How to Test for Bartonella: Methods and Accuracy

Testing for Bartonella is harder than testing for most bacterial infections, and no single test reliably catches every case. The organism hides inside red blood cells at low concentrations, cycles in and out of the bloodstream unpredictably, and grows so slowly in culture that standard methods often miss it entirely. Clinicians typically rely on a combination of serology, PCR, and sometimes tissue examination, each with meaningful blind spots. Understanding what each test actually measures and where it falls short can help you make sense of conflicting results or push for additional workup when an initial test comes back negative.

Why Bartonella Is Difficult to Detect

Most common bacterial infections show up readily on standard blood cultures. Bartonella species do not. They require specialized growth conditions and prolonged incubation that standard hospital microbiology labs do not routinely offer. Classical isolation methods are laborious, and characterizing the organism demands specific techniques that go well beyond what a routine culture provides.¹ This is a core reason why Bartonella endocarditis, for example, has historically been classified as “culture-negative” endocarditis. In one well-known pediatric case, a child presented with clear signs of infective endocarditis on the aortic valve, but repeated blood cultures were sterile; the diagnosis was only established through serology, immunohistochemistry, and PCR on the resected valve tissue.²

Another complicating factor is that Bartonella bacteremia tends to be intermittent. The bacteria do not circulate in the blood at a steady level; they appear, disappear, and reappear over days. A blood sample drawn on a “quiet” day may contain no detectable organisms even in someone with an active infection. Research has shown that testing patients three times within a one-week period significantly improves the chance of catching the bacteremia, with roughly 3.4 times the odds of a positive result compared to a single draw.³ If your doctor suspects Bartonella and the first blood test is negative, that single result does not rule it out.

Serology and Its Limits

The most widely used first-line test is serology, which looks for antibodies your immune system produces in response to Bartonella. The standard method is the indirect immunofluorescence assay, commonly called IFA. In the context of cat scratch disease, one study found that an IFA test for Bartonella henselae had a sensitivity of about 87% and a specificity around 74% at a titer cutoff of 1:64.⁴ Those numbers sound reasonable at first glance, but they mask real-world problems. The sensitivity and specificity shift depending on which commercial kit is used, which lab runs it, and what antibody class is being measured.

IgM antibodies, which are supposed to flag recent infection, perform poorly in many Bartonella assays. One Dutch evaluation found that a commercially available IFA test for IgM had a sensitivity of just 6%, meaning it missed the vast majority of acute cases. In-house IgM assays performed somewhat better but still only reached sensitivities of about 53% for IFA and 65% for ELISA. IgG-based IFA testing did better, at around 67% sensitivity, though the ELISA format for IgG was considerably worse at 28%.⁵ The takeaway is that a negative IgM result should not be used to rule out Bartonella infection, and that IgG-based IFA tends to outperform ELISA for this particular organism.

In suspected endocarditis, antibody titers tend to be much higher. Research on Bartonella endocarditis cases found that IgG titers at or above 1:800 were present in 94% of confirmed patients when tested by enzyme immunoassay.⁶ A large review of 106 endocarditis cases proposed that an IgG titer of 1:800 or above, a positive PCR from a valve or blood specimen, or a positive Western blot should count as a major criterion for diagnosing Bartonella endocarditis.⁷ So serology works better when the infection is deep-seated and the immune response is robust, but it struggles with early, mild, or localized disease.

Cross-Reactivity Can Muddy the Results

One of the most frustrating features of Bartonella serology is cross-reactivity. Antibodies produced against Bartonella can react with antigens from other organisms, and vice versa, leading to false positives or misidentification. The same endocarditis study noted frequent but low-titer cross-reactivity between Coxiella burnetii antibodies and Bartonella henselae antigen in patients who actually had Q fever endocarditis.⁶ Researchers have also documented serological cross-reactions between Bartonella and Chlamydia species. In one study, eight patients originally diagnosed with Chlamydia pneumoniae endocarditis were later shown through adsorption experiments and Western blot to actually have Bartonella infections; the serological overlap had caused the initial misdiagnosis.⁸

Additionally, cross-reactivity between B. henselae and B. quintana means serology often cannot tell you which Bartonella species is responsible. This matters because the species may differ in their clinical behavior and treatment considerations. Specialized testing such as adsorption studies or species-specific PCR is needed to pin down the exact organism, and those tests are not available at most clinical labs.

