Hostname: page-component-cd9895bd7-fscjk Total loading time: 0 Render date: 2024-12-27T07:13:16.722Z Has data issue: false hasContentIssue false

Louse-borne relapsing fever (Borrelia recurrentis infection)

Published online by Cambridge University Press:  01 March 2019

David A. Warrell*
Affiliation:
Nuffield Department of Clinical Medicine, University of Oxford, Oxford, UK
*
Author for correspondence: David A. Warrell, E-mail: david.warrell@ndm.ox.ac.uk
Rights & Permissions [Opens in a new window]

Abstract

Louse-borne relapsing fever (LBRF) is an epidemic disease with a fascinating history from Hippocrates’ times, through the 6th century ‘Yellow Plague’, to epidemics in Ireland, Scotland and England in the 19th century and two large Afro-Middle Eastern pandemics in the 20th century. An endemic focus persists in Ethiopia and adjacent territories in the Horn of Africa. Since 2015, awareness of LBRF in Europe, as a re-emerging disease, has been increased dramatically by the discovery of this infection in dozens of refugees arriving from Africa.

The causative spirochaete, Borrelia recurrentis, has a genome so similar to B. duttonii and B. crocidurae (causes of East and West African tick-borne relapsing fever), that they are now regarded as merely ecotypes of a single genomospecies. Transmission is confined to the human body louse Pediculus humanus corporis, and, perhaps, the head louse P. humanus capitis, although the latter has not been proved. Infection is by inoculation of louse coelomic fluid or faeces by scratching. Nosocomial infections are possible from contamination by infected blood. Between blood meals, body lice live in clothing until the host's body temperature rises or falls, when they seek a new abode.

The most distinctive feature of LBRF, the relapse phenomenon, is attributable to antigenic variation of borrelial outer-membrane lipoprotein. High fever, rigors, headache, pain and prostration start abruptly, 2–18 days after infection. Petechial rash, epistaxis, jaundice, hepatosplenomegaly and liver dysfunction are common. Severe features include hyperpyrexia, shock, myocarditis causing acute pulmonary oedema, acute respiratory distress syndrome, cerebral or gastrointestinal bleeding, ruptured spleen, hepatic failure, Jarisch–Herxheimer reactions (J-HR) and opportunistic typhoid or other complicating bacterial infections. Pregnant women are at high risk of aborting and perinatal mortality is high.

Rapid diagnosis is by microscopy of blood films, but polymerase chain reaction is used increasingly for species diagnosis. Severe falciparum malaria and leptospirosis are urgent differential diagnoses in residents and travellers from appropriate geographical regions.

High untreated case-fatality, exceeding 40% in some historic epidemics, can be reduced to less than 5% by antibiotic treatment, but elimination of spirochaetaemia is often accompanied by a severe J-HR.

Epidemics are controlled by sterilising clothing to eliminate lice, using pediculicides and by improving personal hygiene.

Type
Review
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
Copyright © The Author(s) 2019

Introduction

Louse-borne relapsing fever (LBRF) is a classic epidemic disease, associated with war, famine, refugees, poverty, crowding and poor personal hygiene. After a long history, recorded over many centuries, it is now largely confined to the Horn of Africa, while retaining its potential to cause future epidemics when conditions become conducive. It was a familiar infection in Europe and North America until the end of the 19th century after which it was forgotten. However, the recent surge of refugees from Africa arriving in European countries has brought this fascinating disease back into the view of the medical profession and has stimulated new research into its cause, Borrelia recurrentis, and its vector, the human body louse.

Aetiology [Reference Felsenfeld1]

LBRF is caused by B. recurrentis, a large, loosely coiled, motile spirochaete (family Spirochaetaceae, that also includes Treponema), with tapering ends, 12–22 μm long and 0.2–0.6 μm thick, with an average wavelength of 1.8 μm, an amplitude of 0.8 μm and 8–10 periplasmic flagella [Reference Cutler2]. They divide by transverse binary fission. B. recurrentis can be cultured on chick chorioallantoic membrane, and maintained in rodents [Reference Felsenfeld1]. Strains of immunodeficient mice (SCID lacking B and T cells, and SCID BEIGE lacking B, T, and NK cells) have been proposed as an animal model of LBRF [Reference Larsson3]. B. recurrentis can be cultured in vitro using Barbour-Stoenner-Kelly (BSK-II) medium [Reference Cutler4], BSK-H supplemented with heat-inactivated 10% rabbit serum and modified-Kelly-Pettenkofer (MKP) medium supplemented with 50% fetal calf serum [Reference Marosevic5]. BSK medium supports rapid initial borrelial growth but this is followed by cell deformation and death, whereas MKP medium appears to improve isolation rate, morphology and motility [Reference Ružić-Sabljić6].

Unlike other bacteria, borreliae have a fragmented genome consisting of a linear chromosome, 1–15 linear plasmids and 1–9 circular plasmids. B. recurrentis has the simplest genome of all, composed of one linear chromosome and only seven linear plasmids, and only 990 protein coding genes. It shows low genetic variability [Reference Marosevic5]. Genomes of B. recurrentis and B. duttonii are identical except that in B. recurrentis 30 genes or gene families of B. duttonii are either absent or damaged. This has been cited as evidence that B. recurrentis has a decaying genome and is only a strain or subset of B. duttonii that adapted rapidly to louse-transmission with genome reduction [Reference Lescot7]. B. recurrentis lacks RecA and RadA proteins that are responsible for DNA repair. The average nucleotide identity between the African borreliae, B. crocidurae, B. duttonii and B. recurrentis, is 99%, suggesting that they are merely ecotypes of the same genomospecies ‘B. africana’ [Reference Elbir8].

