L. pneumophila Virulence Factors

A variety of surface structures have been implicated in Lp pathogenesis. Type IV pili modestly promote bacterial attachment to macrophages and epithelial cells,¹¹³ and flagella promote invasion independent of adherence.¹¹⁴ The major outer membrane protein is a porin that also serves as a binding site for complement components and thus mediates opsonophagocytosis.¹¹⁵ The Mip protein is a surface-exposed peptidyl prolyl isomerase that is required for the early stages of intracellular infection and for full virulence in animals,^116–118 whereas the heat shock protein Hsp60 has been shown to enhance epithelial cell invasion.¹¹⁹ The rtxA gene promotes adherence and virulence, although the structure and localization of its protein product are unclear.^120,121 Legionella LPS contains some endotoxic activity, and changes in LPS have correlated with increases in serum resistance, intracellular growth, and virulence.¹²² Finally, the rcp gene, which appears to encode a lipid A modifying enzyme, confers resistance to cationic peptides and promotes macrophage and lung infection.¹⁰⁶

Lp secretes a variety of proteins, degradative enzymes, and putative toxins. The release of proteins by Lp into the extracellular milieu is mediated primarily by a type II secretion system.^81,123–125 Genome sequencing also suggests the existence of type I and type V secretion systems.¹²⁶ Acid phosphatases; aminopeptidases; an RNase; a chitinase; a zinc metalloprotease; monoglycerol, diglycerol, and triacylglycerol lipases; a phospholipase A; a lysophospholipase A; cholesterol acyltransferase; phospholipases C; a peptidyl prolyl isomerase; and a number of novel proteins are all secreted via the Legionella type II system.^127–130 Mutations within the genes encoding the type II secretion system diminish infectivity for macrophages, protozoa, and animals.^129,131 The secreted zinc protease is produced during infection and promotes pathology in the guinea pig model of disease, as well as intracellular infection of some ameba hosts.^127,132 Surprisingly, the type II-secreted chitinase promotes bacterial persistence in the lungs of infected mice.¹²⁸

A Legionella type IV secretion system called Dot/Icm is essential for intracellular replication and virulence in animal models of disease.⁸¹ First, it enhances Lp entry into host cells.^133,134 Second, it is essential for the ability of the Legionella parasite to modulate host vesicular transport to avoid delivery to lysosomes and promote vacuole remodeling by the host vesicles in the early secretory pathway.^95,135,136 Finally, the Dot/Icm system is important for apoptosis and bacterial egress from its spent host cell.^137,138 Mutations in dot/icm loci lead to loss of virulence.^111,139 Many hundreds of proteins are secreted by the Dot/Icm system, and in most cases these “effector” proteins translocate from the Legionella-containing vacuole into the host cell cytoplasm.^81,140 Some of these, including VipA, VipD, VipF, AnkX, LegC3, LegC7 (YflA), and LegC2 (YlfB), have been implicated in bacterial evasion of lysosome fusion.^141–143 Others, such as RalF, DrrA (SidM), LepB, LidA, and SidJ, play roles in the recruitment of endoplasmic reticulum–derived vesicles to the Legionella-containing vacuole.^144–147 The effector SdhA is important for maintaining the integrity of the vacuole in which Lp resides,¹⁴⁸ and SidF is a lipid phosphatase that generates phosphatidylinositol phosphate signatures on the cytosolic surface of the vacuole that are important for the binding of other effectors.¹⁴⁹ The effectors LepA and LepB have been implicated in bacterial egress from the spent host cell.¹⁵⁰ Lp expresses several effectors that function as glucosyltransferases that are capable of inhibiting host cell protein synthesis (elongation factor 1A).^151,152 Although Dot/Icm is generally regarded as essential for Lp infection, a second type IV secretion system known as Lvh is found in addition to Dot/Icm in many strains; however, this system does not appear to deliver effector proteins into host cells but retains the ability to promote DNA conjugation, which could facilitate horizontal transfer of genes encoding effectors.¹⁵³ Interestingly, many of the type II– and type IV–secreted proteins, as well as other mediators of infection (e.g., LpnE, Lpg0971), bear striking sequence similarity to eukaryotic proteins, suggesting that they were acquired by horizontal gene transfer from a eukaryotic host and use similar biochemical activities to facilitate Lp infection.^{126,128,141,154–158}

