Introduction

As developing countries work to eliminate malaria, it is imperative to describe which pathogens are responsible for febrile illnesses that have historically been misdiagnosed as malaria [1]. In the Moshi area of Tanzania, arboviruses (chikungunya and dengue [2]), bacteria (Coxiella burnetii, spotted fever and typhus group rickettsiae) [3] and other zoonotic pathogens (leptospirosis [4], histoplasmosis [5], and brucellosis [6]) were identified as the causes of fever in patients. Studies like these are critical to support the development of diagnostic algorithms to account for other local causes of fever as malaria is eliminated [1], [7]. In Southern Africa, four countries (Namibia, Botswana, South Africa, and Swaziland) are actively engaged in malaria elimination campaigns [8].

In Namibia, the strategy to eliminate malaria has achieved infection rates close to pre-elimination levels and is on track to eliminate malaria in the country by 2020 [9]. With malaria mapped to specific ‘hot spots’, active surveillance has been implemented in the areas of highest incidence with rapid diagnostic tests (RDT) required for all fever cases [10]. The RDT results indicate that almost 90% of the fever cases in these ‘hot spots’ are linked to pathogens other than Plasmodium spp. [9]. At least 15 zoonotic pathogens capable of causing febrile illness have been historically reported in Namibia’s 13 administrative regions [11]. Pathogenic bacteria, notably C. burnetii, Bartonella henselae, as well as, spotted fever and typhus group rickettsiae were highlighted as potentially posing the greatest risk to animal and human health [11]. These zoonotic bacteria have been increasingly noted throughout sub-Saharan Africa, and there is a need to specifically focus on their epidemiology [12]–[14].

Coxiella burnetii, the causative agent of Q fever, is potentially transmissible via blood transfusion [15], [16] or even by ticks [11]. Reported to affect pregnant woman, Coxiella burnetii can cause foetal loss and affect intrauterine growth of the foetus [13]. Spotted fever and typhus group rickettsiae are also common causes of febrile illness in sub-Saharan Africa [17]–[22]. Transmitted mainly by ticks and fleas [23], the pathogenesis of rickettsial infections can involve a spectrum of conditions, from asymptomatic infections to death [18]. Bartonella henselae (cat scratch disease) is usually associated with the bite of an infected flea [14], [24] and has been reported to affect immune-compromised persons such as HIV-infected persons or those with heart valve issues [25]–[27]. Because of the undifferentiated symptoms often associated with Q fever or rickettsial infections, a lack of specific diagnostic tests frequently has led to misidentification with each other [28] or arboviruses [29]–[31].

Rickettsial diseases associated with spotted fever and typhus group rickettsiae as well as Q fever, and cat scratch disease have been recorded in neighbouring countries (South Africa [17]; Botswana [32]; Zambia [33]; Angola [34] and Zimbabwe [35]. However, virtually nothing is known about their presence in Namibia [11]. A single small serosurvey in northern Namibia in the mid-1980s reported a 2% seroprevelence for C. burnetii and evaluated the presence of rickettsiae, but the diagnostics used were not conclusive [36]. No studies have evaluated the presence of B. henselae in Namibia.

To evaluate possible zoonotic bacteria involved with non-malaria febrile illness in Namibia, this study assessed the current exposure of a healthy volunteer cohort to these four zoonotic bacteria and attempted to determine possible risk factors through a questionnaire. Volunteer blood donors are commonly used to evaluate exposure to bacterial pathogens [34], [37]–[40], however, prevalence estimates from these studies are considered conservative because blood donors are normally younger, healthier and screened for other significant pathogens [41], [42]. Volunteer blood donor samples are often used as controls in studies evaluating occupational risk of exposure to zoonotic pathogens [43]–[45].

Methods

Results

Study population

Serum samples (n = 319) were collected by NAMBTS, but 14 did not have accompanying surveys and were excluded. While not all 305 samples were tested for all four pathogen groups, each sample was tested for exposure to at least one pathogen in the study. The mean age of donors was 32.2 years±11.3 years (range 18–64). Almost half of the donors were under the age of 30 (46.1%) and 86.8% of the donors were from the central regions of Namibia (Otjozondjupa, Erongo, Khomas, Omaheke) (Table 1). Eleven of Namibia’s 13 regions were represented in the study sample, even if by only one specimen (Caprivi, Ohangwena, Omusati). No specimens were available for Kavango and Karas regions. Most of blood donors (70.1%) currently lived, or had lived, in rural area or on a farm with exposure to animals in the past 10 years. The majority (81.4%) of those tested lived in urban settings and interacted regularly with animals (97.3%) (domestic and companion animals).

