The regulatory organs of clinical medicine not only hand down and enforce trivial, silly and costly diktats, as a recent blog on this site related, but also impose dangerous constraints on the practice of medicine. Case in point is the College of American Pathologists (CAP) which acts as the surrogate laboratory inspector for the government. CAP’s inspection of our academic training program’s “lab,” which consisted of two microscopes, shut down an entire clinic building’s laboratories in a major academic medical center because, as an untrained and uncertified physician, I was performing gram stains.
I have been doing gram stains since I was taught to perform them 50 years ago in my 7th grade science class; preparing slides is trivial enough that it can be mastered by a 13 year old. Performing gram stains was a requisite competency in my medical school as it has been for virtually all medical students in the 134 years since Christian Gram published his technique in 1884. And I have been interpreting them for over thirty years of medical practice. As an intern thirty years ago when seeing a febrile patient with purulent sputum I would have been severely reprimanded for not having personally done a gram stain and had it available for the attending physician to review. Now it is a cause for action on the order of the breaching of a Level 4 biohazard unit with commensurately severe disciplinary proceedings to follow for the responsible “untrained” and “uncertified” practitioner.
Gram stains offers important, often essential and not infrequently time critical information that may not be obtained from culture or at least not obtained in a timely manner.
1. Gram staining gives immediate results: infectious vs. non-infectious; gram positive vs. gram negative organisms; mycotic vs. bacterial, and does so days before culture results are available.
2. Culture requires viable organisms; gram stain does not. Gram stain works on bacteria that are alive or dead so where an infectious etiology is suspected but only purulent material with non-viable organisms is available a gram stain can still direct treatment.
3. Gram stains can identify the presence of fastidious infectious organisms, especially fungi, that will not culture out on conventional media such as Sabouraud’s dextrose. Mycotic infections can be missed or inappropriately ruled out if the results of cultures are accepted as definitive.
4. It is inexpensive: pan-culture for multiple organisms including fungi will run $500-$1000. Culture may be completely avoided by the gram stain if there is a clear cut clinical correlation; or it can restrict culturing to the class of organisms seen under the scope to determine sensitivity.
Multiple cases from my inpatient service underwrite these attributes. In just the six weeks before the CAP inspection threatened disciplinary action against me the cases below relied heavily on my performance and readings:
1. A 7 day old premature infant in the NICU with KID syndrome, possible sepsis and a localized pustular rash was evaluated; beginning empirical antibiotic treatment would further complicate management of this ventilator dependent baby. The gram stain showed numerous budding yeast as the sole finding and with the clinicopathologic correlation of the rapid appearance of pustules this indicated candida. The contemplated systemic antibiotic therapy was avoided and the parents reassured. Candida was subsequently confirmed by culture but only several days later.
2. A 44 year old man with HIV, toxoplasmosis, mental status changes and cerebral infiltrates on CT scan presented with multiple fungating nodules thought by infectious disease to be either Kaposi’s sarcoma or T-cell lymphoma. A gram stain of scant exudate from these lesions showed dense gram positive cocci making the clinical lesions consistent with the rare staph infection botyromycosis. ID challenged this diagnosis and instructed the primary service to discontinue vancomycin but photos of the gram stain taken with an iPhone convinced them to continue vancomycin before culture subsequently confirmed staph infection as the sole cause of these tumor like nodules.
3. A 65 year old man post-op for glioblastoma on high dose prednisone with mental status changes presented with a rapidly evolving, acneiform facial rash raising the concern for crytptococcal infection. A gram stain from pustules showed a dense infiltrate of gram positive cocci (characteristic of Staph epi), gram positive rods (diphteroids characteristic of P. acnes) and round to oval non-budding yeast (characteristic of Pityrosporum), the classic findings of steroid induced acne avoiding both a biopsy and empirical antibiotic therapy for crypto and permitted his discharge from the hospital the next day.
Again, these were just in the previous 6 weeks of one attending encompassing only 6-9 days of actual inpatient service.
A more critical example from our institution was the early diagnosis of a 50 year old man with a high C-spine fracture from an MVA, in the ICU, septic and rapidly deteriorating. A generalized pustular eruption that I gram stained disclosed candidiasis days before blood culture demonstrated candidemia allowing earlier intervention. And going back in time to when I was expected to perform a gram stain, as a third year resident 30 years ago I evaluated a septic patient transferred in in the middle of the night with multi-organ failure. Blastomycosis was suspected but empirically giving him amphotericin, the only antimycotic available at that time, risked destroying what was left of his kidneys. I called blastomycosis on a gram stain from one of the rare pustules on his leg underwriting the necessity of using amphotericin despite the risk. That diagnosis was subsequently confirmed—but only weeks later by culture.
The exclusion of physicians from performing gram stains has left the microbiology lab techs as the sole authorities on their interpretation. There are many capable lab techs performing and reading gram stains; as noted above preparation and interpretation is typically straightforward as I demonstrated as a 13 year old. But reliance on lab techs should not be taken at face value. My personal observations from the concurrent processing and interpretation of slides with them discloses significant variations in their capabilities particularly in difficult cases. Techs follow the guidelines for performance of the procedure which were developed by bacteriologists for uniformly dense specimens skimmed from culture media. This does not account for variations in specimen thickness or content when obtained from necrotic, infected, inflammed or hemorrhagic tissue where variations in the preparation of slides is necessary.
A case that illustrates this occured following the CAP injunction when I was ordered to discontinue gram stains. I saw an 18 year old boy hospitalized for multiple large ulcers on his leg that began several months previously after he cut himself at the farm where he lived. HIs parents consulted infectious disease and multiple cultures including fungal cultures were obtained, all negative for microorganisms. I performed a touch prep from a biopsy and submitted it to the micro lab for evaluation. At this point I was prohibited from performing gram stains myself and the prep was read by the lab tech as negative. I asked to take a look and pointed out multiple large aggregates of yeast that were missed by the otherwise conscientious technician. Culture was again negative but initiation of itraconazole resolved the ulcers in several weeks confirming the mycotic nature of an otherwise undeterminable fungal species.
If physicians are not performing, interpreting, teaching and supervising those doing gram stains the results will often result in the above scenario. For example, virtually every technician I’ve observed places the stained slide on the stage, immediately applies immersion oil, goes straight to 1000x and evaluates several fields, likely what they were taught. However in a 2 x 1 cm smear of a specimen the evaluation of ten 1000 sq micron fields surveys less than 1% of the sample. As the case above illustrates low and medium power surveys of a specimen that are not performed by the technician can miss infectious infiltrates. I have pointed this to our micro lab techs on a number of occasions when concurrently reading slides. I do not relate this in disciplinary terms but as a teaching opportunity and the techs are universally grateful for such oversight. But I am one physician—and one who is been told to physically stay out of the main lab.
Regulatory diktats have subcontracted these tests to technicians and physicians have lost this skill and have no incentive to push back and take on such regulatory entities as CAP. Those are not good reasons for abdicating responsibility and subordinating what should be a physician performed microscopic exam if the clinician desires or where a practitioner is in the best position to make a clinicopathologic correlation from the results. This is especially true in an academic medical center which should be setting the standard.
CAP’s injunctions resulting in our suspending performance of gram stains and other advanced physician performed microscopy need to be forcefully challenged. Moreover, we need to reinstitute training in these techniques. Gram stain is simpler, more definitive, easier to learn and often critical to the undertaking of acute therapeutic interventions in seriously ill patients—or avoiding undertaking them at all. If CAP or anyone else doesn't think that’s important—just ask the mother of that premie in the NICU.
