0
Communications to the Editor |

Urinary Antigen Test for Pneumococcal Pneumonia Urinary Antigen Test for Pneumococcal Pneumonia FREE TO VIEW

José Domínguez, PhD; Núria Galí, MSc; Silvia Blanco, BSc; Pablo Pedroso, BSc; Cristina Prat, MD; Lurdes Matas, PhD; Vicente Ausina, PhD
Author and Funding Information

Affiliations: Hospital Universitari Germans Trias i Pujol Barcelona, Spain,  Mailman School of Public Health New York, NY

Correspondence to: José Domínguez, PhD, Servei de Microbiologia, Hospital Universitari Germans Trias i Pujol, Carretera del Canyet s/n, 08916 Badalona, Barcelona, Spain; e-mail: jadoming@ns.hugtip.scs.es



Chest. 2001;120(5):1748-1750. doi:10.1378/chest.120.5.1748
Text Size: A A A
Published online

To the Editor:

We read with interest the editorial by Dr. Pesola1 (January 2001) and his comments on our study,2 which was recently featured in your publication. With a view to enabling better interpretation of the results of our study, we wish to clarify some points that we believe Dr. Pesola has misunderstood.

In our study, we evaluated an immunochromatographic membrane test (ICT) to detect Streptococcus pneumoniae in urine samples, in order to assess its utility in the diagnosis of pneumococcal pneumonia. The sensitivity of the test using urine samples from patients with a diagnosis of definite pneumococcal pneumonia (group 1) was 80.4% (41 of 51). In contrast, the sensitivity of the test using urine samples from patients with a diagnosis of presumptive pneumococcal pneumonia (group 2) was 43.7% (7 of 16). The specificity was 97.2%.

In order to make the performance of the test clearer, we explained in the “Results” section that results could be differentiated in three groups according to the color intensity reached: weak, medium, and very intense. Since the two false-positive results gave weak-intensity colored lines, which were difficult to interpret, we proposed to consider these results as negative, thus bringing the specificity to 100%, and maintaining the sensitivity at 74.5% in definite pneumococcal pneumonia.

Dr. Pesola considers that all weak results are difficult to interpret and, therefore, claims that all these results are negative. Citing the five weak line results in group 2, he concludes that the real sensitivity of the test using group 2 was 12.5%.

We don’t agree that all weak results have to be considered negative or that weak results of weak color lines are difficult to interpret. Less colored lines are included in the weak-line group, but only lesser colored lines are very difficult to interpret and can therefore be considered negative. In our study, we obtained a very weak line result from only three patients in group 2. Taking these results as negative, the sensitivity of the test in group 2 would be 25%.

Patients from group 2 were diagnosed by results from Gram’s stain, sputum culture, and/or pneumococcal antigen detection by latex in sputum. The low sensitivity and specificity of these techniques3 is well-known. Problems arise from the difficulty in interpretation due to contamination by the flora of the upper airways, given the possibility that they might contain S pneumoniae.

In our opinion, calculating the sensitivity of the ICT with the results from both group 1 and group 2 patients leads to an underestimation. We believe that it is more probable that the real sensitivity of the test is closer to the sensitivity raised in group 1 than in group 2.

However, we do agree with Dr. Pesola that this test represents a significant step forward in the definitive diagnosis of pneumococcal pneumonia. Yet, additional studies are needed in order to clarify different issues related to the performance of the test, and to assess its utility in the better management of patients.

In our opinion, the ICT test is a rapid and reliable test for detecting S pneumoniae antigen in urine. ICT is a very useful tool for assisting physicians in the diagnosis and focusing of the treatment of patients with pneumococcal pneumonia.

Pesola, GR (2001) The urinary antigen test for the diagnosis of pneumococcal pneumonia.Chest119,9-11. [PubMed] [CrossRef]
 
Domínguez, J, Galí, N, Blanco, S, et al Detection ofStreptococcus pneumoniaeantigen by a rapid immunochromatographic assay in urine samples.Chest2001;119,243-249. [PubMed]
 
Bartlett, JG, Breiman, RF, Mandell, LA, et al Community-acquired pneumonia in adult: guidelines for management.Clin Infect Dis1998;26,811-838. [PubMed]
 

Urinary Antigen Test for Pneumococcal Pneumonia

To the Editor:

Domínquez et al1point out in their excellent article (January 2001) on the urinary antigen test for the detection of pneumococcal pneumonia, that the range of sensitivities to be considered for their test should vary from 44% (7 of 16, Table 2) to 80% (41 of 51, Table 1) at best or during a research environment (efficacy). This was reduced down to 25% (4 of 16, Table 2) to 74.5% (38 of 51) if it is assumed that some of the weak intensity urinary immunochromatographic test (ICT) results are too difficult to determine if positive and should be thrown out so that interpretation would be unequivocal in the community environment (effectiveness). In the editorial,2 I suggested the same upper bound with a lower bound ranging from 12.5% (2 of 16, Table 2) to 74.5% based on excluding all weak intensity urinary ICT test results. As noted in their article, Domínguez et al1 suggest that there are two types of weak intensity lines on the ICT test compared to control. They suggest in their letter that the two weak intensity color lines compared to control are the “less colored lines” and “lesser colored lines.” The latter are more difficult to interpret and can be thrown out to make it easier for users to interpret the results. Therefore, from Table 2 of their article, only three of the weak intensity urinary ICT findings need to be thrown out and not five. This results in a slightly higher range for the lower bound of sensitivities, although it is unclear from Table 2 which weak tests to throw out.

