Heparin-induced thrombocytopenia (HIT) is caused by antibodies targeting platelet factor 4 (PF4)/heparin complexes. The immune response leading to HIT remains perplexing with many paradoxes. Unlike other drug induced reactions, anti-PF4/heparin antibody generation does not follow the classic immunologic response. As Greinacher and colleagues have shown, the primary immune response lacks IgM precedence and class switching, and heparin-induced antibodies can induce HIT by day 5 in heparin-naïve patients.Continued exposure to heparin also is puzzling with a weak or declining secondary immune response. Research by Krauel and colleagues suggests that that there is close interplay among infection, PF4 and the immune system. In 2010 they demonstrated that human and murine PF4 bind to Gram positive (S.aureus, S.pneumoniae, L.monocytogenes) and Gram negative (E.coli, N.meningitidis) bacteria in vitro, with bacterial surfaces acting as polyanions. High dose heparin inhibited this binding and anti-PF4/heparin antibodies from patients with HIT reacted with these PF4/bacterial complexes (S. aureus and E. coli). Using a murine model, they went on to show that polymicrobial sepsis in the absence of heparin led to antibody generation. In a separate study, Krauel and colleagues also showed that PF4 binds specifically to the lipid A component of Gram negative bacteria.

In this analysis, we report on anti-PF4/heparin antibody levels in groups of patients hospitalized for sepsis, as compared to a control group without sepsis. We examined 200 patients with sepsis, retrospectively identified, from a hospital database of anti-PF4/heparin testing done in medical inpatients with thrombocytopenia but low pretest probability of HIT. This included patients with bacteremia (57), fungemia (7) and sepsis without septicemia (136). For comparison, data from 50 patients without sepsis during the same time period was used. Inclusion criteria for all groups were age 18 years and older and antibody testing within 4 days of admission. Exclusion criteria were diagnosis of HIT or heparin allergy, prior hospitalization or heparin exposure within 90 days of admission, cardiopulmonary bypass or orthopedic surgery within 6 months, hemodialysis, active or past malignancy, antiphospholipid syndrome, autoimmune disease or immunosuppressive therapy.

All patients studied were on subcutaneous heparin at prophylactic doses only (i.e. no intravenous use, no therapeutic anticoagulation). UFH use predominated with prevalence of >85% in all groups. Testing was done using a commercially available standardized solid phase enzyme-linked immunoassay (EIA) to detect antibodies (IgG/IgA/IgM) directed against PF4 complexed with polyvinylsulfonate (Genetic Testing Institute, Wisconsin). All assays were performed in the central hospital laboratory according to manufacturer's specifications and measured in optical density (OD) units. The data sets demonstrated continuous unimodal distribution with high OD outliers, indicative of varying immune responses along a continuum. Statistical significance was calculated using independent t-testing with p-value set at 0.05 for significance.

Results showed that patients hospitalized with sepsis have higher anti-PF4/heparin antibody levels. Both patients with bacteremia, and sepsis without bacteremia, had significantly higher OD than patients without sepsis (p<0.05). There was no significant difference between Gram negative and Gram positive bacteremia and antibody levels. This suggests that bacterial cell wall components of both classes have similar antigenicity. Interestingly, patients with fungemia had much lower antibody levels compared to bacteremia and sepsis. Despite the small sample size for fungemia, this difference trended strongly towards statistical significance (p=0.05). The threshold for a positive EIA is currently established at OD>0.400, a value based on sensitivity and set by the manufacturer. When the prevalence of a positive EIA was assessed, 16% patients with sepsis and bacteremia tested positive compared to 4% in the control group. In summary, there is an increased prevalence of anti-PF4/heparin antibodies in patients hospitalized with bacterial but not fungal sepsis. These results support the theory that bacterial infection has a role to play in preimmunization leading to anti-PF4/heparin antibody generation.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.