14. Fever & Antipyretic Therapy in the General Critically Ill Patient (HEAT Trial)

Quick Recap

Cross-cutting protocol — companion to Targeted Temperature Management Post-Cardiac Arrest protocol, which addresses fever specifically in the distinct post-arrest neuroprotection context. This protocol addresses the separate, more common everyday question of whether to treat fever in the general critically ill patient with suspected infection — where the evidence points toward a genuinely different conclusion than the post-arrest context, and where fever's biological role is more ambiguous (and potentially adaptive) than the reflexive clinical instinct to treat it might suggest.

1. Definition

Fever: an actively regulated elevation of core body temperature (typically defined as ≥38.3°C in this literature) mediated by hypothalamic set-point elevation in response to pyrogenic cytokines — a regulated physiological response, mechanistically and conceptually distinct from hyperthermia (passive, unregulated temperature elevation from failure of heat dissipation, e.g., heat stroke, malignant hyperthermia), which requires an entirely different management approach (cross-reference Environmental System protocols for heat-related illness).

The general-ICU fever question addressed by this protocol is distinct from the post-cardiac-arrest fever question addressed in the Targeted Temperature Management protocol: in post-arrest neuroprotection, fever avoidance is robustly, consistently supported (cross-reference that protocol's evidence base) because the injured brain's compromised autoregulation and heightened vulnerability to secondary hypoxic-ischemic injury changes the risk-benefit calculus specific to that context. This protocol addresses the separate, more common question of fever management in the general critically ill patient with suspected infection, without acute brain injury — a population where the evidence, discussed below, points toward a genuinely different, more equivocal conclusion.

2. Pathophysiology

Fever as a plausibly adaptive host response, not simply a harmful epiphenomenon: from an evolutionary and immunological perspective, fever is thought to have persisted as a conserved host response because it confers a net survival benefit — elevated temperature can enhance immune-cell function and directly inhibit growth and spread of certain pathogens, representing a genuine biological cost (increased metabolic demand, oxygen consumption) incurred specifically because the benefit has historically outweighed it. This framing — fever as a plausibly evolved, adaptive response rather than simply a maladaptive symptom to be reflexively suppressed — is the central conceptual lens for understanding why this evidence base has not straightforwardly confirmed a benefit from fever treatment, in contrast to the more intuitive clinical instinct that reducing fever should help.

The competing, also-plausible harm hypothesis: fever imposes genuine additional physiological/metabolic stress (increased oxygen consumption and cardiovascular demand) on an already critically ill patient, and this is the biological rationale underlying the alternative view that fever suppression should free up physiological "reserve" for fighting the underlying illness — both hypotheses are biologically coherent, and this genuine bidirectional plausibility (fever could help or could hurt) is itself the reason this question required rigorous RCT testing rather than being resolved by physiological reasoning alone.

A specific, genuinely counterintuitive observational finding worth holding in mind: a large observational study of approximately 600,000 ICU patients found fever was associated with decreased mortality in an unselected general ICU population — an association, not proof of a protective causal effect, and confounded by the many reasons sicker or different patients might or might not mount a febrile response, but a striking enough finding to further undercut the reflexive assumption that fever is simply harmful and should be treated as a default.

3. Immediate Stabilization (ABCDE) — Not Applicable in the Traditional Sense

This protocol addresses an everyday, non-acute treatment decision rather than acute stabilization; the relevant "checklist" is a decision framework for whether/how to treat fever in a stable, non-post-arrest critically ill patient:

Checklist:

4. Focused History

• Underlying suspected or confirmed source of infection (cross-reference relevant infectious disease protocols)
• Presence of acute brain injury, post-cardiac-arrest status, or other condition where fever avoidance carries separately established benefit (excludes this protocol's general population)
• Patient comfort/symptom burden from fever itself (myalgia, headache, general malaise) — a legitimate, separate indication for antipyretic use independent of any presumed mortality or organ-dysfunction benefit
• Hepatic function (relevant to acetaminophen dosing/safety, given its hepatic metabolism)

5. Comprehensive System-wise Examination

Not a primary examination-driven protocol; relevant findings are the underlying infectious source (addressed via the relevant disease-specific protocol) and the patient's subjective symptom burden from fever, which is a legitimate, independent treatment consideration separate from any presumed physiological benefit.

