Cat-scratch disease and its relatives sit in an unusual corner of infectious disease: caused by a slow-growing, hard-to-culture bacterium that mostly produces a mild illness people never bring to a doctor, yet capable, in a minority of cases, of serious disease in the eye, heart, liver, or brain. This article explains where doxycycline fits into that picture, what the actual evidence shows for each form of Bartonella infection, and why the honest answer is more nuanced than "doxycycline cures cat-scratch disease."

The Bartonella Family: One Genus, Several Diseases

Bartonella henselae, carried by cats and transmitted through a scratch, bite, or flea dirt worked into a wound, causes classic cat-scratch disease: a tender papule at the injury site followed one to three weeks later by swollen, often painful lymph nodes near the scratch, sometimes with low-grade fever and fatigue. In healthy people this is, in the great majority of cases, a self-limited illness. The lymph nodes typically shrink over two to four months whether or not antibiotics are given, and this fact matters for how the treatment decision should actually be framed.

The same organism, and its cousin Bartonella quintana (historically the cause of trench fever in soldiers living in unsanitary conditions, and still seen today among people experiencing homelessness with body lice exposure), can also produce far more serious disease: neuroretinitis with vision loss, granulomatous inflammation of the liver and spleen, encephalopathy, and, in people with weakened immune systems, a proliferative vascular skin and organ disease called bacillary angiomatosis. Both species, along with Bartonella bacilliformis in South America, are recognized causes of culture-negative infective endocarditis, an infection of the heart valves that is fatal without treatment. This spectrum, from a nuisance swelling that resolves on its own to a life-threatening valve infection, is why no single sentence can honestly describe "Bartonella treatment."

Why Doxycycline Suits This Particular Organism

Doxycycline belongs to the tetracycline class, a family of antibiotics originally derived from Streptomyces soil bacteria in the 1940s and 1950s. It works by binding the bacterial ribosome's 30S subunit and blocking protein synthesis. There is something worth pausing on in the fact that some of medicine's most durable and versatile antibiotics were sitting quietly in ordinary dirt long before anyone knew to look for them; it is a small, recurring reminder that the created world has been stocked with more provision than we tend to notice.

Bartonella species are fastidious and largely intracellular, hiding inside red blood cells and the cells lining blood vessels (endothelial cells). This matters clinically because an effective drug needs to reach and act inside those cells, not just in the bloodstream. Doxycycline is highly lipid-soluble, penetrates tissues well, achieves meaningful intracellular concentrations, and crosses into the eye and, to a useful degree, the central nervous system. Those pharmacological properties, rather than any single definitive human trial, are the main reason clinical guidelines lean on doxycycline for the deeper, harder-to-reach forms of Bartonella disease, particularly neuroretinitis, hepatosplenic infection, and endocarditis.

What the Evidence Actually Shows

It is important to separate what is well established from what is reasonable extrapolation.

For uncomplicated cat-scratch lymphadenopathy in an otherwise healthy person, the strongest human evidence involves a different drug. A small randomized, double-blind, placebo-controlled trial published in the Pediatric Infectious Disease Journal in 1998 by Bass and colleagues tested azithromycin in 29 patients and found significantly faster shrinkage of the lymph node in the first month, though the two groups looked similar by later follow-up. This is a modest trial by modern standards, but it is essentially the only placebo-controlled data anyone has for treating garden-variety cat-scratch disease, and it does not involve doxycycline. Guidelines from the Centers for Disease Control and Prevention and pediatric infectious disease references therefore describe uncomplicated disease as usually not requiring antibiotics at all, with azithromycin as an option to consider if the nodes are especially large or painful.

Doxycycline's "first-line" standing applies specifically to the complicated and systemic manifestations, where the case is built less on large randomized trials, which are difficult to run for a rare disease, and more on case series, expert consensus, and the drug's known tissue behavior:

None of these uses appear on doxycycline's original FDA-approved label, which predates the identification of Bartonella as a distinct pathogen in the 1990s. This is off-label use in the ordinary, unremarkable clinical sense: a well-understood drug applied to a disease according to specialty society guidelines and accumulated clinical experience, which is how a great deal of infectious disease treatment actually works when a pathogen is rare and a pharmaceutical company has little incentive to pursue a formal indication.

Dosing, Duration, and Who Should Not Take It

For adults, doxycycline is typically dosed at 100 mg twice daily. Duration varies enormously by syndrome: a couple of weeks for milder systemic disease, four to six weeks for endocarditis, and potentially months for bacillary angiomatosis in a person whose immune system remains impaired. This variation is a good reason the specific regimen and length of treatment should be set by a physician managing the individual case rather than inferred from a general article.

Doxycycline carries real limitations. It is generally avoided in pregnancy, since tetracyclines can affect fetal bone development and cause permanent discoloration of a child's teeth if taken during tooth formation; for a pregnant woman with Bartonella infection, a physician will weigh alternatives such as azithromycin or rifampin. The same tooth-staining concern historically kept doxycycline away from young children, though the American Academy of Pediatrics has since noted that short courses at standard doses do not appear to cause the staining seen with older, longer-acting tetracyclines, and it is now used more freely in children when clinically indicated, always as a considered decision between parent and doctor rather than a reflexive one. Photosensitivity, gastrointestinal upset, and esophageal irritation if taken without enough water are common, manageable, and worth knowing about in advance rather than discovering by surprise.

When No Antibiotic Is the Right Answer

The most important thing a well-informed patient or parent can take from the research is that treating cat-scratch disease is not automatically about which antibiotic to reach for. In a healthy child or adult with a swollen node and no worrying features, watchful waiting, comfort measures, and time are a legitimate, evidence-consistent choice, and it spares the body an unnecessary course of antibiotics. The decision to treat, and with what, should turn on the presence of the features that mark more serious disease: high or prolonged fever, visual symptoms, an abnormal heart valve, or a compromised immune system. That is precisely the kind of judgment call best made between a patient and a physician who can examine the whole clinical picture, not settled in advance by a general reference article. Readers who suspect a cat scratch has led to more than a local swelling, particularly with fever that will not settle or any change in vision, should be evaluated promptly rather than waiting to see if it resolves.

Key takeaway: Doxycycline is the guideline-preferred antibiotic for serious and systemic Bartonella infections such as endocarditis, neuroretinitis, and bacillary angiomatosis, but most ordinary cat-scratch disease resolves on its own and the choice to treat, and with what, belongs to a conversation between patient and physician.