Lyme disease tests are highly accurate in one narrow situation and unreliable in another. Two-tier antibody testing correctly identifies nearly all established infections that meet the CDC surveillance case definition, yet it can miss 50% or more of early infections because antibodies take weeks to appear. The test measures your immune response, not the bacterium itself, which is why timing and prior infection history matter as much as the assay.

What the two-tier algorithm actually measures

The standard approach is indirect serology. It detects antibodies your immune system makes against Borrelia burgdorferi, the spirochete that causes Lyme disease. Nothing in the test binds the bacterium directly, so a result describes your immune history, not necessarily your current infection.

Tier one is an enzyme immunoassay (EIA) or immunofluorescence assay, often run as an ELISA. It reports IgG and IgM antibodies against whole-cell or recombinant antigens, and it is calibrated to be sensitive. Sensitivity-first screening means the cutoff is set low so few true cases are missed, which knowingly produces false positives. A reactive or equivocal tier-one result moves to tier two.

Tier two is the Western immunoblot. It separates Borrelia proteins by molecular weight and looks for antibodies bound to specific bands. The CDC/Association of State and Territorial Public Health Laboratory Directors criteria require at least 2 of 3 designated IgM bands for early disease, or 5 of 10 designated IgG bands, before a blot is called positive. Band patterns matter because some Borrelia proteins cross-react with antibodies to other spirochetes and with certain viral or autoimmune targets.

The two tiers are run in sequence, not in parallel, to raise specificity. A screening test alone over-calls; the blot confirms which antibody specificities are present. Modification of these criteria, such as reading a single band as positive without the full pattern, degrades specificity and produces false positives.

Why early infection defeats the antibody test

Antibody production is a biological process with a lag, and the test cannot report what has not yet been made. In early localized infection, the immune response may be days to weeks behind the rash. This is the central reason a negative test in the first weeks does not exclude Lyme disease.

A landmark prospective study of patients with erythema migrans (the expanding rash) found that acute-phase serology was insensitive early and improved substantially over weeks. In that work, sensitivity of acute-phase testing rose markedly between the earliest and later presentations, and convalescent samples drawn weeks later converted many initially negative patients. The practical lesson: paired acute and convalescent sera, taken roughly 2 to 4 weeks apart, are more informative than a single early draw.

Two-tier testing performs best in later, disseminated disease. In early neurologic or cardiac involvement, and in Lyme arthritis, antibody levels are higher and band patterns more complete, so sensitivity approaches the high 90s in many series. The same test that misses half of early cases can be nearly definitive months in.

The mechanism is straightforward immunology. IgM appears first and fades; IgG appears later and persists. An early sample drawn before the IgG response matures yields an incomplete band pattern. A later sample catches the matured response.

Quantified accuracy across disease stages

Exact sensitivity and specificity figures vary by assay, antigen preparation, and patient population, and they should not be quoted as universal constants. The directional pattern is consistent and well documented.

Clinical scenario Typical two-tier performance Main failure mode
Early localized (rash, days) Sensitivity often around 30% to 50% False negatives from immature antibody response
Early disseminated (weeks) Sensitivity rises, often 70% to 90% Residual false negatives early in this window
Late disease (arthritis, months) Sensitivity commonly above 90% Rare false negatives in treated or immunocompromised patients
Specificity in low-prevalence settings High specificity, but low positive predictive value False positives dominate when disease is rare
Post-treatment antibody check Not a test of cure Antibodies persist for years after successful treatment

Specificity of a properly performed two-tier algorithm is generally reported in the high 90s. That sounds reassuring until you apply prevalence. A test with 99% specificity used in a population where only 1% truly has Lyme disease yields a positive predictive value near 50%. Half of positives would be false. This is the counterintuitive core of Lyme diagnostics: test accuracy depends on who you test.

This is also why guidelines discourage testing patients with nonspecific symptoms and low prior probability. A positive result in that group is more likely to mislead than to help. Bayes' rule, not assay quality, drives the interpretation.

Where the tests go wrong: edge cases

Several situations corrupt interpretation in ways a single number cannot capture.

Prior infection. IgG antibodies can persist for years after successful treatment. A positive IgG blot in someone with remote, treated Lyme disease may reflect old exposure, not active infection. There is no widely used serologic test that reliably distinguishes active from past infection.

