Suspected heparin-induced thrombocytopaenia in pulmonary retrospective cohort

Authors: Guillaume Guimbretière, Clément Dubost, Justin Issard, Virginie Louvain-Quintard, Xavier Jais, Samuel Dolidon, François Stephan, Daniela-Iolanda Ion, Olaf Mercier, Elie Fadel

Abstract

OBJECTIVES

Heparin is given for anticoagulation during and after pulmonary thromboendarterectomy for chronic thromboembolic pulmonary hypertension. Our objective was to add to the limited data available on the incidence, management and outcomes of suspected heparin-induced thrombocytopaenia after pulmonary thromboendarterectomy for chronic thromboembolic pulmonary hypertension.

METHODS

This retrospective single-centre study included consecutive patients with suspected heparin-induced thrombocytopaenia after pulmonary thromboendarterectomy done in 2005–2018. Confirmed heparin-induced thrombocytopaenia was defined as positive findings from both the antibody test and the platelet aggregation test. Patients with versus without confirmed heparin-induced thrombocytopaenia were compared, as well as patients with versus without heparin replacement therapy within the group with unconfirmed heparin-induced thrombocytopaenia. The platelet counts over time were compared to those in controls without suspected heparin-induced thrombocytopaenia.

RESULTS

Heparin-induced thrombocytopaenia was suspected in 86 (6.3%) of 1360 patients and confirmed in 16 (16/86, 19%), all of whom received heparin replacement therapy and survived to intensive care unit discharge. Of the remaining 70 patients, 28 (40%) received heparin replacement therapy and less often experienced bleeding compared to the other 42 (3.6% vs 21.4%, *P = *0.043). Intensive care unit mortality was 17/70 (24.3%) and was lower in the subgroup given heparin replacement therapy (10.7% vs 33.3%; *P = *0.046). Confirmed heparin-induced thrombocytopaenia was associated with a sharp platelet-count drop on Day 5. In unconfirmed suspected heparin-induced thrombocytopaenia, the early platelet-count decline was similar to that in the controls without suspected heparin-induced thrombocytopaenia, but the baseline count was lower.

CONCLUSIONS

Clinical features suggesting heparin-induced thrombocytopaenia after pulmonary thromboendarterectomy for chronic thromboembolic pulmonary hypertension were associated with excess mortality. Relay heparin replacement therapy was associated with lower mortality and fewer bleeding events.

INTRODUCTION

Pulmonary thromboendarterectomy (PTE) is the reference standard for treating chronic thromboembolic pulmonary hypertension (CTEPH) [1, 2]. PTE is technically challenging and requires both cardiopulmonary bypass (CPB) and deep hypothermic circulatory arrest. Unfractionated heparin is widely used for perioperative anticoagulation as a short-acting drug whose effects are easy to monitor and to reverse [1]. However, heparin can induce thrombocytopaenia, which is associated with thrombosis and, less often, bleeding, thereby potentially causing PTE to fail. Heparin-induced thrombocytopaenia (HIT) is suspected when the platelet count drops between 5 and 10 days after heparin initiation and/or when thrombosis occurs.

Thrombocytopaenia is common in the surgical intensive care unit (ICU) and can be due to many factors [3]. Moreover, no clinical or standard laboratory test signs are specific of HIT. Nonetheless, considering American Society of Hematology (ASH) and European recommendations, heparin should be stopped as soon as HIT is suspected with intermediate-probability 4Ts, before the results of confirmatory laboratory tests are available [4, 5]. However, no consensus exists on the best management of HIT specifically after PTE, which carries a risk of pulmonary artery thrombosis, recurrent venous thromboembolism and bleeding. On the other hand, stopping and replacing heparin during the perioperative period based only on a suspicion of HIT may cause patient harm, as the alternative anticoagulants have no reversal agents and may therefore carry a higher risk of bleeding and prolonged hospitalization. Moreover, the alternatives are costlier than heparin. Finally, heparin is superior to the alternatives regarding anticoagulant properties and is therefore routinely re-started after suspected HIT is ruled out [5].

The aim of this study was to describe the incidence, clinical presentation, management and outcomes in patients with suspected or confirmed HIT after PTE.

METHODS

Study design and patients

This single-centre retrospective study was approved by the French Society of Thoracic and Cardiovascular Surgery Ethics Committee (IRB00012919;#CERC-SFCTCV-2022–09-13_23181), which waived the need for informed consent in compliance with French law on retrospective analyses of de-identified health data. We searched our hospital database to identify consecutive patients managed by PTE for CTEPH between January 2005 and December 2018. Among them, we identified those in whom HIT was suspected after PTE. Suspicion of HIT is based on the presence of at least one of the following thrombocytopaenia <50 000 platelets/mm^3^, decrease platelet count >50% in the first week and appearance of a new thrombosis. When HIT is suspected, heparin-platelet factor 4 (PF4) are routinely performed. All patients who underwent PTE during this period, and for whom a PF4 assay was performed, were selected.

