Intended for healthcare professionals

Clinical Review

Managing frostbite

BMJ 2010; 341 doi: (Published 19 November 2010) Cite this as: BMJ 2010;341:c5864
  1. Marc-James Hallam, specialist registrar in plastic surgery1,
  2. Tania Cubison, consultant in burns and plastic surgery1,
  3. Baljit Dheansa, consultant in burns and plastic surgery1,
  4. Chris Imray, consultant vascular and endovascular surgeon and honorary professor (Warwick Medical School)234
  1. 1Queen Victoria Hospital, East Grinstead RH19 3DZ, UK
  2. 2Coventry and Warwickshire County Vascular Unit, University Hospitals Coventry and Warwickshire NHS Trust, Coventry, UK
  3. 3Warwick Medical School, Coventry, UK
  4. 4Centre for Altitude, Space and Extreme Environment Medicine, University College London, Institute of Child Health, London, UK
  1. Correspondence to: mj_hallam{at}

Summary points

  • Prevention, using a combination of appropriate behaviour and equipment, is key

  • Rewarm frostbite as soon as the risk of refreezing is minimal

  • Seek advice from a specialist unit if the injury is potentially severe

  • Consider thrombolysis (with tissue plasminogen activator) in severe injuries presenting within 24 hours of exposure

  • Delay surgery unless there is evidence of compartment syndrome or overwhelming sepsis

  • People who have sustained a cold injury are more susceptible to a future cold injury

Frostbite is defined as the damage sustained by tissues while subject to temperatures below their freezing point (typically −0.55°C); in the broader sense it can include non-freezing cold injuries, where tissues do not freeze but are subject to a sustained and injurious cooling.1 Cold injuries were historically associated with military work in the field, but in the past 20 years the number of civilian cases has increased.2 3 The unusually cold winters in northern Europe in early 2010 caused many cases of frostbite in the United Kingdom, central Europe, and Scandinavia. The severity of frostbite injury is proportional to the temperature, duration of exposure, and amount and depth of frozen tissue.3 The term frostbite covers a wide range of injuries, from minimal tissue damage to substantial tissue necrosis that can result in amputation. Recent advances in treatment mean that frontline medical staff must be able to recognise and treat frostbite effectively.

Sources and selection criteria

We performed a Medline and Google Scholar search with no date limitations using the terms “frostbite”, “cold injury”, and “freezing cold injury” to obtain references that form the basis of this article. The content of this paper was derived from a variety of sources but we included only articles that were well written, had high clinical importance, and (where relevant) were referenced from soundly conducted clinical trials.

What factors predispose to frostbite?

Factors that predispose to freezing and non-freezing cold injury include military activities, winter sports, and homelessness (box 1).4 In a large retrospective epidemiological review of frostbite injuries sustained over 12 years in the Canadian prairies, individual behaviour was an important contributor to overall risk. Forty six per cent of patients were intoxicated with alcohol, 17% had a psychiatric disorder, and 3% had used illicit drugs.5 Fingers, toes, noses, cheeks, ears, and male genitalia were most commonly affected by frostbite in all groups (in that order).6 A large cross sectional study over 19 years by the US army found that men of Afro-Caribbean descent were four times more likely than white men to sustain cold weather injuries.7

Box 1 Predisposing factors

  • General: Unusually cold weather, prolonged exposure to cold, inadequate clothing, inadequate use of appropriate clothing, homelessness, smoking, dehydration, old age, ethnic origin, high altitude

  • Systemic disease: Peripheral vascular disease, diabetes, Raynaud’s disease, sepsis, previous cold injury

  • Psychiatric illness

  • Drugs: β blockers, sedatives, and neuroleptics

  • Trauma: Any immobilising injury, but especially head and spinal injuries and proximal limb trauma that compromises the distal circulation

  • Intoxication: Alcohol and illicit drug use

What pathological mechanisms are involved in the injury?

