Various biomarkers have been investigated with regard to their ability to predict the outcome of allogeneic hematopoietic SCT (HSCT). In this study, we retrospectively reviewed 90 recipients who received HSCT between 2007 and 2011 in our institution, and evaluated the predictive value of the baseline serum C-reactive protein (CRP) levels just before the initiation of conditioning for transplant-related complications after allogeneic HSCT. A receiver-operating characteristic curve revealed that the baseline serum CRP levels had an excellent predictive value for non-relapse mortality (NRM), with an area under the curve of 0.83. The sensitivity and specificity for NRM were 80% and 87%, respectively, with a cutoff of 0.6 mg/dL. With this cutoff value, multivariate analyses revealed that a higher baseline CRP level was an independent risk factor for NRM (HR 6.21, P<0.01), grade III–IV acute GVHD (HR 3.91, P=0.03) and poor overall survival (HR 3.27, P=0.0018). On the other hand, the baseline CRP level did not predict infectious events. These findings suggested that CRP levels before conditioning may be a useful predictive biomarker for poor survival.
C-reactive protein (CRP) is an acute-phase reactant that is mainly produced in the liver. The serum CRP level increases rapidly within 24 h in response to infections or tissue injuries.1 We previously reported that the CRP level before consolidation chemotherapy for AML was useful for identifying patients who were at risk for febrile neutropenia and documented infection (DI), with cutoff values of 0.19 and 0.26 mg/dL, respectively.2 The quantitation of the serum CRP level might make it possible to detect slight inflammation and identify patients at high risk of infectious events. On the other hand, the impact of the serum CRP level in hematopoietic SCT (HSCT) recipients is still controversial.
In this study, we retrospectively assessed the value of the serum CRP level before HSCT for predicting early DI, non-relapse mortality (NRM) and other clinical outcomes after allogeneic HSCT.
Patients and methods
We reviewed the clinical charts of consecutive 125 patients who underwent their first allogeneic HSCT between June 2007 and February 2011 in our institution. Among these patients, 35 patients were excluded as they were receiving i.v antibiotics for fever or DI at the start of the conditioning regimen. Finally, the clinical data of 90 patients were included in this analysis. This analysis was approved by the institutional review board of Saitama Medical Center, Jichi Medical University.
Myeloablative conditioning (MAC) regimens included a combination of CY and either TBI or BU. Fludarabine-based reduced-intensity conditioning (RIC) regimens, such as fludarabine combined with BU or melphalan, were used in elderly or clinically infirm patients. Patients with severe aplastic anemia received fludarabine, CY, and anti-thymoglobulin, with or without low-dose TBI at 2 Gy. Alemtuzumab-containing regimens were used in HSCT from a two or three Ag-mismatched donor.
GVHD prophylaxis consisted of the continuous infusion of CYA or tacrolimus combined with short-term MTX (10–15 mg/m2 on day 1, 7–10 mg/m2 on days 3 and 6 and an optional dose on day 11). Acute GVHD was graded as previously described.3 Prophylaxis against bacterial, fungal and Pneumocystis jiroveci infection consisted of fluoroquinolones, fluconazole or itraconazole, and sulfamethoxazole/trimethoprim or inhalation of pentamidine, respectively. As a prophylaxis against HSV infection, acyclovir was administered from day 7 to day 35, followed by a long-term low-dose administration for VZV reactivation. Pre-emptive therapy with ganciclovir was administered by monitoring CMV antigenemia by the C10/11 method weekly after engraftment. The serum CRP level was measured just before the start of the conditioning regimen and at least twice weekly thereafter, exclusively using a Latex agglutination reaction kit according to the manufacturer’s instructions (Nanopia CRP, Sekisui Medical, Tokyo, Japan; minimum detection level 0.01 mg/dL).
