October 22, 2026 · Science · Inflammation · CRP
A high C-reactive protein result does not prove that a bacterial infection is present or that antibiotics are necessary. In the 2,760-patient ADAPT-Sepsis randomized controlled trial published in JAMA, daily CRP-guided advice failed to shorten antibiotic treatment courses. Daily procalcitonin-guided advice shortened total duration by 0.88 days.
People who receive a raised CRP or high-sensitivity CRP result on an outpatient wellness panel often wonder if the test detects a hidden infection. The trial evidence shows that CRP tracks generalized systemic inflammation rather than identifying a specific bacterial cause. Whether antibiotics are indicated is a clinical judgement. A single blood marker cannot make it.
The Verdict
What the Trial Found When CRP Guided the Decision
Daily CRP-guided treatment advice did not shorten antibiotic therapy compared to standard care in critical illness. Investigators in the ADAPT-Sepsis trial enrolled 2,760 adults with suspected sepsis across 41 intensive care units in the United Kingdom between January 2018 and June 2024. Patients were randomized into three parallel groups: daily procalcitonin-guided advice, daily C-reactive protein-guided advice, or standard clinical care.
Standard care resulted in an average of 10.7 days of antibiotic treatment up to day 28. In the CRP-guided group, patients received a mean of 10.6 days of antibiotics. That represented a difference of 0.09 days, with a 95 percent confidence interval spanning minus 0.60 to positive 0.79 days. Because the confidence interval crosses zero, this change was not statistically significant.
In contrast, the procalcitonin-guided group averaged 9.8 days of antibiotic therapy. That achieved a statistically significant reduction of 0.88 days, with a 95 percent confidence interval from 0.19 to 1.58 days. Even in an intensive care setting with daily blood sampling, a CRP number could not shorten treatment.
| Trial Arm | Cohort Size | Mean Antibiotic Days | Difference vs Standard Care |
|---|---|---|---|
| Standard Care | 918 patients | 10.7 days | Reference group |
| Procalcitonin-Guided | 918 patients | 9.8 days | 0.88 days shorter (95% CI 0.19 to 1.58), significant |
| CRP-Guided | 924 patients | 10.6 days | 0.09-day difference (95% CI -0.60 to 0.79), not significant |
Why the Headline Figure Is Not the Trial Result
Headlines reporting that blood tests allowed clinicians to stop antibiotics nearly two days earlier describe protocol advice rather than actual patient outcomes. A press announcement from the University of Manchester, covered by MedicalXpress, highlighted findings from a companion process-evaluation paper in Anaesthesia. That secondary analysis examined more than 17,000 blood samples and treatment records collected during the trial.
The companion paper evaluated algorithm recommendations rather than bedside prescribing. The procalcitonin protocol generated a discontinuation recommendation nearly two days before the CRP protocol produced one. That gap measures protocol advice against protocol advice. It does not measure the actual duration of medications administered to human patients in hospital beds.
The actual patient outcome in the randomized trial was a 0.88-day reduction with procalcitonin and no significant change with CRP. The publisher returned an access error for the companion report, and the paper is not yet indexed, so none of its internal figures could be checked here. An algorithm's stop recommendation and a patient's last dose of antibiotics are two different measurements. The trial measured the second one.
The Mechanism Behind Why CRP Lagged Behind Procalcitonin
The kinetic profile of C-reactive protein explains why it failed as a daily stopping instrument. The ADAPT-Sepsis researchers noted that procalcitonin concentrations rise earlier following a bacterial stimulus and normalize faster as infection resolves. C-reactive protein is an acute-phase protein that moves slowly. It rises later and clears more slowly during critical illness.
This kinetic delay directly resolves an apparent paradox in published laboratory research. At Longevity Benchmark, our page on high CRP levels reports a published receiver operating characteristic area under the curve of 0.86 for CRP against 0.82 for procalcitonin. That page also records the finding that goes with those two numbers: neither marker reaches sufficient standalone accuracy to confirm or exclude systemic bacterial infection without concurrent clinical evaluation. That AUC metric assesses diagnostic accuracy at a single point in time, measuring how well a baseline test distinguishes bacterial illness from non-bacterial illness.
ADAPT-Sepsis evaluated a different question: treatment-timing utility across consecutive days. A biomarker can possess reasonable diagnostic discrimination at presentation yet perform poorly as a daily progress monitor. CRP concentrations fall slowly, so a level can still read high while an infection is already resolving. A slowly clearing marker cannot tell a physician when it is safe to stop treatment.
- Biological production. CRP rises later than procalcitonin during acute illness.
- Clearance kinetics. CRP clears and normalizes more slowly than procalcitonin.
- Diagnostic non-specificity. Severe tissue injury and major systemic inflammation raise CRP without identifying the cause.