Download:

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

Table 1. Characteristics of the Namibian study population and risk factors associated with exposure to C. burnetii, spotted fever and typhus group rickettsiae.

https://doi.org/10.1371/journal.pone.0108674.t001

Discussion

This was first study since Namibian Independence (1990) to describe exposure to C. burnetii, spotted fever and typhus group rickettsiae and B. henselae in a cohort of healthy Namibians. The results are comparable with past studies in the Southern African region. The observed 26% prevalence for C. burnetii is comparable with the 37% (n = 494) reported in Zimbabwe [35] but quite different that the 2% (n = 113) from Angola [34]. The spotted fever group rickettsiae prevalence of 12% among the donor samples was considerably lower than the 55% (n = 276) reported in Zimbabwe [36], 48% in Angola [34], but similar to the 17% (n = 377) reported in Zambia [33]. The 15% prevalence to R. typhi can only be compared with 5% reported in Zambia [33] and 0% in Angola [34]. Recent B. henselae prevalence rates in South Africa were recently reported at 23% (n = 382) in HIV-infected persons and 10% (n = 42) in clinically healthy volunteers [49] which is considerably higher than the 3% found in the current study.

The global distribution of C. burnetii and its effect on human and animal health is well reviewed [13], [50], [51]. Because zoonotic disease is usually spread to humans via infected livestock [13], the significant regional trend observed between northern (15%) to southern (52%) areas in Namibia could be important as a guide to future public health or disease prevention and control efforts. While the majority of samples were from central regions, the high seroprevelence among apparently healthy blood donors from the southern Hardap region (9 of 15– most of them youth) could indicate a significant transmission zone in Hardap. This finding could potentially impact human and animal health and guide future public health interventions.

Like numerous studies worldwide [40], [43], [52], persons who worked regularly with animals (OR = 2.0), particularly cattle, donkeys and horses (OR = 2.2), may be more at risk in Namibia. Coxiella burnetii has a broad transmission potential especially among occupations that involve inhalation of small cell variants and aerosols or ingestion of unpasteurized milk [13]. While generally noted that chronic C. burnetii infections in cattle, sheep and goats may cause reproductive issues such as abortion, premature, dead or weak offspring, cattle are more often affected than other livestock species [51]. Because most recorded livestock cases in Namibia have occurred in goats [53], the findings of the present study could further focus public health interventions to prevent animal and human infections.

It was surprising to find a significant association between exposure to C. burnetii and a pet recently diagnosed with ‘tick-bite fever’, most likely due to Ehrlichia canis, Hepatozoon canis, or Babesia canis vogeli [54]. While the effect of C. burnetii on the reproductive health of companion animals is not well defined [51], dogs and cats are known reservoirs for C. burnetii in Zimbabwe [35], [55]. Future studies should follow up this possible link between companion animal health and C. burnetii exposure as well the possibility of cross-reactivity to Bartonella species [56].

Rickettsia africae has only been reported in Namibia from unspecified tourist travel studies [18]. While its presence in neighbouring countries indicates a probable presence in Namibia [17], [19], [33], no study has yet addressed the likelihood of spotted fever group rickettsiae among Namibians. Gender was the only significant risk factor involved with males (OR = 3.3) more likely to be exposed as they are possibly more active outdoors on farms or rural areas.

While the study did not differentiate between rickettsial species, spotted fever rickettsiae are normally transmitted by ticks or fleas, of which eight potential vector species have been recorded in Namibia [11]. Rickettsia felis, a spotted fever group rickettsiae transmitted by cat or rat fleas [57], remains to be identified south of Angola, Zambia and Mozambique. To date, the most southerly report in sub-Saharan Africa is from the Democratic Republic of Congo [58]. Given the long history of ports in Southern Africa and the association of R. felis with rat and cat fleas, it is likely a matter of being underreported. The results of this study demonstrate the presence of spotted fever group rickettsiae in Namibia but indicate the need to follow up with molecular identification in order to establish the current epidemiology and confirm the southern to northern trend in exposure which could involve the species composition of vector ticks.

Like spotted fever group rickettsiae, 15% of samples were positive for typhus group rickettsiae, with a significant association with males under 30 years of age, the majority of whom did not regularly work on farms or hunt. The two principal Rickettsia spp. possibly involved are R. prowazekii (louse-borne) or R. typhi (flea-borne) [57]. While R. prowazekii (endemic typhus) is probably not present due to microhabitat conditions and lack of non-migrating populations, R. typhi is most probable. The question, however, of how a pathogen normally found in urban or port cities [57] could be prevalent in a dry, arid climate in Central Namibia needs follow up. The 7–10% seroprevalence of R. typhi among cats in similar climates of Zimbabwe and South Africa indicates that climate may not significantly limit the epidemiology [59]. The observed north-south regional trend observed may not be important, either, as Wessels et al. [36] observed a low agglutination response in 59% of patients in the northern Kavango region, indicating possible exposure to typhus group rickettsiae [11]. Compared with other countries in Southern Africa, the relatively high prevalence among Namibian blood donors demonstrates a need for further follow up on this important pathogenic group.

