Social smile and occult bacteremia.

Abstract

The presence of a social smile as a clinical aid in determining that a child does not have a serious infection was debated in three recent communications.1-3 In the first two reports meningitis was the infection in question and the authors offered only anecdotal observations. Although we are in agreement that most children with meningitis do not have a social smile, we are not aware of any report that provides data confirming this observation. The third report relates social smile to the presence or absence of serious illnesses including meningitis and other illnesses as well. The author cites an earlier study of 312 febrile children in which he was a participant4 where all of 26 febrile children with serious illnesses had absent or impaired social smiles (sensitivity, 100%) and no child who smiled normally had a serious illness (negative predictive value, 100%). However, fewer than 20% of children who did not smile normally had a serious illness (positive predictive value, <20%). In this study only three patients had meningitis and two had bacteremia. Occult bacteremia is a serious bacterial infection that can progress to fulminent septicemia and death or focal infection such as meningitis, although it may subside spontaneously or after antimicrobial therapy. In a study we recently reported5 comparing oral vs. parenteral antimicrobial treatment of children with occult bacteremia, the presence or absence of a social smile was specifically documented as a part of the baseline and follow-up evaluation of all patients. The relationship between social smile and occult bacteremia was not addressed in that report. This analysis is the subject of our report. Methods. The study was a multicenter cooperative investigation with prospective design. Study patients were children 3 to 36 months of age with fever of ≥39.5°C and a white blood cell count ≥15 × 109/l or fever ≥40°C regardless of the result of the white blood cell count, a normal chest radiograph and urinalysis and no focus of infection on examination. Blood culture was obtained and demographic and clinical data were tabulated, including the presence or absence of a social smile, at the initial clinic evaluation. Statistical analysis. Differences of continuous variables (age, degree of fever, duration of fever and white blood cell count) between subjects with and without smiles were analyzed by unpaired t tests. For categorical variables (gender, presence or absence or bacteremia) contingency table analysis with Fisher's exact tests were performed. Statistical significance was defined as P ≤ 0.05. Results. Of 519 patients studied data concerning the presence or absence of a social smile was not recorded for 7 subjects. Of these 7 patients 5 had no growth from their blood cultures and 2 grew Streptococcus pneumoniae. Of the 512 (11%) subjects for whom social smile data were recorded at the initial evaluation, 58 had bacteremia. Social smile was present in 26 of the 58 (45%) bacteremic subjects and in 224 of the 454 (49%) subjects without bacteremia. This difference is not significant (P = 0.58). Of the bacteremic patients social smile was present in 24 of 49 (49%) with S. pneumoniae, 2 of 6 (33%) with Haemophilus influenzae type b, 0 of 2 with Neisseria meningitidis and 0 of 1 with group B Streptococcus. Table 1 shows the percentages by gender and the mean age, mean white blood cell count, mean temperature and duration of fever for subjects with and without a social smile. The only significant difference noted is that subjects without a social smile had a shorter duration of fever than those who had a social smile. Discussion. We conclude that a social smile is commonly present in sick febrile children and that it is equally common in those who have or do not have occult bacteremia. In our study patients the presence of a social smile did not vary with gender, age, presence of bacteremia and white blood cell count; however, it was significantly less common in those with a shorter duration of fever. Assessment of social smile alone has little or no value in predicting occult bacteremia. James W. Bass, M.D., M.P.H.; Robert R. Wittler, M.D.; Martin E. Weisse, M.D. Department of Pediatrics Tripler Army Medical Center Honolulu, HI (JWB) Department of Pediatrics University of Kansas School of Medicine Wichita, KS (RRW) Department of Pediatrics Robert C. Byrd Health Sciences Center of West Virginia University Morgantown, WV