Aplastic anemia (AA) is a rare and heterogeneous disease, characterized by bone marrow failure leading to pancytopenia and empty marrow. Immune-mediated damage of hematopoietic stem cells is the major modality leading to bone marrow aplasia in acquired AA. Our previous cross-sectional study found that the life span of RBCs in severe (SAA) and very severe (VSAA) patients was shortened, but it was close to normal in patients exhibiting a hematologic response (HR) after immunosuppressive therapy (IST).¹ We report here a prospective trial by carbon monoxide (CO) breath test to quantify RBC lifespan in VSAA/SAA, and its changes according to response to treatment.

A total of 22 patients were included in this study: 14 SAA and 8 VSAA (14 males, eight females) with median age 25 (6–46), between November 2016 and November 2017. Pretreatment blood cell counts of the 22 patients were as follows: neutrophils 0.27 × 10⁹/L (0.04–0.9 × 10⁹/L), Hb 71 g/L (42–118 g/L), absolute reticulocytes 12.9 × 10⁹/L (2.5–26.5 × 10⁹/L), and platelets 10 × 10⁹/L (0–16 × 10⁹/L). All the patients had no evidence of hemolysis with indirect bilirubin (IBIL), haptoglobin (HP), lactate dehydrogenase (LDH) in the normal range. Erythrocyte osmotic fragility (EOF) test and Coombs test were also negative. As a control, 13 healthy people (three males, 10 females) were included with a median age of 36 (28–47).

The CO breath tests were obtained with no blood transfusion or 2 weeks after transfusion, no smoking within 24 h, on an empty stomach. An automatic instrument (ELS TESTER, Seekya Biotec Co. Ltd, Shenzhen, China) was used.

The IST treatment and the response criteria were in on-line supplementary material.

Note, RBCs lifespan in untreated patients was significantly shorter than in normal controls (SAA: 49 [25–85] vs. 113 [68–166], P < .001; VSAA: 51 [36–140], p = 0.003). The median baseline value was 51 (24–140). After IST, the lifespan rose overall to 92 days (47–137) (Figure 1A), which could be accounted for almost entirely by the change in patients with HR (96 days 50–137, 17 patients). In no response (NR) patients (five patients), the lifespan of RBCs was unchanged at 51 days (47–98), In complete response (CR) (five patients), good partial response (GPR) (four patients), and partial response (PR) (eight patients), the lifespan of RBCs after IST was no difference than in the normal control (96 days [67–137], p = 0.475; 94 days [79–107], p = 0.267; 103 days [50–126], p = 0.217; respectively). Overall, the patients with HR had a significantly longer RBC lifespan than pre-treatment and NR patients, similar to the normal control. (Figure 1C). The lifespan was prolonged by 52 (23–77) days in CR (p = 0.056), 38 (20–58) days in GPR (p = 0.190), 38 (−7 to 92) days in PR (p = 0.170) and 41 (−7 to 92) days in the overall HR group (p = 0.048), much longer than NR (0 [−89 to 61] days).

We analyzed the impacts of possible factors affecting the RBC lifespan before IST in patients by univariate logistical regression, and found no factor influencing the pre-treatment RBC lifespan, including gender, age, diagnostic classification, interval between diagnosis and treatment, PNH clone, hemoglobin level, absolute neutrophil count, platelet count, lymphocyte count, or absolute reticulocyte count (p > 0.05). No hemolysis signs were found 6 months after IST. We also analyzed the impacts of possible factors affecting the prolonged lifespan after IST. Patients who were male (14 patients) and who had SAA (14 patients) had more prolonged RBC lifespan than female (eight patients) (43 [−7 to 92] vs. 19 [−89 to 79], p = 0.046) and VSAA (eight patients) (42 [0–92] vs. 7 [−89 to 77], p = 0.025).

Linear regression was used to assess the correlation between the lifespan of RBCs and cytokines, including the serum levels of interferon (IFN), tumor necrosis factor (TNF)-α, interleukin (IL)-2, IL-6, IL-8, IL-10, and IL-17. There were no significant correlations between these factors with RBC lifespan before IST or with the change of RBC lifespan after IST (p > 0.05).

Also using linear regression, we found that both the percentage and the absolute count of CD3+CD8+T lymphocytes correlated with the RBC lifespan before IST (p = 0.009, R² = 0.299, and p = 0.010, R² = 0.292, respectively), while other parameters, including percentage or absolute count of T lymphocytes, CD3 + T lymphocyte, or CD3+CD4+T lymphocytes, did not correlate with RBC lifespan. The changes in these factors after IST also showed no significant correlation (p > 0.05).

By multiple linear regression, only the CD3+CD8+T lymphocyte percentage correlated with RBC lifespan before IST (p = 0.008, R² = 0.299). Only HR (p = 0.021; OR = 30; 95% CI, 2.137–421.117) was an independent prognostic factor on the change in RBC lifespan.

We have shown that Levitt's CO breath test is not inferior to the ¹⁵N glycine labeling technique to measure RBC lifespan.² Levitt's formula uses the constant value of 0.7 as an estimated proportion of endogenous CO originated from erythrocytes. Because endogenous CO is produced not only from hemoglobin but also from other heme proteins, the formula may be limited for use in some circumstances such as severe anemia. However, Vreman et al. showed the proportion of endogenous CO originating from RBCs is 85% × 86% = 73.1% (85% of heme come from the RBC destruction and the degradation of heme is the predominant source (> 86%) of endogenous CO.³ Another published paper showed that 10% to 20% of bilirubin does not come from senescent erythrocytes.⁴ Red blood cell lifespan measured by CO breath test showed excellent agreement even in cases of severe anemia and hemolysis with RBC lifespan measured by labeling techniques.⁵

A prior study estimated the lifespan of RBCs in 16 healthy subjects and seven patients with aplastic anemia measured using the ⁵¹Cr method.⁶ 4 (57%) of the seven patients had a shortened RBC lifespan (14.8–18.5 days vs. 21–31 days in the normal value of RBC T_1/2).

Our previous study found that the serum levels of IL-2R and IL-6 were much lower in HR patients than NR patients (IL-2R: 4.3 × 10⁵ U/L vs. 6.5 × 10⁵ U/L, p = 0.006; IL-6: 2.6 [2.0–17.7] ng/L vs. 6.1 [2.0–14.4] ng/L, p = 0.003).¹

In the present study, the CD3+CD8+T lymphocyte percentage correlated with RBC lifespan before IST (p = 0.008, R² = 0.299) after multivariate analysis. CD8+cytotoxic T cells with restricted T-cell receptor (TCR) diversity (oligoclonal T cells) are expanded in AA and secrete proinflammatory cytokines such as IFN-γ and TNF-α, which induce apoptosis of CD34+ cells in part through the Fas-dependent pathway. Systemic exposure to physiologically relevant levels of TNF-α is sufficient to cause acute cytopenias and hemophagocytosis. These could be the mechanism of the shortened lifespan in AA, but we did not observe in the present study a correlation between cytokines and RBC lifespan, even IFN-γ and TNF-α. Because of the small sample size, further studies with a large sample size should be carried out in the future.

In summary, we found the RBCs lifespan in untreated V/SAA patients is significantly shorter than normal and improves with response to therapy. Shortened lifespan of RBCs should therefore be considered in the pathogenesis of anemia in AA.