Publication

• Title: Cold and Room-Temperature Platelets in Cardiac Surgery: The CHIPS Randomized Clinical Trial
• Acronym: CHIPS — Chilled Platelet Study
• Year: 2026
• Journal published in: JAMA
• Citation: Spinella PC, Zantek ND, VanBuren JM, et al; CHIPS Investigator Group. Cold and room-temperature platelets in cardiac surgery: the CHIPS randomized clinical trial. JAMA. Published online August 17, 2026.

Context & Rationale

• Background

  • Cardiopulmonary bypass produces thrombocytopenia and platelet dysfunction, and platelet transfusion is frequently used when bleeding persists during complex cardiac surgery.
  • Conventional room-temperature platelets are stored at 20–24°C with agitation but have a shelf life of only 5–7 days because of bacterial proliferation and progressive loss of function.
  • The short shelf life, together with the time required for bacterial testing or pathogen reduction, contributes to platelet shortages, wastage and the inability of smaller hospitals to maintain an inventory.
  • Refrigeration at 1–6°C reduces bacterial growth and may preserve or enhance immediate haemostatic activity for 14–21 days, but cold-stored platelets have shorter post-transfusion survival and produce smaller circulating platelet-count increments.
  • A preceding 50-patient randomised pilot trial found no significant difference in postoperative chest-drain output between room-temperature and cold-stored platelets stored for up to 7 days; a subsequent non-concurrent cohort supported the feasibility of extending cold storage to 8–14 days.¹
• Research Question/Hypothesis

  • Were cold-stored apheresis platelets non-inferior, or potentially superior, to conventional room-temperature platelets for controlling active bleeding during complex cardiac surgery with cardiopulmonary bypass?
  • If non-inferiority was established, what was the longest acceptable cold-storage duration, up to 21 days?
• Why This Matters

  • Extending platelet storage from 5–7 days to 21 days could materially reduce wastage, improve resilience during shortages and permit hospitals without a conventional platelet inventory to stock platelets for major haemorrhage.
  • Clinical haemostatic efficacy cannot be inferred reliably from laboratory platelet aggregation or circulating platelet recovery, making a large clinical trial essential.
  • Any logistical advantage must be balanced against the possibility of reduced haemostatic durability, increased transfusion requirements or thrombotic complications from activated cold-stored platelets.

Design & Methods

• Research Question: Among actively bleeding paediatric and adult patients undergoing complex cardiac surgery with cardiopulmonary bypass, were cold-stored platelets stored for up to 21 days non-inferior to room-temperature platelets stored for up to 5–7 days for haemostatic efficacy?
• Study Type: Investigator-initiated, phase 3, multicentre, international, randomised, partially blinded, adaptive, Bayesian, non-inferiority, storage-duration-ranging trial conducted at 25 US and 2 Australian hospitals between December 2021 and March 2025.
• Population:
  • Patients undergoing planned complex cardiac surgery with cardiopulmonary bypass who were expected to bleed sufficiently to require platelet transfusion.
  • Initially included patients aged more than 28 days to less than 85 years; viable neonates weighing at least 3 kg were added after the 600-patient interim review.
  • The analysed clinical population comprised patients who actually received at least one study platelet transfusion during surgery or within 24 hours after ICU admission.
  • Important exclusions were platelet transfusion during the preceding 24 hours, platelet count less than 75 × 10⁹/L, congenital bleeding or platelet disorders, antiplatelet or anti-HLA antibodies causing refractoriness, planned postoperative ECMO, ventricular assist device or kidney replacement therapy, and intended whole-blood transfusion for bleeding.
• Intervention:
  • Leucocyte-reduced apheresis platelets stored without agitation at 1–6°C.
  • The maximum permitted storage duration began at 7 days and was adapted after successive cohorts of 200 transfused patients, up to a maximum of 21 days.
  • After approximately 280 patients had been enrolled, new US regulatory guidance permitting storage for up to 14 days led the data and safety monitoring board to approve an immediate increase to 14 days before resuming the adaptive process.
  • Up to eight platelet units, or eight paediatric doses of 10–15 mL/kg, could be administered during the 24 hours beginning with the first study transfusion. The original six-unit limit was increased after the 600-patient safety review.
• Comparison:
  • Leucocyte-reduced apheresis platelets stored with agitation at 20–24°C for up to 5–7 days.
  • The same 24-hour intervention period and maximum of eight units or eight paediatric doses applied.
  • The indication, timing and dose of platelet transfusion were determined by the treating clinicians; transfusion thresholds, other blood products and haemostatic adjuncts were not protocolised.
• Blinding: Patients, most treating clinicians, research staff, data collectors and outcome assessors were blinded. Blood-bank personnel and the clinician physically administering the platelets could not always be blinded because unit temperature and condensation could reveal allocation.
• Statistics:
  • The primary outcome was an adapted five-level perioperative haemostatic efficacy score, ranging from 1 to 5, with higher scores indicating more severe bleeding. The non-inferiority margin was 1 full point.
  • A monotonic Bayesian piecewise-linear model related cold-storage duration to the haemostatic score. A duration was non-inferior if the posterior probability that its mean score was less than the room-temperature mean plus 1 point exceeded 97.5%.
  • A conventional fixed-effect power calculation was not used. The planned sample of 1,000 transfused patients was selected to estimate the maximum non-inferior storage duration; simulations provided 99.1–100% power across the prespecified alternative scenarios, corresponding to beta of 0–0.9%, while controlling the one-sided Type I error at 2.5%.²
  • The principal analysis was a modified intention-to-treat analysis restricted to randomised, treatment-eligible patients who received study platelets. The primary duration-response model excluded patients for whom cold-storage duration could not be defined; per-intervention and other sensitivity analyses were also performed.
  • The single secondary outcome, 24-hour chest-tube output, was tested using a site-stratified van Elteren test with a two-sided alpha of 0.05. Safety, blood-product and laboratory outcomes were exploratory.
• Follow-Up Period: Outcomes were collected for up to 28 days after the first platelet transfusion, with the primary outcome assessed over 24 hours and most safety events assessed over 7 or 28 days.

