Materials and methods

All cases from the pathology archives of the Royal Group of Hospitals Trust, Belfast with a SNOMED diagnosis of actinomyces in the female genital tract between January 2000 and June 2009 were identified. All the cases were diagnosed by one of the authors (WGM) who had SNOMED coded all such cases as actinomyces. The H&E stained sections were reviewed by the authors. Several cases had histochemical stains performed as part of the original workup (Gram stain, Gomori methenamine silver stain and modified Ziehl–Nielson stain). All cases were assessed as to whether they represented actinomyces-like organisms or PAMRAGs. A careful search for actinomyces-like organisms and PAMRAGs occurring within the same granules was undertaken.

Results

Fifteen cases were identified for inclusion in the study (table 1). Age of the patients ranged from 19 to 58 years. Of the six endometrial cases (five endometrial biopsies and one hysterectomy specimen), one contained actinomyces-like organisms (fig 1), one contained PAMRAGs (fig 2) and four had combined granules consisting of actinomyces-like organisms and PAMRAGs. The actinomyces-like organisms consisted of basophilic filamentous organisms, sometimes with a central more eosinophilic core, and the PAMRAGs of refractile radiating eosinophilic club-like formations. The combined granules consisted of a central basophilic core of filamentous organisms surrounded by radiating eosinophilic club-like formations (fig 3A). In some of these combined granules, the basophilic core predominated (fig 3B), while others were largely composed of the club-like PAMRAGs. Gram and silver stains were performed on all the endometrial specimens. The case of actinomyces-like organisms was Gram and silver positive, as was the filamentous core in all the combined granules (fig 4A,B). The case of PAMRAG was negative. Modified Ziehl–Nielson stain was performed in three of the endometrial specimens (one containing actinomyces-like organisms and two containing combined granules); all were negative. In all the endometrial biopsies, there was evidence of endometritis with a mixture of acute and chronic inflammatory cells, including plasma cells. The two cases with tubo-ovarian abscess formation contained combined granules composed of actinomyces-like organisms and PAMRAGs, the former being Gram and silver positive and the latter negative. The vulval specimen (a vulval abscess) contained PAMRAGs embedded within a mass of acute inflammatory cells. All seven cervical specimens (six loop excisions and one hysterectomy) contained PAMRAGs. These were always located within endocervical glands, which were sometimes dilated, in association with inspissated and sometimes inflamed mucus (fig 5). Gram and silver stains (performed in three of the cervical cases and the vulval case) were negative.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Actinomyces-like organisms in the endometrium composed of thin basophilic filaments with a more central eosinophilic core.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Pseudoactinomycotic radiate granules in the endometrium composed of radiating eosinophilic club-like formations.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

(A) Combined granule with central basophilic filamentous organisms and surrounding eosinophilic club-like pseudoactinomycotic radiate granules (PAMRAGs). (B) Combined granule largely composed of basophilic actinomyces-like organisms with surrounding eosinophilic PAMRAGs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

(A) Combined granule where the organisms are Gram positive (Gram stain) and the pseudoactinomycotic radiate granules (PAMRAGs) negative. (B) Combined granule where the organisms are silver positive (Gomori methenamine silver) and the PAMRAGs negative.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Pseudoactinomycotic radiate granules within endocervical gland.

In all of the cases with endometrial involvement by actinomyces-like organisms, PAMRAGs or combined granules, there was a history of IUD usage.

Discussion

Several groups have investigated the composition of and mechanisms behind the formation of PAMRAGs. However, currently both these parameters remain unclear.4 7 8 PAMRAGs may be seen in the presence or absence of an IUD, occur at various sites in the female genital tract and contain a variety of substances including calcium, lipid, neutral glycoproteins, copper, phosphorus, sulphur, chloride and iron. It is likely that PAMRAGs have a variable composition; they may form on a variety of nidi and subsequently develop by encrustation and the accumulation of various elements, such as those listed above. It was initially suggested that the nidus represented material from an IUD4 but subsequently it was shown that PAMRAGs do not always occur in association with such devices. It has also been suggested that the nidus may represent various products of leucocytes. It has been assumed that PAMRAGs do not contain microorganisms but, as we have shown, this is not always the case. In the combined granules we report, we feel the organisms act as the nidus and various elements deposit around these resulting in the formation of PAMRAGs. Evidence for this comes from the histological appearances where the organisms are located centrally within the granules with the radiating eosinophilic club-like projections surrounding these. PAMRAGs have been likened to the Splendore–Hoeppli phenomenon, defined as the presence of eosinophilic radiating material surrounding organic and inorganic substances, and it is likely that these represent the same phenomenon. Interestingly, the association of eosinophilic radiating material with actinomyces and other bacteria, fungi, parasites and inert materials has been described recently in mucocutaneous lesions.9

As stated earlier, it is important to distinguish actinomyces from PAMRAGs in the female genital tract. Both occur most commonly within the endometrium but may be seen at other sites, including in tubo-ovarian abscesses and in the cervix and vulva (discussed below). In the endometrium, both are most commonly found in association with an IUD. The importance of distinguishing between the two is that actinomyces responds to specific treatment, although it is controversial whether antibiotics are required in the absence of signs of pelvic inflammatory disease,2 3 while it has been considered that PAMRAGs are non-infectious and do not require antibiotic treatment. However, our study illustrates that the two may coexist. In fact, judging by our series, combined lesions may actually be more common in the endometrium than actinomyces-like organisms or PAMRAGs alone. It might be expected, given that both most commonly occur in association with an IUD or an abscess, that actinomyces and PAMRAGs might coexist in some cases as separate granules; this has been occasionally reported.7 However, as far as we are aware, the presence of combined granules such as those we describe has not specifically been reported in the female genital tract. Histopathologists need to be aware of this possibility and when they identify PAMRAGs should look carefully at the centre of the granules for organisms and, if necessary, perform Gram and silver stains. Actinomyces organisms are Gram and silver positive while pure PAMRAGs are negative or exhibit non-specific staining with Gram.6 In fact, it may be prudent to undertake Gram and silver stains in all cases of PAMRAGs in order to identify small collections of organisms which are not obvious on histological examination.

One limitation of our study is that we cannot prove unequivocally that the organisms represent actinomyces and not some other morphologically similar Gram and silver positive filamentous organisms, such as nocardia. Culture is required for a definitive diagnosis of actinomyces, although this is often inadequate and is impractical in many cases. Culture is not routinely performed when actinomyces or actinomyces-like organisms are diagnosed by the histopathologist in specimens from the female genital tract. We believe the organisms in our cases to almost certainly represent actinomyces although, in the absence of culture, we cannot confirm this. A modified Ziehl–Nielson stain was negative in several of our cases, helping to exclude nocardia which is usually positive.

To our knowledge, PAMRAGs have not been previously described in the vulva and have only been briefly alluded to in the cervix.8 We identified seven cases of cervical PAMRAGs and one in a vulval abscess. In the latter, it is likely that the mechanism of development is similar to that within tubo-ovarian abscesses. Our cases of cervical PAMRAGs were all located within endocervical glands, which were often dilated and typically contained inspissated and inflamed mucus. It is probable that this inspissated mucus acts as a nidus on which the formation of PAMRAGs can occur.

In summary, we report six cases of combined granules consisting of actinomyces-like organisms and PAMRAGs in the female genital tract. We feel the organisms act as a nidus around which a PAMRAG potentially forms. Although actinomycotic granules and PAMRAGs are distinct entities, our study shows they may coexist. Pathologists should be aware of this and consider the possibility of combined granules when PAMRAGs are identified on histological sections. Gram and silver stains may assist in this regard.