Earthquakes: injury patterns and triage

Twelve hours after an earthquake, a 24-year-old man arrives at the emergency department of a regional hospital. He spent the night trapped under a slab, with his right leg compressed, and was pulled out by the rescue team shortly before dawn. He is talking, his blood pressure is acceptable and pain seems to be the main complaint. The queue behind him has dozens of people with cuts, fractures and frightened children. The triage doctor has to decide in seconds whether that young man who talks and responds is a yellow case or someone about to stop his heart.

Earthquakes produce an injury profile different from other disasters. They mix mass mechanical trauma, prolonged crush, kidney failure and late infections, all concentrated in the first days. This text reviews what studies show about this profile, how triage tools perform in this scenario and which reasoning traps appear when demand exceeds resources.

Key points

• Between 1980 and 2009, earthquakes killed more than 370 thousand people and injured almost 1 million, mostly through the collapse of buildings.
• Soft tissue injuries and limb fractures predominate, but crush defines much of the severity and the consumption of resources.
• Almost half of the patients with crush syndrome need dialysis in the published series.
• Demand concentrates in the first days, especially in the first 24 hours.
• No triage tool is perfect, and in a large Chinese database REMS discriminated severity better than START.
• Infections by resistant germs appear weeks later and change wound care.

The size of the problem

A historical review by Doocy and colleagues gathered the earthquakes between 1980 and 2009 and counted 372634 deaths, 995219 injured and more than 61 million people affected (Doocy and colleagues, 2013). The main mechanism of death and injury was the collapse of buildings (Doocy and colleagues, 2013). The chance of surviving fell as rescue took longer, and in one of the included studies 84 out of every 100 survivors were rescued in the first hour (Doocy and colleagues, 2013). The review also highlights the role of the emergency response and of the health system in reducing deaths and injuries soon after the event (Doocy and colleagues, 2013).

The risk is not the same for everyone. In Doocy's analysis, the very young and the elderly had a higher risk of death and injury (Doocy and colleagues, 2013). This has a direct consequence for triage, because these are precisely the groups that communicate their symptoms worse and that compensate for shock differently from young adults, as already discussed in the text about vital signs that mislead in the elderly.

Knowledge about children is still fragile. A systematic review by Jacquet and colleagues found 2037 articles on earthquake injuries, but only 10 carried quantitative data on the pediatric population (Jacquet and colleagues, 2013). All were retrospective, used different age ranges and classified injuries heterogeneously, which prevents adding the results together safely (Jacquet and colleagues, 2013).

Some numbers from a single catastrophe help give scale. After the Kahramanmaraş earthquakes in Turkey in 2023, a hospital in Adana received 1246 patients under 18 in the first month (Uluöz and Gökmen, 2023). Of these, 560 were admitted to orthopedics (Uluöz and Gökmen, 2023). Another hospital in the region treated 662 victims in the first week, of whom 19.8 out of every 100 died (Sarı and colleagues, 2023).

When patients arrive

Demand after an earthquake has a typical shape: a very early peak and a tail of complications. Abu-Zidan and colleagues describe that the workload is concentrated in the first days, mainly in the first 24 hours (Abu-Zidan and colleagues, 2023). Patients admitted on that first day had more severe crush injuries and more complications (Abu-Zidan and colleagues, 2023).

The Turkish series confirm this pattern. In the study by Sarı and colleagues, 61.0 out of every 100 victims arrived in the first 3 days (Sarı and colleagues, 2023). Among the children of Adana, 42 out of every 100 orthopedic admissions happened in the first 24 hours and 58 out of every 100 in the first three days (Uluöz and Gökmen, 2023). Uluöz and Gökmen conclude that almost all injured children arrive at the hospital in the first days after the earthquake (Uluöz and Gökmen, 2023).

The same group of authors describes the medical response in four phases (Abu-Zidan and colleagues, 2023):

1. Search and rescue.
2. Triage and initial stabilization.
3. Definitive care.
4. Evacuation.

These phases do not happen in a clean sequence. While one neighborhood is still in search and rescue, the city hospital is already in definitive care of the first injured, and patients pulled from the rubble keep arriving for days. That is why the emergency team needs to reassess the situation several times a day, and not only when the plan is activated.