The patients who underwent PTE for CTEPH between 2009 (the start date of the systematic biological data collection in our institution) and 2018, but who did not have suspected HIT, served as controls for assessing platelet count changes, provided that daily measurements were taken for 15 days.

Laboratory testing for heparin-induced thrombocytopaenia

Citrated (0.129 M) plasma samples were obtained from all patients with suspected HIT and were isolated by double centrifugation of blood samples at 2300 g for 15 min at 20°C and stored at −80°C.

HIT antibody (PF4) testing was performed using a commercial enzyme-linked immunosorbent assay (ELISA) with the Asserachrom HPIA kit (Diagnostica Stago, Asnières-sur-Seine, France) that detects the IgG, IgM and IgA isotypes. The result was available within 6 h and was considered positive if absorbance at 450 nm expressed as optical density was ≥0.5, in concordance with the manufacture’s established ranges. The polyspecific ELISA for heparin PF4 antibodies has been reported to have 100% sensitivity [6].

Given the substantial false-positive rate seen with the polyspecific ELISA [6, 7], two functional assays detecting platelet activation by heparin-PF4 antibodies in the presence of heparin are the serotonin-release assay performed by the laboratory of haematology of Tenon Hospital, considered the reference standard for diagnosing HIT, and the platelet aggregation test (PAT) [8] performed by our laboratory of haemostasis. These results were obtained generally between 1 and 2 weeks (supplementary material).

Patients were classified as having HIT if the highly specific serotonin-release assay was positive or the ELISA and PAT were both positive.

Definitions

The 4T score, based on clinical and biological criteria, has been used to classify patients into low, intermediate or high probability of HIT [9]. HIT was defined as immune-mediated HIT. Disseminated intravascular coagulation was the concomitant presence of prothrombin activity <50%, factor V <50% and circulating fibrin degradation products (D-dimers). Bleeding was defined as clinical evidence of bleeding reported by the attending physician or nurse and requiring treatment (blood transfusion, modification of the anticoagulation regimen, surgery). Thrombotic events were defined as evidence of thrombosis on ultrasonography or computed tomography.

Management of patients with suspected heparin-induced thrombocytopaenia

When HIT is suspected, if the 4T score is ≥6 (high probability), heparin treatment is stopped and heparin replacement therapy (HRT) (parenteral danaparoid, argatroban or bivalirudin according to the preference of the senior intensivist) is started without waiting for the results of biological tests. For patients with a 4T score <6, heparin is relayed by HRT if the patient presents new pulmonary embolism, sever thrombocytopaenia <35 10^9^/l or if no alternative diagnosis for thrombocytopaenia is identified by the senior intensivist in charge. Patients for whom HRT has been introduced, heparin has not been reintroduced, and an early switch to anti-vitamin K or new oral anticoagulants has been made.

All patients with suspected HIT underwent ultrasonography of the lower and upper limb veins within 48 h to look for thrombosis.

Data collection

The study data were abstracted from the medical files and recorded on standardized forms. However, platelet counts were available only starting in 2009.

Statistical analysis

Continuous variables were described as mean ±SD if found by the Shapiro–Wilk test to be normally distributed and as median [interquartile range] (range) otherwise. Qualitative variables were described as n (%). Comparisons of quantitative variables were with Student’s t-test if distribution was normal and the Mann–Whitney test otherwise. For comparisons of categorical or binary variables, Pearson’s chi-square test was applied for patient characteristics and the Fisher exact test for outcomes with the Clopper–Pearson method to estimate 95% confidence intervals. No adjustment nor P-value correction has been made; inferences drawn from P-values may not be reproducible.

Values of P smaller than 0.05 were considered significant. The statistical analyses were conducted using R version 4.1.2 (https://www.r-project.org), and the graphs were constructed with GraphPad Prism version 8.3.1 (332) (Dotmatics, Boston, MA).

RESULTS

Patients and diagnosis and management of heparin-induced thrombocytopaenia

Of the 1360 patients who underwent PTE for CTEPH in 2005–2018, PF4 assay was performed in 86 (6.3%) during the postoperative ICU stay (Fig. 1). Tables 1 and 2 report their main features. Of the 86 patients, 16 (18.6%) had the diagnosis of HIT confirmed by positive results of both the ELISA and PAT.