Definition of transplant-related complications and statistical analysis
DI included microbiologically documented and presumed infections based on clinical and/or radiological findings. NRM was defined as death without relapse after HSCT. To be consistent with our previous studies, acute leukemia in first or second remission, CML in first or second chronic phase, and myelodysplastic syndrome (MDS) or myeloproliferative neoplasm without leukemic transformation were defined as standard-risk diseases, and others were defined as high-risk diseases.4 These definitions are based on the previous Japanese studies. For example, the survival of HSCT for MDS-refractory anemia with excess blasts was similar to that for acute leukemia in first remission.5
We assessed the value of the baseline serum CRP level for predicting early DI within 30 and 100 days after HSCT, and NRM within 100 days after HSCT by using the area under the receiver-operating characteristic (ROC) curve. The cutoff serum CRP value was also determined by the ROC analysis. Comparisons between two groups were performed using Fisher’s exact test for categorical variables and the Mann–Whitney U-test for continuous variables. The probability of OS with 95% confidence intervals (CI) was estimated according to the Kaplan–Meier method, and compared using the log-rank test. The cumulative incidences of DI, GVHD and NRM with 95% CI were estimated and compared between groups by using Gray’s test, and death without DI, death without GVHD and relapse were treated as competing events, respectively. In the multivariate analysis, Cox proportional-hazards regression model was used for OS, whereas Fine and Gray’s proportional-hazard model was used for the cumulative incidences of NRM and acute GVHD. Factors with at least borderline significance (P<0.15) in univariate analyses were subjected to multivariate analyses and deleted stepwisely. P<0.05 were considered to be significant. All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, accessed 1 March 2012, at http://www.jichi.ac.jp/saitama-sct/SaitamaHP.files/statmedEN.html), which is a graphical user interface for R (The R Foundation for Statistical Computing, version 2.13.0). More precisely, it is a modified version of R commander (version 1.6–3) that was designed to add statistical functions that are frequently used in biostatistics.
There were 90 patients with a median age of 45 years, including 54 males and 36 females (Table 1). The underlying disease was acute leukemia in more than half of the patients (n=48). Nine patients had MDS. Eleven patients (12%) had high-risk disease. The median serum CRP level before HSCT was 0.19 mg/dL (range 0.01–8.09).
Among the 90 allogeneic HSCT recipients, 89 developed febrile neutropenia and 45 developed DI, including 38 with bacteremia. The most frequently isolated organism was Staphylococcus epidermidis. Grade II–IV and grade III–IV acute GVHD were observed in 33 and 11 recipients, respectively. There were 13 deaths without relapse, including 5 deaths within 100 days after HSCT (early NRM). The cause of early NRM included pneumonia in two, GVHD in two and pulmonary proteinosis in one patient.
Predictive value of the serum CRP level before HSCT
To evaluate the predictive value of the baseline serum CRP level just before the initiation of conditioning for transplant-related complications, we performed ROC analyses and evaluated the area under the curve (AUC). The predictive value for febrile neutropenia was not analyzed, as it was observed in 89 of the 90 recipients. The AUCs were 0.53 (95% CI 0.41–0.66) and 0.52 (95% CI 0.40–0.64) for DI within 30 and 100 days after allogeneic HSCT, respectively. Therefore, the baseline CRP level before conditioning did not seem to be useful for predicting early DI after HSCT, as the AUCs were close to 0.5. On the other hand, the AUC for early NRM within 100 days after HSCT was 0.83(95% CI 0.57–1.00), suggesting that the baseline CRP value could be used to predict early NRM (Figure 1). The sensitivity and specificity were 80% and 87%, respectively, with a cutoff value of 0.6 mg/dL.
Impact of the cutoff value of CRP on clinical outcomes
The patients were divided into two groups according to the baseline serum CRP levels before conditioning with a cutoff of 0.6 mg/dL. The background characteristics were not significantly different between the two groups, although the high-CRP group included more patients with MDS (Table 1). Among the fifteen patients with a high CRP level, four had a recent history of pneumonia or prolonged intestinal mucositis within 3 months before HSCT, six had residual underlying malignant disease, four had oral or anal canal complications and one had pulmonary alveolar proteinosis.
The cumulative incidences of NRM at 100 days and 2 years after HSCT were 1.3% (95% CI 0.1–6.4%) and 10.6% (95% CI 4.6–19.7%) in the low CRP group, and 26.7% (95% CI 7.7–50.5%) and 41.7% (95% CI 15.6–66.2%) in the high-CRP group, respectively (P=0.0015, Figure 2a). In the multivariate analysis, the CRP level was an independent significant variable (HR 6.21, 95% CI 2.17–17.71, P=0.00065; Table 2) after adjusting for the background disease. In addition, the cumulative incidence of NRM was significantly higher in the high-CRP group (P=0.0051) even after patients with MDS were excluded.
