The Most Abused Lab Test in Pediatrics

This is an excerpt from Suffer the Children: Flaws, Foibles, Fallacies and the Grave Shortcomings of Pediatric Care adapted specifically for this blog post.

Often, when a child with a fever is evaluated in a doctor's office or emergency room, a complete blood count or CBC is ordered as part of the workup.  This can be reassuring to parents.  After all, a blood test is an objective test, immune to the human errors we might attribute to the "art" of medicine.  Think again.  In most cases, the CBC is completely unnecessary in the evaluation of fever in children. In many cases it is downright harmful because of the regularity with which it is misinterpreted and leads to the wrong diagnosis and unnecessary treatment.

The history of the use of the CBC is inextricably intertwined with that of Occult Bacteremia, a condition whose mysterious name has nothing to do with the supernatural.  “Bacteremia” refers to the presence of bacteria in the blood stream, whereas “occult” just means “hidden”. Occult Bacteremia is the presence of bacteria in the bloodstream in a well-appearing febrile child, usually between the ages of 3 months and 36 months, with no identifiable focus of infection.

To demonstrate bacteria in the blood stream one must utilize a blood culture: a test whose result can be delayed several days as blood must be incubated in a proper medium before the growth of bacteria can be detected.  Understandably, doctors have wanted a quicker test, as bacteremia can be associated, albeit rarely, with the development of other serious infections such as meningitis.  In the search for a reliable quick test, the first order of business was to look at the statistics on the rate of prevalence of Occult Bacteremia.

In the 1970s, the rate of Occult Bacteremia in febrile children outside the newborn period but less than three years of age ranged from 3% to 10% based on several studies.  One clear discriminator in identifying children at particular risk for this condition was the height of the temperature:  children with a temperature of less than 102.2 degrees Fahrenheit (39 degrees Celsius) were found to almost never have Occult Bacteremia.  The results were very similar when the white blood cell count was utilized with a WBC greater than 10,000 being considered abnormal.  Using this cutoff however, an abnormal white count vastly overestimated the likelihood of bacteremia:  only 5.8% of children would actually be expected to have positive blood cultures when the WBC was greater than 10,000.  A WBC less than 10,000, on the other hand, was highly predictive of the absence of bacteremia with an accuracy of 99.2%.  This may come as no surprise after our earlier examples, because this is another case of a disease with a low prevalence.

What if we tried to be more discriminating, by using the height of temperature in conjunction with an elevated WBC?  In fact, let’s try using 15,000 instead of 10,000 as our cut-off for an abnormal WBC.  In the 1970s, a child under 2 years of age with a temperature greater than 102.2 and a white blood cell count greater than 15,000 had a 10% chance of having occult bacteremia. 
 

Based on this data, several recommended approaches to the highly febrile child with no apparent source of infection were published.  The most influential was a consensus paper published in July of 1993 in the journal, Pediatrics, the official journal of the American Academy of Pediatrics.  This paper established guidelines which included the recommendation that children 3 to 36 months of age with fever greater than 102.2 degrees Fahrenheit and a WBC greater than 15,000 should have a blood culture and be treated with antibiotics at least until the results of the blood cultures are available.

It is important to understand that when recommendations such as these are formulated they are not a direct consequence of the available data but involve the making of judgment calls that take into account economic, political, ethical and socio-cultural factors.  The data available at the time demonstrated that of 10,000 children with a temperature greater than 102.2 ̊̊, 300 would be expected to have Occult Bacteremia (using a rate of 3%).  Of these, only about 25 would go on to have an invasive infection.  The panel thus concluded that it is preferable to submit thousands of children to further blood testing and unnecessary antibiotics than allow 25 from developing a serious infection.  I am not disputing the decision; rather, I want to point out that there is a trade-off.  Other countries, particularly those with fewer resources or at a time where antibiotic resistance is of graver concern, may have easily reached different conclusions.

The net effect was to establish the WBC of 15,000 as an important signpost in the evaluation of fever in children, but many doctors forgot the details of these recommendations or completely misunderstood their limited scope.  The guidelines applied only to children between the age of 3 months and 36 months – older children are not at risk for Occult Bacteremia.  Also, they were specifically geared for children who do not have an otherwise recognizable source of infection.  Many practices routinely obtain CBCs to “rule out Occult Bacteremia” in situations that are completely different from those outlined in the consensus paper wrongly believing that they are adhering to these guidelines.

