19d Other bacterial diseasesRat-bite fevers

Rat-bite fever is an uncommon but well described clinical syndrome. Resulting from a systemic infection following a rat bite, such an illness had been recognized in India for over 2,300 years (Elliott 2007). With the advent of modern microbiology it was found to be caused by two different pathogens that are commonly found in the oropharynx of rodents—Streptobacillus moniliformis and Spirillum minus. Although Streptobacillus moniliformis may be found worldwide, it is said to be found more commonly in North America and Europe whereas Spirillum minus is more common in Asia. Spirillum minus is said by some to be absent in the USA (Centers for Disease Control (CDC) 2005) but in fact, historically, there are several case reports (Pritchard 1990). Common differences in the characteristic clinical features are summarized in Table 19d.1. The two organisms are then discussed separately.

Infection with Streptobacillus moniliformis, gives rise to Rat-Bite fever also known as Streptobacillary Fever, and, when epidemic, Epidemic Arthritic Erythema (Haverhill Fever). It occurs worldwide. Only three outbreaks have been described. It is either caused by the bite of, or similar close contact with, a rat or other infected rodent, or, when epidemic, by the ingestion of water or milk contaminated by rats. Control requires limiting human contact with rats, traditionally by prevention of rat infestation although, of

recent years, in developed countries, the keeping of rats as pets has played an important role.

Streptobacillus moniliformis was first so named in France in 1925, following its isolation from the blood of a laboratory worker and became recognized as one of the two causes of Rat-Bite fever. (The other is Spirillum minis.) In fact published descriptions of disease following a rat bite due to an organism called Streptothrix muris ratti were made in Germany in 1914 and the USA in 1916, and this organism is likely to have been Streptobacillus moniliformis (Elliott 2007). At the same time as the description in France, an epidemic form of the disease, Haverhill Fever, was recognized during an outbreak in the USA that occurred in January 1926 in Haverhill, Massachusetts (Parker and Hudson 1926; Place and Sutton 1934) which was ascribed to milk. Place and Sutton’s account also refers to a similar but much larger outbreak that had occurred the year before in Chester, Pennsylvania, USA that was communicated to them personally by Armstrong and Wood, the State Epidemiologists. In 1983, the only other outbreak ever published occurred among pupils and staff at a boarding school in England. The investigators, having considered unpasteurized milk, eventually concluded that the cause of contamination was probably rat urine in a private water supply from a spring (McEvoy et al. 1987; Shanson et al. 1983). Accounts of sporadic cases occur but are infrequent in the world literature. Elliott has reviewed the English language literature and identified 65 detailed case reports from 1938 to 2005 (Elliott 2007). Graves and Janda have analysed retrospectively data collected on Streptobacillus moniliformis cultures performed by the Microbial Diseases Laboratory of the California State Health Department, USA, between 1970 and 1998, and found 41 cases in all (Graves and Janda 2001).

Streptobacillus moniliformis is a pleomorphic Gram-negative bacterium which may occur as short coccobacillary forms as well as chains and intertwining wavy filaments. Previously believed to be related to the actinobacilli it is now thought to have more in common with the Mycoplasmates (Costas and Owen 1987). It is a facultative anaerobe and slow growing on blood, serum, or other bodily fluid. The optimal condition for isolation is culturing blood and body fluid on blood or tryptose soy agar enriched with 20% rabbit serum and incubated in a microaerophilic atmosphere with 5–10% CO₂ and humidification. Cultures should be held for at least 5 days (Martin and Martin in Topley and Wilson 1998; CDC 2005). The optimal growth temperature is 35–37 degrees C and optimal pH 7.4–7.6. Streptobacillus moniliformis is inhibited by ‘Liquoid’ (sodium polyanethol sulphonate) present in many commercial blood culture media systems such as BACTEC^TM(BD, Sparks, Maryland, USA) and BacT/Alert^TM (Biomerieux, Durham, North Carolina, USA) (Wang and Wong 2007). It is killed at 55° C in 30 minutes or less. Growth in broth produces characteristic ‘puff-ball’ colony formation (Graves and Janda 2001). Cell wall deficient L-Forms may arise spontaneously and give rise to a typical ‘fried-egg’ appearance of colonies in older cultures. As unfavourable outcomes have occurred when diagnosis was delayed, in part because the difficulty of rapidly isolating or identifying Streptobacillus moniliformis, if Streptobacillus moniliformis infection is suspected, in the absence of a positive culture, identification of pleomorphic Gram-negative bacilli in appropriate specimens might support a diagnosis (Washburn 2000; CDC 2005). Historically fatty acid profiles obtained by gas-liquid chromatography, together with characteristic growth were used for rapid identification (Rowbotham 1983) but this has been superseded by PCR based on the 16S rRNA gene sequence (Boot et al. 2002).