Time under the rubble also changes the profile. In a hospital in the earthquake zone, Köroğlu and colleagues assessed 372 of the 2176 patients treated and found 40.86 out of every 100 who had been trapped under rubble (Köroğlu and colleagues, 2023). In another series by Sarı, 36.8 out of every 100 victims had been rescued from rubble (Sarı and colleagues, 2023). Those who are trapped tend to arrive later and with a higher risk of crush injury, which reverses the intuition that the first to arrive are always the most severe.

The profile of injuries

In the review by Doocy, soft tissue injuries such as lacerations and contusions, and fractures were the most frequently reported types of injury, and the extremities were the most affected regions (Doocy and colleagues, 2013). In studies with admitted patients, however, crush injury appears several times as the most common injury (Doocy and colleagues, 2013). The difference is predictable: someone with only a cut is treated and discharged, while crush injury stays in the hospital.

The data from Köroğlu and colleagues show the same pattern. The lower limb was the most affected region, with 29.84 out of every 100 injuries, and 31.45 out of every 100 patients needed admission (Köroğlu and colleagues, 2023). In the study by Sarı, 38.3 out of every 100 were admitted and 17.4 out of every 100 needed the ICU (Sarı and colleagues, 2023).

In children, the distribution changes. Cakir and colleagues reviewed the CT scans of 257 children treated after the earthquake and found event-related injuries in 102 of them (Cakir and colleagues, 2023). The most common injury was head trauma, in 18.7 out of every 100, followed by chest in 15.6, spine in 8.5 and pelvic fracture in 7.4 (Cakir and colleagues, 2023). These numbers are a reminder that a buried child is not a small adult, and the decision about head imaging follows its own criteria, such as those discussed in the text about the PECARN rule in children.

Pediatric orthopedics in Adana gives an idea of the surgical volume. There were 77 children with long bone fractures, 34 of them open (Uluöz and Gökmen, 2023). Fasciotomy was performed in 131 children, 78 of them on both sides (Uluöz and Gökmen, 2023). Of the 31 amputations, 17 were performed in the first 24 hours and 28 in the first week (Uluöz and Gökmen, 2023).

To organize the reasoning, it helps to think of injuries in layers. The first is the immediate injuries from falling structures, which affect the skull, chest, spine, pelvis and extremities. The second is crush injury with compartment syndrome, which depends on the compression time. The third is the systemic complications after release, such as high potassium, shock and acute kidney injury. The fourth is wound infections and hospital-acquired infections in the following weeks. The fifth is the decompensation of chronic diseases due to lack of medication, power or dialysis. Each layer has its own clock, and the team that thinks only about the first tends to be surprised by the others.

Crush injury and kidney

Crush syndrome is the complication that most sets earthquakes apart from other events with many victims. Prolonged muscle compression causes crush injury, and the reperfusion that occurs when the compression is relieved can lead to metabolic changes and rhabdomyolysis, depending on the time until extrication (Abu-Zidan et al., 2023). The authors describe that reperfusion can cause hyperkalemia and cardiac arrest at the moment of release (Abu-Zidan et al., 2023). The topic is covered in depth in the text about potassium at the moment of release.

The scale is large. Abu-Zidan et al. cite a series of 957 victims in Kahramanmaraş in which 7.7 out of every 100 had crush injury and 2.1 out of every 100 had compartment syndrome (Abu-Zidan et al., 2023). The authors point to crush syndrome as the second leading cause of death after direct trauma in the earthquake (Abu-Zidan et al., 2023).

A meta-analysis by Rostami et al. gathered 50 studies published between 1988 and 2023, most from Turkey, Iran and China (Rostami et al., 2026). The pooled proportion of patients with crush syndrome who needed dialysis was 49 out of every 100 (Rostami et al., 2026). Acute kidney injury was the most frequent complication, also at 49 out of every 100 (Rostami et al., 2026). Pooled mortality was 8 out of every 100, with wide variation between regions, from 1 to 26 out of every 100 (Rostami et al., 2026). The authors attribute part of the heterogeneity to inconsistent definitions of the syndrome (Rostami et al., 2026).

