Cardiopulmonary Bypass and Infant Vaccination Titers

We conducted a prospective observational study of patients between 2 and 14 months of age at the University of Virginia that required cardiac surgery with CPB between March 2016 and March 2018. The limits on the age range were to ensure that screened patients had begun routine childhood vaccination. Subjects were recruited if they had received at least the majority (>75%) of age-specific recommended vaccinations as recommended by the Centers for Disease Control and Prevention⁸  and then been verified by documentation in the Virginia Immunization Information System, pediatrician’s record, or from the electronic health record. Subjects were excluded if their surgical procedure was a planned cardiac transplant. Individuals with potential immunodeficiency states, including 22q11 chromosomal deletion or heterotaxy syndrome, were enrolled and underwent subanalysis.

Informed consent was obtained for all subjects. Providers obtained blood samples for antibody assays from the arterial line that was placed at the time of surgery just before going onto CPB. Providers obtained a repeat blood draw the next morning, from the same arterial line, as part of routine morning laboratory draws. The timing of the second sample was variable but was required to fall between 12 and 18 hours postoperatively for standardization purposes so as to not cause unnecessary access to the arterial line. Postoperative blood draws could not extend past 18 hours because the likelihood of receiving exogenous blood products significantly increased, theoretically affecting the results. Titers of antibodies to diphtheria, tetanus, pertussis, Haemophilus influenzae type b (Hib), poliovirus, and hepatitis B were assayed according to our standard central laboratory protocol and reported out as concentrations or titer ratios on the basis of the assay. Thresholds for being considered immunized were reported from each vaccine’s information data sheet, when available. A history of previous natural infection was documented, if present.

Demographic and operative variables were analyzed to assess effect on changes on antibody titers. This included, but was not limited to, baseline patient information, vaccination history, time on CPB, type of surgery, volume of products used to prime the bypass circuit, presence and amount of modified ultrafiltration at the conclusion of the case, and requirements for further blood products within the first 24 hours. All red blood cells used to prime the CPB circuit were washed, leukocyte reduced, and irradiated, as an institutional standard.

Parents or legal guardians had the option to consent to a third blood draw, 3 months later. This blood sample, drawn from a venipuncture, was performed in the clinical setting during a routine postoperative evaluation. The same antibody titers were assessed. Vaccination history was updated if immunizations were provided between the operation and the third blood sample.

Acute effect of CPB on antibody titers was analyzed by using appropriate paired hypothesis testing, looking at central tendencies of each antibody titer to determine statistically significant changes when comparing pre- and post-CPB values. Logarithmic transformation was performed, where appropriate, because of large SDs of some of the antibody titers. Polio titers (including polio 1 and polio 3 antibodies) are reported as dilutional ratios, and thus a test of symmetry was used to determine differences between pre- and post-CPB values rather than discrete paired hypothesis testing. Threshold lines were plotted along with the data to demonstrate where individuals are thought to lose protective immunity.

Multivariate linear regression models were then used to ascertain the effects of the case-specific risk factors as described above on the predictability of the results. Confounders that were controlled in the final analysis or analyzed separately included timing and amount of previous immunizations, presence of native infection, chromosomal anomalies, heterotaxy or known immunodeficiency, ultrafiltration, and time on CPB. Two-sided P values of <.05 were considered statistically significant for the descriptive data as well as being the threshold in the univariate methods to determine if the factor was appropriate for multivariate regression analysis. Polio titers, as a result of being reported out as dilutional ratios, were excluded from regression analysis.

The study was approved by and operated under the guidelines of the Institutional Review Board of the University of Virginia. Funding was provided by Merck Sharp and Dohme Corporation, Merck Investigator Studies Program (No. 53360), although they had no influence on the performance of the study.

Of the 100 subjects who consented to participate, 98 had adequate blood samples to undergo antibody titer analysis. The remaining 2 subjects had early clotting of their samples, preventing any analysis. Baseline demographics for the population are depicted in Table 1. There were no identified patients with 22q deletion syndrome and 3 patients with heterotaxy syndrome.

The median vaccine titer values before and after CPB are displayed in Fig 1 along with threshold lines for immunity. There was no difference between the median pre- and postoperative values in regard to diphtheria (0.23 vs 0.24, P = .16), tetanus (0.49 vs 0.55, P = .66), or Hib titers (log −0.74 vs −0.55, P = .09). Polio titers, as well, did not show any difference between pre- and post-CPB levels (polio 1: P = .80; polio 3: P = .23). There was a significant reduction after CPB for Bordetella (1.04 vs 0.83, P < .001) and hepatitis B (log 2.16 vs 2.06, P < .001) titers.