Several infectivity factors have been localized to the Lp periplasm or cytoplasm. A Cu-Zn superoxide dismutase resides in the periplasm, affording resistance to toxic superoxide anions, and the KatB catalase-peroxidase is necessary for optimal intracellular infection.^159,160 The Legionella phosphoenolpyruvate phosphotransferase and HtrA protein promote intracellular growth and virulence.^161,162 Lp iron acquisition, important for intracellular and extracellular replication, involves, among other things, a secreted ferric iron chelator (the siderophore legiobactin), a secreted pyomelanin with ferric reductase activity, and an inner membrane ferrous iron transporter (FeoB).^163–166

In addition to the identification of eukaryotic-like proteins in the Lp genome, another outcome of sequencing bacterial genomes is the realization that there are large segments of DNA, including plasmids and chromosomal “islands,” which can vary between Lp strains.^{126,167–169} It is possible that these variable regions of the genome will help explain differences in virulence that may exist between strains. Indeed, large deletions of the chromosome that remove individual islands can decrease Lp replication in specific protozoan hosts.¹⁷⁰

Virulence Factors and Pathogenesis of Other Legionella spp.

Relatively little is known about the virulence mechanisms, molecular pathogenesis, and cell biology of infections caused by Legionella spp. other than Lp. With the major exception of Llb, infections caused by the other Legionella spp. are rare and found almost exclusively in severely immunocompromised patients. This implies that these other species lack critical virulence factors that allow ready intracellular multiplication. However, L. micdadei, L. dumoffii, Llb, and other species can multiply normally in guinea pig and other animal macrophages.^171–174 Unlike Lp, non-AJ strain mice are not resistant to infection with a number of other Legionella spp., including L. bozemanae, L. dumoffii, L. micdadei, and Llb.¹⁷⁵ Despite unrestricted growth in otherwise unpermissive macrophages, some non-Lp species fail to promote inflammatory cell death,¹⁷⁶ which could be correlated to a failure to cause disease in nonimmunocompromised people, although this is unstudied. Llb resides within a ribosome-studded phagosome, albeit with markers of late endosomal maturation, unlike Lp, which can block endosomal maturation at an early stage.^171,177,178 In contrast, L. micdadei resides in a smooth phagosome, and morphologic data suggest L. dumoffii grows within the cytoplasm rather than in a phagosome.^{174,177,179,180} Analysis of the genome of several different Llb strains has shown that this species has and utilizes the Dot/Icm system for causing infection, that it is nonflagellated, and that it possesses a wide range of eukaryotic effectors different from those found in Lp.¹⁸¹ Many of the Llb effectors appear to have been derived from plants and soil organisms, suggesting that it has evolved as primarily a soil-adapted bacterium. Experimental virulence of Llb for macrophages appears to be relatively independent of growth phase in contrast to Lp.^171,182

Patterns and Rates of Disease and Mortality

LD occurs in both sporadic and epidemic form. About 65% to 95% of reported cases are not associated with known epidemics of the disease.^30,188 Underreporting of the disease is common because many sporadic cases are treated empirically without diagnostic studies being performed, because of false-negative diagnostic tests, because of underreporting of diagnosed cases, and because only passive surveillance systems are in place to detect disease occurrence. Thus, only approximately 4200 cases of LD were reported in 2011 (13.5/million population) to the CDC. This rate is lower than that reported in France in 2009, of 18.3 cases/million population, and higher than that reported for England and Wales in 2011, 4.2/million (50% acquired abroad). In contrast, prospective studies of both sporadic community-acquired and nosocomial LD have reported far more cases in the United States than would be expected on the basis of the number of cases reported to the CDC. One prospective community-based study in Ohio of adult patients with community-acquired pneumonia requiring hospitalization found that 2.4% of such patients had LD and that the disease incidence was approximately 80/million population per year.¹⁸⁹ When extrapolated to the entire population of the United States, the authors estimated that between 8000 and 18,000 cases of LD occur annually among adults requiring hospitalization for pneumonia. The incidence of LD causing community-acquired pneumonia not requiring hospitalization is not known. One small regional U.S. study estimated that the incidence of LD among outpatients treated for pneumonia was 40 to 280/million population/year.¹⁹⁰ Thus, somewhere between 18,000 and 88,000 cases of LD are estimated to occur per year in the United States, the majority of which are neither epidemic nor hospitalized. A German study of community-acquired LD that used sensitive diagnostic tests found that the annual rate was 180 to 360 cases/million population, which, if extrapolated to the United States, would mean that 56,000 to 113,000 LD cases occur per year in the United States¹⁹¹ and that only approximately 2% to 5% of U.S. cases are currently reported to the CDC. Some geographic regions appear to have more LD than others, such as western Pennsylvania and Ohio in the United States and Catalonia in Spain; whether this is due to true differences in disease incidence or better case ascertainment is uncertain. The incidence of LD in the United States and elsewhere appears to be increasing on the basis of the numbers of cases reported to public health agencies; whether this is due to more widespread use of the urine antigen test, better reporting and surveillance, or to a true disease increase is unknown.¹⁹² Estimates of LD as a cause of community-acquired pneumonia requiring hospitalization in adults range from 0.5% to 10% of all admitted pneumonia cases; an average value is probably approximately 2%, even in geographic regions with excellent diagnostic capabilities.^193–197

