In 2007 and 2008, staff from the Vector-borne Diseases Section of the California Department of Public Health (CDPH) participated in San Francisco’s Project Homeless Connect (SFPHC). Under the auspices of SFPHC medical services, hair, body, and clothing of homeless persons were inspected for lice. Any lice found on the head with the presence of nits were considered to be head lice. Any lice on the body or clothing were considered to be body lice. Most infested persons self-referred directly to the CDPH booth at this event, with the exception of 1 physician referral. Lice were collected by using forceps, identified, sorted by subspecies, and placed in screw-top vials filled with 95% ethanol. Only a portion of the total lice infesting a person were collected for testing. The lice were shipped to the Centers for Disease Control and Prevention (Fort Collins, CO, USA) for detection and identification of Bartonella spp.

Lice were pooled by host and then subspecies. Samples from hosts with >20 lice were further tested individually to obtain an estimate of Bartonella spp. prevalence in the lice. We tested 36 pools of body lice, 7 pools of head lice, 108 individual body lice, and 4 individual head lice. Individual or pooled (2–20 lice/pool) samples were suspended in 250 μL of sterile phosphate-buffered saline and homogenized in an MM300 mixer (Retsch, Newtown, PA, USA) for 8 min. DNA was extracted from the homogenates by using a Mini Kit (QIAGEN, Valencia, CA, USA) and the Blood and Body Fluid Spin Protocol according to the manufacturer’s protocol with a few minor changes. A PCR was performed in 20 μL of the mixtures containing 4–20 ng of the extracted DNA, 20 μL of 2× Ampdirect Plus, 0.5 U of Ex Taq Hot Start Version (Takara Bio, Otsu, Japan), and 1 pmol of each primer. Bartonella DNA was amplified by using gltA (citrate synthase gene) and ftsZ (cell division protein gene) primers as reported (19,20) in a thermalcycler (iCycler; BioRad, Hercules, CA, USA). A strain of B. washoensis was used as a positive control, and sterile deionized water was used as a negative control. Using gel electrophoresis on a 2% agarose gel, we examined the PCR products for 900-bp (ftsZ) and 380–400-bp (gltA) fragments.

The PCR amplicon of each gene was purified by using a QIA quick PCR purification kit (QIAGEN). Direct DNA sequencing of the purified PCR amplicons was conducted by using the BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA, USA) with the specific primers described above on a Model 3130X Genetic Analyzer (Applied Biosystems). Sequence data of each gene were aligned and compared with type strains of Bartonella spp. in GenBank by using MegAlign software (DNA Star Inc., Madison, WI, USA).

In 2007 and 2008, 138 homeless persons had consultations at the CDPH booth at the SFPHC event. Of these persons, CDPH staff observed 33 persons with body lice infestations (23.9%) and 624 lice were collected (mean 18.9 lice/infested host). Head lice infestations were detected in 12 (8.7%) persons and 70 lice were collected (mean 5.8 lice/infested host). Six persons (4.3%) had body lice and head lice infestations.

Bartonella DNA was detected in body lice collected from 11 (33.3%) persons (Table 1) and in head lice collected from 3 (25.0%) persons (Table 2). Nine pools of body lice (n = 2–20, mean infection rate [MIR] 5%) from 9 infested persons and 2 pools of head lice (n = 7–12, MIR 8.3%) from 1 infested person showed evidence of Bartonella DNA. Additional lice from persons with positive pooled samples of body lice (SFB6 and SFB24) were tested individually. Sample SFB6 had 13 (87%) of 15 lice positive for Bartonella DNA. Sample SFB24 had 27 (64%) of 42 lice positive for Bartonella DNA. One of the 4 individual head louse samples (SFH2) showed amplification of Bartonella DNA (Table 2).

Host sample SFB16 showed no amplification of Bartonella DNA in its original test but when an additional 3 pools of 20 lice and 11 individual lice were tested, 1 pool of 20 lice was positive. This host had a massive body louse infestation; 91 lice were collected from his clothing. Host sample SFB27 was also negative in its first test of a pool of 20 lice; 7 additional lice tested afterwards produced a single detection of B. quintana DNA in a body louse (14%).

Samples from 1 person who was co-infested with body lice and head lice were positive for Bartonella DNA by the gltA gene PCR (SFB17, 1 pool of 20 lice) in body lice, but not in the head lice pool (SFH7, n = 4). Samples from another co-infested person were negative for Bartonella DNA in 1 pool of 5 body lice (SFB10). Bartonella DNA was detected in a pool of 12 head lice (SFH3, MIR 8.3%) (Tables 1, 2).

Thirteen (86.7%) and 12 (80.0%) body lice from host SFB6 had Bartonella DNA amplification by gltA and ftsZ, respectively. Twenty-five (59.5%) and 23 (54.8%) of individual body lice from host SFB24 had Bartonella DNA amplification by gltA and ftsZ, respectively. Two samples from hosts SFB6 and SFB24 were sequenced and found to be identical with B. quintana type strain Fuller for both genes. Host sample SFB13 had Bartonella DNA amplification for both genes, and showed a sequence identical to B. quintana type strain Fuller for the gltA product and 99.9% homology to the same type strain for the ftsZ product. One of the individual head lice samples, SFH2, showed positive amplification of the ftsZ and gltA genes. Sequencing showed that this sample was B. quintana type strain Fuller for the gltA product.

Our study has shown that homeless persons in the San Francisco Bay area have body and head lice that harbor B. quintana type strain Fuller. Prevalence of B. quintana in body lice from homeless persons (33.3%) in our study was slightly higher than the prevalence reported by Sasaki et al. in Tokyo, where body lice in 2 (16.7%) of 12 homeless persons were infected with B. quintana (17). Furthermore, similar prevalence of B. quintana infection in body lice has been reported from Russia (12.3%) (21) and Marseille, France (20%) (11).

Although Sasaki et al. detected B. quintana DNA in head lice by using molecular detection methods, their samples came from children in Nepal who also had body lice (18). However, there is no strong evidence that head lice are vectors of this organism between human hosts. Moreover, Fournier and others tested 143 head lice from schoolchildren from 8 countries and found no B. quintana (22) We have detected B. quintana in head lice from persons without a known concurrent body louse infestation. Further work is needed to examine how homeless persons acquire lice and which groups may be predisposed to louse infestation and B. quintana infection.