Transmission

Unlike most borreliae, transmission of B. recurrentis is restricted to one vector, the human body louse Pediculus humanus corporis, and, perhaps, the head louse P. humanus capitis. Although B. recurrentis has been identified in head lice, including those infesting pygmies in the Republic of Congo, outside the currently recognised geographical distribution of LBRF [Reference Amanzougaghene9], transmission by them has not yet been confirmed. Body lice, unlike head lice, retreat from the skin after feeding to hide and lay their eggs in clothing seams rather than on hair shafts. In Addis Ababa, one old man was found to be harbouring more than 21 500 lice in his clothes [Reference Sholdt, Holloway and Fronk10]. Lice are obligate haematophagous human ectoparasites that ingest borreliae in their blood meal [Reference Sangaré, Doumbo and Raoult11]. They are intolerant of deviations in human body temperatures caused by fever, climatic exposure or death, or when infested clothing is discarded. Then, they find a new host to whom borreliae can be transmitted. Coelomic fluid from a crushed louse, or louse faeces infected with B. recurrentis, is inoculated through broken skin, or intact mucous membranes such as the conjunctiva, by scratching. Blood transfusion, needlestick injuries and contamination of broken skin by infected blood are potential causes of nosocomial infections [Reference Bryceson12]. Since lice, unlike ticks, cannot infect their progeny, they do not act as reservoirs. Transplacental infection has been confirmed in a mouse model of B. duttonii infection [Reference Larsson13] and there are reports of congenital infection by B. hermsii and other tick-borne spirochaetes [Reference Fuchs and Oyama14]. There is no known animal reservoir, and so persistence of infection between epidemics can only be through mild or asymptomatic human infections.

Epidemiology and historical background

Human disasters created by war, forced migrations, poverty, famine, breakdown of personal hygiene and seasonal spells of cold, wet weather, promote crowding and increase the risk of infestation by body lice and the transmission of LBRF, louse-borne typhus, trench fever and other louse-borne diseases. LBRF can be identified in historical descriptions of disease epidemics by the repeated recurrences of fever between asymptomatic periods of 4–7 days and by two typical symptoms, jaundice and bleeding. The earliest convincing description of this disease was given by Hippocrates in the 5th century BC in the North Aegean island of Thasos: ‘The great majority (of sufferers) had a crisis on the sixth day, with an intermission of six days followed by a crisis on the fifth day after the relapse.’ Other features typical of LBRF were severe rigors, jaundice, profuse epistaxes and tendency to precipitate abortion [Reference Lloyd15, Reference MacArthur16]. MacArthur has argued convincingly that the ‘Yellow Plague’ that engulfed Europe in 550 AD, in the wake of the Justinian plague, and the famine fevers of the 17th and 18th centuries in Ireland and elsewhere, whose defining feature was jaundice, were predominantly LBRF [Reference MacArthur16].

Recently, a historical genome of B. recurrentis was recovered from the skeleton of a young woman found during the excavation of a graveyard near St. Nicolay's Church in Oslo. Radiocarbon dating suggested that its age was AD 1430–1465. The mediaeval European genome displayed an ancestral oppA-1 gene, and gene loss in antigenic variation sites (variable short and long membrane protein genes) that translated into a genome reduction of 1.2% of the pan-genome, and 5.1–21% of the affected plasmids, perhaps associated with increased virulence but a reduced number of relapses [Reference Guellil17].

In Dublin in 1770, Rutty described ‘a fever altogether without the malignity attending (typhus), of six or seven days duration, terminating in a critical sweat…in this the patients were subject to a relapse, even to a third or fourth time, and yet recovered’ [Reference Rutty18]. In Edinburgh in 1843, Craigie distinguished LBRF from typhus and coined the name ‘relapsing fever’ [Reference Craigie19]. Henderson detailed the differences between the two infections [Reference Henderson20]. In Britain, in the 19th century, LBRF featured prominently in Charles Murchison's treatise on continued fevers. He commented on its intermittent appearance and truly epidemic nature: ‘So completely did relapsing fever disappear from Britain after 1828 that when, after an interval of fourteen years, it again showed itself as an epidemic in 1843, the junior members of the profession failed to recognize it and it was regarded as a new disease’ [Reference Murchison21]. Obermeier saw spirochaetes, now recognised as B. recurrentis, in the blood of febrile patients in Berlin in 1866 [Reference Wright and Maria22]. Transmission by human body lice was proved by Mackie in 1907 [Reference Mackie23].

In the 20th century, from 1903 to 1936, a huge pandemic swept across North Africa, the Middle East and Africa, causing an estimated 50 million cases with 10% mortality. A second epidemic in 1943–46 created 10 million cases [Reference Sangaré, Doumbo and Raoult11].

An endemic focus persists in the Horn of Africa [Reference Sparrow24]. In cold, wet weather, impoverished people with louse-infested clothes crowd together for warmth and shelter. These indigent, malnourished street-dwellers, day workers (casual labourers), usually young men and prisoners, are the most vulnerable to infection. In the Ethiopian highlands there are annual epidemics of thousands of cases coinciding with the rains. Outbreaks have also occurred in Somalia. In Rumbek County, South Sudan, in 1999–2000, there were 20 000 cases with some 2000 deaths, 580 in January 1999 alone [25]. In 1985, in Chavin District if Ancash Province in the Peruvian Andes at altitudes above 3800 m, 60 clinical cases were reported among louse-infested villagers, 36 with B. recurrentis in their blood films [Reference Valdizan, Lopez and Delgado26]. More recently in Calca Province in the Urubamba Valley of Peru, antibodies to B. recurrentis have been found in two of 194 villagers [Reference Raoult27]. The discovery of B. recurrentis in head lice in Congolese pygmies raises the possibility of other undiscovered human reservoirs [Reference Amanzougaghene9].

B. recurrentis infection in African refugees arriving in Europe

Since July 2015, LBRF has been diagnosed in almost 100 mainly young male refugees, who arrived in several European countries, most in Italy and Germany, seeking asylum after travelling from Ethiopia, Eritrea, Somalia and other African countries, usually through Libya [Reference Ciervo28Reference Isenring34]. It is the most frequently reported infection in Eritrean immigrants [Reference Isenring34]. It seems likely that many other cases may have gone undetected and unreported [30]. Most of the patients were from the Horn of Africa, but the duration of their symptoms suggested that the majority had been infected in Libya. However, two cases diagnosed in Turin, Italy were long-term residents who shared accommodation with recently arrived immigrants, suggesting the possibility of autochthonous infection [Reference Antinori31]. Some, from African countries not endemic for LBRF such as Mali, were probably infected during their journey, in crowded transit hostels in Libya, Italy or elsewhere [Reference Grecchi35]. LBRF was largely unknown in Germany and other European countries before its re-emergence after 2015 [Reference Hoch29], but there is now the possibility that it might become re-established in some impoverished and crowded immigrant populations in parts of Europe [Reference Antinori31].