By subgroup analyses stratified according to the conditioning regimen (MAC or RIC), we confirmed that a high baseline CRP level had an adverse impact in both cohorts who received the MAC and RIC regimens. In patients who received MAC regimens, the cumulative incidences of NRM at 2 years after HSCT were 8.2% (95% CI 2.0–20.2%) and 33.3% (95% CI 6.4–64.5%) in the low and high-CRP groups, respectively (P=0.024). In patients who received RIC regimens, the cumulative incidences of NRM at 2 years after HSCT were 14.8% (95% CI 4.5–31.0%) and 58.3% (95% CI 3.3–91.8%) in the low and high-CRP groups, respectively (P=0.025).
In contrast, the cumulative incidences of DI at 30 days and 100 days after HSCT were equivalent between the two groups stratified according to the baseline CRP levels; 38.7% (95% CI 27.6–49.6%) and 49.3% (95% CI 37.5–60.1%) in the low CRP group, and 40.0% (95% CI 15.6%–63.6%) and 46.7% (95% CI 20.1–69.6%) in the high-CRP group, respectively (P=0.99). The incidences of grade II–IV acute GVHD were not significantly different between the two groups; 34.7% (95% CI 24.1–45.5%) and 46.7% (95% CI 20.0–69.7%) in the low and high-CRP groups, respectively (P=0.45). On the other hand, the cumulative incidences of grade III–IV acute GVHD were 9.3% (95% CI 4.1–17.2%) and 26.7% (95% CI 7.7–50.5%) in the low and high-CRP groups, respectively (P=0.065, Figure 2b). A multivariate analysis revealed that a high baseline CRP level was a significant risk factor for grade III–IV acute GVHD (HR 3.91, 95% CI 1.17–13.11, P=0.027, Table 3). In subgroup analyses stratified according to the conditioning regimens, a high baseline CRP level had a significant adverse impact on the incidence of grade III–IV acute GVHD only in recipients who received RIC regimens. In recipients who received RIC, the cumulative incidences of grade III–IV acute GVHD were 3.4% (95% CI 0.2–15.2%) and 50.0% (95% CI 7.7–82.9%) in the low and high-CRP groups, respectively (P<0.001). On the other hand, in recipients who received MAC, the cumulative incidences of grade III–IV acute GVHD were 13% (95% CI 5.2–24.5%) and 11.1% (95% CI 0.5–40.9%) in the low and high-CRP groups, respectively (P=0.83).
In addition, recipients with higher baseline CRP levels had inferior OS at 2 years after HSCT (72.1% [95% CI 58.4–81.9%] vs 34.2% [95% CI 7.5–64.2%], P=0.003, Figure 2c). In the univariate analysis, recipient age >45 years, underlying disease and a baseline serum CRP level >0.6 mg/dL were significantly associated with poor OS. The multivariate analysis revealed that a high baseline CRP level was significantly associated with poor OS (HR 3.27 95%CI 1.22–8.75, P=0.018, Table 4). By subgroup analyses stratified according to the conditioning regimen, we also found that a high baseline CRP level had an adverse impact in both cohorts who received MAC and RIC regimens. Two-year OS was 76.9% (95% CI 58–88.1%) and 44.4% (95% CI 7.9–77.1%) in the low and high baseline CRP groups (P=0.065), respectively, among patients who received MAC regimens, and 63.8% (95% CI 41.3–79.6%) and 22.2% (95% CI 9.6–61.5%) in the low and high baseline CRP groups (P=0.0085), respectively, among patients who received RIC regimens.
We retrospectively analyzed the impact of the serum CRP level on the clinical outcome after HSCT. A baseline serum CRP level >0.6 mg/dL was associated with higher incidences of grade III–IV acute GVHD, high NRM and poor OS.
Several studies have evaluated the impact of elevated CRP levels before the start of a conditioning regimen on HSCT-related complications (Table 5).6, 7, 8 Remberger showed that high CRP (>1.0 mg/dL) appeared to have negative impacts on OS and NRM only in patients who received RIC regimens, and not in those who received MAC regimens.6 Artz7 analyzed recipients of RIC regimens, and suggested that there was a significant correlation between a higher baseline CRP level (1.85 mg/dL) before conditioning and higher incidences of acute GVHD, NRM and poor OS. In Artz’s7 study, DI was frequently observed in the high-CRP group, but this association was not statistically significant. On the other hand, Kanda8 showed that high serum ferritin (>700 ng/mL) and high CRP (0.3 mg/dL) levels were significantly associated with the development of bacterial infections and NRM. Therefore, a high baseline CRP level seemed to have an adverse impact on the clinical outcome after HSCT, but this impact may vary according to the patient characteristics, such as the background disease or the intensity of the conditioning regimen and the cutoff value of the serum CRP level.
