EM@3AM: Cotton Fever

Author: Andrea Nillas, MD (Toxicology Fellow, UT Southwestern Medical Center/Parkland Hospital, Dallas, TX) // Reviewed by: James Dazhe Cao, MD (@JamesCaoMD, Associate Professor of EM, Medical Toxicology, UT Southwestern Medical Center, Dallas, TX); Alex Koyfman, MD (@EMHighAK); Brit Long, MD (@long_brit)

Welcome to EM@3AM, an emDOCs series designed to foster your working knowledge by providing an expedited review of clinical basics. We’ll keep it short, while you keep that EM brain sharp.


A 30-year-old male with history of intravenous drug use (IVDU) presents with fever. He reports last drug use was one hour prior to arrival and admits to reusing the same cotton filter on multiple occasions. He denies recent cough, chest pain, shortness of breath, abdominal pain, nausea, vomiting, or diarrhea. VS include Temp 101.0°F, HR 110, BP 130/80, RR 18, SpO2 98% on room air. On exam, he is diaphoretic and has track marks on bilateral upper extremities. Lung sounds are clear, and there are no heart murmurs. Complete blood count and metabolic panel are unremarkable.

What is the likely diagnosis?

Answer: Cotton fever


  • The term “cotton fever” was first described in 1975 in reference to febrile reactions in intravenous drug users after using drug suspensions that were filtered through cotton balls.1
    • Dissolvable illicit substances such as heroin are commonly strained through cotton or cigarette filters.
    • The practice of “shooting the cottons,” boiling used cotton filters to extract residual narcotic when a user’s supply is depleted, may increase the risk of cotton fever, and is described in early case reports.2
  • Cotton fever is a diagnosis of exclusion and is a self-limiting condition that resolves spontaneously in 24-48 hours.1,3
  • Although well-recognized in the drug-user population, the incidence and prevalence of cotton fever are unclear as few publications or epidemiological data exist.
    • Prevalence may be up to 54% among people who inject opioids as reported in a cross-sectional study of self-reported users.4
    • True incidence is likely more common than recognized by medical community.3,5
  • Risk factors:
    • Drug users who mainly use cotton filters were more likely to report experiencing cotton fever than those who mainly used membrane filters.
    • Duration of opioid use greater than 1 year and type of opioids used are contributing factors.
    • Concomitant crack cocaine injection amplifies risk of cotton fever.4



  • Three proposed mechanisms exist:
    • Pharmacologic theory suggests that extracts of cotton contain small, soluble, heat-stable substances with pyrogenic activity. 6
      • Such identified substances include polypeptides, methyl piperonylate and QAE-2.7
    • Immunologic theory suggests that people have pre-formed antibodies to cotton.4
      • Studies attempting to demonstrate skin reactivity to cotton extracts have been inconclusive and attempts to isolate antibodies have been unsuccessful.8
    • Endotoxin theory uses the knowledge that cotton plants are heavily colonized with endotoxin producing gram-negative rods such as Enterobacter agglomerans (since renamed Pantoea agglomerans).9,10
      • Endotoxins can be released in water due to their solubility. Heating increases the toxin effect, likely by denaturing proteins in endotoxin complexes resulting in more pronounced febrile illness in IVDU.9
  • The endotoxin theory is currently favored as the association between respiratory symptoms and gram-negative contaminants of cotton dust has been established.
    • Additionally, cotton plant extracts have demonstrated ability to induce PMN chemotaxis and complement activation.

Clinical Presentation:

  • Patients may present ill appearing, mimicking sepsis or opioid withdrawal1,11
  • Symptom onset typically about 10-30 minutes post-injection
    • General: Fever, malaise, chills
    • Cardiopulmonary: Dyspnea, tachycardia, palpitations
    • Gastrointestinal: Abdominal pain, nausea, vomiting
    • Musculoskeletal: Myalgia, arthralgia
    • Neuro: Headache6
  • The table below from Zerr et al. (2016) compares vital sign abnormalities and symptom duration between available case reports.5
Age (years)SexIV DrugTime to symptom onset (min)Temperature (°C)HR (bpm)BP (mmHg)WBC at admittance (k/mm3)Time to symptom resolution (hours)
33FPentazocine, methylphenidateA few minutes36110140/904,20012
26FHeroinA few minutes38.6137129/472,0005

Table 1. Presentation vitals for patients with cotton fever. (Source: Zerr et al 2016) BP, blood pressure; F, female; HR, heart rate; IV, intravenous; M, male; NA, not available; WBC, white blood cell.



