Centrifugal Ultrafiltration Method for Rapid Concentration of Legionella pneumophila Urinary Antigen

Detection of Legionella pneumophila antigen in urine by enzyme immunoassay (EIA) or by immunochromatographic testing (ICT) has proven rapid, sensitive, and specific for diagnosing legionellosis (2, 3, 5, 6). Although it requires added processing time, the use of concentrated urine (CU) obtained by selective ultrafiltration significantly improves the sensitivity (1) of antigen detection. We assessed the utility of rapid antigen concentration by centrifugal ultrafiltration (Amicon Ultra-4; Millipore Corporation, Bedford, Mass.) and compared it to passive ultrafiltration (Urifil-10; Millipore Corporation).

We tested 35 patients with confirmed L. pneumophila infection. Diagnosis was achieved by detection of antigen by the EIA method (Binax, Portland, Maine) using 25-fold CU obtained by passive ultrafiltration. In addition, 35 patients with confirmed pneumonia of other etiologies (15 Streptococcus pneumoniae infections, 15 Mycoplasma pneumoniae infections, and 5 Chlamydia pneumoniae infections) and 15 patients with urinary tract infections (10 Escherichia coli infections, 1 Pseudomonas aeruginosa infection, 2 Proteus mirabilis infections, and 1 Klebsiella pneumoniae infection) were examined. These patients were all negative when tested as outlined above.

Prior to the concentration step, all urine samples were initially boiled for 5 min and then centrifuged at 1,000 × g for 15 min to reduce nonspecific reactions (4). Using the passive ultrafiltration method, 4 ml of urine was introduced into a device equipped with an ultrafiltration membrane. This membrane is permanently sealed into the housing, and the back membrane is in contact with an absorbent pad. Water and small molecules penetrate the membrane as a consequence of the absorbent pad, while the Legionella antigen cannot cross the membrane and is progressively concentrated. The centrifugal device combines a low-binding ultrafiltration membrane with a vertical housing. Water and small molecules penetrate the membrane as a consequence of centrifugation, while Legionella antigen cannot cross the membrane and is concentrated. Four ml of each nonconcentrated urine sample was centrifuged to 3,000 × g (MSE Mistral 2000; Sanyo Gallenkap PLC, Middlesex, United Kingdom) for 15 min. When necessary, additional centrifugation processes of 5 min each were used until the 25-fold concentration was obtained.

Initially, all urine samples were concentrated by passive ultrafiltration, and the Legionella antigen was detected by EIA. Then, the urine samples were frozen at −80°C for 3 to 12 months. Afterwards, the urine samples were thawed and concentrated by centrifugal ultrafiltration, and the antigen was detected by ICT. The ICT results for CU obtained by centrifugal ultrafiltration were compared with the previous result obtained by the EIA for CU obtained by passive ultrafiltration. The EIA and ICT were previously compared, and highly comparable results were obtained (2).

The results obtained by using CU processed by centrifugal ultrafiltration were identical to those obtained by using CU processed by passive ultrafiltration for the 35 urine samples with a Legionella diagnosis. The ratios obtained by the EIA method ranged between 4.49 and 41.9 (mean, 16.53; SD, 9.45). Centrifugal concentration did not represent a decrease in the sensitivity of the antigen detection, including samples with a low antigenic content. All CU samples obtained by centrifugal ultrafiltration from the non-Legionella patient specimens were negative by ICT.

The overall agreement between both concentration methods was 100%. The time needed to concentrate urine by passive ultrafiltration was between 1 and 3 h, depending on the physical characteristics of each specimen. Some specimens took over 6 h to reach a concentration of 25-fold. In contrast, the time required in the centrifugal ultrafiltration was only 15 min. A few samples did require an additional 5-min centrifugation step to obtain the desired concentration. In conclusion, centrifugal ultrafiltration is a valid methodology for carrying out L. pneumophila antigen concentration without decreasing the ICT test sensitivity or specificity.