There were no patients after surgery with CPB whose titers fell below the immunized threshold with respect to diphtheria, tetanus, polio, or Bordetella vaccinations.

There were 39 subjects (40%) who were nonreactive to Hib during their prebypass assessment. There were 3 subjects whose titers dropped from immunized status (≥0.15 µg/mL) to unimmunized status (<0.15 µg/mL) after CPB. Likewise, there were 17 individuals (17%) who were considered nonreactive to hepatitis B during their prebypass assessment. Again, only 4 individuals went from immunized (≥10 mIU/mL) to unimmunized status (<10 mIU/mL). There was no overlap between subjects who lost Hib immunity and those who lost hepatitis B immunity.

Multivariate linear regression controlling for suspected confounders demonstrated minimal effect of any variable on the change in pre- and post-CPB titers (Table 2). Increasing time on bypass did have an effect on the tetanus titers (P = .03) and had close to a significant effect on the diphtheria (P = .06) and Bordetella (P = .08) titers. Furthermore, the number of total doses of the vaccine the subject received demonstrated a significant change for tetanus (P = .04) and a nearly significant effect on the change in titers for diphtheria (P = .08), in that the fewer doses one received, the more profound the effect. The volume of ultrafiltration throughout the case as well as blood products administered were not studied in the multivariate analysis because they were not significant in simple univariate analysis, suggesting no effect from those variables.

There were 41 subjects who received only the first dose of the Hib, polio, diphtheria, tetanus, and pertussis vaccines, and of these, 21 had only received one hepatitis B dose. Among subjects receiving one dose of a vaccination, the mean time between hepatitis B delivery and surgery was 8.3 weeks (range: 0.9–18.1 weeks) and was 6.5 weeks for all of the other pathogens (range: 0.9–15.0 weeks). There were 9 individuals who had their only round of immunizations <3 weeks before surgery, and 4 of those individuals had them <2 weeks before surgery.

Of subjects who had only received one dose of vaccine, only Hib demonstrated a difference between mean pre- and post-CPB (log −0.90 vs −0.65, P = .02) and is shown in Fig 2. Diphtheria (0.06 vs 0.07, P = .11), tetanus (0.30 vs 0.46, P = .27), Bordetella (log 0.60 vs 0.55, P = .25), and hepatitis B (log 1.78 vs 1.58, P = .07) were unchanged. Polio titers did not show any difference between pre- and post-CPB levels (polio 1: P = .96; polio 3: P = .22) among patients who had only received one dose of polio vaccine. There were no patients who had received their only round of immunizations <3 weeks before surgery who lost immunity.

The results of multivariate linear regression for subjects who had just received one dose of vaccine are seen in Table 3. Bypass time did have an effect on the observed change in diphtheria titers (P = .007). Otherwise, there were almost no variables, including time between vaccination and surgery, age, weight, and bypass time, that had a significant effect on the demonstrated change in antibody levels among patients who had just received one dose.

Follow-up vaccine analysis done 3 months after surgery was performed in 12 patients and is displayed in Fig 3. Nine of those patients had received additional immunizations between surgery and 3 months. There was no significant difference in antibody titers when comparing pre-CPB, post-CPB, and 3 months after surgery across nearly all measured pathogens. The exception to this was Bordetella, which did demonstrate a significant upward trend between all 3 time points (P < .001).

Currently, there is little guidance on whether adjustment to the usual vaccination schedules is necessary for children with congenital heart disease who will require repair with CPB during infancy. Cardiac surgery with CPB did not markedly impact antibody titers in children under the age of 14 months. Specifically, no individuals who were considered immunized on the basis of their vaccine titers converted to a nonimmunized status for >1 pathogen as a result of CPB, including all individuals who had received their first immunization within 3 weeks of cardiac surgery. The median antibody titers for all pathogens remained above the threshold line for immunity post-CPB. Furthermore, variables associated with background demographics or specifics of the operation had minimal effect on the change in antibody titers.

