Article Text

Download PDFPDF

Original article
Sensitivity of Gram stain in the diagnosis of urethritis in men
  1. Mª Angeles Orellana1,
  2. Mª Luisa Gómez-Lus2,
  3. David Lora3
  1. 1Laboratorio C.E.P, Pontones, Madrid, Spain
  2. 2Departamento de Microbiología Clínica I, Universidad Complutense de Madrid, Madrid, Spain
  3. 3Clinical Research Unit, IMAS12-CIBERESP, Hospital 12 de Octubre, Madrid, Spain
  1. Correspondence to Dr Mª Angeles Orellana Miguel, C/Camino Viejo de Leganés n 36, Madrid 28019, Spain; m.a.orellana{at}jv.e.telefonica.net

Abstract

Background Acute urethritis is among the most common types of sexually transmitted diseases in men. The diagnosis usually requires microscopic evidence of urethritis, but sometimes urethral pathogens are detected in asymptomatic men without such evidence. The aims of this study were to assess the sensitivity of Gram stain in men with urethral symptoms and to relate it to the microorganisms isolated.

Methods Between January 2006 and December 2007, 491 urethral samples were analysed. The authors assessed the presence of leukocytes by Gram stain and tested specifically for Chlamydia trachomatis, Ureaplasma urealyticum, Mycoplasma hominis and Trichomonas vaginalis, as well as analysing the results of conventional culture.

Results The percentages of positive samples as a function of Gram category were two or less polymorphonuclear leukocytes (PMNLs)/high-power field (HPF) 25% (92/364), three to four PMNLs/HPF 32% (18/57) and five or more PMNLs/HPF 54% (38/70). Classing samples with more than two PMNLs/HPF as positive, the sensitivity, specificity and positive likelihood ratio for Gram stain were 38% (95% CI 30 to 46), 79% (95% CI 75 to 84) and 1.8 (95% CI 1.4 to 2.4), respectively. On the other hand, taking as positive five or more PMNLs/HPF, the sensitivity, specificity and positive likelihood ratio for Gram stain were 26% (95% CI 18 to 33), 91% (95% CI 87 to 94) and 2.7 (95% CI 1.8 to 4.2), respectively. The sensitivity of Gram stain to Neisseria gonorrhoeae, Chlamydia trachomatis and Ureaplasma urealyticum were 80% (95% CI 64 to 96), 23% (95% CI 8 to 39) and 11% (95% CI 2 to 20), respectively.

Conclusion The low sensitivity of Gram stain means that negative results do not exclude the presence of urethritis in symptomatic patients.

  • Sexually transmitted infection
  • urethritis
  • sensitivity
  • gram stain
  • bacteriology
  • chlamydia infection
  • bacteriuria
  • antimicrobial resistance
  • diagnosis
  • statistics
  • epidemiology (clinical)
  • biostatistics

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Urethritis is a clinical syndrome characterised by the appearance of mucopurulent or purulent urethral discharge and sometimes dysuria and itching at the urethral meatus. It is the urethral response to inflammation of any aetiology, and notably asymptomatic infections are also common.1–3

Accordingly, urethritis should be suspected when any of the following are present: purulent or mucopurulent discharge from the urethral meatus, positive urine leucocyte esterase test in the absence of signs of urinary infection, >10 polymorphonuclear leucocytes (PMNLs)/high-power field (HPF) (×1000 magnification) in urinary sediment in the absence of signs of urinary infection or five or more cells (PMNLs) per HPF with Gram stain in the direct microscopic examination of urethral discharge.1

Smith et al4 found that there was considerable variation between and within microscopists in the diagnosis of non-gonococcal urethritis (NGU), and this sensitivity was strongly correlated with the grade of urethritis, with a significant proportion of cases of low-grade urethritis falsely diagnosed as negative. In addition, Bradshaw et al5 found less than five PMNLs in 32%, 37%, 38% and 44% of cases of Chlamydia trachomatis, Mycoplasma genitalium, Adenovirus and HSV, respectively, concluding that the threshold of five or more PMNL was not sufficiently sensitive to exclude urethral infection in NGU. Similar conclusions were reached by Janier et al6 who recommend that it should be explored whether Chlamydia trachomatis is present in patients with urethral symptoms, with or without the classic signs and symptoms of urethritis (discharge and presence of PMNLs in the urethra or first-catch urine).

On the other hand, microorganisms considered to be an uncommon cause of urethritis or unlikely to be the causal agent of urethritis, such as Haemophilus sp., Streptococcus sp., Staphylococcus sp. and Candida sp., are sometimes isolated in patients with symptoms. Based on the literature,7–16 it is believed that this could be related to sexual practices (men who have sex with men, oral sex, etc).

The aim of this study was to assess the sensitivity of Gram stain, compared with microorganisms isolated in urethral exudates, and to assess the role of isolated microorganisms considered unlikely to be causal agents of urethritis.

Materials and methods

This cross-sectional prospective study was carried out in a Primary Care Laboratory, which provides diagnostic services to outpatients from the centre of the city of Madrid (Spain).