Streptobacillus moniliformis is carried in the nasopharynx or excreted in the urine of healthy rats (Strangeways 1933). It may be carried in other rodents where, as in the case of mice, it may also give rise to disease, notably septicaemia or arthropathy, which may be fatal. Human cases have been ascribed to exposure to a ferret, a mouse, squirrels, a gerbil and a dog (Elliott 2007). Increasingly humans are exposed to rats as pets in domestic settings or occupationally in pet shops (Downing 2001; Graves and Janda 2001; Grant et al. 2004; CDC 2005; Schachter 2006; Matheson 2007).

The incubation in the original Haverhill outbreak described was 1–4 days (Parker and Hudson 1926) and is rarely longer than 10 days.

These are sometimes differentiated, as clinical entities between sporadic cases, usually with a history of animal exposure, and epidemic cases. It has been suggested that septic arthritis should be separately considered as a third clinical entity (Wang and Wong 2007). However, accounts of the various forms of streptobacillary fever overlap and the account of the original Haverhill outbreak identifies two patients in which organisms were recovered from their swollen knees. Typically, therefore, there is fever followed after a median of 2 days by a symmetrical erythematous, usually maculopapular, rash affecting the palms of the hands, the feet, shins, and ankles, and in about half the cases, the face. Petechial haemhorrages and haemorrhagic vesicles may develop on the hands and feet (Place and Sutton 1934; Elliott 2007). Asymmetrical arthralgia, and also arthritis, affecting usually more than one joint, occurs in most cases (‘Epidemic Arthritic Erythema’). In the Haverhill outbreak wrists and elbows were most frequently involved, followed by knees (Place and Sutton 1934) and in the UK outbreak, knees followed by ankles (McEvoy et al. 1987).

Diagnosis is usually by culture of blood, synovial fluid or lymph nodes. Strains may be distinguished by the electrophoretic profiles of cellular proteins (Costas and Owen 1987).

The list of potential differential diagnoses is lengthy including the many bacterial and viral infections that may give rise to fever and rash (e.g. Epstein Barr Virus, Coxackie Virus, hantaviruses, Rocky Mountain Spotted Fever, meningococcal disease, streptococcal disease, Lyme Disease, leptospirosis, gonorrhoea) as well as auto-immune mediated and reactive arthritides such as Rheumatoid Arthritis and Reiter’s syndrome (Matheson et al. 2008; Wang and Wong 2007). A careful history that identifies a relevant animal exposure may be key.

Penicillin, erythromycin and tetracycline have all been used with evidence of success (Shanson et al. 1983). The currently recommended treatment is intravenous penicillin, 1.2 million units/day for 5 to 7 days, followed by oral penicillin or ampicillin, 500mgs, four times a day, for a further seven days, if improvement is observed. Oral tetracycline (500mgs four times a day) or intramuscular streptomycin (7.5mgs per kilogram body weight), twice daily, are alternatives (CDC 2005).

The occurrence of endocarditis requires dual therapy with both penicillin G, at higher doses than in uncomplicated infections, and streptomycin or gentamicin (Elliott 2007). In septic arthritis, antibiotics than penetrate the inflamed synovium are desirable. This argues against the use of aminoglycosides and it has been suggested that further studies to determine an optimal regime are required. (Wang and Wong 2007).