Early antibiotic treatment. Treating before drawing blood can blunt the antibody response, producing a negative test in a patient who genuinely had Lyme disease. The test then under-represents truth.

Cross-reactivity. Antibodies to other organisms, including other Borrelia species and some viruses, can bind bands on the immunoblot. This is why the full band-pattern criteria exist rather than a single-band shortcut.

Immunocompromise. Patients with weakened antibody responses may seroconvert poorly, yielding false negatives despite infection.

Testing without a compatible clinical picture. Guidelines from the Infectious Diseases Society of America and CDC advise against testing in patients with nonspecific symptoms and no exposure history. The pretest probability is too low for a positive result to be trustworthy.

Common misconceptions, corrected

Misconception one: a negative test proves you do not have Lyme disease. False in early infection, where sensitivity is limited. A negative early test is inconclusive, not exculpatory.

Misconception two: a positive test proves active infection. False. Antibodies persist, and cross-reactivity exists. A positive result must be read against symptoms and exposure.

Misconception three: you can test to confirm cure. There is no validated serologic test of cure. Antibodies linger, so a positive result after treatment is expected and uninformative.

Misconception four: PCR or direct detection replaces serology. Direct tests exist but are not the routine blood test, and their performance depends heavily on sample type and timing. Serology remains the standard blood assay.

Misconception five: specialty labs with unvalidated panels are more accurate. Tests outside the validated two-tier framework, including some antigen or urine assays, have not demonstrated reliable accuracy and are not recommended by major public health bodies.

Red flags that a result is being misused:

A single positive band read as a positive test without full criteria. Testing done for vague symptoms with no exposure and no objective findings. A positive result used alone to justify prolonged treatment. An unvalidated commercial panel offered as definitive. Repeat serology used to monitor response to therapy. A negative early test used to rule out disease outright.

A practical testing timeline

Timing determines what a test can tell you. This table maps the usual course; individual variation is real.

Time since exposure What serology can show Practical note
Days 0 to 7 Usually negative; antibodies not yet detectable A negative result here is uninformative
Weeks 1 to 2 IgM may appear; IgG often immature Early sample frequently negative or incomplete
Weeks 2 to 4 Seroconversion common; two-tier more reliable Convalescent draw improves yield
Weeks 4 to 8 IgG response typically mature Two-tier sensitivity high in disseminated disease
Months IgG stable and often persistent Positive may reflect past, not active, infection
Years after treatment Antibodies may remain for years Not a test of cure

Actionable steps and thresholds

Use these decision rules to interpret testing sensibly. They are general principles, not individualized medical advice; a clinician applies them to a specific patient.

If erythema migrans is present, treatment decisions are typically made clinically without waiting for serology, because early tests are insensitive. If an early test is negative but suspicion remains, obtain a convalescent sample about 2 to 4 weeks later rather than repeating immediately. Do not test patients with nonspecific symptoms and no exposure; the positive predictive value is too low to trust a positive. Read every positive against pretest probability: in low-prevalence settings, confirm the clinical picture before acting. Never use follow-up serology to judge treatment success; antibodies persist regardless of cure. Avoid unvalidated antigen or urine panels; prefer the standardized two-tier algorithm run by an accredited laboratory.

The reconciliation of thresholds is this: sensitivity and specificity are test properties, but positive and negative predictive values are properties of the test plus the patient. A high-specificity test can still mislead when prevalence is low, and a sensitive test can still miss early disease when antibodies lag. Both thresholds must be applied together.

The limits of what a test can tell you

No serologic result stands alone. The test cannot date infection, cannot prove active versus past disease, and cannot confirm cure. Its accuracy is a function of disease stage, pretest probability, and whether the full validated algorithm was followed.

The honest summary is stage-dependent. In late disease with a compatible clinical picture, two-tier serology is strong. In early disease, it is frequently negative and must not be used to exclude infection. In low-prevalence populations, a positive result is as likely to be wrong as right. Accuracy is not one number; it is a conditional property that changes with timing and context, and reading it any other way produces confident conclusions the data do not support.

Reference points for clinicians and patients include CDC surveillance case definitions and testing guidance, FDA statements on unvalidated Lyme assays, and IDSA clinical practice guidelines. These sources anchor the thresholds above; they do not replace individualized clinical judgment.