Statistics from Altmetric.com
The recent paper by Rotimi and colleagues1 does not mention the acridine orange stain2 when comparing staining methods for the identification of Helicobacter pylori. The acridine orange stain uses ultraviolet fluorescence in the identification of bacteria. The typical curved morphology of H pylori can easily be differentiated from other bacteria.3–5 I have used this quick, cheap, and reliable stain in routine histopathology reporting for over 16 years and it has proved to be extremely useful in the identification of H pylori.
Immunohistochemistry is now recognised as the “gold standard” because it is a highly sensitive and specific staining method.1 After the publication of the above mentioned article, 20 consecutive gastric biopsies that were positive for H pylori using the acridine orange stain were also stained using the polyclonal anti-H pylori antibody (Dako, Ely, Cambridgeshire, UK) at a dilution of 1/100. Twenty negative control cases were similarly studied. All 20 cases that were positive with the acridine orange stain were also positive by immunohistochemistry and all negative cases were also negative by immunohistochemistry.
This small study clearly shows that ultraviolet fluorescence of H pylori using the acridine orange stain is highly sensitive and compares equally with the gold standard of immunohistochemistry. The acridine orange stain may not be specific, but the morphology of H pylori is clearly visible down to the single organism (fig 1).
The only disadvantage of the acridine orange stain is that the microscopic needs a fluorescent attachment, which in my laboratory means turning the lever on a Leitz Diaplan microscope to the required position, without the need for a dark room.
The authors reply
In pointing out that we omitted to include acridine orange in our comparison of histological stains for Helicobacter pylori, Dr Haqqani seems to have misunderstood the aim of our study. We sought to compare two recently described staining techniques for which there had been claims of superiority over routine methods with our own previously validated routine stain, the modified Giemsa, and with immunohistology using an anti-H pylori antibody—the acknowledged histological “gold standard”. There was no attempt to be comprehensive and test every possible stain for H pylori. Thus, we also omitted from our study the variants of the Gram stain used by some laboratories—the Brown-Hopps,1 Brown-Brenn,2 and the half Gram3—the simple and inexpensive cresyl fast violet4 and carbol fuchsin5 stains, and the more elaborate Gimenez stain.6 The silver impregnation Warthin Starry stain is also used, but is somewhat inconsistent in our hands. Similarly, we have had problems reproducing the silver based Genta “triple” stain.7 We have no experience of the more recently described carbol fuchsin–alcian blue–haematoxylin and eosin8 and the alcian blue–toluidine blue9 methods, which could have been added to the panel. It is evident that stains for H pylori have become a cottage industry in which laboratories strive to produce some novel tinctorial mélange, with (in many cases) little thought for sensitivity, specificity, ease of use, reproducibility, and cost.
As is apparent, Dr Haqqani has had a long experience of acridine orange as his preferred routine method. Indeed, by introducing this approach before 1985 he anticipated the need for routine histological assessment of H pylori status before many laboratories (including our own), and antedated the first published report of the use of acridine orange for this purpose in 1986.10 We are happy that he finds acridine orange a good method in his hands. Our early experience led us to conclude that it had no particular advantage over the modified Giemsa. Large numbers of organisms within the mucous layer (as shown in the illustration) are readily seen, but we found that scanty numbers of organisms close to the fluorescent gastric epithelium were difficult to discern.11 We also felt it inappropriate to promote a fluorescent staining method as a routine approach when it requires an ultraviolet light source to be fitted to the microscope, notwithstanding the ease with which Dr Haqqani brings this into action.
However, the claim made by Dr Haqqani that acridine orange is “highly sensitive and compares equally with the gold standard” has to be challenged. To take 20 cases declared positive using the acridine orange stain and then find them all positive with the “gold standard” immunostain offers no validation whatsoever. The most insensitive of stains would pass this test. Likewise, no conclusions can be drawn from re-testing 20 acridine orange negative “control” biopsies. An accurate estimate of sensitivity and specificity can only be obtained by testing large numbers of unselected cases verified as positive or negative by non-histological methods. We would like to re-emphasise that in an earlier study based on 520 patients who had their H pylori status validated by urea breath test, biopsy urease test, and culture, the modified Giemsa had a sensitivity of 98.8% and a specificity of 99.2%.12 Although this amply justifies our own confidence in the modified Giemsa stain, these results may not be reproduced by different pathologists serving different populations. As stated previously,11 the choice of stain is a matter of personal judgement and laboratory practice. The most valuable requirement is for diligent, enthusiastic histopathologists who can recognise helicobacters by whichever stain they choose.