LD of children is thought to be uncommon, representing 1% or less of causes of pneumonia in this age group, and generally occurs as a nosocomial disease of immunosuppressed children.¹⁹⁸ Neonates may be at relatively high risk of LD because of the immaturity of their immune systems. Both nosocomially and domestically acquired cases of the disease have been reported in apparently immunologically normal newborns exposed to Legionella-contaminated water in incubators, bath tubs, and during birth.^198–200

Shortly after the 1976 Philadelphia outbreak, nosocomial LD outbreaks were reported in several cities throughout the United States and Europe. Because relatively little was known about the environmental ecology of Lp, or about optimal diagnostic methods, these outbreaks were characterized by long durations, often years in length, and high numbers of cases and fatalities. For example, the LD outbreak at the Wadsworth Veterans Administration Hospital in Los Angeles, California, resulted in more than 250 cases of disease in both patients and visitors over an 8-year period.^16,23,201 A 17-year-long outbreak of unrecognized nosocomial LD occurred at another U.S. hospital, involving many fewer patients.²⁰² Nosocomial LD epidemics continue to occur worldwide, albeit with durations measured in weeks rather than years.^203,204 Nosocomial pneumonia usually only affects a relatively small number of hospitalized patients, with attack rates less than 1% of patients.^16,205 During nosocomial outbreaks of LD, the minority (5% to 11%) of patients with nosocomial pneumonia of all etiologies have been reported to have LD,^16,206–208 although in some explosive outbreaks involving a single ward the attack rates may be much higher.

Community-based LD epidemics continue to occur more than 36 years after the 1976 Philadelphia outbreak, some of them quite extensive. Some recent large outbreaks have been in Murcia, Spain, in 2001, involving up to 700 people, with 6 deaths; Barrow-in-Furness, England, in 2002, involving more than 130 people with six deaths; and Quebec City, Canada, in 2012, involving 180 people with 13 deaths.²⁰⁹ Of note, several recent outbreaks involved people visiting a town center, rather than with being inside a certain building. Long-distance spread of disease from industrial aerosols has been reported from France and Norway.^25,210

LD affecting travelers constitutes up to half of reported cases in some countries. In many cases, a common source outbreak has been found, but for others the cases appear to be sporadic. Multiple common source outbreaks have been uncovered in travelers by a cooperative European reporting system that collects and analyzes LD cases in travelers.²¹¹ Such a well-coordinated traveler’s health monitoring and analysis system is not as well established in the United States, making detection of small travel-associated outbreaks of the disease less likely.^212,213

Despite the immense publicity usually generated by epidemics of LD, sporadic cases of the disease are about fourfold more common than linked cases of the disease. Some sporadic cases are undoubtedly the result of common source exposures. This is especially true of travelers who return to their homes during the incubation period, or while they are ill but still well enough to travel. Evidence for common source links for apparently unrelated cases comes from observations that proximity of residence to wet cooling towers is a risk factor for LD.^214,215 In one study, LD risk was approximately 1.5 times greater for those living within 1 km of a cooling tower than for those living more than 6 km from a tower; the authors estimated that approximately 20% of sporadic LD could be attributed to community cooling tower exposure.²¹⁵ Residential acquisition from drinking and bathing water, which rarely results in more than a single case, accounts for less than 10% to 15% of sporadic LD.^216,217,218

Pontiac fever often causes explosive outbreaks of disease with high attack rates. Attack rates of 70% to 90% have been reported from several epidemics.^{7,185,219–221}

