2026-03-17 オックスフォード大学
＜関連情報＞
- https://www.ox.ac.uk/news/2026-03-17-international-trial-finds-rapid-diagnostic-testing-alone-does-not-reduce-antibiotic
- https://www.thelancet.com/journals/lanprc/article/PIIS3050-5143(25)00104-9/fulltext
プライマリケアにおける急性呼吸器感染症に対する抗生物質処方を安全に削減するためのポイントオブケア検査戦略と通常ケアの比較（PRUDENCE）：13か国で実施された実用的無作為化比較試験 Point-of-care testing strategy versus usual care to safely reduce antibiotic prescribing for acute respiratory tract infections in primary care (PRUDENCE): a pragmatic, randomised controlled trial in 13 countries
Alike W van der Velden, PhD ∙ Prof Samuel Coenen, PhD ∙ Emma Harper, MSc ∙ Marilena Anastasaki, PhD ∙ Prof Sibyl Anthierens, PhD ∙ Femke Böhmer, MD ∙ et al.
The Lancet Primary Care Published: March 4, 2026
DOI:https://doi.org/10.1016/j.lanprc.2025.100104
Summary
Background
Point-of-care testing to guide antibiotic prescribing decisions is promoted for antibiotic stewardship. We investigated whether a point-of-care testing strategy for acute respiratory tract infections could safely reduce antibiotic prescribing in primary care.
Methods
PRUDENCE was a pragmatic, randomised controlled clinical trial in 13 countries. Patients aged 1 year or older were eligible if they presented in primary care practices or long-term care facilities with symptoms of a respiratory tract infection with either cough (lasting <28 days) or sore throat (lasting <14 days) as the predominant symptom and the clinician was considering or had planned to prescribe antibiotics for them. Participants were randomly assigned to a point-of-care testing strategy plus usual care or usual care only. Before random assignment, patients could be tested for SARS-CoV-2. Participants with negative or unknown SARS-CoV-2 status were randomly assigned (1:1:1 or 1:1) according to their predominant symptom and influenza season to either a C-reactive protein (CRP) test, a group A streptococcus test, an influenza test, a group A streptococcus test plus an influenza test, or to usual care. SARS-CoV-2-positive participants were randomly assigned (1:1) to CRP testing or usual care. Randomisation was done via five modules according to predominant symptom, influenza season, and SARS-CoV-2 status and further stratified by setting (primary care or long-term care), age group, and presence of comorbidity. Varying block sizes of two and four were used for 1:1 allocations, with block sizes of three and six for 1:1:1 allocations. The coprimary outcomes were the proportion of participants prescribed antibiotics over 28 days (hypothesising superiority, with at least a 15% reduction in the point-of-care group compared with the usual care group required) and number of days to return to usual daily activities (hypothesising non-inferiority, with the lower bound of a two-sided 95% CI of the hazard ratio exceeding 0·8). The primary analysis population included all eligible, randomly assigned participants for whom data were available, regardless of intervention received, up to the point of any loss to follow-up, withdrawal, or death. The trial was registered with the ISRCTN (ISRCTN13336322) and is complete.
Findings
Between Dec 15, 2021, and Jan 28, 2024, 2642 patients were randomly assigned: 1449 participants to the point-of-care testing strategy and 1193 to usual care. 2433 (92·1%) patients were in primary care (Belgium, France, Georgia, Germany, Greece, Hungary, Ireland, Poland, Spain, and the UK), and 209 (7·9%) were in long-term care facilities (Italy, Portugal, Israel, France, Ireland, and Spain). 2639 participants were included in the primary analysis (1448 [54·9%] assigned to point-of-care testing strategy, 1191 [45·1%] assigned to usual care; 1641 [62·2%] female patients and 998 [37·8%] male patients). Median follow-up was 28 days (IQR 28–28). Antibiotics were prescribed for 561 (47·1%) of 1191 participants in the usual care group and 662 (45·7%) of 1448 in the point-of-care testing group (adjusted risk difference –1·3% [95% CI –4·9 to 2·3]; p=0·47). The median number of days to return to usual daily activities was 4 days (IQR 2–8) for the point-of-care testing group and 4 days (2–7) for the usual care group, with an adjusted hazard ratio of 1·00 (95% CI 0·92 to ∞; p<0·0001). There were 42 reported serious adverse events (26 in the point-of-care testing group and 16 in the usual care group), including five deaths (three in the point-of-care testing group and two in the usual care group), all of which were assessed as being unrelated to the intervention.
Interpretation
A point-of-care testing strategy for respiratory tract infection, which included testing for CRP, group A streptococcus, and influenza, did not reduce antibiotic prescribing when clinicians were considering prescribing or had planned to prescribe an antibiotic. Point-of-care testing is unlikely to be effective as a standalone solution in antimicrobial stewardship.
Funding
Innovative Medicines Initiative Joint Undertaking 2.