Although the precise mechanism of injury is not fully understood, the most recent laboratory research (principally using an animal model) describes a series of progressive, overlapping changes. These may be divided into “pre-freeze phase,” “freeze-thaw phase,” “vascular stasis phase,” and “late ischaemic phase.”2 Cellular injury and ischaemia result from direct and indirect effects of a freezing insult (see figure on

Cellular injury

Direct tissue injury occurs through the formation of intracellular and extracellular ice crystals. Except in cases of very rapid freezing, ice crystals form in the extracellular space, increasing extracellular oncotic pressure. Cells then dehydrate through outward diffusion of water, resulting in intracellular electrolyte disturbance.8 As tissues rewarm, extracellular ice crystals melt and increasing tissue oedema is seen. Freezing also leads indirectly to the release of proinflammatory cytokines.


Freezing causes localised anomalies in blood flow. After exposure to temperatures below freezing the affected extremity exhibits localised vasoconstriction, which reduces blood flow and exacerbates cooling to produce even greater vasoconstriction. Cooling of the vascular contents also causes an increase in viscosity and microvascular damage, which precipitates local transcapillary plasma loss and oedema formation. Endothelial damage encourages the formation of microthrombi, which occlude capillaries and lead to ischaemia.

Whether microvascular clots form mostly during freezing or thawing is unclear, but the rewarming of frozen tissues results in the lysis of ice containing cells and produces a prothrombotic microenvironment. Endothelial damage stimulates degranulation of mast cells and histamine release, which exacerbates oedema formation within tissues.2 9 10

Vasoconstriction may be followed by vasodilatation through the so called hunting response—a normal physiological response thought to protect the extremities from freezing; this occurs at the expense of a drop in core body temperature and may be largely ineffective in locally extreme cold stress.11 Tissue ischaemia may result in widespread and devastating tissue loss. The most severe injuries are often seen in tissues that freeze, thaw, and freeze again. These “freeze-thaw-refreeze” cycles produce progressively severe thrombosis and tissue ischaemia.4

How to evaluate a patient with an apparent cold injury


Seek a history of how and when the cold injury occurred. Pay particular attention to the likely temperature, wind chill, and duration of exposure—factors that will probably affect the severity of injury. Other necessary information (obtain from a collateral source if the patient cannot provide it) includes the patient’s premorbid state, particularly history of peripheral vascular disease and smoking status.9


Cold injuries can present as non-freezing or freezing cold injuries and also as mixed injuries. Mixed injuries will have a dominant component (either freezing or non-freezing cold injury), however, and there will be areas of overlap between deeper, frozen injuries and lesser affected, non-frozen regions.

Freezing cold injuries vary from extremely minor cases (sometimes referred to as frostnip) to grossly frozen limbs. Table 1 lists early and late signs and symptoms. The full extent of injury is often not apparent for some days after injury and close observation is needed during this period.

Table 1

 Early and late features of frostbite

View this table:

In cases of true freezing cold injury, the degree of injury may be assessed by the clinical appearance of the tissues (table 2; figs 1 and 2).

Table 2

 Assessment of degree of freezing cold injury2

View this table:

Fig 1 A typical frostbite affecting the hallux and third left toes showing the initial injury at presentation at base camp on Everest (A), at six weeks (B), and at 10 weeks (C). Note the delayed surgical amputation of the hallux after definitive demarcation and the recovery of the third digit after appropriate management


Fig 2 Typical frostbite injuries in the hands and feet of a climber with mildly haemorrhagic bullae presenting three days after exposure. The bullae were aseptically aspirated and a five day iloporost infusion resulted in a complete recovery

History and examination must ascertain if the injury occurred within the past 24 hours and whether it is mild or severe. Mild cases can be managed locally, but a potentially severe injury merits discussion with a specialist unit and possible transfer. Similarly, if a severe case presents less than 24 hours after injury then urgent transfer to a unit with thrombolysis capabilities must be considered, if it is safe to do so.

