|Home | About | Journals | Submit | Contact Us | Français|
Gram stains of 87 different clinical samples were prepared by the laboratory's conventional methods (automated or manual) and by a new single-slide-type automated staining instrument, GG&B AGS-1000. Gram stains from either heat- or methanol-fixed slides stained with the new instrument were easy to interpret, and results were essentially the same as those from the methanol-fixed slides prepared as a part of the routine workflow. This instrument is well suited to a rapid-response laboratory where Gram stain requests are commonly received on a stat basis.
Performing Gram stains on both original samples and isolated bacteria is an essential procedure in a clinical microbiology laboratory. Several commercial automated instruments are available to expedite the process. We evaluated the AGS-1000 Automated Gram Stainer (GG&B Co., Wichita Falls, TX) by using various sample types to determine if its performance is reliable enough for use in the hospital-based Core Laboratory, where nonmicrobiologist clinical laboratory scientists perform stat Gram stains on all three shifts. This machine utilizes patented electro-optical technology to evaluate the material on the slide to optimize the staining process. For its FDA submission, the AGS-1000 had only been validated for use with heat-fixed slides. The purpose of this study was to determine the performance of the AGS-1000 with methanol-fixed slides. Previous publications demonstrated that an automated Gram stainer could enhance the consistency of stained slides and save technologist time (2, 5).
Three smears were prepared from each of 77 clinical samples representing bone, fluids, respiratory secretions, tissue, vaginal secretions, and wounds, and another set of three smears was made from 10 different positive blood culture broths. One smear from each set was methanol fixed and stained using the laboratory's standard-method Wescor 7320 Aerospray Gram Slide Stainer Cyto-Centrifuge (Wescor, Inc., Logan, UT) or manual Gram staining, and the other two slides were handled separately for this evaluation. One smear was heat fixed by holding the slide to a Bacti-cinerator for 5 to 10 s. The second slide was methanol fixed by flooding the slide with methanol and allowing it to stand for up to 1 min before draining the excess methanol and allowing the slide to air dry. The slides were then labeled with a code and randomized by a person other than those performing the staining and reading so that the latter individuals were blinded to the fixation method.
One methanol- and one heat-fixed smear per specimen were stained using the GG&B Company's AGS-1000 Automated Gram Stainer. Once the operator places the slide vertically into the cuvette and selects the desired protocol, the entire procedure is automated. Protocols included short and long cycles, entered by the operator based on the thickness of the material on the slide. The timing of each reagent can be adjusted if necessary. Stains must be purchased from the company. For each protocol, reagents fill the cuvette in order (crystal violet, iodine, decolorizer, and counterstain, with appropriate washing steps). The front and back surfaces of the cuvette fit closely to the slide so that capillary action helps to wet the slide surface and the process utilizes very low volumes of stain. An optical system sights down the surface of the slide to control the timing for each reagent. When the staining process is complete, the operator removes the slide and blots it dry on paper towels. All reagents are collected into a plastic bottle that is emptied into hazardous waste every few days; the instrument does not require connection to a drainage outlet. A larger bottle could be used to extend the time before emptying is necessary.
All slides were examined with an Olympus compound light microscope at ×10 and then at ×100 under oil immersion. The individual examining the slides did not know which fixation method had been used. The examiner recorded the presence of any visible bacteria, fungal elements, and other cells, including polymorphonuclear leukocytes, red blood cells, and squamous epithelial cells. Slide color was also rated as too light/overdecolorized, too dark/underdecolorized, or just right. The presence of any precipitated stain, excessive debris, or other cells was also noted. A minimum of three separate staining events by both the AGS-1000 and manual methods were timed with a stopwatch.
After both slides for a sample had been evaluated, the results were compared with the electronic record of the original smear. The smear referenced in the database had been methanol fixed and stained in the Wescor Gram Slide Stainer or manually by the laboratory technicians and interpreted by the Clinical Laboratory Scientists in Microbiology as part of the routine culture workup. Statistical analysis was not performed due to the low numbers of comparators.
Of the 87 samples stained, 40 were positive for the presence of microorganisms by the reference method. The numbers of samples showing organisms compared with the total number evaluated were 9 of 9 blood culture samples, 2 of 2 bone biopsy specimens, 3 of 30 fluid samples, 14 of 17 respiratory secretion samples, 5 of 19 tissue samples, 5 of 6 vaginal secretion samples, and 2 of 4 wound samples. Three slides stained on the instrument had discrepancies with respect to the reference results. Two samples yielded rare Gram-positive rods, and one showed rare Gram-variable rods by the reference method that were not observed in the slides stained by the AGS-1000. Discrepant slides were evaluated by an additional expert. In all other cases, the organisms or lack of organisms observed on both the methanol- and heat-fixed slides stained in the AGS-1000 corresponded to the reference slide results.
