for reagent strips in diagnosing spontaneous bacterial peritonitis in cirrhotic patients? An accuracy and costeffectiveness study in Brazil. Hepatology (Baltimore, Md.) 23, 1895–1900. Bonczynski, J.J., Ludwig, L.L., Barton, L.J., Loar, A.S., Peterson, M.E., 2003. Comparison of peritoneal fluid and peripheral blood pH, bicarbonate, glucose, and lactate concentration as a diagnostic tool for septic peritonitis in dogs and cats. Veterinary Surgery 32, 161–166. Butani, R.C., Shaffer, R.T., Szyjkowski, R.D., Weeks, B.E., Speights, L.G., Kadakia, S.C., 2004. Rapid diagnosis of infected ascetic fluid using leukocyte esterase dipstick testing. The American Journal of Gastroenterology 99, 532–537. Campillo, B., Richardet, J.P., Dupeyron, C., 2006. Diagnostic value of two reagent strips (Multistix 8 SG and Combur 2 LN) in cirrhotic patients with spontaneous bacterial peritonitis and symptomatic bacterascites. Gastroenterologie Clinique et Biologique 30, 446–452. Culp, W.T., Zeldis, T.E., Reese, M.S., Drobatz, K.J., 2009. Primary bacterial peritonitis in dogs and cats: 24 cases (1990–2006). Journal of American Veterinary Medical Association 234, 806–913.
Diagnosis
Bacteria cultured
Perforated duodenum and gastric mass Small intestinal mass/perforation Prostatic abscess Neoplasia (mesothelioma) Cranial abdominal mass Duodenal/gastric ulcers (no perforation) GI lymphoma – –
Clostridium perfringens Bacteriodes ureolyticus E. coli – Streptococcus canis – Enterococcus spp. – –
Hosgood, G., Salisbury, S.K., 1988. Generalized peritonitis in dogs: 50 cases (1975–1986). Journal of the American Veterinary Medical Association 193, 1448–1450. Nousbaum, J.B., Cadranel, J.F., Nahon, P., Khac, E.N., Moreau, R., Thevenot, T., Silvain, C., Bureau, C., Nouel, O., Pilette, C., et al., 2007. Diagnostic accuracy of the Multistix 8 SG reagent strip in diagnosis of spontaneous bacterial peritonitis. Hepatology (Baltimore, Md.) 45, 1275–1281. Rerknimitr, R., Lilmmathurotsakul, D., Bhokaisawan, N., Kongkam, P., Treeprasertsuk, S., Kullaveanijaya, P., 2010. A comparison of diagnostic efficacies among different reagent strips and automated cell count in spontaneous bacterial peritonitis. Journal of Gastroenterology and Hepatology 25, 946–950. Ribeiro, T.C., Kondo, M., Amaral, A.C., Parise, E.R., Bragagnolo Junior, M.A., de Souza, A.F.M., 2007. Evaluation of reagent strips for ascetic fluid leukocyte determination: Is it possible for spontaneous bacterial peritonitis rapid diagnosis? The Brazilian Journal of Infectious Diseases 11, 70–74. Tellez-Avila, F.I., Chavez-Tapia, N.C., Franco-Guzman, A.M., Uribe, M., Vargas-Vorackova, F., 2012. Rapid diagnosis of spontaneous bacterial peritonitis using leukocyte esterase reagent strips in emergency department: Uri-quick clini-10SG® vs. Multistix 10SG®. Annals of Hepatology 11, 696–699.