Background Seroprevalence Complicates Interpretation

Even if you get a positive antibody result, it does not necessarily mean you have an active infection. Many healthy people carry low-level Bartonella antibodies from past exposure. A German study found that roughly 30% of healthy individuals had IgG antibodies to B. henselae, and a Korean study found seroprevalence around 15% in healthy adults.⁹¹⁰ In populations with high cat ownership, the background rate tends to be even higher. Low antibody titers, then, are not sufficient evidence of active or prior clinically significant infection. A positive at 1:64 in a cat owner with vague symptoms could easily represent old, resolved exposure rather than current disease.

One way clinicians try to work around this is by checking paired sera: drawing blood at the time of initial suspicion and again two to four weeks later. A fourfold or greater rise in titer between the two samples strongly suggests active or recent infection. In practice, though, convalescent serologic testing is often not performed, particularly in pediatric cases where symptoms may resolve on their own.¹¹

PCR Testing and Why Specimen Type Matters

PCR-based tests detect Bartonella DNA directly rather than relying on the immune response. This gives them an edge in specificity, which in most studies hits 100%: a positive PCR result almost certainly means Bartonella DNA is present in the sample.¹² The catch is sensitivity, and that depends enormously on what kind of sample you send.

A 25-year retrospective analysis of PCR testing for cat scratch disease found that the type of specimen had a bigger impact on whether the test came back positive than which PCR assay was used. Pus aspirated directly from a swollen lymph node had the highest sensitivity at around 96%. Samples from the primary skin lesion (the initial scratch or papule) followed at about 88%. Fine-needle aspirations of lymph nodes reached roughly 85%, while lymph node biopsy specimens dropped to about 73%, and paraffin-embedded tissue samples were the least sensitive at around 59%.¹³ The lesson for patients is worth knowing: if PCR is being ordered on a blood draw and it comes back negative, that does not carry the same weight as a PCR performed on fluid from an infected lymph node.

Blood-based PCR has historically been the weakest link. In the endocarditis review, PCR sensitivity on whole blood was only about 33%, compared to 91% on valve tissue.⁷ The low concentration of circulating bacteria and the intermittent nature of bacteremia are the main reasons. Some labs try to improve blood-based PCR by first running the sample through an enrichment blood culture step, where the blood is incubated in a specialized growth medium for days or weeks before extracting DNA. This pre-amplification step can meaningfully boost the detection rate.

Newer Molecular Approaches

Droplet digital PCR, or ddPCR, represents one of the more promising advances. Unlike standard quantitative PCR, ddPCR partitions the sample into thousands of tiny droplets, each analyzed individually, which lets it detect much lower concentrations of bacterial DNA. In a head-to-head comparison using patient blood and enrichment blood culture samples, ddPCR detected Bartonella DNA in about 47% of samples, while standard qPCR picked up only about 5%.¹⁴ A study comparing the two methods in blood samples also found ddPCR significantly more sensitive, detecting Bartonella in 36% of blood samples where qPCR found nothing at all.¹⁵ The limitation is availability: ddPCR is not offered at most commercial or hospital labs and is still largely a research or specialty reference lab tool.

Enrichment culture combined with molecular detection is another strategy gaining ground. Adding blood supplements to the growth medium before running PCR enhances the bacterial load enough to push otherwise-negative samples over the detection threshold.¹⁶ The turnaround time is longer than a straight PCR draw because the culture step adds days or weeks, but for patients with persistent symptoms and negative initial testing, the wait may be worth it.

Metagenomic next-generation sequencing is an even newer option. Instead of looking for one specific organism, it sequences all the DNA in a sample and computationally identifies whatever pathogens are present. Researchers have used it to detect B. henselae in intraocular fluid from patients with neuroretinitis when standard serology and blood PCR were unhelpful.¹⁷ It remains expensive and is available only at select reference centers, but it can be valuable for unusual presentations where clinicians are not sure what organism to target.

Tissue Examination and Histopathology

When a lymph node is biopsied or excised, pathologists can look for the characteristic tissue pattern of Bartonella infection under the microscope: suppurative granulomas and stellate (star-shaped) microabscesses, often clustered near the outer edge of the lymph node. The Warthin-Starry silver stain has been the traditional method for highlighting the bacteria themselves within tissue. In one study of cat scratch disease patients, the silver stain detected B. henselae in about 52% of cases.¹⁸ That means nearly half of true infections were missed by staining alone. When combined with immunohistochemistry, which uses antibodies to tag the bacteria in the tissue, sensitivity improves. A case report illustrated both tools working in tandem: a patient with negative serology still had clumps of silver-stained bacilli visible in the necrotic foci of the excised lymph node, confirming the diagnosis when blood-based tests failed.¹⁹

Histopathology is best understood as a supporting test rather than a standalone diagnostic method. The tissue pattern is suggestive but not unique to Bartonella, and the silver stain’s sensitivity is mediocre. When tissue is available, sending a portion for PCR in addition to histologic examination gives the best combined yield.

