Leukocytosis and bacteremia When examining a child with acute fever, the most common method is a complete blood count. A lot has been written about this method, but the available literature data are often contradictory, which partly depends on the characteristics of the studied groups of patients. According to Todd, in the differential diagnosis of bacterial infections and non-bacterial diseases, determining the absolute number of polymorphonuclear and nonsegmented neutrophils (ring forms) is more important than the total number of blood leukocytes.
On the contrary, McCarthy believes that the determination of the chils of polymorphonuclear and ring neutrophil forms plays a smaller role than the determination of the total number of leukocytes. However, Morens showed that neither the total number of blood leukocytes nor the leukocyte formula (polymorphonuclear and unsegmented neutrophils) do not assist in the differential diagnosis of bacterial and non-bacterial diseases! But one should not draw conclusions that one of these works is more correct than the other.
The authors simply investigated different populations and answered different questions. So, Todd retrospectively studied hospitalized children with acute illnesses by processing case histories. Moreover, many patients were selected purposefully, taking into account the final diagnosis (bacterial meningitis, osteomyelitis, pyelonephritis, dysentery). Some diseases, such as pneumonia and otitis media, were not included in the analysis, because, according to the author, it is not possible to determine retrospectively what etiology – viral or bacterial – was the disease.
The McCarthy study was prospective; it included 330 children admitted to the emergency room. In addition, McCarthy studied children under the age of 24 months, while Todd examined children of all age groups. In his retrospective study, Morens included 328 children admitted to the Pediatric Infectious Diseases Department. Todd and Morens, as a rule, studied all acutely ill children, and McCarthy studied only children with a body temperature of 40 ° C and higher.
Of these three papers, the work of McCarthy is most relevant to the problem of acute fever without local symptoms and without clinical signs of serious infection. For, according to Todd, out of 244 children with bacterial diseases, 36% had purulent meningitis, 25% had abdominal sepsis (perforative appendicitis, etc.) and 13% had osteomyelitis or septic arthritis.
According to McCarthy, as a criterion for the selection of patients who are at risk for the development of bacteremia or severe bacterial infections, it is advisable to use a leukocytosis level of 15×109 / l and above. According to him, with bacteremia in children of the first 2 years of life with a body temperature of 40 ° C and above, this method has 50% sensitivity and 15% specificity. In other words, 50% of children with bacteremia had leukocytosis of 15×109 / l or more, and 15% of patients with leukocytosis of 15×109 / l or more had bacteremia.
According to Teele, with bacteremia in children of the first 2 years of life with a temperature of 38.3 ° C and (higher, with leukocytosis 15×109 / l, this method was characterized by 79% sensitivity and 7% specificity. For leukocytosis 20×109 / l, this method had only a little more specificity (10%), but at the same time less sensitivity (53%). McGowen showed that when diagnosing bacteremia by seeding blood for leukocytosis 20×109 / l, the method has 12% specificity. In all three studies, a study was made of a general group of children with fever.
If we consider the materials Teele, relating only to children with fever without local symptoms, it will be found that in patients with bacteremia with leukocytosis of 15×109 / l, the method has 100% sensitivity and 11% specificity. The corresponding figures for leukocytosis 20×109 / l were 40% and 12.5%. In Baron’s work on the study of children in the first 2 years of life with a fever with no local symptoms in leukocytosis of 15×109 / l, the method was characterized by 88% sensitivity and 24% specificity.
In children with fever with no local symptoms, leukocyte count is currently most commonly used. The results presented in the table make it possible to draw some conclusions about the use of this test in outpatients. First, with leukocytosis of 15×109 / l and more, this method is more sensitive in the diagnosis of bacteremia in patients with fever without signs of focal infection than in patients with focal manifestations (87-100 against 50-79%).
Secondly, with leukocytosis of 20×109 / l, the sensitivity of the method sharply decreases to 35–50%, and the specificity increases slightly. Thirdly, with leukocytosis of 15–20×109 / l, the specificity of the method is generally low in the whole population. In the works listed in the table, specificity ranges from 7 to 38% and averages 16%.
When examining children with fever without local symptoms, as a rule, there is no need to determine the ESR and platelet count. Tests such as PIEF, SFIS, and latex agglutination are not yet available for clinics, but will continue to be important in the future. Spinal puncture, urinalysis and chest radiography as routine tests are shown only in very young children. In all other children, these studies are performed only for clinical reasons.