Pathophysiology and pathology

The relapse phenomenon

Attacks of relapsing fever end abruptly when specific bactericidal immunoglobulin M antibodies generated by the B1b cell subset lyse spirochaetes in the blood, independently of complement and T cells. Between relapses, spirochaetes may persist extracellularly in spleen, liver, kidneys, eye and other sites. Relapses, accompanied by spirochaetaemia, are explained by antigenic variation, which has been studied in depth in the North American tick-borne pathogen, B. hermsii [Reference Barbour and Hayes36]. Silent gene sequences from an archive stored in plasmids are transposed to one end of an expression linear plasmid where their recombination leads to synthesis of a new variable major outer membrane lipoprotein (vmp) [Reference Barbour and Hayes36]. The new external membrane allows borreliae to evade the host's humoral immune response until antibodies are generated against the new serotypic vmp antigen, explaining the sequential emergence of borreliae expressing different vmps during the course of an untreated infection. Stimulation of the massive release of tumour necrosis factorα (TNF-α) at the start of the Jarisch–Herxheimer reaction (J-HR) to antibiotic treatment in LBRF is also due to vmp [Reference Vidal37]. B. recurrentis is protected from the host's innate immunity by expressing receptors that selectively bind C4bp and C1-Inh, the endogenous regulators of the classical and lectin complement pathway, HcpA protein that, by binding plasmin, decreases C3b deposition and a specific receptor for the serum-derived complement inhibitor of the alternative pathway, CFH [Reference Grosskinsky38]. These mechanisms allow evasion of lysis by complement activation. Another possible protective mechanism is rosetting of erythrocytes around some Borrelia spirochaetes that masks or excludes them by steric hindrance, from host antibody. However, although B. crocidurae and B. duttonii induce rosetting, B. recurrentis do not [Reference Guo39].

Pathophysiology

The spontaneous crisis that terminates untreated attacks, and the J-HR induced by antibiotic treatment, show pathophysiological features of a classic tri-phasic endotoxin reaction. B. recurrentis outer membrane vmps stimulate monocytes to produce TNF-α through NF-κB [Reference Vidal37]. There are transient marked elevations in plasma concentrations of TNF-α, interleukin (IL)-6, IL-8 and IL-1β [Reference Negussie40, Reference Coxon41]. The massive burst of cytokines is triggered by phagocytosis of spirochaetes opsonised by the antibiotic. Penicillin binding results in large surface blebs on the spirochaetes which are then phagocytosed by neutrophils in the blood and by the spleen. In vitro, surface contact with spirochaetes induces mononuclear leucocytes to produce inflammatory cytokines and thromboplastin, causing fever and disseminated intravascular coagulation [Reference Butler42]. Kinins may be released during the J-HR. The profound leucopenia that develops during the reaction reflects sequestration rather than leucocyte destruction. Spirochaetes may be found in liver, spleen, myocardium and brain. Thrombocytopenia rather than vasculitis causes the petechial rash. Cardiorespiratory and metabolic disturbances result from persistent high fever, accentuated by the J-HR or spontaneous crisis [Reference Warrell43].

Pathology [Reference Judge44, Reference Anderson and Zimmerman45]

Spirochaetes are mainly confined to the lumen of blood vessels, but tangled masses occur in splenic miliary abscesses and (Fig. 1C) infarcts as well as within the central nervous system adjacent to haemorrhages. A perivascular histiocytic interstitial myocarditis, found in the majority of fatal cases, may be responsible for conduction defects, arrhythmias and myocardial failure resulting in sudden death. Splenic rupture with massive haemorrhage, cerebral haemorrhage and hepatic failure are other causes of death [Reference Salih46]. There is hepatitis with patchy midzonal haemorrhages and necrosis, meningitis and perisplenitis. Serosal cavities and surfaces of viscera are studded with petechial haemorrhages and sometimes massive pulmonary haemorrhages, reminiscent of leptospirosis. Thrombi are occasionally found occluding small vessels, but the peripheral gangrene, that is a feature of louse-borne typhus (Rickettsia prowazekii infection) [Reference Perine47, has not been reported in LBRF [Reference Bryceson12].

Fig. 1. Ethiopian patients with LBRF. (A) Profuse petechial rash on the trunk in an emaciated patient with complicating infection with Salmonella enterica serovar Typhi (S. Typhi). (B) Subconjunctival haemorrhages and jaundice indicative of hepatocellular damage, thrombocytopenia and coagulopathy. (C) B. recurrentis spirochaetes arrowed (silver stain) in the splenic pulp. (D) Cerebral haemorrhage on the 6th day of illness, a common cause of death in patients with LBRF.

Symptoms and signs [Reference Bryceson12]

The incubation period is 4–18 (average 7) days. The attack starts abruptly with a fever that increases to nearly 40 °C in a few days, accompanied by rigors. Early symptoms include headache, dizziness, nightmares, generalised aches and pains; affecting especially the lower back, knees and elbows; anorexia, nausea, vomiting and diarrhoea. Upper abdominal pain, cough and epistaxis develop later. Prostration and confusion are the rule. The commonest sign is hepatic tenderness (in about 60%) and enlargement (50%). Splenic tenderness and enlargement are less common. Jaundice is found in seven to >70% of patients [Reference Bryceson12]. A petechial or ecchymotic rash, particularly involving the trunk (Fig. 1A), is seen in 2% to 80% of patients [Reference Bryceson12]. It must be distinguished from the maculo-papular or petechial rash of louse-borne typhus. Subconjunctival haemorrhages (Fig. 1B) and epistaxis (25%) are common, haemoptysis, gastrointestinal bleeding and retinal haemorrhages less so. Many patients have myalgia. Meningism occurs in about 40% of patients. Neurological symptoms are less common than in tick-borne relapsing fevers: cranial nerve lesions, monoplegias, flaccid paraplegia and focal convulsions. Untreated attacks resolve by crisis after 4–10 (average 5) days, followed by an afebrile remission of 5–9 days, succeeded by up to five relapses of diminishing severity, during which there may be epistaxis but no petechial rashes.

