Step F: Find the site and control the source

  • The systemic inflammatory response is driven along by persistent infection: you must find the source and remove it. This may involve extensive detective work.

In the early detective phase, a series of cultures were sent as part of a fishing expedition. In addition chest radiography was performed. Following physical examination, which will usually indicate the site of infection, more expensive-extensive tests may have been performed, such as computerized tomography. At this point 95 times out of 100 the source is located and controlled. Control in this situation means treatment or removal of the source of the inflammatory response. This may be necrotic tissue, as in necrotizing fasciitis, feces, as in the case of the ruptured diverticulum, an obstructive ureteric stone, infected central line, intra-abdominal abscess, or just a plain infection of the respiratory or urinary tract. Often source control means surgical intervention, although radiological drainage of abscesses has emerged as an excellent alternative.

But, what if the patient is behaving as if he has sepsis, but there is no obvious source?

This is a very difficult situation. There are many causes of fever without an infective agent (myocardial infarction, pulmonary emobolism, gastrointestinal bleeds, inflammatory bowel disease, hematomas, thrombophlebitis), but these would rarely cause a syndrome akin to septic shock: you must always assume a source, and try to control it. So where do you start?

Firstly, always suspect an iatrogenic source – if the patient has artificial material imbedded in him/her, such as a vascular graft, a tunneled intravenous catheter, mechanical heart valves, joint prosthesis, these should be suspected from the beginning. All invasive devices such as iv lines, present at the time of infection, should be removed/replaced. Otherwise…

Common things are common, we know that respiratory tract infections are common, so this is the first place to start. From medical school you may have learned the adage “pus somewhere, pus nowhere, pus under the diaphragm”: the abdomen is the next place to look and so on. Below is a list of potential sites, and investigations, in order of likelihood and sequence:

Respiratory: lower respiratory tract infection – there may be no evidence on chest x-ray. Further investigations would include sending serological tests for atypical organisms, such as mycoplasma and legionella and performing broncho-alveolar lavage (this tends to be low yield).

Abdominal: an ultrasound of the abdomen will reveal dilated bile ducts if there is biliary obstruction and ascending cholangitis, a necrotic gallbladder (acalculous cholecystitis), intrahepatic or subphrenic collections/abscesses. If there is a perforated viscus, pneumoperitoneum should have been visible on chest x-ray. Ultrasound will also demonstrate splenic abscesses, free fluid (possible perforated bowel), and dilated renal pelvis (urinary tract obstruction). If free fluid is present it should be tapped and sent for culture and gram stain. Ultrasound is a very poor test for imaging the pancreas and bowel. If ultrasound has yielded nothing, the computerized tomography (CT) of the abdomen and pelvis is necessary. Ischemic bowel is characterized by pneumatosis (air in the bowel wall); necrotic lesions in the pancreas are easily visible on CT, as are liver, splenic lesions and fluid collections. Any collection viewed should be tapped and drained.

Urinary tract: urinary cultures, changing or removing catheters and abdominal imaging should demonstrate the diagnosis.

Sinuses: patients with long term nasogastric tubes all develop occulded sinuses, with or without sinusitis. Plain x-rays and CT will demonstrate fluid levels in the sinuses if they are occluded / infected. Conservative treatment is removal of the offending tube and topical anti congestants. Antibiotics and surgical drainage may be required.

Heart: endocarditis is one of the most malignant causes of systemic sepsis. Usually a murmur is audible and there are splinter hemorrhages in the nail beds, hematuria, splenomegaly. The right side of the heart is easily viewed with transthoracic echocardiography (echo). If endocarditis is strongly suspected, then the left side should be viewed with transesophageal echo.

Central nervous system: a brain abscess or meningitis should be considered in cases of sepsis, cause unknown. This includes spreading ear infections. A CT of the head with contrast should performed to look for infection, and to determine if the ventricles are normal in size (i.e. normal intracranial pressure). If so, lumbar puncture should be performed to obtain cerebrospinal fluid (CSF), which is sent for culture.

Other sources: examine the mouth for dental abscesses, the prostate for prostatitis, the ischeo-rectal area for evidence of infection. All intravenous catheter sites should be carefully expected for thrombophlebitis, or subcutaneous pus collection

What else? All intensivists have been in the unenviable situation of patient’s dying on them without the source of sepsis being identified and controlled. There are other potential tests such as radiolabelled white cell scans (where do the white cells collect?) and bone marrow biopsies, but these tend to be very low yield in intensive care.


Copyright 2002 All rights reserved