6. Syndrome Identification — Reframed as Population-Applicability Classification

• General critically ill patient with fever and suspected infection, no acute brain injury, no post-cardiac-arrest status: the population studied in the HEAT trial (Section 11) — current evidence does not support routine antipyretic administration as improving mortality or major clinical outcomes in this population
• Post-cardiac-arrest patient: cross-reference Targeted Temperature Management Post-Cardiac Arrest protocol — active fever prevention is robustly supported in this specific, distinct population
• Acute brain injury (traumatic brain injury, subarachnoid hemorrhage) more broadly: fever has been associated with worse outcomes in observational data in these specific populations (cross-reference relevant Neurology System protocols) — a population where more active fever avoidance may be reasonable, though this protocol's general-ICU evidence base does not directly address this separate population
• Hyperthermia (not fever): an entirely different pathophysiological process requiring active cooling and different management (cross-reference Environmental System protocols), not addressed by this protocol's antipyretic-focused evidence

7. Differential Diagnosis — Fever vs. Hyperthermia and Underlying Cause

Must-not-miss:

• Hyperthermia (heat stroke, malignant hyperthermia, serotonin syndrome, neuroleptic malignant syndrome) — a fundamentally different process requiring active cooling rather than antipyretic pharmacotherapy, since these conditions do not involve an elevated hypothalamic set-point that antipyretics act upon
• An unaddressed, undertreated infectious source — fever management is never a substitute for adequate source control and antimicrobial therapy (cross-reference Sepsis Bundles & Source Control protocol)

Common:

• Standard infectious causes of fever in the critically ill (cross-reference relevant infectious disease protocols)
• Non-infectious causes of fever (drug fever, post-transfusion reaction, thromboembolic disease) — relevant to the underlying diagnostic workup, distinct from the symptomatic fever-management question this protocol addresses

8. Severity/Risk Assessment

Not a severity-scoring protocol; the relevant risk consideration is population-appropriateness (Section 6) — whether the patient falls within the general ICU population where routine antipyretic treatment lacks demonstrated outcome benefit, versus the specific populations (post-arrest, acute brain injury) where a more active fever-avoidance approach carries separately established support.

9. Investigations — Not a Primary Component

Standard infectious workup per the underlying suspected source (cross-reference relevant infectious disease protocols); not a distinct investigational pathway for the fever-management decision itself.

10. Point-of-Care Ultrasound — Not Applicable

11. Evidence-Based Management

HEAT Trial — The Definitive Trial for This Specific Population

• HEAT trial (Young et al., NEJM 2015, n=700, 23 adult general ICUs in Australia and New Zealand): critically ill patients with fever and known or suspected infection randomized to regular intravenous paracetamol (acetaminophen) vs. placebo — specifically excluded acute brain disorders, significant liver dysfunction, post-cardiac-arrest patients requiring temperature control, rhabdomyolysis, and pregnancy, meaning this trial's population and conclusions are specifically about the general, non-brain-injured, non-post-arrest critically ill patient with suspected infection
• Primary outcome (ICU-free days at 28 days): no significant difference between paracetamol and placebo
• Secondary outcomes: no significant difference in mortality or other major clinical outcomes, despite paracetamol producing a measurable reduction in body temperature compared to placebo — this dissociation (temperature reduced, but no corresponding clinical benefit) is the central, clinically important finding
• Rigorous methodology: allocation concealment via indistinguishable packaging, and — a specifically noteworthy methodological strength — all analyses, including post-hoc analyses, were pre-specified and conducted before unmasking of study-group assignments, reducing the risk of the kind of post-hoc, multiplicity-driven false-positive findings discussed as a general concern elsewhere in this library (e.g., cross-reference the primary-vs-secondary-outcome caution in the HFNC vs. NIV protocol's FLORALI discussion)
• Authors' conclusion, stated plainly: early administration of paracetamol to treat fever due to probable infection does not affect the number of ICU-free days — a clean, well-powered, negative result answering a genuinely common, everyday clinical question with important outcome measures