Contents lists available at ScienceDirect
The Veterinary Journal j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / t v j l
Short Communication
Diagnostic accuracy of a urine reagent strip to identify bacterial peritonitis in dogs with ascites E.J. Thomovsky a,*, P.A. Johnson a, G.E. Moore b a b
Department of Veterinary Clinical Sciences, College of Veterinary Medicine, Purdue University, 625 Harrison St., West Lafayette, IN 47907, USA Department of Comparative Pathobiology, College of Veterinary Medicine, Purdue University, 625 Harrison St., West Lafayette, IN 47907, USA
A R T I C L E
I N F O
Article history: Accepted 22 September 2014 Keywords: Bacterial peritonitis Diagnosis Urine dipstick Ascites Purulent exudate
A B S T R A C T
The leukocyte esterase test pad on a urine dipstick has been used as a preliminary test for bacterial peritonitis in humans but has not previously been evaluated in dogs. Here, free abdominal fluid from 60 dogs was tested on the leukocyte esterase test pad and results were compared with culture and microscopic analysis. Depending on the ‘gold standard’ comparator, the dipstick had sensitivity of ~60–75%, specificity of ~91–92%, positive predictive value of ~69%, and negative predictive value of ~87–94%. Based on these data, it appears that the leukocyte esterase test pad is most useful for tentative identification of cases in which bacterial infection is unlikely. Therefore a negative test may aid in re-directing clinician attention to alternative diagnoses in dogs with free abdominal effusion, whereas a positive result implies the necessity for further diagnostic tests. © 2014 Elsevier Ltd. All rights reserved.
Bacterial peritonitis is a commonly diagnosed and life-threatening condition in small animal patients with a reported prevalence of 39% in dogs presenting with peritoneal effusion to a veterinary teaching hospital (Bonczynski et al., 2003) and mortality rates reported from 30% to 68% in dogs (Hosgood and Salisbury, 1988; Culp et al., 2009). Associated clinical signs are variable and may not overtly indicate abdominal disease, making the diagnosis of bacterial peritonitis a challenge for clinicians (Culp et al., 2009). In veterinary medicine, the diagnosis of bacterial peritonitis is made if intracellular bacteria are seen on a smear made from abdominal effusion and/or there is a positive bacterial culture derived from the abdominal effusion (Bonczynski et al., 2003). Results of fluid analysis and cytological examination typically are available after 12– 24 h and culture results take 5–7 days. Therefore, availability of a rapid, inexpensive screening test for bacterial peritonitis would be helpful for practitioners. In human medicine, there are multiple publications describing use of the Multistix 8 SG Reagent Strip (Bayer Diagnostics) to predict spontaneous bacterial peritonitis (Butani et al., 2004; Campillo et al., 2006; Nousbaum et al., 2007; Ribeiro et al., 2007; Araujo et al., 2008; Rerknimitr et al., 2010). These reagent strips were reported to have a high degree of specificity (up to 99.2%) (Nousbaum et al., 2007) and high negative predictive values (as high as 99.4%) (Campillo et al., 2006); therefore a negative test result can almost rule out peritonitis in the patient.
* Corresponding author. Tel.: +1 765 4741107. E-mail address: [email protected] (E.J. Thomovsky). http://dx.doi.org/10.1016/j.tvjl.2014.09.020 1090-0233/© 2014 Elsevier Ltd. All rights reserved.
The aim of this study was to determine the sensitivity and specificity of the Multistix 8 SG reagent strips in diagnosing bacterial peritonitis in dogs presenting with free abdominal fluid. All dogs presenting to the Purdue University Small Animal Veterinary Teaching Hospital with ascites and undergoing abdominocentesis as part of their normal diagnostic workup between June 2012 and August 2013 were eligible for inclusion in the study. Samples were obtained using ultrasound guidance, or blindly via a four quadrant technique while animals were restrained in lateral or dorsal recumbency. At the attending clinician’s discretion, some dogs were sedated. Samples were excluded if the fluid amount obtained by abdominocentesis was too little for cytological examination and culture or if colorimetric testing was altered (e.g. PCV >5%). Prior antibiotic therapy was recorded for each patient but was not an exclusion criterion because antibiotic therapy would be expected to alter all test results equally. Immediately following abdominocentesis, a drop of fluid was applied to the leukocyte esterase test pad. After 2 min, the test pad color was compared to the product color guide provided. Individuals performing the test were either the authors (ET, PJ) or one of the clinicians working in the Small Animal Emergency service. Samples were submitted to the clinical pathology laboratory,1 where cytological examinations were conducted by a board-certified clinical pathologist and a cell count and total protein were determined via analyzer (Abbott Cell Dyne, Abbott Hematology). Cultures were obtained by inoculating cotton swabs with two to three drops
1
Clinical Pathology Laboratory, Purdue University, W. Lafayette, IN, USA.
