Pretransplant Platelet-to-Lymphocyte Ratio Predicts Outcome After Allogeneic Hematopoietic Stem Cell Transplantation

A total of 214 adult patients with hematological diseases who underwent their first allogeneic HSCT were included in this retrospective analysis. All patients were followed up to 1500 days after allogeneic HSCT. We first compared the 4 components of the NLR, MLR, and PLR blood count (absolute monocyte count (/nl), absolute neutrophil count (/nl), absolute lymphocyte count (/nl), and absolute thrombocyte count (/nl)) among healthy individuals. and our patients. Healthy individuals showed statistically significantly higher values ​​for all parameters: absolute lymphocyte count (mean 7.7/nl vs. 5.4/nl; p<0.0001, Fig. 1A), platelet count (mean 247/nl vs. 154/nl; p<0.0001, Fig. 1B), absolute neutrophil count (mean 1.8/nl vs. 1.1/nl; p<0.0001, Fig. 1C) and absolute monocyte count (mean 3.8/nl vs. 0.5/nl; <0.0001, Fig. 1D). NLR (median of patients 2.35 vs. healthy individuals 2.68, p= 0.61, Fig. 1E), MLR (median of patients 0.31 vs. healthy individuals 0.20, p= 0.0001, Fig. 1F) and PLR (median of patients 133 vs. healthy individuals 132, p= 0.82, Fig. 1G) were higher in patients compared to healthy controls, only statistical significance could be demonstrated for MLR.

Figure 1
Figure 1

Comparison of ALC, platelet, ANC and AMC count between healthy individuals (ctrl) and HSCT patients and comparison of NLR, MLR and PLR between healthy individuals (ctrl) and HSCT patients. HSCT patients showed lower ALC (A), Platelet (B.), ANC (C) and AMC (D) count compared to healthy individuals (ctrl). NLR (me), MLRs (F) and public lending right (GRAM) was higher in HSCT patients compared to healthy individuals (ctrl).

Based on your pre-HSCT Disease Risk Index (DRI)9 patients were distributed into 4 risk groups: low risk (n = 23 (10.7%)), intermediate risk (n = 99 (46.2%)), high risk (n = 65 (30.3%) ), very high risk (n = 21 (9.8%)), while 6 (2.8%) patients could not be assigned due to lack of data. There were no significant differences in DRI risk groups with respect to NLR or MLR. Only for PLR were significant differences observed (p= 0.0016): Patients in the intermediate DRI group showed a significantly higher PLR compared to patients in the high DRL group (p= 0.009) or very high DRI group (p= 0.011) (Fig. 2). Analyzing different types of conditioning regimen (MAC vs. sequential including patients with refractory AML vs. RIC) no significant differences were detected between NLR (p= 0.18), MLR (p= 0.63) or public lending right (p= 0.34). Regarding the different diseases, significant differences became evident for MLR (p= 0.044), but not for NLR (p= 0.054) or public lending right (p= 0.094) (Table 1).

Figure 2
Figure 2

Association between NLR, MLR, PLR and Disease Risk Index (DRI). Regarding DRI, no significant differences were observed for NLR (p= 0.12, A) and MLR (p= 0.46, B.). For PLR patients in the intermediate risk group, a significantly higher PLR was observed compared to patients in the high risk group (p= 0.009) and very high risk group (p= 0.011, C).

Taking into account age, sex, HLA compatibility, ECOG, pre-HSCT body mass index (BMI), CD34+ cells (× 106)/kg in graft, GVHD prophylaxis (Alemtuzumab vs. ATG vs. others), HCT-CI score, no statistically significant difference could be detected between high NLR, MLR, PLR groups and low NLR, MLR groups and PLR in univariate analysis.

