Occult Bacteremia in the Post-Pneumococcal Conjugate Vaccine Era: Does the Blood Culture Stop Here?

The issue of “occult” bacteremia was discovered in the 1970s with a realization that some well-appearing febrile children in the toddler age group (3–36 months), who were initially managed as outpatients, had bacteremia.1 This bacteremia was not so much mysterious, but rather unsuspected, because these children appeared well to the clinicians and had either no focal signs of infection or only mild upper respiratory tract illnesses. Due to the fear of the bacteremia leading to invasive bacterial illness, such as meningitis, many practitioners recommended routine blood tests including blood culture as part of the management strategy for these febrile children.2,3 In the 1980s, clinician-scientists from the then-developing field of pediatric emergency medicine took special interest in occult bacteremia. During this time, there was an increasing reliance on urban, academic medical centers to provide health care to inner-city children, and many of these children used the emergency department (ED) for the evaluation of acute febrile illness. Because it was evident that some of these well-appearing febrile children had occult bacteremia, and that some of those children with occult bacteremia went on do develop serious sequelae, many practitioners, especially those practicing in EDs where follow-up was often unreliable, advocated some combination of routine laboratory testing, blood culture, and empirical administration of antibiotics (usually an intramuscular dose of ceftriaxone) for these febrile children.4–6 However, this routine was not without consequences. Administration of antibiotics, without a clear source of infection, altered the ability of the clinician to follow the child’s clinical state and potential need for more directed therapy. In addition, the evolution of certain infections, such as urinary tract infection, could be partially masked by the antibiotics. Concerns about overuse of antibiotics and the increasing incidence of multiple drug-resistant bacteria heightened the controversy in management. One difficulty in studying occult bacteremia was that the incidence of bacteremia was always a moving target. While the original studies focused on pneumococcal bacteremia, which was the most prevalent of those bacteria responsible for occult bacteremia, Haemophilus influenzae type b (HIB) and meningococcus, while less common, were responsible for much of the associated invasive illness. With the development of widespread HIB vaccination, the incidence of occult bacteremia dropped from about 3% to 1.5%–2%.7,8 However, since pneumococcal bacteremia accounted for almost 90% of occult bacteremia, the testing and empirical treatment routine remained the standard of care for many practitioners. By the beginning of the new millennium, approval of the heptavalent conjugate vaccine for pneumococcus (PCV7) was destined to have a marked effect on the management of occult bacteremia. The vaccine was shown to offer protection for about 85% of pneumococcal disease and was 97% effective for prevention of disease caused by vaccine serotypes.9 Furthermore, about 98% of serotypes identified in children with occult bacteremia were included in the vaccine.10 Considering all these factors, it was estimated that the incidence of occult pneumococcal bacteremia would likely fall to less than 0.1% and the incidence of all occult bacteremia would fall to less than 0.25%.11 If indeed the incidence of occult bacteremia would fall to these very low rates, one would need to reconsider the routine laboratory testing of febrile, well-appearing children.12,13 The article by Wilkinson et al.14 in this issue of Academic Emergency Medicine adds to what is becoming a critical mass of growing data validating the marked decline in the incidence of occult bacteremia. In their retrospective study of 8,408 febrile children 3–36 months old who were previously healthy and had no source of infection on physical examination, the occult pneumococcal bacteremia rate was 0.17% (95% confidence interval [CI] = 0.09% to 0.27%), and the overall occult bacteremia rate was 0.25% (95% CI = 0.16% to 0.37%). While encouraging, these data are limited somewhat by the retrospective design. To be sure, a prospective study would be the preferred methodology, for it would allow consecutive enrollment of all eligible subjects, objective and standardized assessment of the child’s clinical appearance, documentation of the immunization status, and more complete outcome data. However, the low incidence of occult bacteremia would require many years of patient enrollment with a management strategy that is becoming quickly outdated. During the period of their study, it was standard policy at the authors’ institution to send blood cultures on well-appearing febrile children in the occult bacteremia age group. Therefore, we assume that most eligible children were included in the study. Still, we do not know how many children (if any) who met the inclusion criteria did not have a blood culture performed and thus were not included in analysis. This may have led to an underreporting of the true denominator (i.e., the number of well-appearing children seen in the ED regardless of whether a blood culture was done) or the true numerator (the number of children with positive blood cultures). The authors point out that sicker-appearing children may have been more likely to have a blood culture obtained, leading to an overestimate of the prevalence of occult bacteremia. This argument highlights another limitation, the lack of a consistent, objective definition of “well-appearing.” Without prospective data collection, the authors were left to deconstruct the term “not well-appearing” to include hospital admission or documentation on the physical exam of ill-appearing, unwell, lethargic, less responsive, or toxic. All in all, the rate of bacteremia reported might not be entirely accurate, due to the methodology used. Despite these limitations, the study does support prior work looking at the decline of occult bacteremia in the post–pneumococcal vaccine era. Recent studies15,16 have shown the incidence of occult pneumococcal bacteremia in febrile children 3–36 months old to have fallen to 0.2%–0.36% and the overall occult bacteremia rate to be 0.41%–0.7%. Furthermore, in the study that was able to stratify the bacteremia rate by vaccination status, none of the 833 children who had at least one immunization of heptavalent pneumococcal vaccine had pneumococcal bacteremia.15 These data, taken together with the data of Wilkinson et al., offer validation that the current rate of occult bacteremia is close to, if not below, 0.5% overall and is below 0.25% for the immunized child. These data should have a profound impact on those practitioners still performing routine testing as a “fever work-up.” Although the cost-effectiveness of routine blood testing for these children is debatable at bacteremia rates above 0.5%, if the rate has dropped below 0.5%, management strategies that involve routine testing and/or empirical administration of antibiotics should be abandoned. To be sure, we need to continue to monitor bacteremia rates in the future. Serotypes that did not play a role in bacteremia may become more common (vaccine switching) and some children will remain unimmunized or incompletely immunized. In addition, other causes of bacteremia, such as invasive meningococcal disease, will be unaffected by the pneumococcal vaccine. Nevertheless, we have likely reached the point where the risks and costs of testing well-appearing young (3- to 36-month-old) febrile children for the presence of bacteremia exceed the potential benefits. Parental pressure for performing unnecessary testing and requesting unnecessary antibiotics for what appears to be routine viral illness remains a major problem today. Many parents have been conditioned to believe that testing and antibiotics are a necessary part of management of fever. This leads to more “fever phobia” and more visits to the practitioner or ED for evaluation of fever. We need to break this vicious cycle and curb the use of unnecessary antibiotics. The ability to put the risk of occult bacteremia and, by extension, the very low risk of sequelae into this clinical context will help with this goal.