Non-freezing cold injury more commonly affects the feet than the hands.12 Patients usually present with prolonged local numbness, sometimes with pain or paraesthesia (or both). After rewarming, the foot rapidly becomes hyperaemic, with local oedema and persistent severe pain. Tissue loss is uncommon, and tissues do not become woody or frozen. Discoloration from vascular stasis or cyanosis is commonly confined to nail beds, and blisters are unusual; when they do occur, the exudate is usually clear and gelatinous and seldom haemorrhagic.1 Disabling pain and neuropathy can persist and may be very difficult to treat.

How should cold injuries be managed in the prehospital phase?

Patients presenting with frostbite may have potentially life threatening conditions, such as hypothermia, or have experienced severe trauma. Prioritise such injuries over the presence of regional cold injury, and start the rewarming of hypothermic patients and stabilisation of traumatic injuries before transport to a health facility (box 2).3 13 Isolated regional cold injuries must be rewarmed only if the risk of refreezing is small—the patient might even have to walk on a frozen limb to receive medical care in a warm environment rather than attempt rewarming in the field. Rubbing a frozen area may cause further damage and must be avoided.14

Box 2 Acute and emergent conditions often seen in patients with frostbite


This life threatening condition is common in people with frostbite, with a reported incidence of 12% in people with a serious cold injury.5 Hypothermia is defined as a drop in core body temperature to below 35°C. Its presentation depends on severity. Initially, the metabolic rate increases to cause tachycardia, tachypnoea, and shivering. In more advanced cases with a core temperature below 28°C, the metabolic rate drops to produce bradycardia, hypoventilation, and reduced consciousness leading to coma and death.13 15


Patients with frostbite commonly have traumatic injuries—in the Canadian study, 19% of patients with frostbite had been involved in a serious traumatic incident.5 Frostbite should therefore be suspected in trauma patients in cold environments, especially those who have been trapped in vehicles, have sustained limb injuries that may compromise distal blood supply, and those who have experienced hypovolaemic shock.

Internet and satellite phones

Patients and clinicians with limited experience of frostbite can now use the internet and satellite phones to access expert advice in remote or difficult situations. A virtual opinion or more specialised advice can be sought from almost anywhere in the world using a combination of digital images and telephone advice.2

How to rewarm an affected area

Once the patient is in a safe environment and freezing cold injury has been diagnosed, rapidly rewarm the affected area(s) (non-freezing cold injuries may be made more florid and have a worse outcome if they are rapidly rewarmed and should be rewarmed slowly in air at 22-27°C16). Warm frostbite injuries in a whirlpool bath containing a mild antiseptic such as chlorhexidine (without the affected region touching the device’s hard sides) at 40-41°C for at least 30 minutes, until all tissues are thoroughly rewarmed and pliable, with a red-purple colour.2 3 9 17 This treatment is thought to preserve the dermal circulation. Continue for 30 minutes twice daily until there is clear demarcation of necrotic tissues or evidence of tissue healing, such as epidermal regeneration and the emergence of normal skin colour. Keep the area warm and dry between treatments. This regimen has become generally accepted since a case series of 38 patients with frostbite treated this way recovered without appreciable tissue loss.17 Foot spas are cheap and readily available alternatives. Although they cannot keep water as hot as 40°C, careful addition of warmer water and monitoring of water temperature make them useful for rewarming frozen hands or feet.

Rewarming can be acutely painful, so give all patients adequate analgesia. Ibuprofen (400 mg orally, every 12 hours) is recommended for its selective antiprostaglandin activity. Aspirin may be less beneficial in frostbite because it blocks all prostaglandins, including some prostacyclins, which may expedite wound healing.2 Sympathectomy can be used to increase blood flow, but vasodilatory drugs such as iloprost (synthetic analogue of prostacyclin I2) have largely superseded the need for this. Recent clinical trials have shown that the phosphodiesterase inhibitor pentoxifylline enhances tissue viability by increasing blood flow and reducing platelet activity and should therefore be considered in severe frostbite.2