Key Results

This trial was not stopped early. It continued to the planned total of 1,000 transfused participants. The adaptive interim analyses altered the maximum permitted cold-storage duration, but the protocol contained no provision for early stopping for success.

• CHIPS met its formal primary criterion: every evaluated cold-storage duration through 21 days had a greater than 99.9% posterior probability of remaining within the 1-point non-inferiority margin.
• Non-inferiority should not be interpreted as equivalence: the model estimated progressively higher, and therefore worse, haemostatic scores with longer storage, while cold-stored recipients received more blood products and underwent more re-exploration.
• No excess in thrombosis, organ failure or mortality was demonstrated, but these were exploratory outcomes with insufficient precision to exclude clinically important safety differences.

Internal Validity

• Randomisation and Allocation: Central web-based randomisation assigned participants 2:1 using variable block sizes. Allocation was concealed from most clinical and research staff, although blood-bank personnel necessarily knew the assignment.
• Post-randomisation Exclusions: Of 1,811 randomised patients, participation was cancelled before surgery for 178, no platelets were ordered for 596, platelets were ordered for 1,037 and 1,000 received a study platelet transfusion. The primary analysis included 989: 650 assigned to cold-stored platelets and 339 assigned to room-temperature platelets.
• Impact of the Modified Intention-to-Treat Population: Conditioning analysis on post-randomisation receipt of platelets can compromise randomisation. However, the proportions ultimately transfused were similar between assignments, and treatment allocation was generally concealed when the decision to transfuse was made.
• Missing Primary Outcomes: The haemostatic efficacy score could not be calculated for 3 cold-stored and 1 room-temperature recipient. A further 8 patients assigned to cold-stored platelets received only room-temperature platelets and were excluded from the duration-response efficacy model.
• Performance and Detection Bias: The administering clinician could detect the cold unit by touch or condensation. Because platelet dosing, additional transfusion and re-exploration were clinician-dependent, incomplete blinding could have influenced important components of the primary and exploratory outcomes. Outcome derivation by blinded personnel reduced detection bias.
• Protocol Adherence: In the cold-stored group, 563/660 (85.3%) received fully adherent platelet transfusions; 89 received both products and 8 received room-temperature platelets only. In the comparison group, 334/340 (98.2%) were adherent; 3 received both products and 3 received cold-stored platelets only.
• Effect of Crossover: The markedly greater crossover in the cold-stored group would tend to dilute a true difference and favour a finding of non-inferiority. The per-intervention analysis was nevertheless consistent with the primary result.
• Baseline Characteristics: Groups were broadly comparable. Median pretransfusion platelet counts were 104 vs 101 × 10⁹/L, ASA class IV accounted for 74.7% vs 75.9%, and median cardiopulmonary bypass durations were 168 vs 157 minutes in the cold-stored and room-temperature groups, respectively.
• Population Heterogeneity: The trial deliberately included neonates, children and adults, different operations and multiple platelet-manufacturing methods. Prespecified analyses found no convincing treatment-effect heterogeneity, but many subgroups were small and unable to exclude important differences.
• Timing: The first transfusion occurred in the operating theatre in 632/660 (95.8%) cold-stored recipients and 329/340 (96.8%) room-temperature recipients, providing appropriate early exposure during active surgical bleeding.
• Dose: Platelet dose was clinician-selected and the maximum permitted dose increased from six to eight units or paediatric doses during the trial. This preserved pragmatism but introduced treatment variability and allowed additional room-temperature platelets when cold inventory was exhausted.
• Separation of the Variable of Interest: Storage conditions were clearly separated at 1–6°C without agitation versus 20–24°C with agitation. The adaptive design generated exposure across 1–21 cold-storage days, but only 38 patients received platelets at the 21-day duration and the estimate at that duration depended on the model borrowing information across storage ages.
• Adjunctive Therapy and Additional Transfusion: Transfusion thresholds and haemostatic adjuncts were not standardised. Total blood-product exposure within 24 hours was 26.0 ± 37.4 vs 19.4 ± 20.1 mL/kg, so additional transfusion may have compensated for a difference in platelet effect while simultaneously contributing to the primary composite score.
• Outcome Assessment: The primary score incorporated bleeding, blood-product use, haemostatic agents, delayed sternal closure and re-exploration. It captured clinically relevant haemostasis but was relatively coarse, partly dependent on clinician behaviour, validated originally in adults and modified with unvalidated weight-adjusted criteria for smaller patients.
• Statistical Rigour: The adaptive Bayesian model, decision thresholds and sensitivity analyses were prespecified. Important qualifications are the wide 1-point non-inferiority margin, the imposed monotonic duration-response assumption, the modified intention-to-treat population and the absence of multiplicity adjustment for numerous exploratory outcomes.