LD mortality rates are highly variable, ranging from less than 1% to as high as 80%, depending on the underlying health of the patient, promptness of specific therapy, and whether the disease is sporadic, nosocomial, or part of a large outbreak.^{30,222–224,225} The lowest mortality rates, around 1%, have been observed in recent large outbreaks of the disease, whereas the highest mortality rates have been reported in untreated nosocomial disease in patients with severe underlying diseases.²²⁶ The average fatality rates for sporadic disease are estimated to be approximately 10% to 15%. Fatality rates of nosocomial disease have declined by more than 50% in the United States over the past 20 years; a similar but less dramatic decrease in death rates of community-acquired cases has also been observed.³⁰ The decline in mortality rates appears to be due to better and faster disease recognition, especially through use of the urine antigen test, and more widespread use of empiric therapy for pneumonia that includes drugs active against Lp.^30,226

Risk Factors

Because of the ubiquity of Legionella spp. in our natural and manmade environments, most people encounter these bacteria frequently. Yet most of those exposed do not get the disease. This is attributed to largely unknown host immune mechanisms and to the lack of sufficient concentrations of virulent Legionella spp. bacteria in small aerosol particles. Even during large outbreaks of LD attributed to cooling tower aerosols, the minority of people in the risk area become ill. The known host risk factors for LD are those that result in decreased local or systemic cellular immunity and those activities that increase the chances of exposure to an infectious aerosol or microaspiration of contaminated water. Also important are environmental and bacterial factors such as relative bacterial virulence, bacterial aerosol stability, organism growth conditions, and factors that facilitate the spread of the bacterium from contaminated water to the host, such as wind direction, relative humidity, and aerosol formation.

Male gender, cigarette smoking, chronic heart or lung disease, diabetes, end-stage renal failure, organ transplantation, immunosuppression, some forms of cancer, and age older than 50 years have all been found to be host risk factors for LD.^{205,217,227,228} The approximately twofold greater risk with male gender may be due to the greater prevalence of cigarette smoking and its complications in males. Cigarette smoking increases risk by approximately twofold to sevenfold. Nicotine directly impairs macrophage restriction of Lp growth, and nicotinic acid, which can be formed from oxidized nicotine, as well as being endogenous, promotes Lp virulence.^229,230 Suppression of local or systemic cellular immunity, and particularly glucocorticoid administration, increases the risk by approximately twofold to sixfold; cytotoxic chemotherapy was a risk factor in one study. Anti–TNF-α therapy has emerged as a significant risk factor for LD, including TNF-α antibodies, soluble TNF-α, thalidomide, and lenalidomide.^231,232 Anti-CD52 therapy can result in severe LD.²³³ Lung, but not gastrointestinal tract, cancer is an LD risk factor.²²⁷ A variety of hematologic malignancies have also been shown to be important risk factors, especially hairy cell leukemia.^206,227 Small single studies of genetic predisposition to LD have shown that polymorphisms in TLR4, TLR5, and mannose-binding lectin produce minor predispositions to the disease^65,72,74; it is unlikely that these polymorphisms account for the majority of human risk for LD, and confirmatory studies are necessary to better define the risks. Animal studies implicate both TLR2 and TLR9 as being important in host immunity, but to date no human data exist on the LD associated with polymorphisms in these TLRs.^66,67 Recent surgery has been an important risk factor for nosocomial disease, probably because of general anesthetic-caused defects in local lung defense, because Legionella-contaminated water is introduced into the respiratory tract in the perioperative or postoperative periods, or both factors combined.^234,235 Alcoholism has been a predisposing condition in only some studies and has never been shown to be a risk factor in multivariate analyses.^{205,206,217,236} Host risk factors for non-Lp LD appear to be similar to those for Lp infection, although data are limited; most patients with these infections are immunosuppressed or have cancer, with the exception of LD due to Llb.^237,238 There appear to be no predisposing host factors for Pontiac fever.