The color and quality of the stain were good for 97.5% of the positive samples fixed by either heat or methanol. One fluid sample (2.5%) appeared overdecolorized on both the methanol- and heat-fixed slides stained in the instrument.
Precipitate of crystal violet in the form of either circles of irregular size or a crystallized pattern was seen on approximately 20% of the slides with both fixation methods. Crystals appeared often on both slides from the same sample. All of the bone and blood samples were positive for organisms, whereas only 10% of the fluid samples had organisms. The appropriateness of stain color also varied by specimen source with all of the positive respiratory, bone, blood, vaginal, tissue, and wound samples demonstrating the correct stain color for the cells present.
Time-motion studies showed the average time required to stain a slide using the AGS-1000 was 4 min 5 s (mostly unattended), whereas staining by hand took an average of 3 min 15 s. True to the product information, we found that one bottle of stain was sufficient to stain 100 slides.
The Gram stain interpretation of patient specimens is a common laboratory procedure with important implications for determining patient treatment (4). In this study, we evaluated the ability of the AGS-1000 to stain methanol-fixed slides (compared to heat-fixed slides, as originally validated by the manufacturer) and produce results equivalent to those of our laboratory standard method. Our data indicate that the AGS-1000 is efficient and produces consistently well-stained slides; it is noteworthy that this machine is ideal in a clinical setting where time is crucial for the delivery of results. We acknowledge that the report is based on a relatively small number of specimens tested in a single laboratory, with only 40 positive smears by conventional Gram staining. Additional studies in other laboratories are merited to confirm the results of this study.
In comparison with a previous study (5), the AGS-1000 stains slides 43 s faster than the automated single-slide stainer. The previous device operated by staining in batches, thus increasing productivity, but this also increased the chance of contamination either through the staining solutions themselves or via transfer from previously processed slides. The AGS-1000 stains one slide at a time, preventing contamination, but this limits the operator to staining only one slide at a time. With the push of a button, the operator can clean the cuvette of the AGS-1000 stainer, decreasing the maintenance efforts.
Previously, the AGS-1000 had only been validated for use with heat-fixed slides. A previous publication, however, referring to manually stained slides (no instrument involved), indicated that methanol-fixed slides produce superior results (3). Heat fixation causes red blood cells to rupture, creating debris which can obscure rare organisms and make slides difficult to interpret (1). With the exception of three slides showing only rare organisms, all of the organisms on the slides evaluated in this study were visualized clearly, regardless of the fixation method used. Therefore, we conclude that the AGS-1000 is effective for staining methanol-fixed slides. Our data suggest that methanol fixation in the AGS-1000 may produce a stain color lighter than that produced by heat fixation. This may be due to the presence of less protein debris from ruptured cells in the background of methanol-fixed slides than on heat-fixed slides, or the methanol-fixed material may affect the electronic sensor.
Overall, the AGS-1000 produced exceptional stain quality that was clear and easy to read. Overdecolorized organisms were seen in both slides prepared from a single fluid sample, but the number of smear-positive fluid samples was too low to detect any trends. Another problem observed during this study was the occasional presence of precipitated crystal violet on the slides. Precipitated stain, however, did not interfere with slide interpretation, as the precipitated stain often had irregular shapes and sizes, making it distinguishable from Gram-positive cocci, which have a uniform morphology. During our evaluation, we found that increasing the number of cuvette-cleaning cycles to once every 10 slides instead of the instrument-prompted schedule of once every 20 slides and also running a cleaning cycle before each shift decreased the precipitate seen on the slides.
While manual staining is faster overall than staining with the AGS-1000, automated staining is advantageous in that it produces uniform results, staining each slide for precisely the same amount of time, and requires less actual hands-on time. In a busy clinical setting, it is beneficial to be able to simply place the slide in the provided cuvette, select the desired protocol, and return 4 min later to retrieve the stained slide. Additionally, automated staining conserves reagents and reduces the lab technicians' contact with both the specimen and the reagents, reducing the opportunity for contamination.
We conclude that staining methanol-fixed slides in the AGS-1000 produces slides that are easy to read and usually correctly colored. Slides prepared from fluids should be observed carefully, as some organisms may be overdecolorized.
We are grateful to Tracy Chang and the Laboratory Assistants in Microbiology for identifying test samples and preparing extra slides for our project.
Published ahead of print on 21 April 2010.