Longitudinal trends of peripheral blood counts in polycythaemia vera and essential thrombocythemia patients in the UK

Abstract There is sparse evidence of how well haematological targets are met in practice for essential thrombocythemia (ET) and polycythaemia vera (PV) patients. Patient data was collected between 2008 and 2020 from two UK NHS Trusts for ET and PV patients. Longitudinal changes in peripheral blood counts, including the proportion of patients meeting peripheral blood count remission, was modelled. Relative risk of cardiovascular‐related events for patients achieving remission within 3‐months was estimated. A total of 620 ET and 429 PV patients were analysed. For high‐risk patients, haematological parameters decreased in the first months of observation then stabilised within normal reference ranges until year 5. Total time spent in peripheral blood count remission was 39.2% for ET and 29.1% for PV. A lower proportion of ET patients reached target platelet counts (48.3%) compared to WBC (79.1%), whilst PV patients were less likely to reach target haematocrit levels (56.9%) compared to platelets (77.3%) or WBC (74.6%). There was no statistically significant association between reaching target blood counts within 3‐months and cardiovascular risk. Complete haematological remission remains a challenging target in managing PV and ET, however this study was unable to show statistically‐significant evidence that this was associated with increased risk of cardiovascular events.

monitoring. Clinically, ET is characterised by an excess platelet count, whereas PV is broadly defined by abnormally raised haemoglobin, haematocrit, and/or red cell mass [2,4,5].
Studies have highlighted the validity of using platelet count as a therapeutic target in ET and haematocrit in PV, primarily with the use of hydroxyurea treatment in lowering platelet counts, haematocrit and thrombotic risk [6][7][8][9]. Treatment is aimed at lowering risk from vascular complications, whilst minimising the risk of disease progression to myelofibrosis or AML [10]. For high-risk patients (age≥ 60 and/or with a history of thrombotic events), the use of cytoreductive therapeutics, anti-platelet drugs, and venesection are standard of care for PV, whilst anti-platelet drugs and cytoreductive treatments are used in patients with ET [3].
Whilst there is significant debate in the literature about what haematological values clinicians should aim for in routine care [10], it is broadly accepted that a target haematocrit < 45% significantly reduces the risk of thrombotic events in PV patients [11], and a target platelet count ≤400 × 10 9 /L as is associated with reduced risk of thrombotic and haemorrhagic events in ET [10]. Additionally, a leukocyte target of ≤10 × 10 9 /L is also recommended for both ET and PV. These thresholds are based on consensus expert opinion by the European LeukemiaNet (ELN) and International Working Group-Myeloproliferative Neoplasms Research and Treatment (IWG-MRT) [12].
Although studies have examined the significance of achieving haematological remission in both ET [13][14][15] and PV [16] on thrombotic risk outside the clinical trial setting, these studies have typically focused on sub-populations exposed to specific medications or have been restricted by shorter lengths of follow-up. This study therefore aims to examine the longitudinal progression of key haematological parameters in high-risk ET and PV patients treated in a real-world setting and examine the extent to which patients can achieve and sustain haematological target values over the course of their disease.

Data source
The data were collected in Chelsea & Westminster Hospital NHS Foundation Trust (CW) and Oxford University Hospitals NHS Foundation Trust (OUH). We obtained anonymised patient-level electronic health records (EHR) between years 2008-2020 for CW, and 2012-2020 for OUH. Both data sources were used to derive one population for the study.  (20 th December 2020). Given the high level of attrition beyond 5-years, longitudinal analysis was restricted to 5-years from cohort entry. Patients with less than 5-years follow-up were included in all analyses.  [17]. Patients were defined as high risk if they were ≥ 60 years at cohort entry and/or had a history of haemorrhagic or thrombotic events (supplemental data in Table 1 for a list of detailed definitions using ICD10 codes). Information on genetic testing relevant to MPNs (JAK2, MPL, and CALR) was not well recorded in the routine patient records and excluded from the main analysis.

Variables
Peripheral blood count remission was defined based on the ELN revised response criteria: haematocrit < 45% (PV only), platelets ≤ 400 × 10 9 / L, and WBC < 10 × 10 9 / L. Absence of leukoerythroblastosis − which is an additional criterion for remission in ET − could not be ascertained from our data and was excluded from the analysis. The occurrence of the following clinical outcomes was ascertained over the course of the study: all-cause mortality, thrombosis, haemorrhage, and cardiovascular disease (supplemental data in Table 1). Incidence of any of these conditions was defined as the earliest recording of a corresponding ICD-10 code as the primary reason for hospital admission.