What Does and Does Not Transfer to a Longevity Panel
The findings from an intensive care unit trial cannot be directly applied to healthy people taking preventive longevity blood tests. Participants in ADAPT-Sepsis had an average age of 60.2 years, were 60.3 percent male, had a mean Sequential Organ Failure Assessment score of 7, and were already receiving intravenous antimicrobial therapy. They were critically ill individuals with life-threatening organ dysfunction.
Outpatient wellness panels measure high-sensitivity C-reactive protein, or hs-CRP, in people who feel well and have no symptoms. The standard CRP assay used in acute illness tracks major systemic inflammatory responses across higher ranges. The high-sensitivity assay used for cardiovascular risk evaluates lower ranges, with bands covered in our guide to hs-CRP reference ranges.
The ADAPT-Sepsis trial does not show that hs-CRP is useless on a preventive panel. It shows that high CRP values cannot confirm bacterial infection or time the cessation of therapy.
| Feature | Critical Care CRP Testing | Outpatient hs-CRP Testing |
|---|---|---|
| Target Population | Patients with suspected sepsis in ICUs | Asymptomatic adults screening for longevity |
| Measurement Range | Standard acute-illness assay | High-sensitivity cardiovascular-risk assay |
| Clinical Objective | Guiding acute antibiotic cessation | Assessing cardiovascular risk and baseline inflammation |
| Sampling Frequency | Daily serial venous draws | Periodic outpatient testing |
28-Day Mortality in All Three Arms
Shortening antimicrobial therapy is beneficial only if it does not increase patient mortality. In the ADAPT-Sepsis trial, all-cause mortality at 28 days was 19.4 percent in the standard care group, 20.9 percent in the procalcitonin group, and 21.1 percent in the CRP group. Procalcitonin met the trial criteria for noninferiority against a predefined 5.4 percentage point margin.
The safety analysis for C-reactive protein yielded an inconclusive finding. The absolute difference in 28-day mortality between the CRP arm and standard care was 1.69 percentage points, with a 95 percent confidence interval from minus 2.07 to positive 5.45 percentage points. Because the upper confidence boundary crossed the 5.4 percent margin, noninferiority could not be proven.
The data show that both biomarker groups carried confidence intervals compatible with roughly a 5 percentage point increase in deaths. The trial aimed for a 1-day reduction in antibiotic duration. Procalcitonin achieved 0.88 days.
What a Raised CRP on a Panel Warrants
A raised CRP level on a commercial screening panel reports inflammation and does not identify its cause. What it warrants depends on the value and on the symptoms alongside it. Our guide to high CRP levels sets out the published bands, including the point above 100 mg/L at which a result calls for prompt contact with the clinician who ordered the test. Elevated readings can stem from acute bacterial infection, severe tissue injury, or major systemic inflammation. A blood test alone cannot identify the anatomical site or microbiological cause of inflammation.
Physicians interpret an unexpected CRP elevation by pairing the number with a physical examination, clinical history, and repeat testing. Our review of chronic inflammation effects discusses how context and repeat testing guide clinical interpretation.
Antibiotic decisions belong entirely to the treating clinician.
What This Trial Could Not Answer
Four questions are left open, and the trial's own design explains why.
- The early reduction did not hold. Secondary endpoints over the initial sepsis period showed reductions of 1.13 days for procalcitonin and 0.71 days for CRP, and neither was maintained to day 28.
- Protocol advice was not binding. The University of Manchester announcement reports that clinicians declined stop recommendations because of existing hospital microbiology guidance setting fixed durations, ongoing severe illness, persistent fever, or raised inflammation markers. The same announcement reports clinicians were more likely to follow the CRP-based advice, which it attributes to that advice more closely matching existing prescribing habits.
- The 10 percent figure is a projection. The authors describe procalcitonin's reduction as equal to about a 10 percent cut in antibiotic use, with possible cost, labour and resistance benefits. Those are their projections, not outcomes this trial measured.
- No outpatient result exists. The trial enrolled no outpatients and no well people, so it reports nothing about CRP on a consumer panel.
Where CRP Sits Among Other Inflammation Markers
CRP reads as a general inflammation marker, and in critical care procalcitonin tracks the same illness faster. Neither marker names the disease process behind the number. The acute, high-range use of CRP in this trial is a different question from the low-range hs-CRP band a preventive panel reports. Our comparison of inflammation blood tests sets the markers side by side, ESR versus CRP covers the older marker clinicians still order alongside it, and hs-CRP versus glycans covers the newer one.
It is not an antibiotic trigger. In critical care, procalcitonin provides superior kinetic tracking for stopping decisions. In outpatient longevity medicine, hs-CRP serves as a baseline vascular metric. Neither marker replaces a thorough medical evaluation by a qualified physician.
Frequently Asked Questions
Does a high CRP mean I need antibiotics?