During the period 2006–2007, a series of consecutive samples of urethral exudates of 773 male patients, over 15 years old, with urethral symptoms was studied. Of these, 282 were excluded from further analysis because the samples were not collected in the Laboratory. In the 491 remaining patients, samples were collected by inserting Dacron swabs 2 to 3 cm into the urethra. Four swabs were taken from each patient, for the following tests:

  1. Gram stain. This was performed for all patients (with or without discharge). The microscopist was always the same person and according to the PMNLs observed per HPF, ×1000 magnification, cases were divided into three categories: G0: two or less PMNLs/HPF; G1: three to four PMNLs/HPF and G2: five or more PMNLs/HPF.

  2. Chlamydia trachomatis detection. Between January 2006 and May 2007, this was performed by immunochromatography using the CHLAMY-CHECK-1 test (GRIFOLS, Aleçon, France). Subsequently, between June and December 2007, it was carried out with nucleic acid amplification by PCR using the COBAS AMPLICOR Analyzer® (Roche Diagnostics, Indianapolis, IN, USA). Both tests were carried out according to the manufacturer's recommendations.

  3. Culture in blood agar plates with 5% sheep erythrocytes, chocolate agar + PolyViteX, chocolate + PolyViteX VCAT3 agar and SGC2 Sabouraud gentamicin chloramphenicol agar (BioMérieux, Lyon, France). Blood agar plates were incubated at 37°C aerobically for 24–48 h. Chocolate VCAT3 agar plates were incubated at 37°C in 5% CO2 for 48 h, while the Sabouraud agar plates were incubated for 48–72 h at room temperature. Neisseria gonorrhoeae and Haemophilus spp. were identified using an API NH system (BioMérieux) and other organisms using a MicroScan® Walkaway system (SIEMENS, Germany). Isolates were considered positive when there was growth of microorganisms known to cause urethritis and/or a pure culture of a microorganism which can be sexual transmitted.

  4. Ureaplasma urealyticum and Mycoplasma hominis were detected using a Mycoplasma IST2 system (BioMérieux) according to the manufacturer's recommendations. The results were considered positive if a count ≥104 CCU/ml (colour changing units) was obtained.

  5. In patients with a partner who had a history of Trichomonas vaginalis infection, direct visualisation was performed with saline solution immediately after sampling. The procedures for the isolation of microorganisms referred to in items 2, 3, 4 and 5 were considered the reference methods.

All tests were performed the same day as collection of the samples, except the PCR to detect Chlamydia trachomatis that was carried out within a week.

Statistical analysis was performed using the SAS System software (V.9.1; IMC Institute). Qualitative data were presented as frequency distributions and percentages. Values of sensitivity, specificity and positive likelihood ratio (LR (+)) were obtained, using 95% CIs for quantifying uncertainty.

Results

The percentage of samples with significant microbiological growth was 30% (148/491). The prevalence of microorganisms isolated is reported in table 1. In 11 samples (2% of the total), two or more organisms/sample were isolated, and the combinations are reported in table 1. The percentages of samples in which at least one pathogen was isolated and the corresponding microorganisms are also reported in table 1 as a function of Gram stain category. Overall, 32 samples were tested for Trichomonas vaginalis, but none of them were positive.

Table 1

Microorganisms isolated as a function of the number of PMNLs/HPF in Gram stain

The results obtained from comparing Gram stain and the reference methods are reported in table 2. In the case of positive cases being defined as those with a Gram stain with >2 PMNLs/HPF and urethritis, that is, isolation of a microorganism that can cause urethritis or a pure culture of a microorganism which can be sexually transmitted, the sensitivity, specificity and positive LR (+) of this stain were 56/148, 38% (95% CI 30 to 46); 272/343, 79% (95% CI 75 to 84) and 1.8 (95% CI 1.4 to 2.4), while considering positive those cases with ≥5 PMNLs/HPF, the figures were 38/148, 26% (95% CI 18 to 33); 311/343, 91% (95% CI 87 to 94) and 2.7 (95% CI 1.8 to 4.2), respectively.

Table 2

Sensitivity, specificity and LR (+) of Gram stain to urethritis, Neisseria gonorrhoeae, Chlamydia trachomatis and Ureaplasma urealyticum

Again considering positive those cases with ≥5 PMNLs/HPF, the sensitivities of the Gram stain for the three most commonly isolated microorganisms were 24/30, 80% (95% CI 64 to 96) for Neisseria gonorrhoeae, 8/34, 23% (95% CI 8 to 39) for Chlamydia trachomatis and 6/55, 11% (95% CI 2 to 20) for Ureaplasma urealyticum.

Discussion

The sensitivity of Gram stain was very low, both for urethritis defined as the presence of >2 PMNLs/HPF (38%) and as ≥5 PMNLs/HPF (26%). On the other hand, the specificity was as high as 91% when defining urethritis as the presence of ≥5 PMNLs/HPF, decreasing to 79% with a threshold of >2 PMNLs/HPF.

