Esophageal Biopsy in Children ‐ Essential, Valuable, or a Waste of Time? It all Depends

Pediatric gastroenterologists spend a lot of time and energy on esophageal biopsies - taking them, reading them, discussing them, sometimes making therapeutic decisions based on them. While there are a number of esophageal conditions diagnosable by esophageal biopsy (Table 1) far and away the most common is reflux esophagitis.TABLE 1: Esophageal mucosal conditions in childrenThe practice of esophageal biopsy is so ingrained in pediatric GI that it's utility is seldom subjected to critical evaluation. Not infrequently, we take esophageal biopsies that don't contribute to management of the patient. Often, biopsies are insufficient for evaluation, or read as normal or equivocal-the scourge of the term ‘chronic nonspecific’! In these circumstances, we use basic clinical common sense, and manage patients according to their symptoms, usually with a trial of medication. Perhaps that is where we should have started in the first place, and omitted the upper endoscopy (EGD) and biopsies entirely. That approach is discussed in this article, in which the utility of esophageal biopsy in patient management is examined. It's less meaningful to consider biopsy outside of the context of, so the real question is whether EGD with biopsy changes management (where ‘with biopsy’ means with multiple biopsies). The issue is examined mostly with reference to GE reflux disease [GERD], but also other esophageal conditions that occur in children. Such an analysis is more meaningful in the real world context of clinical and endoscopic findings. ENDOSCOPIC FINDINGS Some endoscopic findings in the esophagus are considered to be nonspecific and within normal, or have such marked interobserver variability as to be unhelpful for diagnosis. These include erythema, increase or loss of vascularity, and pallor. While magnification endoscopy and other techniques hold promise for identification of subtle changes in the future (1,2), at present it seems that erosions, ie, mucosal breaks, are the main endoscopic findings that are considered definitely abnormal, with good interobserver agreement in reporting (3). In children and adults, GE reflux is far and away the commonest cause of erosions, though they occur also with other conditions (Table 1). For example, fine linear erosions just above the GE junction can be due to forceful vomiting from any cause, even of short duration-an esophageal extension of Mallory-Weiss tears. Of course, in many cases the clinical picture narrows the differential diagnosis. For example, in the context of recurrent substernal pain or heartburn, the presence of erosions has a high specificity for reflux as the cause. Or, in the context of bloody diarrhea or weight loss or pubertal delay, Crohn's disease would likely be the cause. On the other hand, in a patient with epigastric pain or chest pain, the diagnosis might be GERD, eosinophilic esophagitis, peptic ulcer disease, or one of the gastritides or gastropathies (4). Lower grades of esophagitis often manifest as discrete erosions on the tops of folds in the distal esophagus. More severe erosive disease can present diagnostic problems - for example, in confluent erosive disease, hemorrhage, pus and debris may camouflage landmarks, and thereby hide Barrett's mucosa (5-6). Esophageal infections may also present in this way, especially in the immune-compromised host (7). HISTOLOGY We often get poor quality tissue from the esophagus, in part because the esophagus is somewhat difficult to biopsy, it's axis being parallel to the endoscope, but also because small biopsy forceps are often used in pediatrics-and these skid off the mucosa more easily than larger forceps. These problems can be avoided with careful biopsy techniques involving turning into the mucosa and applying suction (7,8), and by using large cup (‘jumbo’) forceps. These yield biopsies of greater surface area (though not necessarily greater depth) for morphometry than smaller forceps. Their larger size allows for easier mounting of biopsies, orientation and cutting in the core of the biopsy block, facilitating morphometric evaluation of changes such as elongation of papillae. Large biopsies are less compromised by crush artefact. These biopsies are standard of care in many adult GI units - with minor, basic precautions, they can be also taken in children with safety (8). The quality and quantity of tissue influences diagnostic utility (7-9). FOR REFLUX ESOPHAGITIS Reflux esophagitis is a patchy lesion (9,10). Therefore, it's not surprising that even when erosive esophagitis is present, biopsies may be normal, or show mild through severe changes. While sampling ‘error’ can be minimized by obtaining more tissue-more biopsies, larger biopsies-from multiple sites, in reflux esophagitis there remain some qualitative limitations. For example, one recent study (11) in adults showed that some 20% of biopsies for GERD were insufficient for analysis; but even with tissue that was