Median time from surgery to HIT suspicion was 4 days (range 0–43). Of the 86 patients with suspected HIT, 21 (24.4%) had thrombosis, consisting of deep vein thrombosis (n = 8), pulmonary embolism (n = 8), distal arterial thrombosis (n = 5), continuous hemofiltration membrane thrombosis (n = 1) or thrombosis of tricuspid valve (n = 1).

Heparin was replaced by danaparoid in all 16 (100%) patients with confirmed HIT and in 28 (40%) patients with suspected but unconfirmed HIT. The reasons for switching to HRT are reported in Table 3. Danaparoid replacement was performed in 16/42 (38.1%), 18/34 (53%) and 10/10 (100%) patients with 4T scores <4, 4 or 5, and ≥6, respectively. HIT was confirmed in 4 of the 42 patients whose 4T score was below 4. One patient presented with a recurrence of pulmonary embolism without thrombocytopaenia. The other three patients had isolated thrombocytopaenia, with thrombocytopaenia in the immediate postoperative period, without a drop in platelets. Danaparoid was given for a mean of 8.9 ± 2.7 days in the patients with confirmed HIT and 5.6 ± 6.1 days in those with unconfirmed HIT (*P = *0.048).

Outcomes

Tables 4 and 5 reports the main outcomes. The only significant difference was for in-ICU mortality, which was significantly higher in the group with suspected but unconfirmed HIT than in the group with confirmed HIT (24.3% vs 0%, *P = *0.034).

In the analysis confined to the 70 patients with suspected but unconfirmed HIT, both bleeding and in-ICU death were significantly less common with versus without HRT (3.6% vs 21.4%, *P = *0.043 and 10.7% vs 33.3%, *P = *0.046, respectively).

Platelet count changes

Platelet counts were compared among groups with confirmed HIT, unconfirmed suspected HIT and a control group of 407 patients (Fig. 2). Baseline platelet counts were at the lower end of the normal range in the control and confirmed HIT groups, while the suspected but unconfirmed HIT group had slightly lower counts. In the control group, platelet counts declined moderately in the first 3 days post-surgery before gradually exceeding baseline levels. The confirmed HIT group experienced a similar initial decline but had a sharp drop to below 100 10^9^/l by Day 5. The suspected but unconfirmed HIT group also had an initial decline from a lower starting point, leading to significant thrombocytopaenia, with counts returning to the normal range around Day 10. Table 6 shows that, compared to the control group, the confirmed HIT group had significantly lower platelet counts on the day of surgery and Days 6–13, with no significant differences on Days 1–5.

DISCUSSION

The main findings from our study are that all 16 patients in whom suspected HIT was confirmed survived, whereas, in contrast, nearly a quarter of the patients with suspected but unconfirmed HIT died. Importantly, within this last group, HRT was associated with significantly lower mortality.

Most epidemiological data on HIT for bypass surgery are reported in the context of cardiac surgery, with rates close to 3% [10–12]. In our study confined to PTE, only 6.3% and 1.2% of patients had suspected and confirmed HIT, respectively. One possible explanation is that HIT is more common in females than in males [13]. Females made up 79% of our cohort.

CPB for surgery is often followed by thrombocytopaenia and thrombosis, even in the absence of HIT [14], making the diagnosis of HIT challenging. The 4T score is used to assess the risk of HIT, thereby guiding decisions about whether to stop heparin. The goal is high sensitivity to ensure that heparin is stopped in all patients with HIT. In our study, four patients with confirmed HIT had scores below 4, despite evidence of a negative predictive value >99% in this situation [9]. Furthermore, for scores of 4–5 and of 6 or more, the positive predictive values were only 14% and 64%, respectively [9]. The 2018 American Society of Haematology guidelines recommend heparin discontinuation and replacement in all patients with 4T scores of 4 or more [5]. In our cohort, only 28 of the 44 patients with scores ≥4 had HRT. Moreover, in the group with 4T scores below 4, 16 (38.1%) received HRT, including 6 whose ELISA was negative. The most common reason for HRT in this group was new pulmonary embolism, sever thrombocytopaenia <35 10^9^/l and no alternative diagnostic for thrombocytopaenia. The presence of exposure to heparin in the last 3 months also prompts us to initiate treatment, without being a sufficient reason for doing so.

Laboratory tests play a crucial role in diagnosing HIT. The ELISA for heparin-PF4 antibodies is 100% sensitive but has low specificity. A platelet function test must therefore be obtained if the ELISA is positive [15, 16]. In one study, a positive ELISA for heparin-PF4 antibodies combined with a positive PAT had 100% sensitivity and over 95% specificity [17]. In contrast, in a study that used the heparin-induced platelet activation test as the reference standard, this test was positive in 39 of 122 ELISA-positive serum samples, and the PAT was positive in only 27 of these 39 samples [18]. In another study, the PAT was less sensitive than the ELISA and heparin-induced platelet activation test [19]. These discrepancies may be related to differences in the protocol applied for performing the PAT. The possible limited sensitivity of the PAT is of concern.