Although specific vaccine antibody titers have historically not been tested before this study, it is known that other markers of a competent immune system, including total lymphocyte counts, complement, and total serum immunoglobulins, do not change significantly during CPB.⁹  This is contrasted with a known significant cytokine-mediated inflammatory response that occurs in children while undergoing CPB that seemingly can stimulate the immune response.^3,4  Despite this, previous studies in which researchers have looked directly at changes in specific subtypes of immune cells after CPB give mixed conclusions as to whether CPB stimulates or suppresses humoral immunity.^10–12 

There is little previous work evaluating how CPB affects circulating protective antibodies in children. The only available data have examined the use of palivizumab, an exogenously administered humanized murine monoclonal anti-F glycoprotein antibody that has been shown to be protective against respiratory syncytial virus. In a population of children undergoing cardiac surgery with CPB, a 58% reduction in palivizumab titers has been demonstrated when compared with preoperative levels.²  These data have led to the recommendation from the American Academy of Pediatrics that redosing of palivizumab should occur soon after bypass in patients undergoing treatment courses but does not recommend delaying administration of the antibody before an operation.¹³ 

Palivizumab is a form of immunoprophylaxis in which passive antibody is delivered. This differs from vaccinations that purposely stimulate a native immune response with the goal of developing memory and long-term immunity. Thus, it is interesting that patients who received only 1 round of immunizations, even at a close interval to surgery, did not demonstrate an appreciable change. Vaccine antigen, on administration, binds with naive B cells in the lymph nodes or spleen, beginning an extrafollicular reaction that produces a large amount of antibody from plasma cells nearly immediately after antigen exposure.^14,15  Although these antibodies are overall low affinity, this mechanism resulting in large production of antibodies could suggest why minimal change was demonstrated in antibody titers because of CPB even in the youngest infants.

A germinal center reaction follows the extrafollicular reaction and results in the production of memory B cells. This process occurs as early as 10 to 14 days after antigen exposure, peaking at 4 weeks and resulting in robust and long-lasting immunity soon after initial vaccine administration.^14,15  Consequently, even if CPB depletes circulating antibodies, as suggested by the palivizumab studies, enough plasma cell and memory B-cell response can exist early on to replete antibody levels and potentially preserve immunity.

There were individuals who started out below the immunized threshold for certain vaccines, specifically hepatitis B and Hib. Those individuals remained nonimmunized after CPB. This percentage of patients who were nonresponders falls near the reported rates of individuals who do not mount a response to those vaccines^16,17  but makes individual interpretation of the CPB effect on those 2 vaccines somewhat challenging. This is coupled with the fact that the only patients who lost immunity as a result of CPB were with those vaccines specifically.

Survey data have suggested that upwards of 70% of most major cardiac surgical programs across the country do not have formal policies surrounding vaccination immediately before or soon after CPB. Owing to the lack of data or previous work on this topic, programs that do have formal policies range anywhere from 1 to 6 weeks pre-CPB and 1 to 6 weeks post-CPB in limiting vaccine administration, suggesting significant variability surrounding a population that is known to be at risk for severe infectious disease.^7,18  In our study, we suggest that alteration of the standard vaccination schedule for children requiring CPB in infancy is unnecessary. More studies in this area will be helpful in confirming this conclusion and are essential in protecting this vulnerable group.

There are several limitations to this study that affect the broader applicability. First, antibodies and antibody titers are just 1 marker of immunity and were chosen because of their ease of measurement and lack of previous data surrounding them in pediatric heart surgery. Still, reporting antibody thresholds as a function of complete immunity, although standardized, is a simplified binary result, whereas actual immune responses are continuous with respect to antibodies and more complex with respect to other aspects of the immune system. Second, immunodeficiency, especially surrounding children with heterotaxy and 22q deletion syndromes, is prevalent in this population. Only 3 patients with heterotaxy and none with 22q deletion consented to the study, making absolute interpretation of the results among those patients challenging. Targeted studies, looking specifically at this population, could shed light on the immune status after CPB in these patients. Finally, only 12 individuals consented to the third blood draw, making complete interpretation of the long-term effect on immunity challenging to complete.

Infant vaccine antibody titers are minimally altered by CPB, whether the initial or multiple doses of the immunizations had been administered. Titer declines were not associated with any demographic factors or modifiable surgical variables. This suggests that children expected to undergo congenital heart surgery with CPB should be vaccinated using the recommended vaccination schedule.

CPB

cardiopulmonary bypass

Hib

Haemophilus influenzae type b