E.J. Thomovsky et al./The Veterinary Journal 202 (2014) 640–642
Table 1 Disease diagnoses for dogs included in the study. Presenting disease Neoplasia Ascites (secondary to right sided heart failure or right atrial mass) Peritonitis (secondary to GI perforation) Ascites (secondary to PLE/PLN) Pancreatitis Nephrotic syndrome Renal failure Duodenal or gastric ulcers without perforation Mesenteric torsion Hepatic fibrosis GDV Bile peritonitis Septic uroabdomen Prostatic abscess Open
Number of dogs enrolled 12 13 7 6 3 2 2 1 1 1 1 1 1 1 8
GI, gastrointestinal; PLE, protein-losing enteropathy; PLN, protein-losing nephropathy; GDV, gastric-dilatation-volvulus.
of abdominal fluid and placing them in culture medium prior to submission to the diagnostic laboratory.2 Upon arrival in the diagnostic laboratory, samples were inoculated onto blood agar, MacConkey agar, anaerobic blood agar and then placed into thioglycolate enrichment broth. Cultures were handled according to standard microbiology procedures. Patient signalment and final diagnosis plus abdominal fluid test results were recorded. Leukocyte esterase test results were considered positive if the reading was ‘small,’ ‘moderate,’ or ‘large’ and negative if ‘trace’ or ‘negative’ was identified. Samples were considered positive if intracellular bacteria were observed within neutrophils or if there was positive growth during culture of the fluid. The observers recording dipstick results were blinded to final diagnosis because the laboratory results were not available at the time the dipstick analysis was conducted. The results of the leukocyte esterase dipstick test were compared to the cytology and culture results for calculation of sensitivity, specificity, positive predictive value, and negative predictive value. Separate calculations were conducted using different reference standards for diagnosis of bacterial peritonitis as follows: (1) only the results of cytology; (2) only the results of culture; and (3) either the results of cytology or culture. A total of 60 dogs of a variety of breeds and ranging in age from 8 months to 14 years (mean 8 ± 3.1 years) were enrolled in the study. The dogs presented for a range of diseases (see Table 1) that were divided into the following cytological categories: non-septic exudates (7), neoplastic effusions attributed to lymphoma (6), modified transudates (25), pure transudates (10), and septic exudates (12). In all comparisons, the leukocyte esterase test pad on the reagent strip had a specificity of 91.1% (CI: 78.8–97.5%) or greater depending on the comparison being conducted (see Table 2 for more details). The highest sensitivity for the reagent strips (75%; CI: 42.8–94.5%) was found when comparing reagent strip results with the positive cytology results (see Table 2). Table 3 presents more detail on the relationship between the dipstick results and the type of effusion determined through cytological examination. A total of 18 cases were culture-positive. Nine of these cases were also positive on the reagent strip: seven were diagnosed as
641
septic effusions based on cytology and the remaining two were neoplastic effusions. The specificity of the leukocyte esterase test pad on the Multistix 8 SG Reagent Strips ranged between 91.1% and 95.3% depending on the standard used for diagnosis of bacterial peritonitis. These specificities were similar to those found in the human literature when the reagent strip was used to test patients with abdominal effusion for spontaneous peritonitis. Specificities reported in the human literature were 98% (Campillo et al., 2006), 96% (Ribeiro et al., 2007), 99.2% (95% CI: 98.8– 99.6) (Nousbaum et al., 2007), 98.5% (95% CI: 94–99.7%) (Araujo et al., 2008), 94.5% (Rerknimitr et al., 2010), and 97.7% (95% CI: 93–98%) (Tellez-Avila et al., 2012). In the human literature, peritonitis was defined as polymorphonuclear (PMN) cell counts ≥250 cells/μL, which is consistent with using cytology results in which an exudate would have a total nucleated cell count ≥250 cells/μL. The specificity from our study comparing the reagent strip results to cytology results was 91.7% (CI: 80.0–97.7%). The sensitivity for the leukocyte esterase test pad in this study when compared