key: cord-0009162-eg7r71kp
authors: Doern, Gary V.
title: The role of the laboratory in the etiologic diagnosis of gastrointestinal infections
date: 1989-07-15
journal: Clin Microbiol Newsl
DOI: 10.1016/0196-4399(89)90046-9
sha: ae27d6930b6a1d2360d8dea0e38bc74e0903a001
doc_id: 9162
cord_uid: eg7r71kp

nan

Acute gastrointestinal illnesses are extremely common in the United States. It is estimated from various epidemiologic and patient demographic factors, that Americans experience 1.8 to 2.1 acute gastrointestinal illnesses per person per year (1) . A large, but unknown percentage of these illnesses is due to microorganisms. Table 1 contains a list of selected bacteria, protozoa, and viruses that are recognized as causes of symptomatic gastrointestinal infections. Intentionally excluded from this list are numerous agents that either do not occur in the United States or do so very infrequently. In addition, certain bacteria that have been associated with, but not proven to be the cause of symptomatic gastrointestinal disease (e.g., Citrobacter spp., Edwardsiella spp., and Pseudomonas spp.), have been omitted. Even with these exceptions, the list is very long.

It should be recognized that laboratory methods (albeit, in some cases experimental methods) now exist for detecting the presence of all agents listed in Table 1 in clinical specimens representative of gastrointestinal infections. Theoretically, laboratories could attempt to rule out the presence of each of these organisms in all such specimens. In reality, this elimination is not done for many reasons, the two most important of which are the exorbitant cost of supplies and reagents necessary to perform the requisite analyses and the excessive amount of technologist time required.

Most general diagnostic microbiology laboratories in the United States routinely examine representative specimens from patients with diarrheal illness only for the presence of Salmonella spp., Shigella spp., and Campylobacter spp. This limited testing is quite appropriate at least for nonhospitalized patients and can be accomplished by using a differential gram-negative agar medium (e.g., MacConkey or EMB), semiselective 

Picoma/parvovirus-like agents gram-negative agar medium (e.g., Hektoen or XLD), and a Campylobacter selective medium. In addition, it has been recommended that a highly selective gram-negative agar medium (e.g., bismuth sulfite or brilliant green) be used together with gram-negative enrichment broth (e.g., GN, selenite, or tetrathionate (2) . Finally, use of broth enrichment for Campylobacter has been advocated (2). Little objective evidence supports the routine use of a highly selective medium, such as a bismuth sulfite or brilliant green, when processing fecal specimens. Similarly, broth enrichment for Campylobacter on a routine basis is probably not necessary (3). Selective broth enrichment for gramnegative bacilli also may be superfluous (4) . First of all, many laboratories are not staffed to accommodate subculturing enrichment broth 4 to 6 h after inoculation, as is recommended with GN broth. Second, at least in symptomatic patients, routine use of gramnegative enrichment broth may not result in significantly increased rates of enteric pathogen detection (4) . The primary role of enrichment broth is to facilitate detection of small numbers of Salmonella and Shigella, such as might be present in fecal specimens from asymptomatic carriers. However, two factors argue for routine use of gramnegative broth enrichment: i) at least one laboratory accrediting agency, the College of American Pathologists, indirectly mandates its use, and ii) accurate and rapid detection of Salmonella and/or Shigella in fecal specimens is now possible by performing a latex agglutination test directly on enrichment broth (5). These issues notwithstanding, it would seem appropriate that individual laboratories assess their own experience with gram-negative enrichment broth when making judgments as to its utility.

Having addressed detection of Salmonella spp., Shigella spp. and the enterocolitis-associated Campylobacter spp., the obvious question is when should a laboratory attempt be made to detect the other agents listed in Table  1 ? The laboratorian's most common answer to this question is, of course, ~'upon request by the clinician." This response, however, totally ignores the fact that in many, if not most, cases, there exists little clinical basis for suspecting these agents. Exceptions to this rule would be i) individual patients with strongly suggestive epidemiologic histories and a clinical disease compatible with a particular agent(s) or ii) patients seen in the midst of an outbreak of known etiology. These two circumstances are, however, uncommon. As a result, it is often not possible for the clinician to determine which microorganism(s) in Table 1 is most likely present and thus specifically request that analyses be performed to detect it.