Conclusion on Internal Validity: Overall internal validity is moderate. The large randomised design, concealed allocation, blinded outcome derivation and consistent sensitivity analyses support the formal conclusion of pragmatic non-inferiority, but crossover, post-randomisation population selection, a generous margin, incomplete clinician blinding and non-standardised transfusion practice limit claims of equivalence or comparable safety.

External Validity

• Age Range: CHIPS included a broad population from neonates to adults aged almost 85 years; 29.8% of cold-stored recipients and 33.5% of room-temperature recipients were younger than 12 years, while approximately one-third were aged 65 years or older.
• Clinical Setting: The findings apply most directly to active bleeding during complex cardiac surgery with cardiopulmonary bypass in centres with established transfusion services.
• Excluded High-Risk Groups: Patients with platelet counts below 75 × 10⁹/L, congenital bleeding disorders, platelet refractoriness, planned postoperative ECMO, ventricular assist devices or kidney replacement therapy were excluded. Efficacy in these particularly high-risk groups remains uncertain.
• Platelet Products: Multiple apheresis collection platforms, plasma and additive solutions, pathogen-reduced and non-pathogen-reduced products were included, improving applicability across apheresis systems.
• ABO Compatibility: At least one ABO-incompatible platelet product was given to 50.5% of cold-stored recipients and 37.4% of controls. Subgroup findings were consistent, but this imbalance complicates interpretation of laboratory platelet recovery.
• Geographical Applicability: All sites were in the US or Australia. The findings do not directly establish non-inferiority for pooled whole-blood-derived platelet products commonly used in other healthcare systems.
• Other Indications: CHIPS does not establish efficacy for prophylaxis in hypoproliferative thrombocytopenia, where sustained circulating platelet survival is important, or for trauma, gastrointestinal or obstetric haemorrhage, liver transplantation, non-cardiac surgery or bleeding without cardiopulmonary bypass.
• Resource-Limited Settings: A 21-day shelf life could have major logistical value in smaller or remote hospitals, but implementation requires validated cold-storage, transport and manufacturing systems not evaluated as clinical interventions in this trial.

Conclusion on External Validity: Generalisability is strong for leucocyte-reduced apheresis platelets used therapeutically during complex cardiac surgery in comparable systems. It is substantially more limited for other platelet products, prophylactic transfusion, non-cardiac haemorrhage and patients with profound thrombocytopenia or planned mechanical organ support.