Necrotic or potentially non-viable tissues are at increased risk of infection. Give tetanus toxoid and broad spectrum prophylactic antibiotics to patients with any risk of infection of necrotic tissue.17 Leave simple non-tense areas of clear blistering intact because their rupture may increase the risk of infection. Tense or haemorrhagic blisters may be carefully aspirated (fig 2), but only in facilities that can maintain rigorous asepsis, because infection has been the cause of early radical amputations and death in the past.2 3 17 Complex cases, such as patients who may have been exposed to bacteriological contamination during transit or those at risk of hospital acquired infections such as meticillin resistant Staphylococcus aureus, are best left with as much intact skin as possible until after discussion with a specialist unit. Areas of blistering require light bandaging with loosely applied and non-airtight dressings. Affected limbs should be raised to prevent further tissue oedema and venous stasis. Prevent patients with affected lower limbs from bearing weight as much as possible to protect fragile tissues from mechanical stress and ischaemia.

Early surgical debridement is contraindicated in almost all patients because it can take weeks for definitive demarcation of non-viable tissues to occur, and tissues that initially looked unsalvageable may recover. Exceptions to this include freeze-thaw-refreeze injuries, complicating limb trauma, infection, and compartment syndrome.2 Even if radical debridement or amputation(s) will be necessary, it is normally best to delay surgery until at least six to eight weeks after injury because surgical trauma to proximal tissues can result in poor wound healing. Specialist advice is essential before any earlier surgery.2 18 19

Re-evaluation after rewarming and appropriate management often shows that surgery is not needed, despite an initial suspicion that tissues would be unsalvageable. When definitive demarcation of necrotic tissue does occur, the area may be left to auto-amputate, especially if the patient is frail or has serious comorbidities; definitive (delayed) surgical debridement should be planned in consultation with a specialist surgical unit.


The precise severity and depth of tissue injury are usually difficult to evaluate, so early surgical debridement is contraindicated in most cases of frostbite.3 Several investigations have been proposed for predicting the need for and extent of debridement—plain radiography, laser Doppler studies, digital plethysmography, and infrared thermography. None of these investigations can accurately predict outcome in isolation, but they may be useful adjuncts to clinical examination in difficult or severe cases.

Technetium 99 (Tc-99) triple phase scanning (fig 3 and magnetic resonance angiography (fig 4) seem to be the most useful imaging modalities. A large retrospective review of 92 patients with severe frostbite showed that Tc-99 scans obtained during the first few days of injury accurately indicated the level (toe, hand, or arm) of amputation in 84% of cases.20 Case reports have also shown that these scans can accurately assess injury and help determine the level of, or need for, surgical debridement.21 22 Case reports suggest that magnetic resonance angiography is superior to Tc-99 because it allows direct visualisation of occluded vessels, effectively images surrounding tissues, and may show a clearer demarcation of ischaemic tissues, although this has not been confirmed by large scale studies.23


Fig 3 (A) Technetium 99 scans of the hands of a patient with frostbite. The terminal digits have reduced signal (especially in the left hand), suggesting that substantial tissue necrosis has occurred. (B) Clinical picture after a five day iloprost infusion showing the close correlation between the initial technetium 99 scans and the subsequent clinical appearance


Fig 4 Digital subtraction angiography showing the thrombotic occlusion of the terminal digits in a major frostbite injury of the hand (A) and reperfusion after urgent thrombolysis (B). Reproduced, with permission, from Sheridan and colleagues14

The routine use of complex imaging techniques in minor frostbite is not justified. These investigations are best reserved for severe frostbite, the few patients who might need early surgery, or those where thrombolysis is being considered; in such cases discuss the proposed management with a specialist unit first (fig 5).