I once attended a tall, stocky 14 year old who was dispatched to the hospital by his Pediatrician; his mother frantic because she had just been told that the child had “an infection in his blood”.  From the young man’s appearance I was immediately doubtful of the accuracy of this assessment.  It turns out that the young man had started having an achy throat the night before, and had started running a fever during the night.  Earlier that morning he had a headache and was having pain with swallowing his breakfast.

Love, Redemption, Treachery

 Peter Palmieri's engaging novel of medical suspense,

The Art of Forgetting is the summer's hottest new read

He went to his pediatrician who diagnosed a Strep Throat infection based on a positive rapid strep test.  But then the pediatrician also obtained a CBC which showed a White Blood Cell count of 17,000.  He told the mother that the teen had “Occult Bacteremia, a serious blood infection”, and referred him to the hospital.  I carefully explained to the mother that her son did not actually have a blood infection and, though I would end up angering the referring pediatrician (who thereafter stopped sending me patients), I wrote a prescription for Penicillin to treat his throat infection and immediately sent the family home.

There were so many foibles in this case that it is hard to know where to start.  First off, Occult Bacteremia is a concern for children under 36 months of life, not teenagers.  Even if we invoked the 1993 guidelines in this setting, this young man would not be suspected of having this condition.  Furthermore, even in 1993, most children under 36 months of age did not actually have Occult Bacteremia; only about 3% did.  But what of the elevated WBC of 17,000?  Since the CBC is not part of the routine evaluation of Strep Throat infections, this test is simply indecipherable.  What is the expected WBC in a child with Strep Pharyngitis?  Who knows!  I tried researching this question but was unable to find any concrete answer.  To my knowledge, there never has been a study on the height of the WBC in children with strep throat infection because it is an irrelevant test for this condition.

Unfortunately, I’ve been involved in scores of similar instances where children were diagnosed with having “an infection in the blood” on the sole basis of the WBC:  an impossible assessment. I would therefore like to take this opportunity to launch an appeal to all pediatricians:  please stop telling parents that children have a blood infection on the basis of an elevated white blood cell count!

In the 1970s, two bacteria accounted for the majority of Occult Bacteremia: Streptococcus Pneumonia (65% to 70% of cases) and Haemophilus Influenza B (10% to 20% of cases).  Much has changed since the 1970s.  In 1990 a Haemophilus Influenza B vaccine was introduced, and in 2000 one for Streptococcus Pneumoniae became part of the vaccination schedule for all infants.  The result has been a steep drop in the occurrence of this disease in the last twenty years.  Whereas in the 1970s studies reported that 3% to 10% of children with fever had occult bacteremia, more recent studies indicate rates below 1%.  The CBC is even less reliable as a predictor of occult bacteremia than it was when the 1993 guidelines were issued.  The current consensus is that well appearing children with high fever do not require a CBC or a blood culture.  Many pediatricians, however, seem to be stuck in the early 1990’s.

There is one test that is indispensable in the evaluation of high fever in boys under one year of age and in girls under 36 months of age: the evaluation of a urine sample for evidence of a Urinary Tract Infection (UTI).  Of the “hidden” infections, this is by far the most common: one that can have serious repercussions if not identified and treated early.  But pediatricians prefer to get a CBC than a urine sample.  In the office, obtaining a blood sample for a CBC is a simple procedure that requires little skill: a needle is used to poke a finger and a few drops of blood are obtained.  On the other hand, to get an adequate, uncontaminated urine sample in a child who is not yet toilet trained requires introducing a tiny catheter through the urethra and into the bladder.  This is a procedure that requires some degree of skill, patience and time.  Many doctors avoid it altogether by referring children to the Emergency Room if they believe that a urine specimen is needed.  But they happily poke holes in children’s fingers.

This reminds me of the joke of a gentleman who, late one night after having too much to drink, is seen on all fours under a lamppost feeling his way around the pavement.  A Good Samaritan stops and asks if he can be of assistance.  The inebriated man explains that he’s lost his keys.  The kind stranger at once crouches down and joins the search.  After some ten minutes of rummaging, the Samaritan asked the other man if he was sure he’d dropped the keys in that area.  “Oh no”, says the carouser, “I dropped them in that dark alley across the street”.

“Then why in good heavens are you looking for them here?”

 “Well”, the tipsy man said, “the light’s much better under the lamppost”.

Some doctors would rather embark in a fruitless search with blood work than looking in the right place, simply because obtaining the indicated test is inconvenient.  They’d rather avoid the dark alleys.