Strengths & Limitations

• Strengths:
  • Largest randomised clinical evaluation of cold-stored platelets to date.
  • Phase 3, multicentre and international recruitment with paediatric and adult representation.
  • Innovative adaptive design directly evaluated storage duration rather than treating all cold-stored platelets as a single product.
  • Prespecified Bayesian operating characteristics, independent safety oversight and no early stopping for success.
  • Broad inclusion of apheresis platforms, storage media and pathogen-reduction methods.
  • Excellent completeness of the primary outcome among transfused participants.
  • Clinically relevant bleeding, transfusion, laboratory and safety outcomes with consistent sensitivity analyses.
• Limitations:
  • The 1-point non-inferiority margin was large relative to a five-level bleeding score and could accept an entire category of worse bleeding.
  • The adapted primary score was coarse, partly influenced by clinician decisions and not validated in its modified paediatric form.
  • Platelet indications, transfusion thresholds, other blood products and haemostatic adjuncts were not standardised.
  • The principal analysis was not a complete intention-to-treat analysis of all 1,811 randomised participants.
  • Protocol adherence was lower in the cold-stored group, with 14.7% receiving at least some room-temperature platelets.
  • Only 38 patients received platelets at the exact 21-day duration; inference depended on the prespecified duration-response model.
  • The clinician administering platelets could be unblinded.
  • The greater re-exploration and blood-product exposure with cold-stored platelets were exploratory and therefore important but not definitive safety or efficacy findings.
  • The trial was not powered to establish equivalence for mortality, thrombosis or other uncommon adverse events.
  • Whole-blood-derived platelets and most non-cardiac indications were not studied.

Interpretation & Why It Matters

• Primary Interpretation

  Within contemporary pragmatic cardiac-surgical practice, apheresis platelets stored at 1–6°C for as long as 21 days remained within the prespecified 1-point haemostatic non-inferiority margin compared with platelets stored at room temperature for 5–7 days.
• Not Proof of Equivalence

  The findings do not show identical efficacy. At 21 days, the model-estimated haemostatic score was 0.25 points worse with cold-stored platelets, with a 95% credible interval of 0.06 to 0.46, although this remained well inside the 1-point margin.
• No Demonstrated Superiority

  Despite laboratory hypotheses that refrigeration might enhance immediate haemostasis, no cold-storage duration was superior for the primary outcome or 24-hour chest-tube output.
• Operational Importance

  A 21-day platelet product could reduce expiry, improve supply-chain resilience and make therapeutic platelet transfusion available in hospitals unable to maintain conventional room-temperature stock.
• Clinical Caution

  Greater re-exploration, greater blood-product exposure and smaller platelet-count increments mean that adoption should include haemovigilance, protocolised transfusion practice and monitoring of bleeding, product use and thrombotic events.

Controversies & Other Evidence

• Non-inferiority Margin: A 1-point margin on a five-category scale is clinically generous. With a control mean near 3, the formal criterion could accept a cold-stored mean approaching 4, corresponding conceptually to movement from moderate towards severe bleeding.
• Evidence of Duration-Related Deterioration: The primary model estimated progressively worse haemostatic scores as storage approached 21 days. The 21-day difference of +0.25 points had a 95% credible interval entirely above zero, indicating probable modest deterioration even though the 1-point non-inferiority criterion was comfortably met.
• Coarse and Behaviour-Dependent Outcome: Heterogeneous platelet indications, a relatively coarse modified bleeding scale and clinician-determined transfusion may obscure true differences between products. The accompanying editorial therefore interpreted the trial as establishing non-inferiority under current practice rather than proving that 21-day cold-stored platelets are universally equivalent.³
• Uncertain Platelet Indication: Contemporary international guidelines acknowledge limited high-quality evidence for platelet use in active cardiac-surgical bleeding and recommend against transfusion in non-thrombocytopenic cardiovascular surgery without major haemorrhage.⁴ CHIPS appropriately enrolled actively bleeding patients, but the absence of protocolised indications means that some transfusions may have been unnecessary.
• Re-exploration and Additional Blood Products: Re-exploration occurred in 4.2% vs 1.8%, and cold-stored recipients received more platelet, plasma and total blood-product volume. These internally consistent findings challenge an uncomplicated assertion of equivalent haemostasis, although they were exploratory and vulnerable to clinician behaviour and incomplete blinding.
• Possible Explanations: Cold-stored platelets may be cleared from circulation more rapidly, incorporated more rapidly into developing thrombi, produce laboratory results that prompt additional transfusion, or have genuinely lower sustained haemostatic efficacy. CHIPS could not distinguish these mechanisms.
• Platelet Recovery and Function: Room-temperature platelets produced much greater platelet-count increments and somewhat greater improvement in viscoelastic clot strength. A preceding mechanistic randomised study similarly found better reversal of platelet inhibition with room-temperature platelets, although substantial donor variability existed.⁵
• Infusion Temperature: Multiple platelet units at 1–6°C may contribute to perioperative cooling. The trial did not report warming the products or quantify any temperature effect, while even mild perioperative hypothermia can increase bleeding. This becomes operationally important if cold-stored platelets are adopted widely.³
• Trauma Evidence: In the phase 2 CRISP-HS trial, early administration of a single cold-stored platelet unit was feasible but did not significantly reduce 24-hour mortality: 5.9% vs 10.2%; absolute difference −4.3 percentage points; 95% CI −12.8 to 3.5; P=0.26.⁶ A later Bayesian reanalysis estimated an 89.1% posterior probability of reduced 24-hour mortality, which remains suggestive rather than definitive.⁷
• Traumatic Brain Injury Evidence: CRISP-TBI found no improvement in six-month Glasgow Outcome Scale-Extended distribution with cold-stored platelets: OR 1.58; 95% CI 0.71 to 3.54; P=0.27. A lower rate of craniotomy or craniectomy was observed, but this was not accompanied by improved functional outcome.⁸
• Limits of Extrapolation: Short circulating survival may be acceptable when treating acute surgical bleeding in patients with intact thrombopoiesis, but it may be disadvantageous for prophylaxis in chemotherapy-associated or marrow-failure thrombocytopenia.
• Subsequent Evidence: CHIPS was published on August 17, 2026; no post-CHIPS confirmatory randomised trial, meta-analysis or guideline incorporating its results is yet available.