No. An elevated C-reactive protein result does not mean you need antibiotics. In the 2,760-patient ADAPT-Sepsis trial, daily CRP-guided advice changed total antibiotic duration by only 0.09 days (95% CI -0.60 to 0.79), which was not statistically significant. That trial enrolled critically ill adults with suspected sepsis across 41 UK intensive care units who had a mean SOFA score of 7. These patients already had intravenous antibiotics running within 24 hours of critical-care admission, with at least 72 hours of treatment expected. The clinical question being tested was when to stop existing therapy, not whether an outpatient should start it. Antibiotic decisions belong to the treating clinician.
Does a CRP test show infection?
A CRP test measures systemic inflammation, but it cannot independently confirm or rule out an infection. Elevated levels occur with acute bacterial infection, severe tissue injury, or major systemic inflammation. As this site notes from published diagnostic literature, neither CRP nor procalcitonin reaches sufficient standalone accuracy to confirm or exclude systemic bacterial infection without concurrent clinical evaluation. The ADAPT-Sepsis trial enrolled 2,760 adults across 41 UK intensive care units between 1 January 2018 and 5 June 2024 to evaluate daily management rather than initial diagnosis. A high reading must always be evaluated in clinical context alongside physical signs and symptoms by a treating clinician.
Can a CRP test detect pneumonia?
A CRP test cannot diagnose pneumonia on its own. Diagnosing an acute respiratory illness requires concurrent clinical evaluation, physical examination, and diagnostic imaging. As published medical literature indicates, neither CRP nor procalcitonin reaches sufficient standalone accuracy to confirm or exclude systemic bacterial infection without concurrent clinical evaluation. In the ADAPT-Sepsis trial, which enrolled 2,760 critically ill adults across 41 UK intensive care units, patients already had intravenous antibiotics started within 24 hours of critical-care admission with at least 72 hours of treatment expected. The marker was evaluated as an instrument for stopping antibiotics, never as a tool to diagnose pneumonia or any other infection.
Can a CRP test detect cancer?
A CRP test is not a diagnostic test for cancer, and the ADAPT-Sepsis trial did not evaluate malignancy. The trial studied 2,760 critically ill adults with suspected sepsis in 41 intensive care units to determine whether daily biomarker advice could safely reduce antibiotic duration. The trial dataset records no outpatient, well-person, or consumer-panel results for CRP, and it provides no data on cancer detection or screening. For readers examining panel results, our guide to inflammation blood tests details what the inflammation markers covered on this site do and do not distinguish. Any concerns regarding cancer screening require direct medical consultation and clinical evaluation with a physician.
Is a high CRP dangerous?
An elevated CRP level is not a diagnosis in itself, but severe elevations require urgent clinical attention. As outlined on this site's review of high CRP levels, a blood test dangerous level generally exceeds 100 mg/L, pointing to acute bacterial infection, severe tissue injury, or major systemic inflammation. In the ADAPT-Sepsis trial, 28-day all-cause mortality was tracked as the primary safety outcome, recording 19.4% in standard care, 20.9% with procalcitonin, and 21.1% with CRP. Those figures reflect critically ill intensive care unit patients with suspected sepsis, not an outpatient reading a wellness panel. Any elevated test result warrants clinical evaluation by a physician.
What is the difference between CRP and procalcitonin?
C-reactive protein and procalcitonin differ primarily in their biological kinetics during acute illness. Procalcitonin rises earlier and normalizes faster than CRP. In the ADAPT-Sepsis trial, which randomized 918 patients to procalcitonin-guided advice, 924 to CRP-guided advice, and 918 to standard care, this kinetic difference led to distinct treatment outcomes. In the secondary initial sepsis period, procalcitonin advice cut antibiotic duration by 1.13 days (95% CI 0.58 to 1.68) and CRP advice cut it by 0.71 days (95% CI 0.16 to 1.26), though neither reduction was maintained to day 28. This divergence reflects a timing property rather than a difference in baseline diagnostic accuracy.
Does this trial mean the hs-CRP on my longevity panel is useless?
No. The ADAPT-Sepsis trial does not show that hs-CRP on a longevity panel is useless. The trial enrolled no outpatients and no well people; it enrolled 2,760 critically ill adults with suspected sepsis admitted to intensive care units. Its clinical question was whether daily blood testing could safely guide when to stop intravenous antibiotics in critical illness, not whether a baseline marker has predictive utility in healthy individuals. Outpatient wellness panels use a high-sensitivity assay (hs-CRP) measuring baseline cardiovascular risk across low concentrations, whereas critical care uses a standard acute-phase assay across high concentrations. The trial simply reports nothing about consumer panel testing.
How high can CRP go?
CRP has no fixed ceiling, and the ADAPT-Sepsis trial did not publish the thresholds its daily protocols used. On the reference bands this site publishes by result range, a level above 100 mg/L counts as a dangerous result and points to acute bacterial infection, severe tissue injury or major systemic inflammation. The number on its own still does not name a cause, so clinicians read it alongside symptoms, examination and repeat testing. Our guide to high CRP levels sets out the full bands.