Hypothermia during microsurgical head and neck reconstruction and incidence of venous thromboembolism

Title: Hypothermia during microsurgical head and neck reconstruction and incidence of venous thromboembolism

Authors: Saadoun R, Guerrero DT, Bengur FB, Moroni EA, Surucu Y, Smith RE, Esper SA, Whitehurst SL, Artman J, Veit JA, Kubik M, Sridharan S, Solari MG

Journal: JAMA Otolaryngology–Head & Neck Surgery. February 2025

PMID: 39636654

Key takeaways

• Venous thromboembolism (VTE) occurred in 3.2% (35/1078); free-flap pedicle thrombosis occurred in 2.2% (24/1078).

• Intraoperative hypothermia (core temperature <36°C) was common (76.5%) and associated with higher 30-day VTE odds when sustained beyond 30 minutes.

• Adjusted VTE odds were higher with sustained hypothermia versus no hypothermia: 30 to <120 minutes aOR 3.82 (95% CI 0.99–14.07); ≥120 minutes aOR 3.55 (1.05–11.95) (*not a strict dose-response; estimates are similar with wide CIs.)

• Hypothermia was not associated with free-flap pedicle thrombosis (overall OR 0.61, 95% CI 0.26–1.43).

Background

Head and neck free tissue transfer involves long operations and substantial thrombotic risk. Intraoperative temperature is modifiable, yet hypothermia remains common.

Objective

Assess whether intraoperative hypothermia (core temperature <36°C) is associated with 30-day VTE and/or free-flap pedicle thrombosis requiring revision after head and neck free tissue transfer.

Methods

• Design and setting: Retrospective cohort study; tertiary academic center.

• Study period: January 1, 2012 to August 31, 2023.

• Sample size: 1078 analyzed.

• Demographics (overall): Mean age 61.3 years; 67.2% male; mean Caprini score approximately 6.4 to 6.6.

• Inclusion criteria: Head and neck free tissue transfer cases within the study period.

• Key exclusions: Cases with postoperative hematoma requiring revision (to reduce confounding of venous compromise).

• Exposure definition (temperature):

  • Core temperature recorded every minute.

  • Temperatures harmonized to a “bladder-equivalent” core temperature using published offsets.

  • Data binned into consecutive 30-minute blocks using the median of minute-level values.

  • Hypothermia episode defined as any 30-minute block with median <36°C.

  • Duration groups based on continuous hypothermia:

    • No Hypothermia (<30 min)

    • 30 to 120 minutes of hypothermia

    • ≥120 minutes

• All patients warmed under institutional warming protocol during surgery: Systemic warming with warmed IV fluids and convective warming from pre-induction through PACU handoff.

• Outcomes (30 days):

  • VTE: deep vein thrombosis and/or pulmonary embolism.

  • Pedicle thrombosis: clot in pedicle artery or vein confirmed at revision surgery.

• Statistics:

  • Univariable and multivariable logistic regression.

  • Covariates highlighted in primary VTE model included Caprini score, chemoprophylaxis regimen, and surgery duration.

  • Alpha 0.05; no multiplicity adjustment.

Results

• Hypothermia: 76.5% (825/1078) had at least one 30 minute period with median core temperature below 36°C.

• VTE

  • Incidence: 3.2% (35/1078).

  • Timing: median postoperative day 5 (IQR 4.0–10.5).

  • Univariable associations:

    • Any hypothermia episode: OR 3.36 (95% CI 1.02–11.1).

    • Heparin prophylaxis vs enoxaparin: OR 3.16 (1.33–7.55).

  • Multivariable model (adjusted for Caprini score, chemoprophylaxis, surgery duration):

    • Hypothermia >30 to <120 minutes: aOR 3.82 (0.99–14.07).

    • Hypothermia ≥120 minutes: aOR 3.55 (1.05–11.95).

    • Heparin 5000 IU three times daily vs enoxaparin 30 mg twice daily: aOR ~3.01 (1.24–7.30).

• Free-flap pedicle thrombosis

  • Incidence: 2.2% (24/1078).

  • Timing: median postoperative day 2 (IQR 1.0–3.0).

  • Arterial thrombosis: 0.5% (5/1078).

  • Association with hypothermia: none detected (overall OR 0.61, 95% CI 0.26–1.43).

Conclusion

Intraoperative hypothermia during head and neck free tissue transfer was associated with higher 30-day VTE risk, without an observed association with free-flap pedicle thrombosis.

Strengths

• Large, procedure-specific cohort with granular minute-level temperature data.

• Clinically important outcomes with pedicle thrombosis and VTE examined

• Predefined exposure processing that emphasized sustained, continuous hypothermia.

Limitations

• Retrospective design limits causal inference.

• Residual confounding likely (case complexity, transfusion, vasopressor exposure, fluid balance, and unmeasured illness severity may track with hypothermia).

• Exposure captured uninterrupted hypothermia duration; intermittent hypothermia burden (overall time spent under 36°C ) was not the primary metric.

• Chemoprophylaxis findings are vulnerable to confounding by indication (e.g., renal dysfunction, bleeding risk, or protocol drift).

Clinical relevance

Hypothermia is already associated with worse perioperative outcomes (bleeding/transfusion, surgical site infection, and cardiac complications in other surgical populations). This study is another reminder that in long head and neck free-flap cases, teams should actively avoid prolonged hypothermia to improve outcomes. Target core temperature ≥36°C, minimize donor-site and skin exposure, employ multi-team approach to reduce operative time, and consider sterile forced-air warming over exposed regions when feasible.

Editorial Notes

• Hypothermia may be a marker rather than a mediator: even with adjustment for case duration and Caprini score, unmeasured factors (transfusion, vasopressors, tumor factors, nuances in comorbidities, etc.) may explain part of the VTE signal.

• Does intermittent hypothermia (repeated dips) confer similar risk as continuous hypothermia? Does duration or degree of hypothermia matter more?

  • Would an area-under-the-curve metric below 36°C correlate more strongly with VTE?

• Chemoprophylaxis signal could use deeper context:

  • What were the institutional criteria for selecting heparin versus enoxaparin?

  • How often were doses delayed, held, or changed due to bleeding, reoperation, renal dysfunction, or epidural use?

• At first glance, the lack of association with pedicle thrombosis seems counterintuitive, but systemic VTE and pedicle thrombosis often arise from different mechanisms and time courses (early local/technical vs later systemic). Other studies on VTE have similarly found that systemic VTE and flap VTE do not always correlate.