Finally, the L-Form of S.moniliformis is not penicillin sensitive (although it is sensitive to tetracycline). This has given rise to a suggestion that penicillin may encourage the development of L-forms and thereby prolong the course of the illness (Roughgarden 1965). In practice, however, this does not seem to have been an important consideration and some consider that the L-form is, in fact, not pathogenic (Elliott 2007).

The duration of illness ranges from 2–32 days (median 16 days). At least one relapse occurs in about half the cases and more than one relapse in 17%. Mortality has been absent in epidemics of Haverhill Fever but severe complications, including endocarditis, pneumonia, metastatic abscesses, and anaemia have been described in sporadic cases. This was largely before the advent of antibiotic treatment (Roughgarden 1965; Taber and Feigen 1979). Thus Rat-Bite Fever is said to have a case fatality rate of 7–10% when left untreated (Washburn 2000) with a disproportionately higher mortality in infants and patients with endocarditis. Recently two fatal cases were identified, after their death, by the CDC’s Unexplained Deaths and Critical Illnesses (UNEX) Project. Both were previously healthy women, aged 52 and 19 years respectively (CDC 2005).

A combination of the infrequency of sporadic cases of streptobacillary fever, their generally good response to treatment and the absence of fatalities in Haverhill Fever, means that information on pathogenesis is scanty.

Post mortem examination of a recent fatal case showed findings suggestive of systemic infection including disseminated intravascular coagulopathy and inflammatory cell infiltrates in the liver, heart, and lungs (CDC 2005). Endocarditis is also well described (Balakrishnan et al. 2006).

Sporadic cases occur worldwide. In a case series from California, males and females were affected with approximately equal frequency, ages ranged from 6 weeks to 92 years, the mean age was just under 19 years and the median age was just over 10 years. Sixty eight per cent were children less than 16 years of age (Graves and Janda 2001). In the series based on the published literature, 77% of patients reported were male and ages ranged from 2 months to 87 years (Elliott 2007). In the two case series only 46% (Graves and Janda 2001) and 66% (Elliott 2007) respectively, gave a specific history of being bitten by a rat.

Three outbreaks have been described (Parker and Hudson 1926; Place and Sutton 1934; McEvoy et al. 1987). These include the first to be described which was at Haverhill, Massachussetts, which gave the epidemic form of the disease its name of ‘Haverhill Fever’ (Parker and Hudson 1926). Eighty six cases, aged from 8 months to 54 years, occurred. They comprised 51 females and 35 males. Except for two patients, they were drawn from 39 families, comprising 231 persons (attack rate 36%). Their onsets occurred between 2–29 January 1926, peaking on 16 January. They were largely among a population of Lithuanian, Italian, Polish or Jewish extraction, in a circumscribed district, 1,500 by 500 metres, of Haverhill, then a manufacturing city with a population of 35,000. Of the cases 97% had a fever and 93% a rash. Streptococcus moniliformis was isolated from the blood of 11 cases. Suspicion fell on one small on-farm dairy that supplied, directly or indirectly, every case and whose entire milk supply was distributed to the small area of the city that corresponded to the residences of the cases. Once the milk from the dairy was pasteurized, only one further case occurred who gave a history of, nevertheless, having drunk the milk unpasteurized. By contrast the water supply to the area was the same city reservoir that supplied the whole town.

Suspicion of the milk was further emphasized by the experience of an outbreak in Chester, Pennsylvania, in May and June 1925 of which the Haverhill investigators were aware and an account of which they included in their own published account (Place and Sutton 1934). Here 400 cases were discovered of which 92% received their milk from the same bottling plant in the city. About 20% of persons who received milk from this plant had the disease compared with 0.5% of the population of Chester as a whole.

In the most recent outbreak which was at a boarding school (McEvoy 1987), the school was situated on a 100-acre site in a rural area near a market town in England. Within the grounds are extensive gardens, stables, and a farm which, at the time, provided the school with unpasteurized milk. There were about 700 people on the site, including 500 schoolgirls, aged 8–19 years, of which 370 were boarders and 130 attended daily.