Supporting and Contextualizing Evidence

• Paracetamol's antipyretic effect in critically ill patients is itself modest: observational and mechanistic data indicate only a limited temperature reduction (approximately 0.25°C) with paracetamol in this population — a smaller effect than commonly assumed, and relevant context for interpreting why a clinically meaningful outcome difference might not be expected even if fever reduction were beneficial in principle
• A separate, smaller RCT (febrile ICU patients with SIRS, APACHE II ≥10): compared scheduled IV paracetamol vs. treatment only if temperature reached 40°C — found no differences in inflammatory biomarkers (CRP, TNF-α, IL-1), ICU length of stay, mortality, or infectious complications between groups, concluding that "allowing fever to take its natural course does not appear to have detrimental effects" and that antipyretic therapy "may not be indicated in all ICU patients" — a smaller but directionally consistent finding reinforcing HEAT's conclusion
• A large observational study (~600,000 ICU patients): found fever was associated with decreased mortality in the general, unselected ICU population — an association (not a demonstrated causal protective effect, and subject to substantial confounding) but a genuinely striking finding that runs directly counter to the intuitive assumption that fever is simply a harmful sign requiring correction, and worth holding in mind as further context for why this evidence base has consistently failed to demonstrate benefit from routine antipyretic treatment
• A theoretical, biologically plausible concern regarding potential harm from fever suppression in specific infections: paracetamol has been shown (in non-ICU populations) to potentially prolong the course of certain infections (varicella, malaria, rhinovirus) — raising a theoretical concern that suppressing an adaptive host response could, in specific infectious contexts, be counterproductive, though this has not been specifically, definitively demonstrated as clinically significant in the critically ill ICU population itself

Practical Synthesis

In the general critically ill patient with fever and suspected infection (excluding acute brain injury and post-cardiac-arrest status), current evidence does not support routine antipyretic administration as improving mortality, ICU-free days, or other major clinical outcomes — despite paracetamol producing a measurable, if modest, temperature reduction. This stands in genuine, important contrast to the robust fever-avoidance recommendation in the post-cardiac-arrest population (cross-reference Targeted Temperature Management Post-Cardiac Arrest protocol), illustrating that fever management is not a one-size-fits-all clinical question but one where the specific patient population and underlying condition meaningfully changes the appropriate approach. The most defensible, evidence-supported indication for antipyretic use in the general ICU population is patient comfort/symptom relief, not a presumed mortality or organ-dysfunction benefit — a legitimate, separate rationale that does not require the outcome-based evidence this protocol addresses to justify its use. Clinicians should resist the reflexive instinct to treat fever simply because it is present, and should instead treat fever selectively, based on patient comfort or the presence of a population-specific indication (acute brain injury, post-arrest status) rather than as a default response to an elevated temperature reading.

12. Organ Support — Not Applicable

13. Disease-Specific Therapy

• Paracetamol (acetaminophen): reasonable for patient comfort/symptom relief from fever in the general critically ill population, without an expectation of mortality or major outcome benefit; standard dosing with attention to hepatic function
• NSAIDs and surface cooling: no clear evidence of differential effect compared to paracetamol in the HEAT trial's specific comparisons; not separately, more strongly supported as superior alternatives for the general fever-management question addressed here
• Active fever prevention/treatment remains specifically, robustly indicated in the post-cardiac-arrest population (cross-reference Targeted Temperature Management Post-Cardiac Arrest protocol) — a genuinely different clinical scenario from the general population this protocol addresses

14. Consultation Matrix

15. Monitoring Framework

• Clinical: temperature trend as one component of overall clinical trajectory assessment, not a standalone treatment target in the general ICU population
• Escalation triggers: not applicable in the traditional sense; persistent or worsening fever should prompt reassessment of the underlying infectious source and treatment adequacy, not simply more aggressive antipyretic dosing

16. ICU Bundle Checklist

17. Complications — Not Applicable in the Traditional Sense

Theoretical concern: possible prolongation of certain infections (varicella, malaria, rhinovirus) with paracetamol-mediated fever suppression, demonstrated in non-ICU populations — not definitively established as clinically significant in the critically ill ICU population specifically, but worth holding as a theoretical caution against reflexive fever suppression.

