Introduction

In June 1915, a British military doctor on the western front of the First World War reported the first case of recurrent fever in a soldier who presented with a headache, dizziness, and severe pain in the lower back and leg. Additional cases with the same clinical features were reported among soldiers in the trenches and it began to be referred to as “trench fever” [1]. The causative agent, Rickettsia quintana, was isolated from a patient in Mexico City by Vinson in 1961 and was reclassified in the genus Rochalimaea. In 1993, Rochalimaea merged with Bartonella [2, 3]. The vector for transmission is the human body louse (Pediculus humanus corporis), the gastrointestinal tract of which is colonised by B. quintana. It is shed in lice faeces, and penetrates the human body through damaged skin, entering the bloodstream [4, 5]. The typical cycle of fever occurred at five-day intervals (hence its other name, the “five-day fever”), resulting in prolonged disability so that affected soldiers were unfit for at least two months, many of them suffering from chronic fatigue [2]. Morbidity rates were not revealed by the authorities at the time, and no deaths were reported [2, 3]. The clinical manifestations of B. quintana infection range from asymptomatic to a severe, life-threatening infection. These include, in addition to trench fever, lymphadenopathy, bacillary angiomatosis, and chronic bacteraemia and endocarditis [5, 6]. In the absence of any appropriate treatment, the potential for relapse is due to the existence of an intraerythrocytic phase, as erythrocytes host bacteria during their life cycle [7]. Since the 1990s, B. quintana has been recognised as a re-emerging agent in homeless populations as a result of unsanitary living conditions [8–11].

The presence of trench fever can be traced back to the military in Europe before the First World War, when B. quintana DNA sequences were found in dental pulp from Napoleonic soldiers in Vilnius (1812) [12], and in bone from soldiers in Kassel (1813–814) [13]. Indeed, dental pulp with high vascularity has been the definitive key to diagnosing bacteraemia caused by microorganisms in paleomicrobiology [14]. Molecular biology has confirmed B. quintana bacteraemia occurred between 200 and 4,000 years ago, through the identification of B. quintana in dental pulp [15]. There has been no comprehensive study of the prevalence of the pathogen over the past two millennia in Europe. In order to perform such a study, we reviewed all paleomicrobiological data regarding B. quintana and combined this with the study of 400 additional specimens collected from 145 individuals over 20 centuries.

Materials and methods

Results

A total of 400 ancient teeth from 145 individuals collected in nine European sites with different times of burial were analysed using qPCR. All negative controls were negative, and qPCR detected B. quintana DNA in all sites. The results indicated that a total of 78 civilians from six sites had 17.9% positivity, and 67 soldiers from three sites had 20.1% positivity. The positive detection at the Besançon site was the oldest, with data from the 1st– 4^th centuries (Table 1). Seven of the 17 archaeological sites had data from the 18th– 19^th centuries (Tables 1 and 2), where the prevalence of B. quintana DNA was 3/16 (18.8%) among civilians and 24/120 (20%) among soldiers; the difference in the prevalence was not significant (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Biomolecular results of B. quintana infection in ancient dental pulp.

https://doi.org/10.1371/journal.pone.0239526.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Bartonella quintana detection in ancient specimens from previous studies.

https://doi.org/10.1371/journal.pone.0239526.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Comparison of infected populations of 18^th–19^th centuries.

https://doi.org/10.1371/journal.pone.0239526.t003

Discussion

The results reported here were authenticated by the fact that we chose ancient teeth, preferentially monoradicular teeth with a close apex, no traumatic lesions, and an absence of dental caries that helped minimise any risk of external contamination [16]. A positive control was not used in the PCR experiments because it could be a source of contamination; the negative controls remained negative [18]. To screen a total of 400 teeth, we used real-time PCR as previously described to identify B. quintana [18–20], and in this work, each experimental step was performed in different rooms of a new building, where these gene sequences had not previously been used. The number of teeth per individual was variable in each burial site due to availability, and we used as many teeth per individual as possible to increase the chance of detection. Indeed, the first report of using dental pulp for ancient septicaemia diagnosis indicated that the number of teeth per individual infected with Yersinia pestis was 1/4 teeth, 1/2 teeth, and 3/3 teeth [14]. Furthermore, two teeth from one individual reported that the first was infected with Y. pestis and the second had co-infection of Y. pestis and B. quintana [18]. Accordingly, in several instances, co-detection of B. quintana with another deadly pathogen has been reported, such as Yersinia pestis, Rickettsia prowazekii, indicating that the detection of B. quintana does not preclude that of other pathogens in any archaeological site [12, 18, 23].