Pregnant women are especially susceptible to severe disease and premature labour, and still births are frequent. In tick-borne relapsing fever caused by B. duttonii, intrauterine growth retardation, placental damage and inflammation, impaired fetal circulation and maternal anaemia have been described. Spirochaetes frequently cross the placenta, resulting in congenital infections [Reference Larsson13].

Clinical in refugees arriving in Europe

Clinical findings and results of laboratory investigations in 55 refugees [Reference Antinori48] have been compared with those in 62 Ethiopian patients studied in Addis Ababa in the late 1960s [Reference Bryceson12]. Symptoms such as fever, headache, myalgias, abdominal pain and vomiting were common in both groups, but, among the refugees, jaundice was more often reported (51% vs. 34%). Bleeding (8.9% vs. 23%), meningism (5.5% vs. 39%), J-HRs (62% vs. 100%) and fatalities (1.8% vs. 4.8%) were less common. Levels of C-reactive protein [median 284 (55–440.9) mg/dl] and procalcitonin [13.93 (0.95–62.1) ng/ml] were raised in the refugees [Reference Antinori48]. Twenty percent of the refugees had raised serum creatinine concentrations [2.4 (0.9–4.7) mg/dl], indicating renal impairment [Reference Antinori48]. Body lice were recovered from 22% of the refugees and in others there were scratch marks suggesting infestation [Reference Antinori48].

Children: In Sheshamane, Ethiopia, children younger than 15-years-old were compared with adults [Reference Ramos49]. Clinical features in children resembled those in adults but were generally less severe and less frequent. Headache (40%), dizziness (39%), abdominal cramps (17.4%), vomiting (23.8%), cough (27.6%), musculoskeletal pain (30.5%), petechial rash (1.9%) and bleeding (3.8%) were all less common in the children [Reference Ramos49]. In a later study from the same hospital, fever, headache, dizziness and musculoskeletal pains were said to be the commonest symptoms [Reference Ramos50]. A study of infants and children in Arsi Region, Ethiopia, found that the common clinical features were fever (100%), headache (84.5%), chills (74%), abdominal pain (51%), epistaxis (20%), hepatomegaly (26%), splenomegaly (14%), petechial rash (34%) and jaundice (10%). Pneumonia (14%) and central nervous system involvement (10%) were common complications. J-HRs occurred in 61%. Case fatality was 1.9% [Reference Borgnolo51].

Prognosis

Case fatalities between 30% and 70% have been reported in untreated patients during major historic epidemics, but in treated cases, on average, 2–6% will die [Reference Bryceson12]. In an outbreak in Arsi Zone, Ethiopia in 2016, the case-fatality was 13% [Reference Nordmann52]. Reported case fatalities in children range from 1.9% to 5.5%. In one series of 154 children (<15 years) in Ethiopia, overall case fatality rate was 2.4%, less than in adults (13.2%) [Reference Ramos50].

Severe louse-borne relapsing fever

Clinical features associated with a bad prognosis include coma; shock; hyperpyrexia; myocarditis with acute pulmonary oedema [Reference Parry53]; acute respiratory distress syndrome; hepatic failure; ruptured spleen and haemostatic failure from thrombocytopenia liver damage and disseminated intravascular coagulation leading to intracranial (Fig. 1D), massive gastrointestinal, pulmonary or peripartum haemorrhage [Reference Bryceson12, Reference Warrell43, Reference Ramos50]. Complicating co-infections such as dysentery, salmonellosis, typhoid, typhus, tuberculosis, bacterial pneumonia, visceral leishmaniasis and malaria increase mortality [Reference Bryceson12, Reference Anderson and Zimmerman45].

Spontaneous crisis and J-HR

An impending crisis on about the fifth day of the untreated illness, or a J-HR about 1 to 2 h after antibiotic treatment, is signalled by restlessness and apprehension, followed by distressingly intense rigors lasting 10 to 30 min [Reference Bryceson12, Reference Warrell43]. During this chill phase, temperature, respiratory and pulse rates, and blood pressure rise steeply, with associated delirium, gastrointestinal symptoms, cough and limb pains. Fatal hyperpyrexia may occur. The flush phase, characterised by profuse sweating, a fall in blood pressure, and a slow decline in temperature, may last for many hours. During this period, patients may collapse and die if they stand up or may develop progressive and intractable hypotension, especially if they suffer acute myocardial failure attributable to borrelial myocarditis [Reference Parry53]. Treatment with intravenous tetracycline carries the highest risk of provoking a J-HR, reaching 100% in some studies [Reference Bryceson12]. Low-dose or slow-release penicillin causes fewer reactions but may not prevent relapses (see below). In children, J-HRs are less common.

Laboratory investigations [Reference Bryceson12]

Circulating spirochaete densities may exceed 500 000/mm3 of blood. Patients commonly have a moderate normochromic anaemia with neutrophil leucocytosis. The spontaneous crisis and J-HR are marked by leucopenia. Thrombocytopenia is usual and there is a mild coagulopathy (raised prothrombin time and INR) with evidence of increased fibrinolysis (increased fibrinogen degradation products or D-dimer). Raised serum concentrations of aminotransferases, alkaline phosphatase, direct and total bilirubin and low albumin suggest hepatocellular damage. Mild renal impairment is common. The cerebrospinal fluid shows a lymphocyte or neutrophil pleocytosis without detectable spirochaetes.

Electrocardiographical (ECG) evidence of myocarditis includes prolongation of the QTc interval, T-wave abnormalities and ST-segment depression with transient acute right heart strain after the J-HR and various arrhythmias [Reference Parry53]. Chest radiographs are usually clear but may show pulmonary oedema or pneumonic consolidation.

