Abstract
Introduction Kidney injury is a common finding in patients with myeloproliferative neoplasms (MPN) such as CMML. Its etiology is multifactorial. In most cases, leukemia induced kidney injury is related to tumor lysis resulting in increased uric acid, direct leukemia kidney infiltration, chemotherapy-induced toxicity, and preexisting medical comorbidities. In CMML, kidney involvement is observed in more than 50% of the cases. Monocytosis, a distinct feature of CMML, can affect kidney function in several ways including lysozyme induced nephropathy. Here we aim to better understand the effect of white blood cell count (WBC) and absolute monocyte count (AMC) on kidney function in patients diagnosed with CMML.
Methods A retrospective analysis performed on a cohort of CMML patients at Upstate Medical University Hospital between January 2014 to June 2025. The subjects were at least 18 years old and diagnosed with CMML. To examine the effect of WBC and AMC separately on creatinine (Cr), a linear mixed‐effects model was used while adjusting for uric acid (UA), hypertension, and diabetes. Cr, AMC, and WBC were natural‐log transformed to normalize their distributions. PROC MIXED (SAS v9.4) was used with restricted maximum likelihood (REML) estimation, specifying a random intercept for each subject to account for within‐individual correlation of repeated measures. A locally estimated scatterplot smoothing (LOESS) curve was used to visualize the relationship between both predictors and creatinine separately, and to identify non-linear trends or inflection points. Frequency of relevant mutations for overall population and median survival were reported.
Results A total of 56 patients were included. Mean age at diagnosis was 76.4 years old (SD 9.75), 38 patients (68%) were males, and the majority were white (88%). Median of included variables at diagnosis was as follows: Cr 1.09 mg/dL, AMC 3.51 K/uL, WBC 16.3 k/uL, UA 6.09 mg/dL, Hemoglobin 9.99 g/dL, platelets 89 K/uL, bone marrow blasts 2.4%. Occurrence of relevant mutations in the cohort were as follows: TET2 (31.9%), SRSF2 (25.8%), RAS (21%), ASXL1 (14%), RUNX1 (9%), JAK2 (8%), DNMT3A (8%), TP53 (2%), NPM1 (3%). 11% lost follow up by time of data analysis. Median survival duration was 7.2 months (IQR 23.4). After adjusting for covariates, one unit increase in the log-transformed WBC count was associated with a 13.1% increase in mean Cr and was statistically significant (95% CI [9.04%, 17.24%], p < 0.001), and a one unit increase in the log-transformed AMC was associated with a 9.69% increase in mean Cr and was statistically significant (95% CI [4.89%, 14.70%], p < 0.001). On the LOESS-smoothed curve of Cr plotted against WBC, the only statistically significant infliction point was Cr 2 mg/dL and WBC 21.1 K/uL; from the start of the dataset to WBC 21.1 K/uL, a one unit increase in the log-transformed WBC was associated with an 5.56% increase in mean Cr (95% CI [0.66%, 10.71%], p=0.026). While on the LOESS-smoothed curve of Cr plotted against AMC, the only statistically significant infliction point was Cr of 2 mg/dL and monocytes of 5.25 K/uL; from the start of the dataset to AMC 5.25, a one unit increase in the log-transformed AMC was associated with an 16.84% increase in mean Cr 95% CI [3.98%, 31.30%], p < .05).
Conclusion In our study, increase in Cr up to 2 mg/dL was more affected by increments of AMC than WBC, a one unit increase in (AMC up to 5.25 K/uL) was associated with higher effect than (WBC up to 21.1 K/uL) on mean Cr (16.84% vs 5.56%). No other significant infliction points were identified. While leukocytosis is currently an indication of treatment in CMML, our assumption is that monocytosis might be a better predictor for early signs of organ involvement in CMML, particularly nephropathy. Larger studies are needed to best test this theory.