adequate, there was an interobserver variability of 34% and intraobserver variation of 10% in diagnosing the presence and severity of reflux esophagitis on pre-op biopsies, and 21% and 6% in post-op biopsies. While several abnormalities have been described in reflux esophagitis, histological criteria for reflux esophagitis vary among even expert pathologists. For example, although much store is set by the presence and number of intraepithelial eosinophils, they are not pathognomonic for reflux esophagitis, being present also in infectious esophagitis (candida, herpes), idiopathic eosinophilic esophagitis, drug-associated esophagitis, and Crohn's disease (9). The density of eosinophils on biopsy does not correlate with the severity of reflux (12,13), and they may persist in the epithelium after antireflux surgery (11). For these reasons, and because they lack sensitivity for reflux, some expert pathologists use scoring systems that don't include eosinophils as a diagnostic criterion for reflux (1,11,12) (Table 2). Intraepithelial lymphocytes, while more sensitive than eosinophils or neutrophils in erosive esophagitis, similarly are relatively insensitive indicators for esophagitis in nonerosive reflux disease (12).TABLE 2: Histologic features of reflux disease in a specific study11An additional strike against histology for GERD is that there is poor correlation between esophageal histology and symptoms, endoscopic findings, or pH monitoring, both qualitatively and quantitatively in children (14-16). Recent data show that ultrastructural abnormalities are discernible even in nonerosive disease, and that these heal with PPI treatment (17), but this requires transmission electron microscopy. Other data suggest that elongation of the papillae, and basal cell hyperplasia are sensitive indicators for reflux disease in biopsies taken from 2 cm above the Z-line, and at the Z-line itself (12); however, there were no controls involved, and these may be normal findings (7). Taking biopsies at the Z-line might be useful when metaplasia or H. pylori infection is suspected (18), but it is the gastric columnar side of the biopsy that is of interest in these circumstances. Utility in patient management: A recent study in adults (19) showed that an EGD altered antireflux therapy only in about one-third of cases, in which either stricture dilatation was performed, or Barrett's esophagus was present, or severe esophagitis was found. The authors recommended that EGD not be used to tailor proton pump inhibitor (PPI) therapy-if the response to PPI is poor or absent, the dose should be increased and EGD performed only in case of persistent non-response, or alarm symptoms, or risk for Barrett's, or need for chronic therapy. Recent practice guidelines from the US and Canada (20,21) support the approach of empiric therapy for GERD at the outset. This approach is well-illustrated by the entity of nonerosive reflux disease [NERD], in which patients may have symptoms, but no changes at endoscopy or on biopsy. Because not all esophageal symptoms are due to acid reflux, histology might be useful in cases that don't respond to PPI treatment - these patients may have ‘functional heartburn’, ie, a motility disorder, non-acid reflux, which might require different treatment (22). FOR BARRETT'S ESOPHAGUS Barrett's esophagus [BE] is an uncommon disorder in children, but an important one not to miss or overcall. The diagnosis carries major implications for the patient - risk of malignancy, need for endoscopic surveillance, and possibly hefty premiums for health insurance and life insurance. Even though the diagnosis of BE may be suspected by the presence of pink mucosa extending into the tubular esophagus at endoscopy, confluent erosive disease or hiatal hernia can also mimic this appearance. In any endoscopic scenario, tissue will make or rule out the diagnosis (5,6,23). The diagnosis is made definitively in adults and children only by the presence of goblet cells containing acid mucin that stains with Alcian blue at pH 2.5 (5,6,23) False-positive and -negative diagnosis of BE may result from poor identification of landmarks and poor targeting of biopsies, too few biopsies, failure to use specialized stains, goblet cells missed or overcalled. Sometimes the diagnosis of BE is missed or cannot be ruled out because the esophagus is severely inflamed, and biopsies yield poor tissue or just debris. In this case, repeat EGD with biopsy after a trial of a proton pump inhibitor [PPI] is important to be determine whether BE is present. Utility in patient management: In established BE, endoscopic surveillance is indicated; for dysplasia or cancer, further therapies are indicated. In the absence of BE, surveillance endoscopy is not indicated. Therefore, histology is essential for patient management when this disorder is suspected or present. FOR EOSINOPHILIC ESOPHAGITIS [EE] This is a specific disorder characterized by the presence of high eosinophil counts [>15-20/hpf] in the esophageal epithelium, typically through