What's actually different about it
A 3rd-generation test looks for IgG antibodies against HIV. Antibodies take time to appear — typically three to four weeks, sometimes longer. That delay is the window period, and it's why older guidance says 12 weeks.
A 4th-generation test adds p24 antigen to the panel. p24 is a viral capsid protein that shows up in the blood before antibodies do — often within two to three weeks of exposure. Catching the infection at the antigen stage rather than the antibody stage is what removes roughly half the waiting time.
The test doesn't become more accurate in some general sense. It becomes able to answer the question earlier, because it's looking at a signal that arrives earlier.
Where the 45-day number comes from
CDC publishes a range rather than a single figure: a 4th-generation antigen/antibody lab test using blood from a vein usually detects HIV 18 to 45 days after exposure. Forty-five days is the outer edge of that range, not the typical case.
BHIVA in the UK uses 45 days as the standard testing window for laboratory blood tests, and arrived there after reviewing when the small number of late-detecting infections actually turn positive. UK guidance shortened it from an earlier, more conservative figure precisely because the data supported a shorter window.
So 45 days isn't an arbitrary round number someone picked. It's where the tail of the detection curve flattens out — the point past which retesting finds almost nobody new.
Why 28 days isn't the answer
People often assume that if the window is 18–45 days, then day 28 — roughly in the middle — is "basically fine." It's a reasonable guess and it's the wrong one, for a specific reason.
The distribution isn't symmetric. Most infections become detectable well before the midpoint; the range's upper end exists to catch a long, thin tail. Cutting the window in half doesn't halve the certainty — it puts you in the region where most of the remaining undetected cases live.
In practice: a negative at 4 weeks is genuinely good news and materially lowers the probability. It just isn't the same statement as a negative at 45 days, and if you're going to test once and be done, 45 days is the number that lets you be done.
When 45 days is not the end
Four situations genuinely require a later test, and they're worth knowing because a clinician will ask about all four:
PEP was started after the exposure. Antiretrovirals suppress the virus, which delays both antigen and antibody detection. Standard practice is to test at the end of the 28-day course and again at 3 and 6 months.
You're taking PrEP or any HIV medication. Same mechanism, same consequence.
Your immune system is suppressed by illness or treatment — chemotherapy, immunosuppressants after transplant, advanced kidney or liver disease.
There was a new exposure during the window. The clock starts over. This is the most common reason a "45-day negative" isn't final, and the one most people can control.
What the result does and doesn't tell you
Negative at 45 days, no exceptions: you don't have HIV from that exposure. There's no further test that would add information. If you test again and it's negative again, that was never going to be anything else.
Reactive (positive) on a 4th-generation screen: this is not a diagnosis. Screening tests are tuned to miss nothing, which means they flag some samples that turn out negative on confirmatory testing. The lab runs confirmation on the same sample. Wait for that result before drawing any conclusion, and don't buy another rapid test in the meantime — it can't answer the question the confirmatory test is answering.
Negative but you had symptoms at week 3: symptoms don't change the window. Acute HIV symptoms, when they happen, look like every other viral illness — fever, sore throat, rash, swollen glands. Around half of people never have any. Symptoms are not a reason to extend testing beyond the standard window, and they're not a reason to shorten it.