We note that a strength of our study is that the large number of samples tested were all collected in the laboratory and processed immediately. Further, the microscopic examination was performed by the same person, an experienced microscopist, to avoid interobserver variability that has been reported by some authors.4

A causative agent of urethritis was isolated in 30% of all the samples analysed. This figure was 25% of the samples without leucocytes in the Gram stain, increased to 32% when the number of observed leucocytes was three to four PMNLs/HPF and further to 54% for five or more PMNLs/HPF. On the other hand, in 46% of the samples with ≥5 PMNL/HPF, no pathogens known to cause urethritis were detected. This could be due to the low sensitivity of the Chlamydia trachomatis diagnostic technique used until June 2007 or to the presence of other microorganisms, such as Mycoplasma genitalium or Adenovirus, that would not be detected by the routine series of tests used. We note that the series of tests used was restricted to those which were available in our laboratory, and this is one of the limitations of the study.

As for particular microorganisms, we did not detect Trichomonas vaginalis in any samples. In contrast, some authors indicate that this parasite is a major cause of NGU and detected it in >70% of men whose partners were infected, using multiple clinical samples (semen, urethral exudate and urine), culture and PCR,17 18 but, similar to our findings, Maeda et al19 did not find positive isolation by culture methods. As seen in table 1, 10% (3/30) of Neisseria gonorrhoeae, 59% (20/34) of Chlamydia trachomatis and 84% (46/55) of the cases in which Ureaplasma urealyticum was isolated, no leucocytes were detected in the Gram stain. These results are consistent with those of Geisler et al20 who found no evidence of urethral inflammation from Gram stain in 12% of Chlamydia trachomatis and 5% of Neisseria gonorrhoeae infections. On the other hand, Neisseria gonorrhoeae was the microorganism most often associated with the presence of leucocytes in the Gram stain.

Defining urethritis as the presence of ≥5 PMNLs/HPF in the Gram stain, the microorganisms considered to be uncommon causal agents of urethritis, such as Haemophilus spp., Candida spp., Enterobacteriaceae, Streptococcus agalactiae, Mycoplasma hominis and Shaphylococcus aureus, should indeed not be regarded as causative agents of urethritis. It is, however, also true that if only the cases with ≥5 PMNLs/HPF in Gram stain are considered as an indicator of urethritis, 76% (26/34) of the cases of urethritis caused by Chlamydia trachomatis, 20% (6/30) of those due to Neisseria gonorrhoeae and 89% (49/55) of Ureaplasma urealyticum would not have been diagnosed. Accordingly, our results indicate that the absence of leucocytes in the Gram stain does not exclude the diagnosis of urethritis.

In Hong Kong, Yu et al,21 in a study of prevalence of urethral infection in 274 asymptomatic male patients, found 36 patients with NGU and two positive gonococcal cultures. Among the cases of NGU, 17%, 23% and 14% were positive PCR for Chlamydia trachomatis, Mycoplasma genitalium and Ureaplasma urealyticum, respectively, and, further, 14 of the patients positive for Chlamydia trachomatis by PCR had no signs or symptoms of NGU, demonstrating that urethral infection can be present in a significant number of asymptomatic patients.

The sensitivity of Gram stain for Chlamydia trachomatis (23%) we found is similar to the figure of 29% obtained by Janier et al6 but lower than that of Iwuji et al22 who reported a sensitivity of 73% and specificity of 71% considering the presence of ≥5 PMNLs/HPF. This difference in results for Chlamydia trachomatis could be due to the population under study; in the case of Iwuji et al, all patients were HIV positive and 93% men who have sex with men.

On the other hand, we found a considerably lower sensitivity for Ureaplasma urealyticum than that obtained by Janier et al6 (11% vs 33%). This low sensitivity of Gram stain may be due to the high rate of colonisation in asymptomatic sexually active individuals (5%–20%)23 and/or failure to discriminate between the different subtypes of Ureaplasma parvum. The latter possibility is related to the methodology we used, that is, we were not able to distinguish between Ureaplasma urealyticum, which is known to cause urethritis, and Ureaplasma parvum, which might be responsible for colonisation only.24

We conclude that the relatively low sensitivity of Gram stain means that bacteriological analysis of urethritis should be performed in all patients with symptoms, even when leucocytes are not observed. This statement is also applicable to asymptomatic patients, when the study of urethritis is indicated, by previous studies demonstrating the presence of urethritis-producing microorganisms in the patient.21 It is, however, necessary to conduct further research to clarify the significance of the isolation of microorganisms whose role in the production of urethritis remains unclear.

Key messages

  • The sensitivity of Gram stain was very low.

  • The absence of leucocytes in the Gram stain does not exclude the diagnosis of urethritis.

  • The bacteriological analysis of urethritis should be performed in patients even when leucocytes are not observed.

  • We did not observe the presence of leucocytes in the Gram stain when microorganisms considered unlikely to be causal agents of urethritis were isolated.

References

Footnotes

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.