Activities that increase the chances of exposure to Legionella bacteria in water increase the risks of disease acquisition. Recent overnight travel, use of well water in the home, recent plumbing work inside the home, disruptions of water supply resulting in “brown” water at the water tap, living in a water distribution network with older plumbing, and using an electric water heater all increase the acquisition risk of community-acquired LD.^{215,217,218,239,240} Other community, recreational, or travel-related activities increasing the chances of acquiring the disease include the use of, or proximity to, whirlpool spas or hot water spring spas; living in close proximity to a cooling tower; or being near decorative fountains.^{214,215,228,241,242} Rarely reported risks are near-drowning including during tsunamis^{10,43,242,243} and delivery by water birth.²⁴⁴ Exposure risk factors for LD due to Llb are much different than for Lp infection, with handling of potting soil, being near dripping irrigated hanging plants, cigarette smoking, and not washing hands after use of the soil being the major risk factors.²⁴⁵

A wide range of nosocomial exposures can result in LD. Almost all involve delivery of Legionella-contaminated water into the respiratory tract and include the use of tap water filled or rinsed: nebulizers, humidifiers, oxygen humidifiers, ventilator tubing, and nasogastric feedings or lavages.^246–249 Proximity to a contaminated decorative fountain has transmitted nosocomial LD in several outbreaks.^26,250 LD acquired during dental procedures, a rare event, is the result of inhalation of dental drill cooling water aerosols.²⁵¹ Consumption of Legionella-contaminated ice can also be a risk factor for nosocomial LD.²⁵² These risk factor are in addition to exposure to Legionella-contaminated air originating from a cooling tower.^253–255 Rare cases of nosocomial Legionella wound infection have resulted from irrigation or bathing of wounds in Legionella-contaminated water.^256,257 Nosocomial transmission has occurred in hospitals, nursing homes, assisted living facilities, a dentist office, and other types of health care facilities.

Modes of Transmission

LD is transmitted from the environment to humans by inhalation of an infectious aerosol.²⁵⁸ In an unknown fraction of cases, microaspiration of contaminated water into the lungs is the mode of transmission.^249,259 Finally, massive aspiration of contaminated water into the lungs during near-drownings is an unusual but reported mode of transmission of the disease.^{10,43,242,243}

Multiple examples of exclusive aerosol transmission of LD exist, especially in epidemics having a cooling tower, water spa, water fountain, or water mister as the source of disease.^{1,241,254,255,260–265} In these cases, proximity to the aerosol generator, duration of exposure, and presence in an area downstream of the contaminated device have all been found to be risk factors for disease acquisition. Of note, when reported, either consumption of water at the epidemic site has not occurred in the majority of disease victims, or the drinking water of the outbreak site has been culture negative for Lp.

The data supporting microaspiration of water as a major mode of transmission are less convincing, but in some specific reports the evidence is compelling. These data include examples of nosocomial disease in patients whose major risk factor was nasogastric tube irrigation with tap water and interruption of nosocomial outbreaks by substituting sterile water for tap water for drinking and nasogastric tube irrigation.^{249,259,266,267} Whether microaspiration is the major transmission mode for nosocomial disease is controversial and unproven.²⁶⁸ Experimental animal data are unhelpful in this regard because both aerosol delivery and tracheal instillation produce disease.^269,270

Rare reports of peritonitis or bowel abscesses caused by Lp have led to speculation that oral ingestion may be a mode of disease transmission.^271–273 In all these cases, Lp pneumonia or empyema occurred concurrently, making unclear which organs were the ones primarily infected. In contrast to the relative ease of producing lethal guinea pig pneumonia by the aerosol or intratracheal route, huge amounts of Lp delivered by oral gavage are cleared rapidly without causing severe pneumonia,²⁷⁴ although another experimental study gave somewhat different results.²⁷⁵ It is unlikely that oral ingestion of Legionella-contaminated water is more than a minor mode of transmission of the disease.

Outbreak Investigation

Prompt notification of public health authorities of any strongly suspected or confirmed case of LD is critically important for detecting epidemics of the disease and is legally required in many regions. What appears to be only a single case of the disease may be part of an epidemic or the index case. Alert physicians have sometimes detected pneumonia clusters that led to discovery of both emerging and long-standing epidemics of LD.^1,241

In addition to reporting nosocomial LD to public health authorities, medical institutions should embark on extensive investigation of even a single case of nosocomial LD.²⁷⁶ This is because more cases may have occurred previously or will appear subsequently. Retrospective review of nosocomial pneumonia cases for a 3- to 6-month period may yield more cases. Prospective laboratory testing of all patients with nosocomial pneumonia for LD can also be useful in finding more cases, for a 3- to 6-month period. Hospital physicians should be notified to consider LD when making diagnostic and therapeutic decisions in patients with nosocomial pneumonia until it is clear that any ongoing nosocomial outbreak has been eradicated.