Fig 5 Algorithm for the management of acute frostbite. HDU=high dependency unit

Acute intravascular thrombolysis

Evidence is emerging that patients who present within 24 hours of an acute, severe frostbite injury, in whom major tissue loss is predicted should be considered for acute intra-arterial or intravenous thrombolysis.14 This treatment aims to clear the intravascular thromboses and restore arterial perfusion of the affected limb(s), thereby partially or completely salvaging the tissue. Risks are bleeding, infection, or failure. Contraindications to thrombolysis with tissue plasminogen activator will be familiar to most doctors, but in acute frostbite the possibility of coexisting trauma must always be considered.

Patients must be managed in a high dependency ward or an intensive care setting within a major vascular unit that regularly undertakes acute peripheral thrombolysis. A diagnostic angiogram must be performed and an intra-arterial catheter inserted close to the site of thrombosis. Giving tissue plasminogen activator within 24 hours of injury reduced digital amputation rates from 41% to 10% (P<0.05) in one retrospective study in a single institution.24 Reperfusion of ischaemic tissue can cause tissue oedema, and if this occurs within a confined space, interstitial pressures may rise and result in a compartment syndrome that requires urgent fasciotomy.

What other treatments might follow rewarming?

Several adjunctive treatments have been proposed but the evidence base to support their use is variable.25 26 Topical aloe vera is widely used in North America (for its antiprostaglandin effect) in superficial frostbite. Iloprost and buflomedil are powerful vasodilators, which are used in Europe for more severe cases.2 20 Anecdotal reports advocate hyperbaric oxygen, but no trials have been performed. Vacuum assisted closure of wounds may be useful after amputation and the technique merits study.

What are the long term sequelae of cold injury?

The most commonly encountered long term problem is that of chronic pain, probably as a result of vasomotor dysfunction. Patients have sometimes asked for otherwise healthy digits to be amputated to alleviate the pain,9 which is often unresponsive to conventional analgesia and may be lifelong. Some (mostly anecdotal) evidence suggests that amitriptyline (usually the treatment of choice for the pain after non-freezing cold injuries) may be of benefit. Patients with refractory pain may need referral to a specialised pain service.

Many reports detail the association between the development of complex regional pain syndromes and cold sensitivity in frostbitten tissues. A study that followed 30 patients with severe frostbite injuries for 11 years showed that 53% had cold hypersensitivity, 40% had numbness of the fingers, and 33% had reduced sensitivity of touch. The authors speculate that these sequelae are secondary to thermophysiological changes in people with an increased tendency to vasospasm.27 Thus, a single episode of frostbite can result in cold sensitisation (inability to tolerate cold temperatures in the affected area) and predispose to cold injury in the future. Cohort studies have shown that people who have had frostbite injuries have a significant lifelong risk of further injury.28 All patients who have had frostbite therefore need preventive advice (box 3).

Box 3 Prevention of cold injuries

Cold injuries can be prevented only by total avoidance of the cold but the following advice is useful if cold exposure is anticipated:

  • Wear protective clothing (multiple loose layers provide insulation)

  • Avoid tight clothing that constricts body parts, reducing blood flow

  • Stay dry and avoid prolonged cold exposure

  • Avoid the wind or protect against it

  • Wear an insulated hat, covering the ears

  • Wear gloves to protect the fingers, and in extreme cold, insulated mitts

  • Wear insulated boots

  • Maintain adequate nutrition and hydration

  • Avoid alcohol and smoking

  • Use supplemental oxygen above 7500 m

  • Use chemical or electrical hand and foot warmers

Other chronic morbidities

Freezing injuries may produce localised osteoporosis and in severe cases subchondral bone loss. These changes are thought to be secondary to vascular damage, so more profound bony changes may be an indicator of severity. Changes are sometimes evident a month after injury but often need longer to develop: by 16 months radiographs may show multiple lucencies in the subchondral bone.9 In children the damage may be more serious, with undergrowth of affected bone, irregularities in the articular surfaces, and a propensity to early arthritis.29 30

Areas that have been frostbitten are susceptible to chronic ulceration and are prone to malignant squamous transformation, similar to Marjolin’s ulcers in old burn scars.31