The first case of fever was reported on 11 February 1983, and a further 24 cases presented the next day. Cases continued to occur in children and staff members, with a peak on 14 February, when 86 people became ill. The onset of the last known cases was 21 February by which time a total of 304 cases had occurred.

Analysis of these 304 cases by day of onset of symptoms suggested a common source for the outbreak, with exposure probably continuing over several days in early February. It is likely that the earliest cases, which were recognized retrospectively, had occurred on 6–8 February. All cases had lived on or visited the site of the school, and there were no reported cases of secondary transmission.

A random sample of 230 pupils was interviewed in greater detail. There were significant associations between the development of illness and the consumption of cold milk (Relative Risk = 3.3, p = 0.02) and cold water (Relative Risk = 10, p=6 x 10^-5) at the school. No other exposures were associated with illness.

Milk and water consumption were examined further, with data collected from the whole at risk population. Although both milk and water drinking remained significantly associated with the development of illness, water was associated independently of milk, but milk was not associated independently of water.

The major source of supply to the school was mains chlorinated water. However, water was also supplied by a spring. In addition, the sewage system had been transferred from the old filter-bed method to the mains system 2 weeks before the onset of the outbreak. The filter-bed and cesspit were situated on high ground above the spring. Inspection of this area revealed the presence of rats. Moreover, water could have filtered back into the spring pond, where rats were also seen. The spring water fed a well in an old courtyard. The well water passed through a chlorine-dosing unit and calorifiers to a pressurized storage unit in the school basement, where it was then used to feed the hot water system. Inspection revealed no domestic hot water return to the calorifiers. It was reported to be impossible to raise the water temperature above 50°C. Despite the presence of the chorine-dosing unit there was no evidence of chlorine in the water feeding the calorifiers. An inspection of the raw milk supplied from the school farm, by contrast, showed no evidence that the milk could have been contaminated with rat urine.

S.moniliformis was isolated from the blood cultures of 4 boarders who had returned home and were subsequently admitted to hospital (Shanson et al. 1983). The organism was not cultured from the samples of milk or water or from rats trapped in the school grounds after 16 February.

The attack rate was 49% (304/700). The duration of illness ranged from 2 to 32 days (median 16 days). All cases had fever and 95% had a rash. The distribution of the rash, unlike the joint involvement (which affected 97%), tended to be symmetrical.

Over half had 2 or more joints involved and a third 3 or more joints. Altogether, 56% of patients experienced one recurrence, 11% two recurrences, and 6% had three.

The conclusion that water was the vehicle of infection was strongly supported by epidemiological evidence and by circumstantial evidence that opportunity existed for the consumption of water from a spring infected with rats.

The hot water supply was probably contaminated intermittently between 5–12 February. During the first 2 weeks of February, operations on the building site involved digging around the foundations and the weather was exceptionally stormy. These factors may have disturbed the local rat population, causing it to contaminate the spring pond. The mains water supply was shut off for frequent short spells during this period, and drinking water could have been drawn from the hot taps.

There was evidence that the ‘hot’ water was cool, and heavy demands on the supply would have caused the temperature to fall further. A higher attack rate which was observed among the boarders (54% v 20% in day pupils) can be explained by their greater exposure to contaminated water. The total absence of cases outside the school suggests that the mains water was not contaminated.

Up to 50% of rats carry the organism as do, rarely, other rodents (Strangeways 1933). Sporadic cases usually occur following rat bites (‘Rat-Bite Fever’) but may occur following living or working in a rat infested building. Increasingly humans are exposed to rats as pets in domestic settings or occupationally in pet shops (Downing 2001; Graves and Janda 2001; Grant et al. 2004; Centers for Disease Control 2005; Schachter 2006; Matheson 2007). In California between 1970 and 1998, exposure to pet rats was reported in 54% of cases where animal exposure was recorded with exposure to rats in school, presumably largely in laboratories, contributing a further 14% of exposures (Graves and Janda 2001). Outbreaks may result from the contamination by rats of milk (Parker and Hudson 1926; Place and Sutton 1934) or water (McEvoy 1987).

