Seven patients (1 in 1940 and 6 in 1944) had a clinical diagnosis of typhus and a histopathologic diagnosis of typhus on the basis of FFPE tissue blocks. All samples were of CNS tissues (cerebral cortex, pons, medulla oblongata, and cerebellum); 1–3 different CNS regions were available per patient (Table). All but the case from 1940 (patient 6) originated from Langenhorn hospital in the northern part of Hamburg.

Figure 1

Figure 1. Hematoxylin and eosin staining of typical typhus nodules in brain of typhus patients during World War II, Hamburg, Germany, 1940–1944. Most nodules were found in the pons and medulla oblongata. A)...

Figure 2

Figure 2. Immunohistochemical analyses of nodule cell compositions from typhus patients during World War II, Hamburg, Germany, 1940–1944. Tissue sections were incubated with specific antibodies and visualized with immunoperoxidase (brown) or immunophosphatase (blue)...

Typical typhus nodules (Figure 1) were found in all cases. Lesions varied in size and location; most were seen in the pons or medulla oblongata (Table). No intracellular rickettsiae were found in Giemsa-stained sections. Immunohistochemical stains with CD20 antibody indicated B cells were only rarely present in perivascular regions and were not found in nodules. In contrast, T cells (CD3-positive cells) contributed prominently to nodule cell composition, consisting of ≈60% CD8+ T cells and ≈40% CD4+ T cells (Figure 2, panels A–C). CD177+ neutrophils were found in nodules rarely, and when present, these neutrophils were scattered mostly throughout the brain parenchyma (D. Tappe, unpub. data). In contrast, a strong signal was seen for CD68+ microglia and macrophages in and around the nodules (Figure 2, panel D). No iNOS+ cells were found (D. Tappe, unpub. data).

The panrickettsial qPCR, typhus group rickettsiae–specific rpr331 qPCR, and R. prowazekii–specific nested qPCR results indicated all samples were negative for rickettsial genomic material. However, the nested prsA qPCR indicated tissue sections of the pons of 2 patients (patient 3 and patient 5; Table) were positive for R. typhi DNA. Positive qPCR results were obtained again with the pons of these patients when the extractions and qPCRs were repeated. All negative controls (FFPE blocks from patients with unrelated conditions and water) were negative. We sequenced the 2 amplicons (201 bp in length, including sequences of the outer primers; patient 3 GenBank accession no. MH553441), and BLAST analyses (http://www.blast.ncbi.nlm.nih.gov) showed sequence identity with R. typhi strains B9991CWPP (GenBank accession no. CP003398), TH1527 (GenBank accession no. CP003397), and Wilmington (GenBank accession no. AE017197) (Technical Appendix Figure).

Cases of typhus resulting from war, population displacement, poverty, and overcrowding are usually attributed to epidemic louseborne R. prowazekii (1). We show successful amplification of R. typhi DNA for 2 out of 7 patients with fatal typhus and encephalitis during World War II by a specific nested qPCR with FFPE CNS tissue samples stored for >70 years. Because of this finding, the diagnosis of murine typhus was made retrospectively for these 2 patients. For the remaining 5 patients, no rickettsial DNA could be amplified in the limited tissue materials available, possibly because of poor DNA quality. Nested PCRs are more sensitive than conventional PCRs, which could explain why the panrickettsial qPCRs and typhus group–specific qPCRs did not amply genomic DNA, whereas the nested PCR did with samples from 2 patients. Because nested PCRs are prone to contamination, we took several precautions: we included several negative controls (which remained negative throughout the study), omitted the positive control, and ran the nested PCR as a qPCR.

No intracellular bacteria were found by conventional light microscopy in Giemsa-stained brain sections. We could only assess CNS tissues because other tissue types had not been archived; the reason for the absence of these tissues is unclear. Microscopy results negative for rickettsial bacteria in brain tissue sections of typhus patients are in line with old reports; the doubtless detection of rickettsiae on microscopic examination of epidemic typhus patient samples was notoriously difficult (10,11). We did not attempt immunofluorescence testing of tissue sections. The postmortem detection of tissue nodules in brain sections during microscopic examination has been regarded as specific for epidemic typhus. Such typhus nodules were first described in the skin of patients with rash during epidemic typhus (typhus exanthematicus) in 1913. Shortly thereafter, during World War I, these nodules were found postmortem in the brains of epidemic typhus patients. The nodules were shown to be most prominent in the medulla and pons, also as seen in our study, followed by the basal ganglia, cortex, hippocampus, and cerebellum (11–13). In the 1910s, typhus was correctly described as a discontinuous vasculitis that could also occur in the brain involving lesions that form in endothelial cells of the intima of capillaries and small arteries. Inflammatory nodules always arranged around a blood vessel and were composed of lymphocytes, plasma cells, neutrophil granulocytes, and glia cells (12,13). In contrast with autopsy data of epidemic typhus, only rare reports of histopathologic findings of murine typhus exist. Typical rickettsial vascular injury and perivasculitis with lymphocytes, monocytes, and macrophages had been described in the CNS of murine typhus cases (14,15). In our study, immunohistochemistry confirmed the histopathologic findings of all these old reports. We also differentiated the T-cell subsets, showing that most T cells are CD8 positive. Using histopathology and immunohistochemistry, we found no obvious cell composition differences between nodules positive and those negative for R. typhi DNA. Thus, a similar histologic architecture and composition of typhus nodules in murine and epidemic typhus can be assumed.

The 7 cases examined in our study represent a minor fraction of the typhus cases seen in Langenhorn Hospital. The hospital had opened a ward specifically for typhus patients in May 1942, and 320 cases were seen during the following 2 years (16). Forced laborers from Russia who had typhus had also been admitted to this hospital (17).

Murine typhus is often a mild illness but can become more severe in refugee camps (18), and fatal and severe CNS cases of murine typhus have been described (14,19,20). Murine and epidemic typhus can be discerned neither histopathologically nor clinically (except perhaps by their expected respective severities and associated mortality rates, the presence of vectors, or flea or louse bite reactions on the skin). Eschars are usually absent in both forms of typhus. Serologic discrimination of the 2 species by cross-absorption and Western blotting (21) and molecular differentiation are confined to reference laboratories and only became available half a century after the cases were investigated.

The natural reservoir for R. typhi is rats, and unhygienic conditions that occur during times of civil disturbance promote the expansion of rat populations and their fleas and, thus, also the likelihood of persons acquiring murine typhus. Fleas remain infected for life, and their lifespan and feeding behavior are not altered by R. typhi infection (22). However, the R. typhi bacterium has also been shown experimentally to replicate in the body louse (22). Similar to infection with R. prowazekii, lice infected with R. typhi become red due to the rupture of their epithelial gut linings caused by the multiplication of rickettsiae, which reduces their lifespan. Thus, we speculate that R. typhi might also be transmitted by body lice under conditions of civil disturbance previously shown to favor the spread of R. prowazekii.

In conclusion, we show evidence that R. typhi played a role during local outbreaks or epidemics of typhus that were classically attributed to R. prowazekii. Therefore, more cases of typhus should be investigated molecularly to determine the type of rickettsial pathogen involved. Whether R. typhi was transmitted during World War II by lice remains to be elucidated.