Diagnosis

Microscopy

The possibility of rapid bed-side diagnosis makes LBRF a satisfying disease for the clinician. Thick and thin blood films should be taken while patients are febrile and stained with Giemsa, May-Grünwald Giemsa, Wright, Wright-Giemsa, Field's, or Diff-Quick stains, or examined under dark-field. Positivity thresholds of thin and thick smear blood are respectively estimated at 105 and 104 spirochaetes per millilitre of blood [Reference Hovette54]. A two-stage centrifugation concentration method has been described [Reference Larsson and Bergström55]. Quantitative buffy coat technique (acridine orange) is also possible. The higher and more persistent spirochaetaemia in LBRF makes microscopic diagnosis more reliable than in the other borrelioses. Exflagellating Plasmodium vivax microgametes may be mistaken for spirochaetes (‘pseudoborreliosis’) [Reference Berger and David56], but microfilariae are far too large to cause confusion.

Polymerase chain reaction (PCR)

An important break-through has been the development of a multiplex real-time PCR (MR-TPCR) method, to differentiate the four main Borrelia species in Africa [Reference Elbir57]. It targets the 16S rRNA gene (detecting all four species); glpQ gene (B. croidurae); recN gene (B. duttonii/B. recurrentis) and recC gene (B. hispanica). The assay has a 100% sensitivity and specificity for B. duttonii/B. recurrentis, but could not discriminate between these two species because of their very close genetic and genomic proximity that suggests they may be a single species [Reference Lescot7, Reference Elbir8]. PCR detected 100 copies, proving to be more sensitive than the 103–105 borreliae/mL visible by microscopy. Among infected immigrants to Europe, PCR has proved a valuable method for confirming the species diagnosis of B. recurrentis [Reference Hoch29Reference Hytönen32], and has detected some microscopy-negative cases [Reference Hoch29, Reference Antinori48]. PCR has been successfully introduced at a point-of-care laboratory in rural Senegal for diagnosis of B. crocidurae infections [Reference Sokhna58].

Serology

Sera from patients with LBRF may give positive reactions with Proteus OXK, OX19 and OX2, which might suggest the diagnosis. False-positive serological responses for syphilis are found in in 5–10% of cases. Serology has generally proved unreliable and non-specific, but it has been improved by the use of the glpQ gene as a recombinant antigen [Reference Porcella59], or monoclonal antibodies (to B. crocidurae) [Reference Fotso Fotso60]. However, serology lacks sufficient specificity, is not commercially available, and may fail to detect acute infections [Reference Porcella59, Reference Fotso Fotso60].

Differential diagnosis

In a febrile patient in or from Africa, who has all the classic features of LBRF – jaundice, petechial rash, epistaxis, hepatosplenomegaly, thrombocytopenia, coagulopathy and elevated serum aminotransferases – severe falciparum malaria is the most urgent differential diagnosis. In the Horn of Africa, yellow fever and other viral haemorrhagic fevers such as Rift Valley Fever and viral hepatitis, rickettsial infections, especially louse-borne typhus which occurs in mixed epidemics with LBRF, must be considered. If there is evidence of acute kidney injury, leptospirosis is more likely. Co-infection of LBRF with leptospirosis was diagnosed by PCR in a refugee from East Africa who arrived in Italy [Reference Cutuli61]. Trench fever (Bartonella quintana), transmitted by lice, can also cause episodic recurrent fever with headache and pains in the shins, but it lacks the bleeding and jaundice of LBRF. In endemic areas, complicating bacterial infections, particularly typhoid, or coinfection with malaria should not be forgotten [Reference Bryceson12, Reference Anderson and Zimmerman45]. In refugees diagnosed in Europe, P. falciparum malaria, sepsis, leptospirosis and meningitis have been cited as leading differential diagnoses [Reference Antinori48].

Treatment

Antibiotics

Complete cure with prevention of relapses is achievable with a single oral dose of 500 mg tetracycline or 500 mg erythromycin stearate. However, vomiting is so common that parenteral treatment is more dependable. A single intravenous dose of 250 mg tetracycline hydrochloride or, for pregnant women and children, a single intravenous dose of 300 mg erythromycin lactobionate (children 10 mg/kg body weight) is effective. In mixed epidemics of LBRF and louse-borne typhus, a single oral dose of 100 mg doxycycline is effective [Reference Perine and Teklu62]. Benzyl penicillin (300 000 units = 80 mg), procaine penicillin with benzyl penicillin (600 000 units) and procaine penicillin with aluminium monostearate (600 000 units), all by intramuscular injection, are often effective but may fail to prevent relapses [Reference Warrell63]. Long-acting preparations clear spirochaetaemia slowly and the J-HR is protracted. Some experienced clinicians prefer to use a low initial dose of penicillin (adult dose, 100 000–400 000 units by intramuscular injection) in severe cases and pregnant women because they believe that the incidence and severity of the J-HRs will be less. In a randomised clinical trial of three regimens of intramuscular procaine penicillin and one of oral tetracycline in Gondar, Ethiopia, the incidence of J-HRs increased with increasing doses of penicillin, from 5.1% with 100 000 units to 31.1% with 400 000 units, and was 46.6% in patients treated with tetracycline [Reference Seboxa and Rahlenbeck64]. Since fatalities (3.3%) were associated with J-HRs, the authors recommended that treatment be initiated with low dose penicillin, despite relapse rates that decreased from 45% to 9.4% from the lowest to highest doses of penicillin. There were no relapses after tetracycline treatment as had been described earlier [Reference Bryceson12]. The combination of penicillin on the first day of treatment, followed by tetracycline on the next day, deserves further study [Reference Salih and Mustafa65]. Chloramphenicol is effective in a single dose of 500 mg by mouth or intravenous injection in adults. It may not be available in some Western countries.

Preventing the J-HR

Antibiotics, such as tetracycline, that rapidly eliminate spirochaetes from the blood and prevent relapses, often induce a severe, and rarely fatal, J-HR. Antibiotic treatment cannot be withheld in view of the high untreated mortality, especially as severe spontaneous crises, that occur in a large proportion of LBRF cases on or after the fifth day of fever, may also prove fatal. There is no conclusive evidence that the apparently milder reaction following slow-release penicillin, compared to tetracycline is any less dangerous [Reference Warrell63, Reference Seboxa and Rahlenbeck64].

