Is it necessary to perform a biopsy in erosive esophagitis?
C.E. Pope II (Seattle)
Endoscopic biopsy can provide added information to that obtained visually through the endoscope. Tissue removed serves as a permanent record of the state of the
mucosa. Biopsy specimens may be examined at leisure, and by different observers. The risk to the patient of endoscopic biopsy with standard forceps is very small; bleeding post biopsy is very unusual unless the platelets or clotting factors are markedly depressed. Obtaining biopsies lengthens the procedure and the cost of disposable forceps (or processing of reusable forceps), and the cost of processing and interpreting the biopsies is not inconsiderable. This latter fact makes it essential to make certain that the information obtained by the endoscopic biopsy is needed.
When the issue of whether endoscopic biopsy is needed in erosive esophagitis is raised, many would answer in the affirmative. After all, is it not good medical practice to document observations (even those made through the endoscope) with objective proof of what was seen? The author would like to present the point of view that biopsy is only necessary in a few circumscribed circumstances; that biopsy in the usual case of erosive esophagitis is a waste of time and money.
A large portion of my opinion is based on the idea of discrimination in diagnosis [1,2]. That is, if one knows what the condition is with one form of test, then it is redundant to obtain other tests to confirm a diagnosis which has already been established. What will be the most common cause of erosive esophagitis? Clearly, reflux disease will be, by far and away, the most common cause. What are the endoscopic features of gastroesophageal reflux (GER) disease? Most authorities agree that erythema or friability are too nonspecific, and subject to too much observer variation [3]. More specific for reflux disease are longitudinal erosions which are usually located on the tops of the esophageal folds. Often the white erosions are surrounded by a rim of erythema. Biopsy of such a lesion will uniformly reveal polymorphonuclear leukocytes infiltrating the lamina propria, as well as an increase in the basal cell layer and elongation of the dermal papilla [4-6].
When the erosions became more confluent, as is found in more severe reflux disease, the endoscopic appearance becomes less diagnostic and the chance that biopsy may be useful rises. If the endoscopic appearance shows a confluent group of erosions and exudate in the lower tubular esophagus, often with an indistinct or absent Z line, reflux disease is still the best possibility, especially if this appearance is associated with a hiatus hernia as recognized endoscopically. A patulous G-E junction which gapes open increases the probability of reflux damage. If the gastroesophageal junction is viewed from below by retroflexing the instrument, absence of a valve-like structure is found in many patients with severe reflux. (See page 126).
If, however, the same endoscopic appearance is seen in the middle of the tubular esophagus, then biopsy is probably warranted. Samples should be taken not only of the eroded areas, but also distally from the intact mucosal zone below the area of inflammation, as this will usually be found to show columnar (Barrett's) epithelium. In this case, biopsy is to be used not only to document the presence of erosions and inflammation, but also to confirm the presence of columnar or metaplastic epithelium. Knowledge of the presence of this type of epithelium will set into motion other questions of evaluation and therapy covered elsewhere in this book.
The patient's clinical history can be of a great deal of help in deciding whether or not to biopsy an erosive lesion of the esophageal mucosa. If the patient has been
taking certain pills (tetracyclines, quinidine, potassium chloride), and then develops sudden odynophagia, then endoscopy will usually show an eroded area in the mid-esophagus [7]. Yet again, it is necessary to ask whether biopsy is warranted, as it will show only nonspecific inflammation and will not add to the clinical diagnosis of pill esophagitis. In fact, the history is so specific that even endoscopy may not be necessary as the diagnosis can be firmly established on clinical grounds.
Are there other conditions which can mimic esophagitis from GER? An interesting paper suggested that candidal esophagitis (even in the nonimmunocompromised host) can present with a confluent eroded appearance, which can be difficult to distinguish from severe GER damage [8]. The presence of hyphae in the biopsy (best seen on silver or PAS stain) confirms the diagnosis and requires a change in therapeutic management. It seems likely that some such cases are missed even when biopsied, as the fungal organisms are difficult to recognize on ordinary hematoxylin and eosin (H and E) stains.
Erosive esophagitis in the immunocompromised host is one situation in which biopsy can be very helpful [9]. Involvement of the mucosa with herpes simplex virus (HSV) or cytomegalovirus (CMV) can produce not only vesicles and ulcers, but confluent esophagitis as well. With the use of suppressive mediums such as acyclovir and gancyclovir, such infections are becoming less common. However, in the untreated host, they are still encountered. In bone marrow patients, the presentation is often atypical with the patient complaining of nausea and vomiting instead of the more usual clinical manifestation of esophageal inflammation such as odynophagia or dysphagia [10]. Biopsies should be taken not only from the sides of the lesions (for herpes virus), but deep at the base of the erosions in order to sample CMV. Culture and immunostaining of the biopsies will show the viral lesions more often than standard histological changes, such as nuclear and cytoplasmic inclusions.
One situation in which endoscopic biopsy would be indicated in erosive esophagitis, is during investigations involving new drug therapy or the results of antireflux surgery. In this special situation, the need for biopsy is not to sort out the individual clinical situation, but rather to document the results of a clinical trial. The ability to pass around coded biopsies to several different investigators during such a trial allows a degree of objectivity which is difficult to reach with endoscopy alone.
There are other situations in which the role of biopsy of erosive lesions of the esophageal mucosa are less well defined. Damage to the mucosa by radiotherapy or chemotherapy may show characteristic changes on biopsy, but the clinical situation will still be the most useful diagnostic aid.
Does biopsy of erosive esophagitis, when treatment has not caused the lesions to regress, offer anything? Possibly, although even with biopsy results the usual clinical response will be to increase acid peptic therapy (switching to a proton-pump inhibitor if H2 blocking agents had been the prior form of therapy). Only in the failure of the latter type of therapy would it be worthwhile to check for an unexpected fungal or viral pathogen.
To summarize, it would appear that routine biopsy in erosive esophagitis would seem not to be a worthwhile strategy, as it increases time and cost of the procedure without increasing the diagnostic yield. Exceptions to this strategy would be in the
immunocompromised host or when the erosive esophagitis is located in an atypical location. If biopsy is used in the patient suspected of viral or fungal invasion, then special stains and cultures should be performed.
References
2. Edwards DAW. Discriminative information in diagnosis. Proc R Soc Med 1971:64:676-677.
7. Eng J, Sabanathan S. Drug-induced esophagitis. Am J Gastroenterol 1991;86:1127-1133.