We observed significant differences in our new defined groups when we investigated the clinical outcome parameter:

Patients in the high PLR group required fewer red blood cell transfusions (mean 6, median 8.3, range 0–100 vs. mean 13.1, median 10.0, range 0–50; p= 0.000), fewer platelet transfusions (mean 8.5; median 6.0; range 0-115 vs. mean 14.5; median 10.5; range 0-79; p= 0.000) and less fresh frozen plasma (mean 0.27; median 0; range 0–8 vs. mean 1.36; median 0; range 0–34; p= 0.000). Analyzes of infection and ICU treatment also showed statistically significant differences: patients in the high PLR group had fewer episodes of fever > 38.5°C (days: mean 2.61, median 2.0, range 0-15 vs. mean 3.8, median 2.0, range 0-26; p= 0.002) and received fewer different antibiotics (mean 1.6, median 1.0, range 0–6 vs. mean 2.1, median 2.0, range 0–6; p= 0.005). 42 patients were sent to the ICU, 14 (33.3%) of these were assigned to the high PLR group compared with 28 (66.7%) patients matched to the low PLR group (p= 0.017). 22 (10.3%) of 214 patients died during hospitalization, 6 (27.3%) in the high PLR group and 16 (72.7%) patients in the low PLR group (p= 0.024). The low PLR group showed significantly higher pre-HSCT ferritin levels (mean 2,294; median 1,477; range 140–10,663 vs. mean 1,337; median 754; range 12–10,000 ng/mL; p= 0.004). Significant differences in platelet transfusion were observed between the high NLR group and the low NLR group (mean: 9.8; median: 6; range: 0-79 vs. mean: 13.3; median: 9 ;range: 0-115; p= 0.03) (see Table 1).

NLR, MLR, and PLR did not result in differences with respect to length of hospitalization, granulocyte reconstitution time (>0.5/nl), development of acute graft-versus-host disease (aGVHD), cytomegalovirus reactivation, Epstein-Barr-Virus post-HSCT or C-reactive protein (CRP) pre-HSCT (Table 1).

130 (60.7%) patients died within the 1,500-day observation period after allogeneic HSCT, 63 (29.4%) had a relapse of the underlying disease, while 73 (34.1%) patients died within remission. Overall survival 1,500 days after allogeneic HSCT was 39.2% (84/214), and median survival was 543 days. The 1,500-day OS was significantly higher in the high PLR group (52/107 (48.5%) vs. 33/107 (30.8%), p= 0.001). Also OS at 1 year (75/107 (70.0%) vs. 51/107 (47.6%)); p= 0.0005) and OS at 2 years (62/107 (57.9%) vs. 36/107 (33.6%); p= 0.0001) were significantly higher in the high PLR group. No significant differences were observed between the NLR and MLR groups with respect to OS (Fig. 3A, B and C day 1500 OS post-HSCT).

figure 3
figure 3

Association between NLR, MLR, PLR, and overall survival (OS), both relapse and non-relapse mortality (NRM). (A), (B.) Y (C) show the association between OS (observation time 1500 days (= 4.1 years) after allogeneic HSCT) for NLR, MLR, and PLR, while patients with PLR > 133 showed significantly higher OS compared with patients with PLR < 133 (p= 0.001). (D), (me) Y (F) show the association between the 3 markers and relapse, while patients with PLR > 133 showed a significantly lower relapse compared to patients with PLR < 133 (p= 0.016). (GRAM), (H) Y (me) show the association between NLR, MLR and PLR and NRM, while again patients with PLR > 133 showed a significantly lower NRM compared to patients with PLR < 133 (p= 0.02).

We observed a significantly lower incidence of relapse in the high PLR group 1500 days after allogeneic HSCT (p= 0.016). No significant differences could be detected at any other time for NLR, MLR or PLR (Fig. 3D, E and F day 1500 after HSCT relapse).

NRM was significantly higher in the low PLR group after 1,500 days: 44/107 (41.1%) vs. 29/107 (27.1%); p= 0.022). No significant differences were observed using the NLR or MLR score (Fig. 3G, H, I day 1500 NRM).

Additionally, we look at causes of death by focusing on GVHD, relapse, sepsis, or other reasons. However, no significant differences were observed (see Table 1).

In a multivariate analysis focusing on OS, relapse, and NRM adjusted for NLR, MLR, PLR, DRI, conditioning regimen, GVHD prophylaxis, CD34 cells/kg, HCT-CI, and age, PLR was found to be an independent prognostic factor. for OS (hazard ratio [HR] 0.56, p= 0.009). Three other independent prognostic factors for OS were GVHD prophylaxis, HCT-CI, conditioning regimen, and age for NRM (Table 2).

Table 2 Multivariate analysis for OS, Relapse and NRM.

Leave a Comment