Cutaneous Anthrax

At the end of October 2001, a 51-year-old woman presented to the emergency room with a localized skin ulceration surrounded by prominent edema on the forehead. She complained that the lesion had appeared recently and was enlarging. The event took place in the middle of a postal anthrax crisis in the Hamilton/Trenton area of New Jersey, and the clinical differential diagnosis immediately included anthrax. However, 2 factors made the case uncertain regarding the etiology: (1) she worked as a clerk in an accounting office but she was not a postal worker, and (2) she reported that just recently she had done a lot of cleaning of her house for a major family event. It was possible that an insect or a spider had bitten her during that time.

A swab of the patient's skin ulceration grew coagulase-negative Staphylococcus species but did not grow Bacillus anthracis, for which it was specifically examined. In the cultures, we searched for nonhemolytic ground-glass colonies that would retain their shape when manipulated and be nonmotile in the testing media.1 The Gram-stained smear of the skin ulceration did not reveal any Gram-positive, spore-forming rods.

The patient was started on antibiotic therapy; however, she came back 3 days later because the skin lesion and edema appeared to enlarge further and cervical lymphadenopathy became noticeable. At that time, the lesion presented as a cutaneous eschar surrounded by edema. The eschar swab rendered growth of coagulase-negative Staphylococcus species, light growth of diphtheroids, and rare Streptococcus viridans. No Gram-positive, spore-forming rods were detected on the Gram stain of the smears, and there was no growth of B anthracis in cultures. Nasal swabs of both nostrils were performed, and blood cultures were obtained from 2 sites; the culture results from both of those modalities were negative for B anthracis.

Finally, since the eschar was large and persistent and the patient was concerned, a surgeon removed the eschar at viable borders after approximately 7 days of the clinical course. The Centers for Disease Control and Prevention (CDC) initially refused to accept the tissue for examination, since no definite epidemiologic link had been established. Therefore, the tissue was examined on site. Both swabs of the removed tissue and fresh fragments of that tissue were negative for B anthracis in smear preparation and tissue cultures. At the same time, as a confirmation of the validity of the procedures, the laboratory detected positive cultures with morphologic characteristics of B anthracis in some patients among the more than 1000 postal workers examined.2 The tissue was examined as routine hematoxylin-eosin–stained sections, and additionally as sections with special stains for bacterial microorganisms (modified Gram, acid-fast bacilli) and fungal microorganisms (periodic acid–Schiff, Gomori methenamine silver).

The morphology of the lesion appeared to be consistent with an eschar related to B anthracis infection. The histologic examination of the eschar and surrounding tissue showed coagulative necrosis of superficial dermis and epidermis, with prominent edema of underlying viable dermis, frequent focal hemorrhages, and intense mononuclear inflammatory infiltrates around the small vessels and around some adnexal structures (Figure 1). Acute inflammatory cells were rarely noted, except for necrotizing sebaceous glands, where neutrophils were frequent. No liquefaction or abscess formation was seen. There was a sharp demarcation zone between the superficial necrotic and deeper edematous viable tissue, which was best visualized in the periphery of the lesion. Occasional islands of regenerating epidermis were present under a necrotic layer of eschar tissue, superficially lining the prominently edematous viable tissue (Figure 2). Lack of granulation tissue formation and presence of epidermalization of the viable tissue under the eschar could be an explanation for the fact that often no permanent scar is formed in the process of the eschar healing. The mononuclear perivascular and periadnexal inflammatory infiltrates were predominantly lymphocytic, plasmacytic, and histiocytic (Figure 3). The lymphoid infiltrates seemed somewhat atypical; however, they appeared reactive although no germinal centers were seen. The vessels showed degenerated endothelial cells. Foci of thrombosis were noted.

The Gram stains of the tissue, performed with different levels of intensity, revealed a focal group of partly stained Gram-positive, spore-forming, rod-shaped microorganisms, suggestive of B anthracis. With this morphologic impression of anthrax, frozen tissue and paraffin blocks were accepted by and sent to the CDC (CDC DASH# 2002705960/69) through the New Jersey State Laboratory (NJ001-10, Township of Hamilton case 1025DF04). The CDC laboratories issued a report of “positive histologic staining for B anthracis antigen” in the paraffin block and “PCR positive” in frozen tissue. The fact that no viable microorganisms of B anthracis could be detected by tissue cultures is not unusual, but rather typical in a patient treated with antibiotics. In fact, treatment of cutaneous anthrax with antibiotics decreases mortality from 20% to less than 1%.3