Thrombosis is the main complication of HIT, whereas bleeding is less common. In our cohort, both events were more common in the group with confirmed HIT than with unconfirmed suspected HIT, but the differences were not statistically significant. This finding is probably ascribable to the small sample size. In an earlier study, thrombosis occurred in 50% of patients with versus 16% without confirmed HIT (*P = *0.0006) [11]. Our group with confirmed HIT less often had manifestations consistent with another diagnosis than did the group with suspected but unconfirmed HIT, although the difference was not statistically significant.

The platelet count decline seen early after CPB with deep hypothermic cardiac arrest is ascribable to a combination of mechanical platelet damage, intraoperative blood loss and platelet activation by the pulmonary artery media exposed by the endarterectomy [20]. This decline occurred over the first 3 days after surgery in our control group. The group with confirmed HIT exhibited the typical platelet count drop between Days 6 and 10 after surgery [10, 21]. HIT was probably suspected when the drop occurred, with the subsequent platelet count recovery being a response to heparin discontinuation and clearance from the body. In the group with suspected but unconfirmed HIT, the low platelet counts were due to the usual decline, as seen in the controls, occurring against a background of pre-existing thrombocytopaenia. Notably, there are two distinct kinetics of thrombocytopaenia based on heparin exposure. In our study, the HIT+ population exhibited kinetics associated with first-time exposure to heparin, resulting in delayed thrombocytopaenia. In contrast, the HIT− population likely demonstrated kinetics of prior heparin exposure; however, 25% of both HIT+ and HIT− patients reported exposure within the past 3 months. The biological diagnosis of HIT is complex and can be influenced by several factors, including CPB, deep hypothermia, time since surgery and platelet antibody consumption. It is possible that true HIT diagnoses were missed in the HIT− group, particularly among those with prior heparin exposure. Ultimately, we found that immediate postoperative thrombocytopaenia was associated with the worst prognosis.

In-ICU mortality was 17/70 (24.3%) in the group with unconfirmed HIT. None of the 16 patients with confirmed HIT died before ICU discharge. In a cohort of 93 patients diagnosed with HIT during the postoperative period of heart valve surgery, the in-hospital mortality rate was 1% [12]. One possible explanation to the mortality in the group with unconfirmed HIT is that the alternative diagnoses, causing the symptoms responsible for the HIT suspicion, may have been more severe than HIT. Moreover, no cause to these symptoms was identified in over a quarter of patients in this group and, consequently, no aetiological treatment could be given. Importantly, despite the absence of HIT confirmation, not only mortality but also bleeding was less common when heparin was stopped and replaced. This finding raises questions about the sensitivity of the PAT, as discussed above. However, in a large study of 894 patients with suspected HIT including 232 with confirmed HIT, combining the ELISA and the PAT done on platelet-rich plasma had a low false-negative rate of 0.7% [7]. Also, the absence in our platelet-count evaluation of the typical drop on Day 5 or 6 in the group with unconfirmed suspected HIT argues against a substantial proportion of these patients having undiagnosed HIT. The serotonin-release assay has better diagnostic performance than the ELISA and PAT but is not widely available due to its complexity and use of radioactive materials.

Limitations

The study has several limitations, including its retrospective design, which may have led to missing data, particularly regarding complications. The total number of included patients was small, as this is a rare condition. Consequently, the protocol for managing HIT remained variable depending on the urgency of the situation and the experience of the physicians involved. However, we aimed to report our practice as accurately as possible in the hope that it will help the medical community define a dedicated protocol for CTEPH patients. Additionally, the platelet activation test (PAT) used to confirm positive ELISAs is not 100% sensitive, meaning some cases of HIT may have been missed. Despite these limitations, this is the first report on HIT management in a cohort of pulmonary endarterectomy patients who are typically exposed to heparin prior to the procedure and also at risk for thromboembolic disease.

CONCLUSIONS

After PTE, HIT was associated with a good prognosis when diagnosed and treated by heparin discontinuation and replacement therapy. Patients with an unconfirmed suspicion of HIT, in contrast, had substantial rates of mortality. These rates were lower in the subset given HRT. The reasons for this finding remain unclear. However, until further data are available, routine heparin discontinuation and replacement as soon as HIT is suspected after PTE might improve patient outcomes.

Supplementary Material