with cytology results was 75% (CI: 42.8– 94.5%). This is very similar to or in some cases greater than the sensitivities found in the human literature that were reported as: 45.7% (Campillo et al., 2006), 86% (Ribeiro et al., 2007), 45.3% (36.6–54.3%) (Nousbaum et al., 2007), 80% (95% CI: 58.7–92.4%) (Araujo et al., 2008), 80% (Rerknimitr et al., 2010), and 77.5% (95% CI: 64–88%) (Tellez-Avila et al., 2012). Several hypotheses were advanced by these authors to explain the low sensitivities. Some authors proposed that leukocyte esterase test pads were negative in the presence of low numbers of neutrophils (Nousbaum et al., 2007; Araujo et al., 2008). Others (Campillo et al., 2006) suggested that the sensitivity was limited by the need for leukocyte lysis – which typically occurred with chronicity – to produce a positive leukocyte esterase test pad result. Looking critically at the cases of peritonitis diagnosed by either culture or cytology, there were three cases in which the reagent strip was erroneously negative in cases of bacterial peritonitis. Two dogs had intestinal perforations and the third dog had a prostatic abscess with secondary bacterial peritonitis. No definite reason was determined for the false negative results in these cases but it is possible that the fluid was tested from these dogs so early in the course of disease (i.e. shortly after perforation or abscess rupture) that leukocyte lysis was limited, leading to a false negative result as suggested by Campillo et al. (2006). We suggest that using the reagent strip as an initial diagnostic test to determine if a patient might have bacterial peritonitis is worthwhile because reagent strips are readily available, minimal in cost, require no special training or skills, and testing is easily conducted with a single drop of peritoneal fluid. With its high specificity and negative predictive value, the leukocyte esterase test pad on the reagent strip can be used by practitioners to determine how intensely bacterial peritonitis should be suspected. In general, if the reagent strip is positive, follow up testing is mandatory, including cytology of the fluid, paired glucose and/or lactate testing between the fluid and blood, and/or imaging/exploratory surgery of the abdomen. If the reagent strip is negative, the likelihood that the patient has bacterial peritonitis is low but the authors would recommend re-testing of the reagent strip at a future date and/or follow-up testing if the patient’s condition does not improve.
Conflict of interest statement
2 Indiana Animal Disease and Diagnostic Laboratory, W. Lafayette, IN, USA; accredited by the American Association of Veterinary Laboratory Diagnosticians (AAVLD).
None of the authors of this paper has a financial or personal relationship with other people or organizations that could inappropriately influence or bias the content of the paper.
642
E.J. Thomovsky et al./The Veterinary Journal 202 (2014) 640–642
Table 2 Detailed results of the performance of the leukocyte esterase test pad on the Multistix 8 SG Reagent Strips. Cytological analysis
Positive leukocyte esterase on reagent strip Negative leukocyte esterase on reagent strip Prevalence of bacterial peritonitis Sensitivity of reagent strip Specificity of reagent strip Positive predictive value reagent strip Negative predictive value reagent strip
Culture results
Cytology or culture results
Positive
Negative
Positive
Negative
Either positive
Both negative
9/60 3/60
4/60 44/60
9/60 6/60
4/60 41/60
11/60 6/60
2/60 41/60
12/60 = 20% 75% (CI:42.8–94.5%) 91.7% (CI:80.0–97.7%) 69.2% (CI:38.6–90.9%) 93.6% (CI:82.5–98.7%)
15/60 = 25% 60% (CI:32.3–83.7%) 91.1% (CI:78.8–97.5%) 69.2% (CI:38.6–90.9%) 87.2% (CI:74.3–95.2%)
17/60 = 28% 64.7% (CI:38.3–85.8%) 95.3% (CI:84.2–99.4%) 84.6% (CI:54.6–98.1%) 87.2% (CI:74.3–95.2%)
Positive cytology results indicate the presence of intracellular bacteria diagnosed by a clinical pathologist. Positive culture results indicate bacterial growth under either aerobic or anaerobic conditions. Positive for the combined cytology and culture results mean that bacterial growth was found under aerobic or anaerobic culture and/or intracellular bacteria were seen during cytological examination.