The limited scope of this editorial does not permit a discussion of when the laboratory might consider searching for each of the organisms listed in Table 1 . However, several general observations apply in nearly all instances. The f'n'st concerns screening fecal specimens for the presence of fecal leukocytes with the aim of distinguishing gastrointestinal infections with an inflammatory component from those that are largely noninflammatory. A fecal leukocyte determination is a rapid, simple, and inexpensive procedure. Once a satisfactory level of skill has been achieved, a strong case can be made for using this test as a routine preliminary screen on all stool specimens. The exception would be those instances in which mixed infections occurma positive fecal leukocyte exam largely rules out noninvasive organisms such as enterotoxigenic Escherichia coli, Vibrio cholerae, gastrointestinal viruses, and protozoa such as Giardia lamblia and Cryptosporidium spp. As a result, a positive fecal leukocyte test may obviate the need to perform diagnostic procedures aimed at detecting these organisms.

A second general consideration concerns the importance of meaningful interaction between the laboratory and the clinician when fecal specimens are processed for agents other than the routine Salmonella, Shigella, and Campylo. bacter. It is the laboratory's responsi-bility to convey to the clinicians which organisms the laboratory has the expertise and/or materials to detect and those they do not. Clinicians must also be made aware of the complexity and cost of the analyses that are required to detect in fecal specimens the presence of the agents in Table 1 . The limitations of existing test methods should be made known together with the usual length of time required to generate positive and negative results. When asked, clinicians should be able and willing to provide the reasons for their request that a particular agent(s) be sought in the laboratory. If such reasoning is incorrect, an effort should be made to educate the clinician.

Another general issue is that the laboratory often has information unknown to the requesting physician that should impact on what is done with fecal specimens. For example, assume a diarrheal stool specimen arrives in the laboratory with gross evidence of blood. The specimen was obtained from an inpatient hospitalized for management of hemolytic-uremic syndrome (HUS). The laboratory knows this diagnosis either because it is written on the requisition slip or because the laboratory has knowledge of laboratory values that define the patient as having HUS (i.e., microangiopathic hemolytic anemia, thrombocytopenia, and acute renal failure). The patient's physician has probably not seen the fecal specimen, and as a result, may not know that the patient has grossly bloody diarrheal stool. This specimen definitely should be examined for enterohemorrhagic E. coli, particularly if the patient is afebrile. The physician should be contacted and this possibility discussed. If the physician cannot be reached, the laboratory should do the test even though it has not been requested.

Finally, one simple rule can be applied to all of the agents in Table 1 . When their prevalence exceeds the prevalence of the least common agent of the three that are routinely sought, i.e., Salmonella spp., 5higella spp., and Campylobacter spp., then the laboratory should seriously consider rou-tinely screening for that organism. Two agents for which this rule may often apply are Clostridium difficile in hospitalized patients receiving antimi-crobi~ therapy and rotavirus in very young children during the winter months. In addition, while in the midst of outbreaks, it may be appropriate to search routinely for agents that otherwise would be considered selectively.

The question is, of course, how does a laboratory ascertain the prevalence of something for which it is not searching routinely? One approach is to interact with other laboratories in your geographic area, particularly those with similar patient populations. Shared experiences are often illustrative. In addition, information derived from published surveillance studies can, in some cases, be extrapolated to a particular laboratory setting. Finally, it may be appropriate on occasion, to perform routine surveillance for a specific organism over a defined time period, expressly to establish the prevalence of that organism. Obviously, surveillance of this type should be restricted to those organism(s) likely to be common enough to warrant the effort and expense such surveillance studies require.

A cost effective and effective approach to the diagnosis and management of acute diarrhea

Laboratory diagnosis of bacterial diarrhea

Evaluation of four enrichment media for isolation of Campylobacter jejuni

Is enrichment broth necessary for stool cultures?

of a latex test with traditions/methods for detection of Salmonella and Shl&ella in emichraent broths