  • Detailed history taking may reveal recent drug use and use of cotton filters.
  • To assess for signs of IVDU and serious bacterial infection, physical exam should include the following:
    • Cardiopulmonary: Evaluate for murmurs or focal lung findings
    • Abdomen: Benign or have diffuse tenderness without rebound or guarding
    • Musculoskeletal: Joint tenderness may be present
    • Skin: Evaluate for skin rashes or nail findings suggestive of endocarditis, or track marks which may raise suspicion for IVDU
  • Routine laboratory tests are of limited use and may be completely normal or have variable abnormalities including leukocytosis or elevated liver enzymes.6,12
  • Urine drug screens may detect the opioids codeine, heroin, and morphine. False negatives may result if adulterated or if incorrectly marketed as heroin but contain fentanyl analogues. Opioids such as oxycodone or fentanyl may require targeted immunoassays for detection.
  • Blood cultures should be obtained to rule out bacteremia but will have no growth.
  • Imaging should be tailored to patient presentation and clinical suspicion (i.e. Chest X-Ray for pneumonia, neuroimaging for central nervous system infection, echocardiogram for cardiac vegetations).12

Differential Diagnosis:

  • Bacteremia
  • Cellulitis/Abscess
  • Endocarditis
  • Epidural abscess
  • HIV
  • Meningitis
  • Opioid withdrawal
  • Osteomyelitis
  • Pneumonia
  • Sepsis
  • Viral syndrome


  • Serious bacterial infection should be excluded before diagnosing cotton fever.
  • The mainstay of treatment is supportive care as cotton fever is self-limiting.
    • As needed IV fluids, antipyretics, empiric antibiotics.
  • Patients should be educated about safe practices to reduce morbidity and mortality from IVDU.12
    • E.g., Needle exchange programs, substance use disorder treatment referrals, sexually transmitted disease screening.


  • Place in for observation for at least 24 hours to rule out endocarditis or bacteremia.
  • Patients with negative blood cultures and appropriate follow up may be discharged if symptoms have improved.12

Key points:

  • Cotton fever is a diagnosis of exclusion.
  • The exact mechanism is not fully elucidated but endotoxin contaminants are currently favored to be the causative agents.
  • Symptom onset occurs within minutes post injection and will self-resolve within 24-48 hours.
  • Febrile patients with a history of IVDU and poor follow up should be placed in for observation to await negative growth on blood cultures.
  • Management is supportive care as symptoms are self-limiting.


  1. Bryson, PD. Comprehensive Reviews in Toxicology: For Emergency Clinicians. Washington, DC: Taylor&Francis;1996.
  2. Shragg T. “Cotton fever” in narcotic addicts. JACEP. 1978;7:279-280.
  3. Xie Y, Pope BA, Hunter AJ. Cotton fever: does the patient know best?. J Gen Intern Med. 2015;31(4):442-444.
  4. Mezaache S, Briand-Madrid L, Laporte V, et al. Correlates of self-reported cotton fever experience among people who inject opioids. Subst Use Misuse. 2020;55(6):1021-1027.
  5. Zerr AM, Ku K., Kara A. 2016. Cotton Fever: A Condition Self-Diagnosed by IV Drug Users. J Am Board Fam. 2016;29(2):276–279.
  6. Harrison DW, Walls RM. “Cotton fever”: a benign febrile syndrome in intravenous drug abusers. J Emerg Med. 1990;8(2):135-139.
  7. Merchant JA: Byssinosis. In: Baum GL, Wolinsky E (eds). Textbook of pulmonary disease. 3rd Boston: Little, Brown: 1983:789-800.
  8. Torka P, Gill S. Cotton fever: an evanescent process mimicking sepsis in an intravenous drug abuser. J Emerg Med. 2013;44(6):e385-387.
  9. Rylander R. The role of endotoxin for reactions after exposure to cotton dust. Am J Ind Med. 1987;12(6):687-97.
  10. Ferguson R, Feeney C, Chirurgi VA. Enterobacter agglomerans–associated cotton fever. Arch Intern Me 1993;153(20):2381-2.
  11. Francis MJ, Chin J, Lomiguen CM, Glaser A. Cotton fever resulting in Enterobacter asburiae ID Cases. 2020;19:e00688.
  12. Feldman JA, Nentwich LM. Complications of Injection Drug Use. In: Wolfson AB. Harwood-Nuss’ Clinical Practice of Emergency Medicine 6e. Wolters Kluwer; 2015.


Leave a Reply

Your email address will not be published. Required fields are marked *