Person to person transmission does not occur.

Preventing the infestation by rats of human dwellings and workplaces is the key to control. Nevertheless a number of recent cases have arisen as a result of keeping or supplying rats as companion animals. It is important, therefore, that those who keep rats as companion animals or those who supply them are aware of the diagnosis as prevention of severe disease can be greatly assisted by early recognition and treatment. The CDC advise protective gloves when handling rats or cleaning their cages. They advise further that children under 5 years of age should be supervized by adults to prevent bites and hand to mouth contact and that, in the event of a rat bite, the wound should be promptly cleaned and disinfected. The efficacy of antimicrobial prophylaxis is unknown (CDC 2005).

Disease clusters must prompt the systematic search for a common source.

Spirillary Fever was first described in a rat in Bombay (Mumbai), India, by Henry Van Dyke Carter, a distinguished physician and pathologist who was the original illustrator of Gray’s ‘Anatomy’, in 1888 (Hiatt and Hiatt 1995). Spirillum minus became recognized, with Streptobacillus moniliformis as one of the two causes of Rat-Bite fever. Accounts of cases have been rare and although they occur worldwide (Pritchard 1990) it has been suggested that they are more common in Asia (Booth and Katz 2002; CDC 2005).

Spirillum minus is a minute (0.2 microns wide by 2–5 microns in length) spirillar bacterium which has 2–3 rigid spirals of 0.8–1.0 micron wavelength and one or more flagella attached at each pole. It may be stained with methylene blue and with Giemsa. Its true taxonomic place is unknown (Skirrow 1990). It is aerobic and can be cultured only in vivo in mice and guinea pigs (Pritchard 1990).

Spirillum minus is found in rats, mice and guinea pigs and does not seem to be pathogenic in animals.

The incubation period is generally 7–21 days but may extend to months (Pritchard 1990).

These can be differentiated clinically from Streptobacillary Fever (Holmgren 1970). Redness and swelling are visible at the wound site initially. This may be healed before the onset of systemic disease but Spirillum minus is characterized by re-opening and eventual ulceration of the wound even after it has initially healed. Systemic disease is heralded by paroxysmal fever accompanied by lymphadenopathy and dark-red eruptions on the skin. Myalgia and arthralgia, more pronounced on the side of the body affected by the bite, occur. Arthritis seldom occurs.

The organism is present in the local lesion or lymph nodes but may occassionally be demonstrated, with difficulty, by dark field microscopy. Subcutaneous inoculation of fluid or lymph nodes in guinea pigs is followed by a chancre and enlargement of regional lymph nodes. In both of these sites Spirillum minus may be demonstrated (Pritchard 1990). Unlike Streptobacillary Fever, leucocytosis is usually absent in Spirillosis (Holmgren 1970).

Parenteral penicillin is the drug of choice although may cause a Jarisch-Herxheimer reaction. Erythromycin or chloramphenicol may all be used in patients allergic to penicillin. Sulphonamides are ineffective.

Attacks last 3–4 days but, in the absence of treatment can recur, usually at a regular interval, for months or even years. Case fatality rate is 2–10% (Pritchard 1990).

Sporadic cases occur worldwide. Outbreaks are not described. Wild rats, living in close proximity to man, were examined in Vancouver in 1950 and 14% carried Spirillum minus. (Pritchard 1990). Carriage rates in London in the 1980s in trapped rats were 25% (Martin 2005). Carriage by laboratory rats, mice and guinea pigs is also described. Disease is exclusively transmitted to humans by the bite of an infected animal in contrast to Streptobacillus moniliformis where transmission by the contamination of milk or water by the urine of infected animals is described.

Person to person transmission does not occur.

Preventing the infestation by rats of human dwellings and workplaces is the key to control. As with Streptobacillus moniliformis, the taking of general precautions to avoid being bitten when handling rats, including those kept in laboratories or as pets, is important. Meticulous local wound care of any bites that do occur is also necessary. The efficacy of antimicrobial prophylaxis is unknown but is sometimes suggested (Pritchard 1990).