Investigation of both community-acquired and nosocomial outbreaks of LD requires a thorough epidemiologic investigation. The epidemiologic investigation will help generate hypotheses concerning the source of the outbreak and demonstrate risk factors using controlled studies. Environmental testing without concurrent epidemiologic investigation can lead to misleading findings, even when molecular fingerprinting is used to compare clinical and environmental isolates.^277,278 It is crucially important to have only a qualified laboratory perform environmental studies because not all environmental microbiology companies are competent in the proper collection, processing, and culturing of water specimens for Legionella spp. bacteria; certification is required in some countries, with a voluntary certification program available in the United States.²⁷⁹ It is crucially important to obtain Legionella bacteria clinical isolates as part of an epidemiologic investigation; this allows molecular fingerprinting and comparison of clinical and environmental isolates. Environmental sampling protocols for outbreak investigation are available.^279–281

Rough clues to the environmental source of an LD epidemic can be found in the pace of the outbreak and its geographic distribution. Explosive outbreaks involving tens to hundreds of people over a several-day period are most often due to a contaminated massive aerosol generator, usually a cooling tower, but sometimes a whirlpool spa or misting device. Potable water associated outbreaks may produce as many cases, but generally over a much longer period of time, such as many weeks or months. Potable water–related outbreaks are confined to a single building or building complex with common plumbing. Cooling tower–related epidemics often affect both building visitors and those within several hundred meters up to a 10-km distance from the tower.²⁵ Interior aerosol-generating devices, such as recreational spas and misters, cause disease only in building visitors.

Environmental Decontamination for Outbreaks

All aerosol-generating sources implicated or highly suspected of being a source of epidemic LD should be taken out of operation as soon as possible; this generally includes cooling towers, hot tubs, and other such devices. Implicated potable water sources such as plumbing systems should be shut down if possible. A number of guidelines exist for emergency disinfection of such sources and may differ according to local regulations or guidelines.^{276,281–283} Long-term remediation can be complex and requires expert engineering and public health advice.

Extrapulmonary Infections

Extrapulmonary infections are rare and usually occur as metastatic complications of pneumonia in immunocompromised patients. Metastatic infection has been reported almost exclusively in immunocompromised patients, or patients with fatal LD, who may develop abscesses and other infections of the brain, spleen or extrathoracic lymph nodes, and skeletal and myocardial muscles.^303–307 Other reported sites of metastatic infection have been the intestines and liver, the kidneys, the peritoneum, the pericardium, vascular shunts and grafts, bone marrow, joints, surgical wounds including prosthetic heart valves and aorta, native heart valves, the perirectal area, and the skin and subcutaneous tissues.^{271,308–319} Both reported cases of culture-confirmed native valve endocarditis were in immunocompromised patients with LD.^319,320 In some of these cases the onset of symptoms of the metastatic infection preceded the recognition of pneumonia by several days, and in other cases the metastatic infection presented days to weeks subsequent to onset of the pneumonia. In some cases, the metastatic infection site was the only evidence of relapse of infection. Three cases of metastatic infection have been reported in apparently previously healthy patients,^305,321,322 but otherwise immunosuppression is found in such patients. Local extension of a thoracic empyema into the soft tissues of the chest has been reported after thoracentesis.³²³

Rare cases of primary infection not preceded or accompanied by pneumonia have also been reported. Some of these appear to be the result of direct inoculation of Legionella-contaminated water into various tissues, usually in hosts with immune compromise. Some infections were introduced by bathing postoperative patients with contaminated tap water, by the use of therapeutic baths, and by inadvertent tap water irrigation of the mediastinum after esophageal perforation. In other cases the route and source of infection are unknown. Culture-proven sites of such infections have been surgical or other wounds, subcutaneous tissues, prosthetic and native joints, prosthetic heart valves, the mediastinum, and the respiratory sinuses.^{256,257,317,322,324–335} In contrast to the case of native valve endocarditis, almost all patients with prosthetic valve endocarditis were neither immunosuppressed nor had LD.^{324,325,328,329,336–338} Native and prosthetic joint infections have been reported uniformly in immunocompromised patients with preexisting arthritis, with one exception of a patient who was not immunocompromised but did have underlying arthritis.^{317,322,333–335,339} With one exception all joint infections occurred in the absence of LD. PCR-positive, but culture-negative, meningoencephalitis has been reported in patients without LD.³⁴⁰