This is the first report of possible exposure to B. henselae in Namibia. The low seroprevalence (3%) among blood donors is lower than other studies among healthy volunteers in South Africa (10%) [49]. Also known as ‘cat scratch disease’, Bartonella henselae appears to be common in Southern Africa with prevalence rates of 21–23% reported among domestic and wild felines [60] and 14% in dogs [61] in Zimbabwe and South Africa. While the low prevalence in the study population indicates the presence of Bartonella spp. in Namibia, lack of exposure to infected fleas or the dry, arid climate may reduce actual risk among residents in the central regions. Future studies should evaluate the presence of B. henselae in fleas, companion animals, and potentially patients with undifferentiated febrile illness in the coastal areas around the cities of Luderitz, Walvis Bay and Swakopmund. As in other countries, cross-reactivity could have occurred with other Bartonella spp. that may be present in the region [17] as well as C. burnetii [48].

One of five samples was positive for exposure to two or more pathogens. These results probably indicate that persons tested had either experienced an active infection or had an exposure to the pathogen which was significant enough to stimulate antibodies. While the ELISA kits used were based on antigens which differentiated between spotted fever and typhus group rickettsiae [62], [63], cross-reactivity has been recorded among spotted fever rickettsiae species [64] and between Bartonella and C. burnetii [56].

The exposure to bacterial pathogens among volunteer blood donors was a significant component in this study as it may be important for the safety of the national blood supply. While it is unlikely that spotted fever and typhus group rickettsiae would survive the blood storage process [65], B. henselae [15] and C. burnetii [16] are known to survive in donated blood units and can be transmitted via transfusion. A transfusion of C. burnetii into an immunocompromised person, either pregnant or HIV-infected (19% HIV-infection rate in Namibia [66]), could result in an abortion in the pregnant woman or death of the HIV-patient.

One exciting aspect of this study was the use of a Primo-star iLED scope for determination of IFA results. The Primo-star iLED microscope was designed for research laboratories in resource-limited settings, where cost, efficiency, and ease of use are important [67]. This microscope has been shown to have equivalent accuracy to that of international standards and is now recommended by the World Health Organization as a standalone option for field diagnosis of tuberculosis [47].

In regards to limitations, it is difficult to identify true regional trends and establish risk factors from a small, non-random sample. Also, because most Namibian blood donors are males with farm/rural experience from the central regions, it was not possible to gather a more heterologous population for a true population-based epidemiological assessment. The cross-sectional approach to use only one sample from each donor meant that the benefits of comparing acute and convalescent samples from febrile inpatients were not present [3]. While these limitations require a broader, population-based investigation in order to potentially impact public health policy, the study demonstrated that these pathogens are present among healthy persons and raises the profile of the issue of vector-borne bacterial pathogens as a source of non-malaria fevers. Without Western Blot, molecular confirmation of DNA present in the blood, or cultivation of viable pathogen from specimens, this study can only state broadly that Namibians are being exposed to these groups of pathogenic bacteria. Additionally, samples were only screened at 1∶16 and 1∶256 titers for antibodies to C. burnetii, and 1∶64 for B. henselae but end point titers were not determined. While indicating a rather high seroprevalence of exposure to C. burnetii, future work is needed to fully characterize the serologic status of the population in regards to possible acute and chronic infections.

Conclusions

This study establishes that Namibians, particularly those involved in farms or rural life, are being exposed to pathogenic, vector-borne bacteria. Given Namibia’s priority to eliminate malaria [9], it will not be possible to upgrade national diagnostic algorithms to account for other causes of febrile illness without active clinical surveillance. While most of these bacteria can be identified via symptoms and simply treated using tetracycline-class antibiotics, it is critical for health professionals to begin differentiating between these four types of pathogens. The possibility of missing an acute infection that could become chronic could cause serious complications for either the patient (R. conorii) [29], the foetus of an infected pregnant woman [13] or a person receiving a blood transfusion (C. burnetii) [16]. The low prevalence of B. henselae among blood donors may actually be a positive finding, indicating that this pathogen may not be as important as Q fever and rickettsiae in the health of Namibians.

Knowing that healthy Namibians are encountering these pathogens, future studies should concentrate on three main areas: 1) Determine the frequency of febrile illnesses caused by vector-borne bacterial pathogens; 2) Further describe the occupational risks factors for different professions, such as farmers, abattoir workers, and veterinarians, and; 3) Systematically map the occurrence of each pathogen in humans and animals, determine the risk of transmission by different vector species, and develop public health interventions to reduce vector populations or transmission. Without vectors, most of these bacterial pathogens (except C. burnetii) will not spread [14].

Acknowledgments

The research team would especially like to thank the blood donation team at NAMBTS for their hard work in recruiting volunteers for this study as well as processing and storing the samples. Special thanks to Mr. John Pitman for editorial assistance.