The purpose of this case study is to highlight ways in which medical algorithms can be integrated into the clinical decision making process.  This case study is the first of a series focused on sepsis, a common but complex clinical problem.  Use of algorithms can help healthcare providers diagnose sepsis early. Throughout our discussion, appropriate algorithms are suggested to help clinicians with diagnosis, risk prediction, prognostics, management of sepsis, and management febrile neutropenia when it develops following cancer therapy.

Case History

Patricia is a 55-year-old female with a history of anal squamous cell carcinoma. The cancer was resected but additional chemoradiation was recommended by her oncologist.

After a recent course of chemotherapy she presented with fever, diarrhea and rectal pain. A CBC showed pancytopenia with neutropenia (the absolute neutrophil count was 100 per microliter). She was admitted for observation since the cause of the diarrhea was unknown.

One day after admission the patient’s fever had gotten worse and she became hypotensive. She showed signs of confusion. The clinical diagnosis of sepsis was made, blood cultures were taken and antibiotics started. The patient’s blood pressure continued to fall and she was diagnosed with septic shock.  She was admitted to the ICU where she was treated aggressively. Although there was a transient decline in renal function, no other organ failures occurred. Blood cultures isolated a single pathogen susceptible to the initial antibiotic therapy. Her condition improved as the neutrophil count started to recover and the patient was able to be discharged from intensive care.


Sepsis is a relatively common clinical problem. While septic patients may share certain findings, there may be differences depending on the patient, the underlying pathogen and its source.

One category of sepsis affects cancer patients following cancer therapy. Cancer therapy can be toxic to the bone marrow and result in pancytopenia. The fall in white blood cells can negatively impact host defenses, making the person susceptible to infection during the period of neutropenia. These patients may develop fever, resulting in the label of febrile neutropenia.

Risk Factors for Predicting Febrile Neutropenia Following Chemotherapy

The first task when identifying a patient with febrile neutropenia is to determine whether the person is infected, and, if so, where?

Risk Score for Invasive Bacterial Infection in Pediatric Patients with Cancer and Febrile Neutropenia

Features that distinguish a patient with febrile neutropenia who is septic from others patients with sepsis are:

  1. The patient is vulnerable with an underlying malignancy and often other serious comorbid conditions.
  2. The risk for infection is increased during the period of therapy-associated bone marrow suppression, and the risk decreases as the bone marrow recovers.
  3. The prognosis for a septic cancer patient who is admitted to the ICU is often poor.

Probability of Mortality Model for Predicting Outcome for Cancer Patients Treated in the Intensive Care Unit

Model for Predicting Mortality in a Cancer Patient After 72 Hours in the Intensive Care Unit (ICU)

Prediction of Mortality from Bacteremic Sepsis in ICU Patients

Not all patients with febrile neutropenia are infected. One task of the oncologist is to identify a subgroup of patients who are at low risk for infection who can be discharged home and followed as an outpatient.

Multinational Association for Supportive Care in Cancer (MASCC) Risk Index for Identifying Low Risk Febrile Neutropenic Cancer Patients

Criteria for Early Discharge of Pediatric Cancer Patients with Febrile Neutropenia

Factors from the IDSA Fever and Neutropenia Guidelines Favoring a Low Risk for Severe Infection in a Cancer Patient with Neutropenia

Take Home Messages

  • Sepsis is a common but complex clinical problem.
  • A cancer patient with febrile neutropenia can develop a life-threatening infection.
  • Some patients with febrile neutropenia are low risk and can be managed differently from those at greater risk.
  • Medical algorithms can help to diagnose these patients early so that effective therapy can be started before serious complications develop.