What raises the risk to the kidney? In the study by Köroğlu et al. with 33 patients with crush injury, the frequency of acute kidney injury rose with the number of compressed limbs: 35.7 out of every 100 with one limb, 66.7 with two and 100 with three (Köroğlu et al., 2024). Being trapped for more than 6 hours was also associated with higher risk (Köroğlu et al., 2024). These two data points fit into simple triage questions: how many limbs were trapped and for how long.

Dialysis is a scarce resource in a disaster. When almost half of patients with crush injury need it (Rostami et al., 2026), the region's dialysis capacity becomes a bottleneck, added to the patients who were already on dialysis before the earthquake. The text about dialysis when power and water run short deals with this second group, and the text about rhabdomyolysis and acute kidney injury details the pathophysiology.

Some signs suggest crush injury with systemic risk and deserve to be actively sought: a limb compressed for hours, especially for more than 6 hours, more than one trapped limb, tense edema, pain out of proportion or loss of sensation in the limb, dark or scanty urine, and rhythm changes on the monitor right after release. None of them alone defines the diagnosis, but two or three together already change the patient's priority and the vigilance over them.

How triage works in an earthquake

Mass triage starts from the principle of doing the greatest good for the greatest number, and the most used algorithms, such as START and SALT, are described in the text about START, a simple and fast algorithm. The earthquake tests these algorithms in a particular way, because the patient with crush injury may walk little, breathe well and have acceptable perfusion on the first assessment, and still be at high risk.

Peng and colleagues compared five triage methods in a database of 29523 earthquake victims in western China between 2008 and 2018 (Peng and colleagues, 2021). The curves were calculated based on in-hospital death, injury severity score above 15 and need for ICU, and the area under the curve was 0.750 for START, 0.737 for CareFlight, 0.835 for REMS, 0.736 for T-RTS and 0.797 for TEWS (Peng and colleagues, 2021). REMS had the largest area under the curve, which the authors attributed to the larger number of indicators assessed, and there was no clear difference among the other four methods (Peng and colleagues, 2021). The authors note that prospective study of triage in earthquakes is very difficult, given the sudden nature of the event and ethical and logistical issues, and that most of the evidence comes from simulations and later analyses (Peng and colleagues, 2021).

Method	What it assesses, in general terms	Area under the curve in the Chinese database
START	Walking, breathing, perfusion and response to commands	0.750 (Peng and colleagues, 2021)
CareFlight	Following commands, walking, pulse and breathing	0.737 (Peng and colleagues, 2021)
REMS	Score with vital signs, consciousness and age	0.835 (Peng and colleagues, 2021)
T-RTS	Score with breathing, pressure and consciousness	0.736 (Peng and colleagues, 2021)
TEWS	Alert score with vital signs and mobility	0.797 (Peng and colleagues, 2021)

The table does not justify swapping one algorithm for another without training, but it shows that no method captures the severity of the earthquake on its own. Under-triage and over-triage have different consequences, and the text on over-triage and under-triage discusses this balance.

The Turkish experience proposes additions. Yilmaz and colleagues revised the management algorithm used in the adult emergency department after Kahramanmaraş based on START and SAVE, a secondary assessment system (Yilmaz and colleagues, 2024). They describe START-A as a way to speed up triage and pain control, and SAVE as a reassessment step that uses consciousness and limb severity scales to decide on admission, surgery, transfer and discharge (Yilmaz and colleagues, 2024). The authors highlight the use of blood gas analysis and bedside ultrasound during this assessment (Yilmaz and colleagues, 2024).

In practice, triage in an earthquake calls for a few extra questions beyond the algorithm:

1. Were you trapped? For how long?
2. How many limbs were crushed?
3. How is your urine?
4. Do you have kidney disease, diabetes or another condition that depends on medication or a device?
5. Who will reassess this patient, and when?