the length of the esophageal body (9,24). In contrast, when eosinophils are present as a feature of reflux esophagitis, they are present in fewer numbers and only or prevalently in the distal esophagus. In addition, in EE, typically, there is no basal cell hyperplasia or elongation of papillae, features of reflux disease. However, in some cases, there are high eosinophil counts and features of basal cell activation. In these cases, the diagnosis may be reflux esophagitis, or eosinophilic esophagitis complicated by reflux - an unsurprising circumstance, given the presence of esophageal dysmotility secondary to EE. In these circumstances, intraesophageal pH study may help, but given that it is not always reliable/reproducible, a therapeutic trial, ie, a trial of proton pump inhibitors vs trial of corticosteroids may determine the diagnosis. The clinical presentation is sometimes suggestive of EE, but there is much overlap with GERD (24). Some endoscopic features are strongly suggestive of the presence of EE, such as small-caliber esophagus, corrugated or ringed esophagus, proximal stricture, or ‘white specks’ or papules in patches or through the length of the esophagus (24,25) - the latter resemble candida infection, and are eosinophilic abscesses. Vertical lines or furrowing is a sign of esophageal inflammation, less specific to EE. The esophagus may also be normal at endoscopy. It appears that in pediatrics, we may have misinterpreted the significance of intraepithelial eosinophils. Given their poor sensitivity as indicators of reflux, it may be that esophageal eosinophils more often represent food allergy or a form of eosinophilic esophagitis than reflux; and the treatment is different. Utility in patient management: EE is a biopsy diagnosis, and this governs treatment. Corticosteroids, inhaled or systemic, are the main treatment for EE, and generally should be given on the basis of a tissue diagnosis. CONCLUSIONS It seems self-evident that for esophageal disorders with specific histologic characteristics with specific management, for which biopsy is sensitive and specific, EGD with biopsy is essential. These include Barrett's esophagus, eosinophilic esophagitis, or rarer conditions. In contrast, for GERD, the sensitivity and specificity of histology are relatively poor, even in expert hands, and the diagnosis often can be imputed from other approaches. Pope argued that for adults ‘biopsy in the usual case of erosive esophagitis is a waste of time and money’ (10). His ‘usual case’ is an adult presenting with typical symptoms of reflux, ie, heartburn, regurgitation, etc. In this case, one could justify not doing an endoscopy at all, just giving the patient a trial of PPI, and performing EGD and biopsy only in case of a chronic need for drug. This approach is supported by recent guidelines for adults (20-23). The reason to perform EGD in the chronic relapsing case, is to verify the diagnosis as GERD (rather than peptic ulcer disease, for example), and to rule out Barrett's (6,26). Then the patient can be committed to long-term PPI or antireflux surgery. Therefore, for reflux esophagitis without Barrett's, there's little evidence that esophageal histology makes a difference to initial clinical care decisions, beyond ruling out other conditions. Severity is determined by symptoms and endoscopic changes (27). Most important is the patient's response to an adequate dose of antireflux medication. These practices make sense for pediatrics as well, and reflect what many practitioners are doing. The difference in pediatrics is that heartburn is much less common than in adults and reflux often presents with upper abdominal pain, or atypical manifestations, ie, there are fewer ‘usual’ cases. Nevertheless, in cases of suspected GERD, EGD with biopsy can be used in the same way, to make a diagnosis in case of poor response to PPI, and to have a diagnosis in those cases requiring chronic therapy. The most common scenario would be to determine whether the diagnosis is GERD, or EE, or a combination. Prior to endoscopy, in some cases, it is more helpful for patients to come off drug at least a week or two before endoscopy, so that there can be positive findings (28). There is no evidence that in the absence of erosions, abnormal squamous histology has any sequelae or bearing on patient outcome. In other words, in the asymptomatic patient on or off treatment, there is no indication to perform ‘follow-up’ endoscopy to look for residual inflammation. The same is the case for eosinophilic esophagitis. Finally, a key factor is the diligence and skill the endoscopist brings to the procedure. This involves identifying landmarks, obtaining adequate quantity and quality of tissue, and communicating to the pathologist key information such as what drugs the patient is taking, the location of biopsies in relation to landmarks, and the differential diagnosis based on clinical and endoscopic findings. With this approach, the utility of esophageal histology as a patient care tool is maximized.