Pre-treatment with oral prednisolone may prevent the J-HR of early syphilis, but neither an oral dose of 3 mg/kg prednisolone given 18 h beforehand, nor an infusion of 3.75 mg/kg betamethasone, prevented the J-HR in LBRF [Reference Warrell66]. Hydrocortisone in doses up to 20 mg/kg [Reference Warrell66], paracetamol [Reference Butler42] and pentoxifylline [Reference Remick67] failed to prevent the J-HR. However, meptazinol, an opioid antagonist/agonist, diminished the reaction when given in a dose of 100 mg by intravenous injection [Reference Teklu68, Reference Wright69]. A polyclonal ovine Fab anti-TNF-α antibody infused for 30 min before treatment with intramuscular penicillin, effectively prevented or diminished the J-HR [Reference Fekade70], but, unfortunately, has not been made available for use in the endemic areas. Recombinant human IL-10 was not effective [Reference Cooper71].

Supportive treatment

During spontaneous crisis or J-HR, hyperpyrexia must be actively prevented with antipyretics, vigorous fanning and tepid sponging. The prolonged ensuing flush phase poses dangers of hypotensive shock and postural hypotension and so patients must be nursed lying in bed for at least 24 h after treatment. Most patients are dehydrated and relatively hypovolaemic and adults may need 4 litres or more of isotonic saline intravenously during the first 24 h. Infusion should not be excessive, but carefully monitoring, by observing jugular venous or central venous pressures. Borrelial myocarditis predisposes to acute myocardial failure during the flush phase when there is a demand for a high cardiac output to sustain blood pressure in the face of systemic vasodilatation. Warning signs are symptoms of acute myocardial failure, a rise in central venous pressure above 15 cm H2O, and prolonged ECG QTc interval [Reference Parry53]. One mg of digoxin given intravenously over 5–10 min has proved effective in this situation [Reference Parry, Bryceson and Leithead72]. Diuretics may worsen the circulatory failure by causing relative hypovolaemia in the presence of the intense vasodilatation. Oxygen should be given during the reaction, particularly in severe cases. Patients with prolonged prothrombin times should be treated with vitamin K. Heparin is not effective in controlling coagulopathy and should not be used. Complicating opportunistic infections (typhoid, salmonellosis, bacillary dysentery, tuberculosis, typhus, visceral leishmaniasis, malaria) must be anticipated and treated appropriately.

Prevention and control of epidemics

For B. burgdorferi, an effective vaccine has been developed for dogs, but not yet for humans. However, there has been no interest in developing vaccines against relapsing fever borreliae.

Breaking louse transmission is essential for the control of an epidemic. Infested clothing should be deloused using heat (>60 °C) or washing at 52 °C for 30 min. Patients should be bathed with soap. Lice have developed some degree of resistance to the most commonly used topical pediculicides, including 10% dichloro-diphenyl-trichloroethane, lindane, 1% malathion, 2% temephos, 1% propoxur and 0.5% permethrin [Reference Sangaré, Doumbo and Raoult11]. Head lice should be removed by washed or shaving although their role in LBRF is unproven. Separating infested clothes from wearers for 10 days starves lice to death at any ambient temperature [Reference Barker and Barker73].

Author ORCIDs

David A. Warrell, 0000-0001-7792-1615

Acknowledgements

I am grateful to the many friends and colleagues with whom I studied LBRF in Ethiopia between 1968 and 1994, notably Anthony Bryceson, Eldryd Parry, Helen Pope (Watkins), David J. Wright, Bayu Teklu, Daniel Fekade and Kyle Knox; and the late Charles Leithead, Peter L. Perine and Jemal Abdulkadir.