Distinguishing Between Bartonella Species

Most routine serological tests cannot reliably distinguish B. henselae from B. quintana because the two species share so many surface antigens. This is more than an academic concern: B. henselae is typically associated with cat exposure and cat scratch disease, while B. quintana is historically linked to louse-borne transmission and is more common in people experiencing homelessness. Knowing which species you are dealing with can shape treatment decisions and epidemiological follow-up.

Specialized PCR assays have been developed specifically for this purpose. Duplex PCR assays can simultaneously detect and differentiate B. henselae and B. quintana in a single reaction, achieving detection limits as low as 10 copies per reaction when applied to heart valve tissue.²⁰ Another group developed TaqMan probe-based and high-resolution melting assays that were about 100 times more sensitive than conventional PCR for detecting and differentiating the two species.²¹ These methods are available at specialized reference laboratories but are not part of routine clinical panels at most hospitals.

For less commonly encountered species, the diagnostic picture is even thinner. B. bacilliformis, which causes Carrión disease in parts of South America, is typically diagnosed by thin blood smear in endemic areas, but the smear is notoriously insensitive. In a study of children in Peru, only 3% had positive blood smears while about 25% were positive by PCR performed on dried blood spots.²² Dried blood spots are easier to collect and transport than fresh blood in remote settings, making them a practical option for field surveillance.

Why Combining Tests Improves Your Odds

Given that each test has gaps, the most effective strategy is using more than one method simultaneously. Serology can be positive when PCR is negative, and the reverse is also true. In a study of veterinary personnel in Spain, for instance, some individuals who tested positive by PCR were seronegative to all Bartonella antigens tested by IFA.²³ This discordance makes sense biologically: PCR detects current bacterial DNA, while serology detects the immune response, which may lag behind or never develop at detectable levels in some individuals. Research on dogs with suspected Bartonella found that running serology, PCR, and enrichment culture at the same time enhanced overall diagnostic sensitivity compared to any single method.²⁴

A study of cat scratch disease diagnosis found that the best performance came from requiring at least two of the following to be positive: serologic, epidemiologic (cat contact), histological, and molecular criteria.¹² This composite approach reduced both false positives and false negatives. In practice, if you have a swollen lymph node after a cat scratch and your doctor is considering Bartonella, pushing for both serology and PCR on the best available specimen (ideally aspirated pus or a biopsy sample if one is taken) gives the strongest diagnostic picture.

Sample Handling and Storage

A detail that rarely gets discussed with patients but can affect results is how the sample is handled between the draw and the lab. Research on a nested PCR assay for Bartonella endocarditis found that prolonged storage of serum specimens at standard freezer temperatures significantly reduced the test’s ability to detect bacterial DNA. The same study noted that ongoing antibiotic treatment did not affect test performance, which is reassuring for patients who have already started empiric therapy before results come back.²⁵ If your samples are being shipped to a distant reference lab, it is worth confirming how the lab wants them stored and transported.

Testing Unusual Fluid Types

Bartonella can cause disease in sites beyond the lymph nodes and bloodstream. Ocular involvement, including neuroretinitis and posterior uveitis, is one of the better-recognized complications. Standard blood testing may be negative while the organism is active in the eye. In these cases, PCR on aqueous humor or vitreous fluid samples obtained during an ophthalmologic procedure has proven useful. One early case report demonstrated B. grahamii DNA in intraocular fluids from a patient with neuroretinitis, establishing the diagnosis when blood-based methods had not.²⁶ More recently, PCR on aqueous humor and metagenomic sequencing of intraocular fluid have been used to confirm ocular bartonellosis in cases where serology and clinical presentation left the diagnosis uncertain.²⁷

Western blot testing also deserves a mention as an adjunct. One study comparing Western blot to IFA for Bartonella henselae found that the Western blot format had higher sensitivity and diagnostic accuracy.²⁸ In the endocarditis review, Western blotting had perfect sensitivity among the confirmed cases, outperforming both IFA and PCR on blood.⁷ Western blot is not as widely available as standard IFA or ELISA, but it is worth asking about when initial serology returns equivocal results, particularly if endocarditis is on the differential.