18. Escalation & De-escalation — Not Applicable

19. ICU Discharge Criteria — Not Applicable

20. Documentation & Medicolegal Checklist

• Rationale for antipyretic use (patient comfort vs. any other specific indication) documented
• Population-appropriateness (general ICU vs. post-arrest/acute brain injury) explicitly considered and documented where relevant
• Underlying infectious source investigation and treatment documented independently of the fever-management decision

21. Key Guidelines

• No major guideline recommends routine antipyretic treatment of fever in the general critically ill patient based on an expectation of mortality or outcome benefit, consistent with HEAT trial findings; guidance specific to post-cardiac-arrest fever prevention remains separately, robustly established (cross-reference Targeted Temperature Management Post-Cardiac Arrest protocol)

22. Landmark Trials

23. Controversies

• The evolutionary/adaptive framing of fever versus the intuitive clinical instinct to treat it: this protocol deliberately foregrounds the biological plausibility that fever may be an adaptive host response (enhanced immune function, pathogen growth inhibition) rather than simply presenting the HEAT trial's negative result as an isolated statistical finding — understanding why a seemingly obvious clinical intervention (reduce fever, presumably reduce physiological stress) failed to show benefit is more instructive than the negative result alone, and is directly analogous to several other topics in this library where a mechanistically plausible intervention did not translate into demonstrated clinical benefit.
• Whether fever avoidance genuinely differs by population, or whether the general-ICU null finding might eventually be shown to also apply less strongly to acute-brain-injury populations: current practice maintains a meaningfully different approach for post-cardiac-arrest and acute brain injury populations (active fever prevention) versus the general ICU population (no expectation of benefit from routine treatment) — this distinction rests on separate evidence bases specific to each population rather than a unified theory of fever's role across all critical illness, and the biological plausibility for why brain-injury populations might specifically benefit from fever avoidance (impaired autoregulation, heightened vulnerability to secondary injury) while the general population does not is reasonably well-articulated, but the populations have not been directly, head-to-head compared within a single trial framework.
• The observational fever-mortality association's causal interpretation remains genuinely unresolved: the ~600,000-patient finding of fever being associated with lower mortality is provocative but cannot establish that fever itself is protective, given the substantial potential for confounding (patients capable of mounting a robust febrile response may differ systematically, in ways relevant to prognosis, from those who cannot) — this protocol presents the finding as noteworthy context rather than as established causal evidence.

24. References

1. Young P, Saxena M, Bellomo R, et al; HEAT Investigators. Acetaminophen for fever in critically ill patients with suspected infection. N Engl J Med. 2015;373(23):2215-2224.
2. Young PJ, Saxena MK, Bellomo R, et al. The HEAT trial: a protocol for a multicentre randomised placebo-controlled trial of IV paracetamol in ICU patients with fever and infection. Crit Care Resusc. 2012;14(4):290-296.
3. Young P, Hodgson C, Dulhunty J, et al. End points for phase II trials in intensive care: recommendations from the Australian and New Zealand Clinical Trials Group consensus panel meeting. Crit Care Resusc. 2012;14(3):211-215.
4. Antipyretic therapy in critically ill patients with established sepsis: a trial sequential analysis. PLoS One. 2015;10:e0117279.
5. Suzuki S, Eastwood GM, Bailey M, et al. Paracetamol therapy and outcome of critically ill patients: a multicenter retrospective observational study. Crit Care. 2015;19:162.
6. Randomized trial of the effect of intravenous paracetamol on inflammatory biomarkers and outcome in febrile critically ill adults. [journal citation per source].
7. Young PJ, et al. Fever in critical illness: a large observational study. 2012.
8. Jeffries S, et al. Acetaminophen may prolong certain viral and parasitic infections [non-ICU population data]. 2015.
9. The Washington Manual of Critical Care, 4th ed. 2025 — fever and infection chapter.
10. ICU Protocols: A Step-wise Approach, 2nd ed. — relevant sepsis/fever content.