References

1.Felsenfeld, O (1971) Borrelia: Strains, Vectors, Human and Animal Borreliosis. St. Louis, Missouri, USA: Green, 180.Google Scholar
2.Cutler, SJ et al. (1997) Borrelia recurrentis characterization and comparison with relapsing-fever, Lyme-associated, and other Borrelia spp. International Journal of Systematic Bacteriology 47, 958968.Google Scholar
3.Larsson, C et al. (2009) A novel animal model of Borrelia recurrentis louse-borne relapsing fever borreliosis using immunodeficient mice. PLoS Neglected Tropical Diseases 3, e522.Google Scholar
4.Cutler, SJ et al. (1994) Successful in-vitro cultivation of Borrelia recurrentis. Lancet 343, 242.Google Scholar
5.Marosevic, D et al. (2017) First insights in the variability of Borrelia recurrentis genomes. PLoS Neglected Tropical Diseases 11, e0005865.Google Scholar
6.Ružić-Sabljić, E et al. (2017) Comparison of MKP and BSK-H media for the cultivation and isolation of Borrelia burgdorferi sensu lato. PLoS ONE 12, e0171622.Google Scholar
7.Lescot, M et al. (2008) The genome of Borrelia recurrentis, the agent of deadly louse-borne relapsing fever, is a degraded subset of tick-borne Borrelia duttonii. PLoS Genetics 4, e1000185.Google Scholar
8.Elbir, H et al. (2014) African relapsing fever Borreliae genomospecies revealed by comparative genomics. Frontiers in Public Health. 2, 43.Google Scholar
9.Amanzougaghene, N et al. (2016) Head lice of pygmies reveal the presence of relapsing fever Borreliae in the Republic of Congo. PLoS Neglected Tropical Diseases 10, e0005142.Google Scholar
10.Sholdt, LL, Holloway, ML and Fronk, WD. The epidemiology of human pediculosis in Ethiopia. Special Publication, Navy Disease Vector Ecology and Control Center, Naval Air Station, Jacksonville, Florida 32212 (Fig. 26, page 76), pp. 159.Google Scholar
11.Sangaré, AK, Doumbo, OK and Raoult, D (2016) Management and treatment of human lice. Biomed Research International 2016, 8962685.Google Scholar
12.Bryceson, ADM et al. (1970) Louse-borne relapsing fever. A clinical and laboratory study of 62 cases in Ethiopia and a reconsideration of the literature. Quarterly Journal of Medicine 39, 129170.Google Scholar
13.Larsson, C et al. (2006) Complications of pregnancy and transplacental transmission of relapsing-fever borreliosis. Journal of Infectious Diseases 194, 13671374.Google Scholar
14.Fuchs, PC and Oyama, AA (1969) Neonatal relapsing fever due to transplacental transmission of Borrelia. Journal of the American Medical Association 208, 690692.Google Scholar
15.Lloyd, GER (ed). (1983) Hippocratic Writings (Translated by Chadwick J and Mann WN). London: Penguin Books, 9598.Google Scholar
16.MacArthur, W (1957) Historical notes on some epidemic diseases associated with jaundice. British Medical Bulletin 13, 146149.Google Scholar
17.Guellil, M et al. (2018) Genomic blueprint of a relapsing fever pathogen in 15th century Scandinavia. Proceedings of the National Academy of Sciences U S A 115, 1042210427.Google Scholar
18.Rutty, J (1770) A Chronological History of the Weather and Seasons, and of the Prevailing Diseases in Dublin. London: Robinson and Roberts.Google Scholar
19.Craigie, D (1843) Notice of a febrile disorder which has prevailed at Edinburgh during the Summer of 1843. Edinburgh Medical and Surgical Journal 60, 410418.Google Scholar
20.Henderson, W (1844) On some of the characters which distinguish the fever at present epidemic from typhus. Edinburgh Medical and Surgical Journal 61, 201225.Google Scholar
21.Murchison, C (1862) A Treatise on the Continued Fevers of Great Britain. London, Parker, and Bourn, 290409.Google Scholar
22.Wright, DJ and Maria, B (2011) Ich bin ein Berliner. The contributions of early British colonial and German scientists to the elucidation of the nature of spirochaetes. Clinical Microbiology and Infection 17, 484486.Google Scholar
23.Mackie, FP (1907) The part played by Pediculus corporis in the transmission of relapsing fever. British Medical Journal 2, 17061709.Google Scholar
24.Sparrow, H (1958) Étude du foyer Éthiopien de fièvre récurrente. Bulletin de l'organisation. mondiale de la Santé 19, 673710.Google Scholar
25.ProMED-mail (International Society for Infectious Diseases). Borreliosis, relapsing fever – Sudan (South). ProMED-mail 1999; 19990511.0767; http://www.promedmail.org. Accessed 27 Oct 2018.Google Scholar
26.Valdizan, M, Lopez, J and Delgado, A (1985) Borreliosis en las alturas del Peru. Diagnóstico (Peru) 15, 913.Google Scholar
27.Raoult, D et al. (1999) Survey of three bacterial louse-associated diseases among rural Andean communities in Peru: prevalence of epidemic typhus, trench fever, and relapsing fever. Clinical Infectious Diseases 29, 434436.Google Scholar
28.Ciervo, A et al. (2016) Louseborne relapsing fever in young migrants, Sicily, Italy, July-September 2015. Emerging Infectious Diseases 22, 152153.Google Scholar
29.Hoch, M et al. (2015) Louse-borne relapsing fever (Borrelia recurrentis) diagnosed in 15 refugees from northeast Africa: epidemiology and preventive control measures, Bavaria, Germany, July to October 2015. Eurosurveillance 20(42), 15.Google Scholar
30.European Centre for Disease Prevention and Control. Louse-borne relapsing fever in the EU 17 November 2015 Stockholm: ECDC; 2015; http://ecdc.europa.eu/en/publications/publications/louse-borne-relapsing-fever-in-eu-rapid-risk-assessment-17-nov-15.pdf accessed 27-20-18.Google Scholar
31.Antinori, S et al. (2016) Louse-borne relapsing fever among East African refugees in Europe. Travel Medicine and Infectious Diseases 14, 513514.Google Scholar
32.Hytönen, J et al. (2017) Louse-borne relapsing fever in Finland in two asylum seekers from Somalia. APMIS Acta Pathologica, Microbiologica, et Immunologica Scandinavica 125, 5962.Google Scholar
33.Ackermann, N et al. (2018) Screening for infectious diseases among newly arrived asylum seekers, Bavaria, Germany, 2015. Eurosurveillance 23, 111.Google Scholar
34.Isenring, E et al. (2018) Infectious disease profiles of Syrian and Eritrean migrants presenting in Europe: a systematic review. Travel Medicine and Infectious Diseases 25, 6576.Google Scholar
35.Grecchi, C et al. (2017) Louse-borne relapsing fever in a refugee from Mali. Infection 45, 373376.Google Scholar
36.Barbour, AG and Hayes, SF (1986) Biology of Borrelia species. Microbiological Reviews 50, 381400.Google Scholar
37.Vidal, V et al. (1998) Variable major lipoprotein is a principal TNF-inducing factor of louse-borne relapsing fever. Nature Medicine 4, 14161420.Google Scholar
38.Grosskinsky, S et al. (2010) Human complement regulators C4b-binding protein and C1 esterase inhibitor interact with a novel outer surface protein of Borrelia recurrentis. PLoS Neglected Tropical Diseases 4, e698.Google Scholar