The last question is the most often forgotten. Disaster triage is dynamic, and a patient classified as yellow on arrival may become red after release or rehydration. The text on triage tools and what each one sees shows the same principle in another context.

Infections and wounds in the following weeks

The second wave of an earthquake is infectious. Hasanoglu and colleagues followed 225 survivors with musculoskeletal injuries admitted between February and April 2023 (Hasanoglu and colleagues, 2024). The most frequently isolated germ was Acinetobacter baumannii, in 49.4 per 100 cultures, followed by enterococci, in 28.6 (Hasanoglu and colleagues, 2024). Among the germs isolated in wound cultures, 76 per 100 were multidrug-resistant, and in deep tissue cultures, 58 per 100 (Hasanoglu and colleagues, 2024).

In the same study, 74 patients, or 32.8 per 100, had surgical site infection, and 80, or 35.6 per 100, went through the ICU (Hasanoglu and colleagues, 2024). Overall mortality was 2.7 per 100 (Hasanoglu and colleagues, 2024). The authors call attention to the fact that, contrary to what is expected in post-traumatic infections, gram-negative germs predominated (Hasanoglu and colleagues, 2024).

A meta-analysis by Xue and colleagues gathered 16 studies with 2974 strains isolated from earthquake victims (Xue and colleagues, 2025). The five most frequent pathogens were Escherichia coli, Acinetobacter baumannii, Pseudomonas aeruginosa, Enterobacter cloacae and Staphylococcus aureus (Xue and colleagues, 2025). Resistance rates to several commonly used antibiotics were high (Xue and colleagues, 2025).

In practice, this changes the reasoning at a few points. A rubble wound should be treated as contaminated, with careful cleaning and debridement. Fever in the second or third week after the earthquake calls for culture before presuming which germ it is. The local resistance profile should guide the empirical choice, not the habit inherited from other traumas. And probes, catheters and long hospital stays add the risk of hospital-acquired infection to a patient who already arrived with contaminated wounds. The series by Hasanoglu shows how these layers add up: besides surgical site infections, there were hospital-acquired urinary infections in 12 patients and bloodstream infections in 13 (Hasanoglu and colleagues, 2024).

The parallel with floods is useful: contaminated water and soil also change the profile of infections, as discussed in wounds and floodwater.

Reasoning traps in the earthquake setting

The disaster environment favors mental shortcuts. Some appear often:

• The patient who talks seems stable. In crush injury, release can be the moment of greatest risk (Abu-Zidan and colleagues, 2023).
• The dramatic wound steals attention. Open fractures draw the eye, while the closed, tense limb progresses to compartment syndrome.
• The quiet child is read as calm. Head trauma was the most common injury on pediatric CT scans (Cakir and colleagues, 2023).
• Late arrival is read as lower severity. Those trapped longer have a higher risk of kidney injury (Köroğlu and colleagues, 2024).
• The first classification becomes final. Without reassessment, the yellow patient who worsens disappears in the queue.

These traps are variations of the same error: trusting the initial impression too much. The pretest probability changes in an earthquake, because the context raises the chance of crush injury, kidney injury and infection in patients who, on another day, would be low risk. The same reasoning applies to the blast patient who looks well and deserves observation, the topic of the text on who looks well and should be observed with judgment.

There is also a cost for those providing care. Deciding who gets the first bed, the first dialysis or the first surgery weighs on the team, and the text on what studies show about moral injury in health professionals discusses this impact.

Frequently asked questions

What is the most common injury in an earthquake?

In reviews, soft tissue injuries and extremity fractures are the most reported, but among inpatients crush injury often appears as the main injury (Doocy and colleagues, 2013).

Why can a patient with crush injury worsen after being rescued?

Because reperfusion of the released muscle can cause metabolic changes, hyperkalemia and cardiac arrest (Abu-Zidan and colleagues, 2023).

How many patients with crush syndrome need dialysis?