Summary

• CHIPS randomised 1,811 patients before surgery and analysed 1,000 who received study platelets during active bleeding after complex cardiac surgery with cardiopulmonary bypass.
• Cold-stored apheresis platelets stored at 1–6°C for up to 21 days met the prespecified 1-point non-inferiority criterion for haemostatic efficacy, with a greater than 99.9% posterior probability of non-inferiority at every storage duration.
• No cold-storage duration was superior, and 24-hour chest-tube output was similar: 8.9 vs 8.4 mL/kg; difference in medians 0.4 mL/kg; 95% CI −1.0 to 1.5; P=0.83.
• Cold-stored recipients had smaller platelet-count increments, received more blood products and underwent more re-exploration for bleeding: 4.2% vs 1.8%; absolute difference 2.5 percentage points; 95% CI 0.4 to 4.6.
• The trial supports 21-day cold storage as a potentially transformative inventory strategy for therapeutic apheresis platelets in cardiac surgery, but does not prove equivalence, superiority or safety in other indications.

Overall Takeaway

CHIPS provides the strongest clinical evidence to date that cold-stored apheresis platelets can remain acceptably haemostatic for up to 21 days when used for active bleeding during complex cardiac surgery. Its major potential is logistical and supply-related rather than superior haemostasis; the signals of greater transfusion exposure, re-exploration and modestly worse modelled efficacy at 21 days justify cautious implementation and continued clinical evaluation.

Bibliography

• 1.Strandenes G, Sivertsen J, Bjerkvig CK, et al. A pilot trial of platelets stored cold versus at room temperature for complex cardiothoracic surgery. Anesthesiology. 2020;133(6):1173-1183.
• 2.Zantek ND, Steiner ME, VanBuren JM, et al. Design and logistical considerations for the randomized adaptive non-inferiority storage-duration-ranging Chilled Platelet Study. Clin Trials. 2023;20(1):36-46.
• 3.Hess AS. A world with platelets stored cold for 21 days: chill and tender. JAMA. Published online August 17, 2026.
• 4.Metcalf RA, Nahirniak S, Guyatt G, et al. Platelet transfusion: 2025 AABB and ICTMG international clinical practice guidelines. JAMA. 2025;334(7):606-617.
• 5.Kogler VJ, Miles JA, Özpolat T, et al. Platelet dysfunction reversal with cold-stored vs room temperature-stored platelet transfusions. Blood. 2024;143(20):2073-2088.
• 6.Sperry JL, Guyette FX, Rosario-Rivera BL, et al; CRISP-HS Study Group. Early cold stored platelet transfusion following severe injury: a randomized clinical trial. Ann Surg. 2024;280(2):212-221.
• 7.Gerard J, Sperry JL, Guyette FX, et al. Bayesian analysis of the early cold stored platelet transfusion following severe injury randomized clinical trial. Trauma Surg Acute Care Open. 2026;11(3):e002145.
• 8.Neal MD, Okonkwo DO, Guyette FX, et al; CRISP-TBI Study Group. Early cold-stored platelet transfusion following traumatic brain injury: a randomized clinical trial. Ann Surg. 2025;281(5):796-805.