In 1915, the history of trench fever was marked by the report of new clinical features from soldiers involved in trench warfare [2, 3]. In 2005, authentic evidence showed that B. quintana had caused bacteraemia in humans for more than 4,000 years, through analysis of DNA from dental pulp collected from Peyraoutes in France [21]. This pathogen was also later identified in this specimen from different burial sites, including Venice (15^th–16^th centuries) [19], Douai (18^th century) [20], Vilnius (1812) [12] and Bondy (11^th–15^th centuries) where an individual had coinfection with Y. pestis [18]. In addition to dental pulp, B. quintana was detected in ancient bone (1813–1814) [13] and coprolites (14th) century [22]. The bacterial confirmation in ancient pulp indicates that individuals had bacteraemia before their death, but it may not have been the cause of death, supporting the fact that no deaths have been recognised as being caused by trench fever [2, 5, 24]. A survey of 930 homeless people in Marseille revealed that 5.3% were blood culture positive for B. quintana [8], and asymptomatic chronic bacteraemia could be maintained for 78 weeks [25]. In 2004, B. quintana was found in the dental pulp of a homeless patient who had had bacteraemia in the previous six months [26].

Most of the reports of B. quintana in dental pulp come from our laboratory, with nearly half of burial sites being located in the 18^th–19^th centuries; 7/14 sites were in France, 4/14 were in Italy, and the rest came from Ukraine, Russia, and Lithuania. This can be explained by the availability of samples in these periods, and the conditions in which our research team in France operate made it highly favourable to receiving these teeth from archaeological centres. The wide geographical and temporal distribution suggests that this bacterial infection was common in historic European populations. In the medical literature, the “five-day fever” and the Moldavia fever in the 19^th century had clinical signs similar to trench fever but laboratory techniques to confirm this were lacking at that time [27].

Millions of soldiers were contaminated around the world during the two World Wars. After the end of each war, however, incidence dropped dramatically but still sporadically persisted in some countries [5, 24]. It was noted that Napoleonic soldiers buried in Vilnius, Lithuania (1812) were exposed to body lice containing B. quintana, and approximately 20% and 8.6% of soldiers were infected with B. quintana and R. prowazekii, respectively [12]. An investigation of ancient bones from 18 soldiers from a mass grave inhumed in the winter of 1813–14 in Kassel, Germany, revealed that 16.7% of the soldiers had the infection [13]. In our study, only 6.7% of the soldiers from the Kaliningrad site tested positive, in contrast to the higher positive rates from Dax and Sevastopol. This can be explained by the use of only one tooth per individual for the Kaliningrad samples. The catacombs in St. Lucia, which represented the ancient Christian monument of the Late Roman period [28], had the similarity of also using one tooth, with the lowest positive proportion (3.6%). During wartime, soldiers were generally crowded close together in unsanitary conditions for prolonged periods of time. Moreover, despite knowing the role that lice played in transmission and their presence in clothes, there were no effective methods of disinfection available [2, 3].

In order to ensure we were working with a uniform time period, we chose the 18^th–19^th centuries, which had a greater number of both populations. The incidence in civilians and soldiers was 18.8% and 20%, respectively, and there was no significant difference. This suggests that this bacterial infection was relatively common throughout human society, without any detectable difference in lice infestation and clothing hygiene, although the sample size of citizens was small (18 citizens), which could affect the comparison. The epidemiology of trench fever involves interhuman transmission via vectors, such as body louse infection, or other transmission that may involve contact with reservoirs, insects bites [29], cat bites [30] or others that have yet to be defined and need further elucidation. Humans are not the only reservoirs of B. quintana; it is also found in cat fleas [31] and monkey fleas [32], and some studies have indicated other animals as known reservoirs through the detection of B. quintana in domestic cats [33], dogs [34], rhesus macaques [35, 36], and Japanese macaques [37]. Following an experimental study on cat fleas, it was shown that this pathogen was absorbed via the gastrointestinal tract and released into faeces [38]. Another study showed that the Pedicinus obtusus louse was postulated as an efficient vector of transmission between rhesus macaques [36]. Homelessness was defined as a high-risk factor for B. quintana infection as a consequence of inadequate hygiene [8, 10, 11, 25]; however, some subpopulations have relatively high exposure rates, such as blood donors, with a rate of between 27% and 51% [9, 39, 40], and healthy people, with a rate of between 11.2% and 25% [40–43]. There is no evidence to show that these people lived in unhygienic living conditions, such as being homeless, or that they had come into contact with body lice, suggesting the existence of some underestimated factors that cause a bacterial infection. The new findings on transmission, transmitted vectors, reservoirs, and widely infected populations contribute to the understanding of the infection from the past to the present; paleomicrobiological studies as this one helping to clarify complex host-pathogen interactions in past human populations, as previously reported [44, 45].

Conclusions

Dental pulp is conducive to the investigation of B. quintana bacteraemia in ancient populations, and it is relatively easy to collect and is well protected inside the tooth. Previous studies analysed in this study showed the presence of B. quintana infection 4,000 years ago and between the 1^st and 19^th centuries. Soldiers are considered the main target of this bacterial infection in human history, but in this study, we saw no significant difference between civilians and soldiers during the 18^th and19^th centuries.

Supporting information

Acknowledgments

Our thanks go to Michael Decker (University of South Florida, Department of History, Tampa, Florida, USA) for his information on the samples. Thanks also go to the Pontifical Commission for Sacred Archaeology–Inspectorate for the Catacombs of Eastern Sicily for the permission to work on other samples from the Catacombs of Saint Lucia of Syracuse, Italy.