The idea to throw out weak intensity lines on the urinary ICT also came up regarding the calculation of specificity results for group 3 of their data. The initial specificity of 97.2% (69 of 71) was raised to 100% when the two false-positive urinary ICT results were discarded due to a weak intensity color on the sample line that was difficult to interpret. Presumably the latter two were of the lesser colored lines and not less colored lines.

The problem with denoting two different weak intensity lines for the urinary ICT and throwing out one set is there is no proof in their article that doing so is valid. One would have to mix up the two types and instruct blinded judges to separate one type from the other. If the reproducibility and accuracy were there from this type of study, I might buy into the idea. Otherwise, it could be argued that the separation was one of convenience to maximize the end result (both specificity and sensitivity) of the study, and as long as no tests were done to challenge this, everyone would blindly go along. Most likely, the authors are correct and will come out with data to clearly show the difference to the reading audience on the differences between the two weak intensity lines. Until then, it is arbitrary, and I will stay with the current range of sensitivities.2

It should be noted that the strength of the findings of Domínguez et al1 have nothing to do with the possible less-than-ideal sensitivity of their data. It is the purported specificity of almost 100%. Although the urinary ICT result can be positive secondary to Streptococcus oralis and Streptococcus mitis, these organisms do not cause pneumonia.,3 This translates clinically into suggesting that if the test result is positive in someone with pneumonia, the disease is present. The real question is,“ Are their really no false-positive findings such that a diagnosis can be made every time the test result is positive in the presence of pneumonia?” Only time and more data will answer this question.

Domínguez et al1 also note that their group 2 patients with a presumptive diagnosis of pneumococcal pneumonia seen in Table 2 might not all be due to Streptococcus pneumoniae, leading to the possibility of overreliance on this data to determine sensitivities that would keep them lower. This may be true as the phrase presumptive suggests but the diagnostic criteria for the S pneumoniae diagnosis are not too shabby. All diagnoses in Table 2 were supported by either a positive sputum culture finding and/or latex agglutination finding for detecting S pneumoniae in the sputum and chest radiography findings consistent with pneumonia. Of the sputum culture results that were positive (13 of 16), all except one were supported by at least one other positive test result consistent with S pneumoniae (excluding radiography). Two indicators of the same diagnosis significantly increase the likelihood of being correct. Of the latex agglutination test results positive for S pneumoniae in the sputum (12 of 14), only two were not supported by at least one other test result (excluding radiography) consistent with S pneumoniae. The false-positive rate for latex agglutination detection of S pneumoniae in sputum has been described as about 6%,,4 making it possible but unlikely that some of the positive results from this test are not true, particularly if other tests are consistent with the same diagnosis. In addition, the sensitivity for latex agglutination detection of S pneumoniae in the sputum of patients with S pneumoniae has been described as 86%,,4 which is identical to the sensitivity seen in Table 2 of this study (12 of 14). Therefore, the diagnosis of S pneumoniae in the group 2 patients in the study is fairly solid and a tribute to the authors in being wise enough to attempt to document this pneumonia by a variety of methods.

Finally, we all agree that this study is a significant step forward in the attempt to make a diagnosis of S pneumoniae in a timely matter. I congratulate the authors on a very interesting study.

References
Domínguez, J, Galí, N, Blanco, S, et al Detection ofStreptococcus pneumoniaeantigen by a rapid immunochromatographic assay in urine samples.Chest2001;119,243-249. [PubMed] [CrossRef]
 
Pesola, GR The urinary antigen test for the diagnosis of pneumococcal pneumonia.Chest2001;119,9-11. [PubMed]
 
Bisno, AL, Rijn, IVD Classification of streptococci. Mandell, GL Bennett, JE Dolin, R eds.Principles and practice of infectious disease2000,2100-2101 Churchill Livingstone. Philadelphia, PA:
 
Holmberg, H, Krook, A Comparison of enzyme-linked immunosorbent assay with coagglutination and latex agglutination for rapid diagnosis of pneumococcal pneumonia by detecting antigen in sputa.Eur J Clin Microbiol1986;5,282-286. [PubMed]
 

Figures

Tables

References

Pesola, GR (2001) The urinary antigen test for the diagnosis of pneumococcal pneumonia.Chest119,9-11. [PubMed] [CrossRef]
 
Domínguez, J, Galí, N, Blanco, S, et al Detection ofStreptococcus pneumoniaeantigen by a rapid immunochromatographic assay in urine samples.Chest2001;119,243-249. [PubMed]
 
Bartlett, JG, Breiman, RF, Mandell, LA, et al Community-acquired pneumonia in adult: guidelines for management.Clin Infect Dis1998;26,811-838. [PubMed]
 
Domínguez, J, Galí, N, Blanco, S, et al Detection ofStreptococcus pneumoniaeantigen by a rapid immunochromatographic assay in urine samples.Chest2001;119,243-249. [PubMed] [CrossRef]
 
Pesola, GR The urinary antigen test for the diagnosis of pneumococcal pneumonia.Chest2001;119,9-11. [PubMed]
 
Bisno, AL, Rijn, IVD Classification of streptococci. Mandell, GL Bennett, JE Dolin, R eds.Principles and practice of infectious disease2000,2100-2101 Churchill Livingstone. Philadelphia, PA:
 
Holmberg, H, Krook, A Comparison of enzyme-linked immunosorbent assay with coagglutination and latex agglutination for rapid diagnosis of pneumococcal pneumonia by detecting antigen in sputa.Eur J Clin Microbiol1986;5,282-286. [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543