In the meta-analysis of 50 studies, the pooled proportion was 49 out of every 100 (Rostami and colleagues, 2026).

Does START work in an earthquake?

It works, but it is not perfect. In a database of 29523 victims, REMS had the largest area under the curve, and the other methods were close to each other (Peng and colleagues, 2021).

When do patients arrive at the hospital?

Most in the first days, with a concentration in the first 24 hours (Abu-Zidan and colleagues, 2023), and a portion keep arriving after days under the rubble.

Why are wound infections different?

Because gram-negative and multidrug-resistant germs, such as Acinetobacter baumannii, predominated in series of earthquake survivors (Hasanoglu and colleagues, 2024).

Take-home messages

• Building collapse accounts for most deaths and injuries in earthquakes.
• Soft tissue injuries and extremity fractures predominate, but crush injury concentrates the severity.
• Time trapped and number of compressed limbs are mandatory questions in triage.
• Almost half of patients with crush syndrome need dialysis.
• No triage tool is enough on its own, and reassessment is part of the algorithm.
• Weeks later, infections by resistant germs require culture and attention to the local profile.

References

1. Abu-Zidan FM, Idris K, Cevik AA. Prehospital management of earthquake crush injuries: A collective review. Turk J Emerg Med. 2023. PMID 38024191.
2. Cakir IM, Sengul I, Bekci T, et al. A needful, unique, and in-place evaluation of the injuries in earthquake victims with computed tomography, in catastrophic disasters! The 2023 Turkey-Syria earthquakes: part I. Rev Assoc Med Bras (1992). 2023. PMID 37585997.
3. Doocy S, Daniels A, Packer C, et al. The human impact of earthquakes: a historical review of events 1980-2009 and systematic literature review. PLoS Curr. 2013. PMID 23857161.
4. Hasanoglu I, Kayaaslan B, Ozdemir G, et al. Infections after musculoskeletal injuries in earthquake survivors. Jt Dis Relat Surg. 2024. PMID 39189579.
5. Jacquet GA, Hansoti B, Vu A, et al. Earthquake-related injuries in the pediatric population: a systematic review. PLoS Curr. 2013. PMID 24761308.
6. Köroğlu M, Karakaplan M, Barakat M, et al. Predictive factors for acute kidney injury and amputation in crush injuries from the Kahramanmaraş earthquakes. Ulus Travma Acil Cerrahi Derg. 2024. PMID 38967526.
7. Köroğlu M, Karakaplan M, Ergen E, et al. The initial response of a local hospital in the earthquake zone during the February 6, 2023 Kahramanmaraş earthquakes: Injuries and challenges. Acta Orthop Traumatol Turc. 2023. PMID 38454212.
8. Peng Y, Hu H, Hu H. Assessment of earthquake casualties and comparison of accuracy of five injury triage methods: evidence from a retrospective study. BMJ Open. 2021. PMID 34625415.
9. Rostami P, Jafari Rouhi A, Hajebrahimi R, et al. Crush injury syndrome in earthquakes: a systematic review and meta-analysis on its frequency and complications. BMC Emerg Med. 2026. PMID 41928063.
10. Sarı H, Özel M, Akkoç MF, et al. First-Week Analysis after the Turkey Earthquakes: Demographic and Clinical Outcomes of Victims. Prehosp Disaster Med. 2023. PMID 37165847.
11. Uluöz M, Gökmen MY. The 2023 Turkey Earthquake: Management of 627 Pediatric Musculoskeletal Injuries in the First Month. Children (Basel). 2023. PMID 38002824.
12. Xue W, Chen W, Wang J, et al. Distribution and antibiotic resistance of pathogens in post-earthquake traumatic infections: a systematic review and meta-analysis. BMC Infect Dis. 2025. PMID 41168685.
13. Yilmaz S, Tatliparmak AC, Karakayali O, et al. February 6th, Kahramanmaraş earthquakes and the disaster management algorithm of adult emergency medicine in Turkey: An experience review. Turk J Emerg Med. 2024. PMID 38766417.