CUMULATIVE EFFECT of ABNORMAL BIOMARKERS on the RISK of MORTALITY and THROMBOEMBOLIC EVENTS in PATIENTS with COLD AGGLUTININ DISEASE.
No Thumbnail Available
All Authors
Hill, Q.A.
Roth, A.
Barcellini, W.
Broome, C.M.
Bozzi, S.
Carita, P.
Msihid, J.
Karaouni, A.
Rubio, J.
Zebachi, S.
LTHT Author
Hill, Quentin
LTHT Department
Oncology
Haematology
Haematology
Non Medic
Publication Date
2024
Item Type
Conference Abstract
Language
Subject
AGED , ANAEMIA, HAEMOLYTIC, AUTOIMMUNE , BENCHMARKING , COMORBIDITY , COHORT STUDIES , CONTROLLED CLINICAL TRIALS AS TOPIC , DIAGNOSIS , DRUG THERAPY , HEALTH AND CARE RECORDS , WOMEN , FOLLOW-UP-STUDIES , MEN , MORTALITY , SMOKING , RETROSPECTIVE STUDIES , THROMBOEMBOLISM , BILIRUBIN , HAEMOGLOBINS
Subject Headings
Abstract
Background: Published evidence suggests that patients with cold agglutinin disease (CAD) are twice as likely to die or experience thromboembolic events (TE) than matched non-CAD referents. An increased risk of mortality and TE with abnormal biomarkers has been observed in patients with CAD. However, there is a paucity of data on the cumulative effect of abnormal biomarkers on mortality and TE in the CAD population. Aim(s): This study assessed the cumulative effect of abnormal values of hemoglobin (Hb), bilirubin, and lactate dehydrogenase (LDH) on mortality and TE risks in patients with CAD. Method(s): A retrospective cohort study was conducted using the US Optum's de-identified Market Clarity electronic health records (EHRs) and claims data (2007-2022). Patients entered the CAD cohort on the date of first encounter with a diagnosis of autoimmune hemolytic anemia or CAD (ICD-10 diagnosis code) recorded in the claims or CAD terms recorded in EHRs. Two metrics (M1 and M2) were built to estimate the cumulative extent of abnormal biomarkers (defined as <10g/dL for Hb and >ULN for bilirubin and LDH): M1 represents the frequency of abnormal values over duration of follow-up and M2 represents the proportion of abnormal values across available laboratory measures (Figure). The potential impact of these metrics on the risk of mortality and first TE (hazard ratio [HR], 95% confidence interval [CI]) was evaluated using three types of models: Cox proportional hazards models (CM), Cox models with time-dependent covariates (TDCM), and shared random-effects joint models (JM) using age on index, sex, history of TE, smoking status, index season, and Charlson comorbidity index score as covariates. Result(s): The study included 876 patients with CAD. Using M2, increase in cumulative abnormal values of Hb enhanced the risk of mortality (HR [95% CI]) across all models: CM (3.21 [2.00-5.14]), TDCM (6.87 [3.99-11.85]), and JM (6.05 [3.61-10.16]). Similarly, the risk of first TE increased with CM (1.46 [1.03-2.08]), TDCM (2.41 [1.57-3.69]) and JM (2.37 [1.57-3.57]). Using M1, mortality risk (HR [95% CI]) increased with CM (1.02 [1.01-1.04]) and TDCM (2.89 [1.77-4.73]), and numerically increased with JM. Further, the risk of first TE increased with JM (2.09 [1.02-4.30]) and numerically increased with CM and TDCM for more abnormal values of Hb. Using M1 and M2, more abnormal values of bilirubin increased the risk of mortality (HR [95% CI]) with TDCM (1.93 [1.13-3.29] and 1.46 [1.02-2.08], respectively) and numerically increased the risk of first TE by using M2 with TDCM and JM. Using M2, more abnormal values of LDH increased mortality risk (HR [95% CI]) with JM (1.50 [1.01-2.22]) and numerically increased with TDCM; besides the risk of first TE increased with JM (1.49 [1.03-2.15]). Importantly, the percentage of patients among the CAD cohort (N = 876) with no laboratory data during follow-up was high (bilirubin: 19% and 31%; LDH: 50% and 59.5% for mortality and first TE analyses, respectively), which could explain the large CIs and the convergence issues (Figure). (Figure present) Conclusion(s): Increase in cumulative abnormal values of Hb enhanced the risk of mortality and TE in patients with CAD. More abnormal values of bilirubin and LDH also increased the risk of mortality but with less consistency across methods. Similar but less marked effects were observed on the risk of first TE for bilirubin and LDH. These results highlight the potential for early and chronic control of complement activation and the resulting hemolysis in CAD, to manage the risk of mortality and first TE.
Journal
HemaSphere