Table 3 Detailed results of the dipstick analysis compared to cytological effusion type. Cytological effusion type
Dipstick results
Septic exudate
3/12 false negatives
Non-septic exudate
3/7 false positives
Neoplastic Pure transudates Modified transudates
1/6 false positives 0/6 false positives 0/25 false positives
Dog 1 2 3 1 2 3 1 – –
References Araujo, A., Lopes, A., Michalczuk, M.T., Stifft, J., Nardelli, E., Escobar, G., Rossi, G., Alvares-da-Silva, M.R., 2008. Is there yet a place for reagent strips in diagnosing spontaneous bacterial peritonitis in cirrhotic patients? An accuracy and costeffectiveness study in Brazil. Hepatology (Baltimore, Md.) 23, 1895–1900. Bonczynski, J.J., Ludwig, L.L., Barton, L.J., Loar, A.S., Peterson, M.E., 2003. Comparison of peritoneal fluid and peripheral blood pH, bicarbonate, glucose, and lactate concentration as a diagnostic tool for septic peritonitis in dogs and cats. Veterinary Surgery 32, 161–166. Butani, R.C., Shaffer, R.T., Szyjkowski, R.D., Weeks, B.E., Speights, L.G., Kadakia, S.C., 2004. Rapid diagnosis of infected ascetic fluid using leukocyte esterase dipstick testing. The American Journal of Gastroenterology 99, 532–537. Campillo, B., Richardet, J.P., Dupeyron, C., 2006. Diagnostic value of two reagent strips (Multistix 8 SG and Combur 2 LN) in cirrhotic patients with spontaneous bacterial peritonitis and symptomatic bacterascites. Gastroenterologie Clinique et Biologique 30, 446–452. Culp, W.T., Zeldis, T.E., Reese, M.S., Drobatz, K.J., 2009. Primary bacterial peritonitis in dogs and cats: 24 cases (1990–2006). Journal of American Veterinary Medical Association 234, 806–913.
Diagnosis
Bacteria cultured
Perforated duodenum and gastric mass Small intestinal mass/perforation Prostatic abscess Neoplasia (mesothelioma) Cranial abdominal mass Duodenal/gastric ulcers (no perforation) GI lymphoma – –
Clostridium perfringens Bacteriodes ureolyticus E. coli – Streptococcus canis – Enterococcus spp. – –
Hosgood, G., Salisbury, S.K., 1988. Generalized peritonitis in dogs: 50 cases (1975–1986). Journal of the American Veterinary Medical Association 193, 1448–1450. Nousbaum, J.B., Cadranel, J.F., Nahon, P., Khac, E.N., Moreau, R., Thevenot, T., Silvain, C., Bureau, C., Nouel, O., Pilette, C., et al., 2007. Diagnostic accuracy of the Multistix 8 SG reagent strip in diagnosis of spontaneous bacterial peritonitis. Hepatology (Baltimore, Md.) 45, 1275–1281. Rerknimitr, R., Lilmmathurotsakul, D., Bhokaisawan, N., Kongkam, P., Treeprasertsuk, S., Kullaveanijaya, P., 2010. A comparison of diagnostic efficacies among different reagent strips and automated cell count in spontaneous bacterial peritonitis. Journal of Gastroenterology and Hepatology 25, 946–950. Ribeiro, T.C., Kondo, M., Amaral, A.C., Parise, E.R., Bragagnolo Junior, M.A., de Souza, A.F.M., 2007. Evaluation of reagent strips for ascetic fluid leukocyte determination: Is it possible for spontaneous bacterial peritonitis rapid diagnosis? The Brazilian Journal of Infectious Diseases 11, 70–74. Tellez-Avila, F.I., Chavez-Tapia, N.C., Franco-Guzman, A.M., Uribe, M., Vargas-Vorackova, F., 2012. Rapid diagnosis of spontaneous bacterial peritonitis using leukocyte esterase reagent strips in emergency department: Uri-quick clini-10SG® vs. Multistix 10SG®. Annals of Hepatology 11, 696–699.