39.Guo, BP et al. (2009) Relapsing fever Borrelia binds to neolacto glycans and mediates rosetting of human erythrocytes. Proceedings of the National Academy of Sciences U S A 106, 1928019285.Google Scholar
40.Negussie, Y et al. (1992) Detection of plasma tumor necrosis factor, interleukins 6, and 8 during the Jarisch–Herxheimer reaction of relapsing fever. Journal of Experimental Medicine 175, 12071212.Google Scholar
41.Coxon, RE et al. (1997) The effect of antibody against TNF alpha on cytokine response in Jarisch–Herxheimer reactions of louse-borne relapsing fever. Quarterly Journal of Medicine 90, 213221.Google Scholar
42.Butler, T (2017) The Jarisch–Herxheimer reaction after antibiotic treatment of spirochetal infections: a review of recent cases and our understanding of pathogenesis. American Journal of Tropical Medicine and Hygiene 96, 4652.Google Scholar
43.Warrell, DA et al. (1970) Cardiorespiratory disturbances associated with infective fever in man: studies of Ethiopian louse-borne relapsing fever. Clinical Science 39, 123145.Google Scholar
44.Judge, DM et al. (1974) D Louse-borne relapsing fever in man. Archives of Pathology 97, 136170.Google Scholar
45.Anderson, TR and Zimmerman, LE (1955) Relapsing fever in Korea; a clinicopathologic study of eleven fatal cases with special attention to association with Salmonella infections. American Journal of Pathology 31, 10831109.Google Scholar
46.Salih, SY et al. (1977) Louse-borne relapsing fever: I. A clinical and laboratory study of 363 cases in the Sudan. Transactions of the Royal Society of Tropical Medicine and Hygiene 71, 4348.Google Scholar
47.Perine, PL et al. (1992) A clinico-epidemiological study of epidemic typhus in Africa. Clinical Infectious Diseases 14, 11491158.Google Scholar
48.Antinori, S et al. (2017) Diagnosis of Louse-Borne relapsing fever despite negative microscopy in two asylum seekers from Eastern Africa. American Journal of Tropical Medicine and Hygiene 97, 16691672.Google Scholar
49.Ramos, JM et al. (2004) Characteristics of louse-borne relapsing fever in Ethiopian children and adults. Annals of Tropical Medicine and Parasitology 98, 191196.Google Scholar
50.Ramos, JM et al. (2009) Louse-borne relapsing fever in Ethiopian children: experience of a rural hospital. Tropical Doctor 39, 3436.Google Scholar
51.Borgnolo, G et al. (1993) Louse-borne relapsing fever in Ethiopian children: a clinical study. Annals of Tropical Paediatrics 13, 165171.Google Scholar
52.Nordmann, T et al. (2018) Outbreak of louse-borne relapsing fever among urban dwellers in Arsi zone, central Ethiopia, from July to November 2016. American Journal of Tropical Medicine and Hygiene 98, 15991602.Google Scholar
53.Parry, EH et al. (1970) Some effects of louse-borne relapsing fever on the function of the heart. American Journal of Medicine 49, 472479.Google Scholar
54.Hovette, P et al. (2001) Value of Quantitative Buffy Coat (QBC) in borreliasis-malaria co-infection. Médecine Tropicale (Marseille) 61, 196197.Google Scholar
55.Larsson, C and Bergström, S (2008) A novel and simple method for laboratory diagnosis of relapsing fever borreliosis. Open Microbiology Journal 2, 1012.Google Scholar
56.Berger, SA and David, L (2005) Pseudo-borreliosis in patients with malaria. American Journal of Tropical Medicine and Hygiene 73, 207209.Google Scholar
57.Elbir, H et al. (2013) Multiplex real-time PCR diagnostic of relapsing fevers in Africa. PLoS Neglected Tropical Diseases 7, e2042.Google Scholar
58.Sokhna, C et al. (2013) Point-of-care laboratory of pathogen diagnosis in rural Senegal. PLoS Neglected Tropical Diseases 7, e1999.Google Scholar
59.Porcella, S et al. (2000) Sero-diagnosis of louse-borne relapsing fever with glycerol phosphodiesterase (GlpQ) from Borrelia recurrentis. Journal of Clinical Microbiology 38, 35613571.Google Scholar
60.Fotso Fotso, A et al. (2016) Monoclonal antibodies for the diagnosis of Borrelia crocidurae. American Journal of Tropical Medicine and Hygiene 94, 6167.Google Scholar
61.Cutuli, SL et al. (2017) Lice, rodents, and many hopes: a rare disease in a young refugee. Critical Care 21, 81.Google Scholar
62.Perine, PL and Teklu, B (1983) Antibiotic treatment of louse-borne relapsing fever in Ethiopia: a report of 377 cases. American Journal of Tropical Medicine and Hygiene 32, 10962000.Google Scholar
63.Warrell, DA et al. (1983) Pathophysiology and immunology of the Jarisch–Herxheimer-like reaction in louse-borne relapsing fever: comparison of tetracycline and slow-release penicillin. Journal of Infectious Diseases 147, 898909.Google Scholar
64.Seboxa, T and Rahlenbeck, S (1995) Treatment of louse-borne relapsing fever with low dose penicillin or tetracycline: a clinical trial. Scandinavian Journal of Infectious Diseases 27, 2931.Google Scholar
65.Salih, SY and Mustafa, D (1977) Louse-borne relapsing fever: II. Combined penicillin and tetracycline therapy in 160 Sudanese patients. Transactions of the Royal Society of Tropical Medicine and Hygiene 71, 4951.Google Scholar
66.Warrell, DA (1969) Respiratory, Circulatory and Metabolic Disturbances in Relapsing Fever (Dissertation). Oxford, UK: University of Oxford, 246 pp.Google Scholar
67.Remick, DG et al. (1996) Pentoxifylline fails to prevent the Jarisch–Herxheimer reaction or associated cytokine release. Journal of Infectious Diseases 174, 627630.Google Scholar
68.Teklu, B et al. (1983) Meptazinol diminishes the Jarisch–Herxheimer reaction of relapsing fever. Lancet 1, 835839.Google Scholar
69.Wright, DJ (1983) Endogenous opioid withdrawal in the Jarisch–Herxheimer reaction. Lancet 1, 11351136.Google Scholar
70.Fekade, D et al. (1996) Prevention of Jarisch–Herxheimer reactions by treatment with antibodies against tumor necrosis factor alpha. New England Journal of Medicine 335, 311315.Google Scholar
71.Cooper, PJ et al. (2000) Recombinant human interleukin-10 fails to alter proinflammatory cytokine production or physiologic changes associated with the Jarisch–Herxheimer reaction. Journal of Infectious Diseases 181, 203209.Google Scholar
72.Parry, EH, Bryceson, AD and Leithead, CS (1967) Acute hemodynamic changes during treatment of louse-borne relapsing fever. Lancet 1, 8183.Google Scholar
73.Barker, SC and Barker, D (2019) Killing clothes lice by holding infested clothes away from hosts for 10 days to control louseborne relapsing fever, Bahir Dah, Ethiopia. Emerging Infectious Diseases 25, 304310.Google Scholar
Figure 0

Fig. 1. Ethiopian patients with LBRF. (A) Profuse petechial rash on the trunk in an emaciated patient with complicating infection with Salmonella enterica serovar Typhi (S. Typhi). (B) Subconjunctival haemorrhages and jaundice indicative of hepatocellular damage, thrombocytopenia and coagulopathy. (C) B. recurrentis spirochaetes arrowed (silver stain) in the splenic pulp. (D) Cerebral haemorrhage on the 6th day of illness, a common cause of death in patients with LBRF.