A patient’s perspective

In January 1999, I was involved in a mountaineering accident in the French Alps that tragically claimed the life of my climbing partner. After being trapped by severe weather for five days, I was rescued and taken to hospital in Chamonix. On arrival, my hands and feet were frozen, literally solid, and when thawed out it was obvious that there was much permanent damage. The doctors were hopeful that they could save parts of my hands and feet but I was warned that it would take a long time before the final outcome was known. For me this uncertainty was one of the hardest factors during what was a very troubled period. After only 10 days, though, I became increasingly unwell and developed septic shock. While I was unconscious the decision was taken to amputate.

Learning to live my life without hands or feet has been the most incredible, challenging, difficult, exciting, and rewarding experience. When I started out I was dependent on others for every daily task. Now I can walk, cook, drive, run, ski, snowboard, sail, climb mountains again, and I have a beautiful baby daughter. Life couldn’t be better.

Jamie Andrew, Edinburgh

Areas for future research

  • Formal evaluation of the use of intra-arterial tissue plasminogen activator (including the timing of treatment and the potential use of ultrasound accelerated thrombolysis)

  • Formal evaluation of hyperbaric oxygen and other adjunctive treatments

  • Larger long term follow-up studies to look at outcomes in patients with cold injury

  • Improved prevention strategies, including developments in protective equipment; for example, recent developments in extreme altitude climbing boots are now rated down to −60ºC and step-in crampons help reduce finger exposure to cold

Tips for non-specialists

  • Prevention is key

  • Treat any serious or life threatening conditions in the field, but do not rewarm frostbitten tissues until there is no possibility of refreezing

  • Establish the likely severity of the injury and seek advice from a specialist unit for any potentially severe injury

  • Rewarm mainly non-freezing cold injuries slowly, but rapidly rewarm mainly freezing cold injuries at 40-42°C for a minimum of 30 minutes in a whirlpool device or foot spa with chlorhexidine solution; continue treatment twice daily until improvement is seen

  • Areas of simple blistering should be left intact, lightly dressed, and elevated; the patient must not bear weight on frostbitten tissues

  • Give all patients with tissue loss broad spectrum antibiotics and a tetanus vaccination to prevent infection; consider treatment adjuncts such as thrombolysis

  • Appropriate clinical management often results in the recovery of tissues that initially looked unsalvageable, so early surgical debridement is rarely beneficial

  • Provide all patients with lifestyle advice, including immediate smoking cessation and the need to avoid refreezing the affected area

Additional educational resources

Resources for healthcare professionals
  • Johnson C, Oakley EHN. Cold climates. In: Johnson C, Anderson SR, Dallimore J, Winser S, Warrell DA. Oxford handbook of expedition and wilderness medicine. Oxford University Press, 2008:579-606

  • Imray CHE, Freer L. Frostbite. In: Auerbach PS. Wilderness medicine. 6th ed. Elsevier [forthcoming]

  • McIntosh SE, Hamonko M, Freer L, Grissom CK, Auerbach PS, Rodway GW, et al. Consensus guidelines for the prevention and treatment of frostbite. Wilderness Med J [forthcoming]

Resources for patients
  • British Mountaineering Council (—Internet frostbite service run by David Hillebrandt, Paul Richards, and Chris Imray

  • Andrew J. Life and limb. Portrait Publishers. 2005


Cite this as: BMJ 2010;341:c5864


  • Special thanks to EHN Oakley (head of survival and thermal medicine for the Institute of Naval Medicine) whose sound advice and collection of frostbite literature was invaluable in constructing this article.

  • Contributors: All authors were involved in all stages including conception and design of the paper, subsequent revisions, and final approval of the article. CI provided photographic and radiological images and is guarantor.

  • Competing interests: All authors have completed the Unified Competing Interest form at (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Commissioned; externally peer reviewed.

  • Patient consent obtained.


View Abstract