Statistical analysis
The patient characteristics at cohort entry were summarised for all patients using the median and interquartile range (IQR; continuous variables) and counts and percentages (categorical variables). To account for potential data entry delays, a window of 30-days before or after their initial MPN diagnosis was chosen. If more than one value of a laboratory measurement were recorded within ±30 days, the average of those values was chosen.

RESULTS
A total of 620 ET and 429 PV patients were included across both data sources (
tively over the 5-years and were included in the longitudinal analysis (Table 1). Following patients over time, higher average blood count measurements were observed in both diseases close to cohort entry which tended to stabilise at a constant level after 3 to 12 months (Table 2 and Figures 1 and 2).

Essential thrombocythemia
Among high-risk ET patients, the most pronounced decline was

Polycythaemia vera
Initially higher levels of FBC measurements around cohort entry were also observed in high-risk PV patients (Table 2 and Figure 2 Figure 3D). This further increased to 71.5% for platelets and 69.0% for WBC. 3). Reaching peripheral blood count remission was generally associated with reduced risks of developing the outcome. The only exception to this was haemorrhage, where patients that were in remission for haematocrit and platelets had a (not statistically significant) higher risk. of < 400 × 10 9 /L at 12-months was just 26%, whereas those achieving a WBC count < 10 × 10 9 /L at 12-months was 89%. This corroborates our finding that meeting the platelet target in routine care is more challenging than meeting the target leukocyte count in ET.

DISCUSSION
Although these findings highlight the challenge in achieving peripheral blood count remission in ET and PV in routine care, the extent to which meeting these thresholds translates to improved outcomes for patients is less clear. In agreement with previous studies of both ET and PV, our results highlight how those patients achieving complete remission within first 3-months demonstrated no statistically significant reduction in the risk of thrombosis or bleeding [13][14][15][16]. It should be noted in one study however, PV patients who were able to reach the platelet target count demonstrated a statistically significant reduction of between 30% and 90% in the risk of thrombosis or bleeding [16]. No statistically significant reduction in risk of either thrombosis or bleeding was observed for those achieving haematocrit or leukocyte targets.
Likewise, whilst studies in ET did not show significant reductions in risk of thrombosis for those achieving platelet counts of < 400 × 10 9 /L, there was a significant reduction in thrombosis [13,14] and mortality [14] for those who met the leukocyte threshold of < 10 × 10 9 /L.

STRENGTHS AND LIMITATIONS
The strengths of this study namely include the provision and retrospective analysis of longitudinal data in a large, real-world UK patient cohort diagnosed with ET and PV, which is otherwise scarce in present literature. Our analysis utilises up to 20,000 observations from over 1,000 patients and explores the extent to which peripheral blood count remission targets are met over the course of the disease.
However, the study is not without its limitations. There are clinical and molecular similarities between PV and ET, which can lead to misrepresentation of disease pathophysiology and under-or overestimation of treatment effects [23]. The observational nature of the study did not make it possible to examine how specific treatments were associated with outcomes, however this study focuses only on those high-risk patients reflecting the natural treated progression, where treatment is assumed to be based on best clinical practices as the time.
The lack of recorded fluctuation in hematological markers after the first year of observation may also be attributed to the induction of apparent disease remission in these patients [2]. Coverage of JAK2 mutation status was limited and it's possible that those PV patients without a recorded JAK-2 mutation status could be negative. Furthermore, the use of retrospective data from EHR means it is difficult to account for patients that may be lost to follow-up, although the use of multilevel models would account for data attrition over time.

CONCLUSIONS
We have provided one of the first descriptions of real-world data to show differences in laboratory haematological progression over a five- year follow-up period in high-risk PV and ET patients. Additionally, we have provided a detailed overview of the extent to which peripheral blood count remission was achieved in those patients during a seemingly 'stable' period of the disease. This study highlights the challenges in achieving complete haematological remission. However, the analysis was unable to show any statistically significant increases in thrombotic or cardiovascular related risk for those patients not achieving remission within 3-months.

ACKNOWLEDGMENTS
This work uses data provided by patients and collected by the NHS as part of their care and support. We believe using patient data is vital to improve health and care for everyone and would, thus, like to thank all those involved for their contribution. The authors would like to thank Neil Copping for their invaluable contribution to this project. This work was supported by Bristol Myers Squibb. only process data in respect of projects that have identified patient benefit. As data were anonymised at source, ethical approval was not required.

AUTHOR CONTRIBUTIONS
All authors designed the study. S.D and N.L carried out the data acquisition. All authors analysed and interpreted the data.