In this study, a high baseline CRP level of at least 0.6 mg/dL was significantly associated with the development of grade III–IV acute GVHD, which may have contributed to inferior survival. High baseline CRP levels may reflect inflammatory environments induced by remnant infections, which could not be detected even with the thorough examinations before HSCT, or the tumor itself including minimal residual disease. Previous reports suggested the relationship between inflammatory cytokines and the induction of GVHD. For example, tumor necrosis factor-α, IL-1, and IL-6 are released following the conditioning regimens, and have a primary role in activating T cells.9 The pathogenesis of GVHD is related not only to a donor–recipient allo-immune response, but also to an inflammatory environment, such as that induced by tissue injury due to the conditioning regimen. Therefore, the presence of an inflammatory environment as detected by a high baseline CRP level might induce severe acute GVHD. In a subgroup analysis, a high baseline CRP level appeared to have an adverse impact impact on the incidence of severe acute GVHD only in patients who received RIC regimens. The effect of a baseline inflammatory environment might be masked by the inflammation induced by toxic conditioning regimens in those who received MAC regimens.
On the other hand, we did not find a significant association between baseline CRP levels and the incidence of DI, in contrast to the findings in our previous study in patients who were receiving consolidation chemotherapy for AML. There are several possible explanations for this discrepancy. First, the conditioning regimen in the current study induced severe neutropenia and mucosal damage, and the development of these post-treatment risk factors for infectious complications might lead to the less significant predictive value of the baseline CRP level. Second, only 58 of the 90 patients in this study, who underwent HSCT, were in remission of the underlying disease at the measurement of the baseline serum CRP level. Therefore, residual tumor cells might have affected the CRP level and impaired its specificity.10 Recently, procalcitonin has been investigated as a biomarker for the diagnosis of bacterial infections.11 Therefore, the serum CRP level in combination with procalcitonin level might be more suitable for predicting infectious events after HSCT.
In conclusion, the current findings suggest that a high baseline CRP level just before conditioning could predict higher incidences of grade III–IV acute GVHD, higher NRM and inferior OS. On the other hand, the baseline CRP level was not a useful predictor of infectious events. Although this study includes several limitations owing to the retrospective study design, small population and heterogeneous background of the patients, measurement of the serum CRP level is easy and inexpensive, and therefore further studies are warranted to investigate risk assessment and management strategies based on the baseline CRP level in HSCT recipients.
Pepys MB, Baltz ML . Acute phase proteins with special reference to C-reactive protein and related proteins (pentaxins) and serum amyloid A protein. Adv Immunol 1983; 34: 141–212.
Sato M, Kako S, Oshima K, Sato K, Terasako K, Kimura S et al. Prediction of infectious events by high-sensitivity C-reactive protein level before undergoing chemotherapy for acute myeloid leukaemia. Scand J Infect Dis 2010; 42: 97–101.
Przepiorka D, Weisdorf D, Martin P, Klingemann HG, Beatty P, Hows J et al. 1994 Consensus Conference on Acute GVHD. Bone Marrow Transplant 1995; 15: 825–828.
Kanda J, Saji H, Fukuda T, Kobayashi T, Miyamura K, Eto T et al. Related transplantation with HLA-1 Ag mismatch in the GVH direction and HLA-8/8 allele-matched unrelated transplantation: a nationwide retrospective study. Blood 2012; 119: 2409–2416.
Nakai K, Kanda Y, Fukuhara S, Sakamaki H, Okamoto S, Kodera Y et al. Value of chemotherapy before allogeneic hematopoietic stem cell transplantation from an HLA-identical sibling donor for myelodysplastic syndrome. Leukemia 2005; 19: 396–401.
Remberger M, Mattsson J . C-reactive protein levels before reduced-intensity conditioning predict outcome after allogeneic stem cell transplantation. Int J Hematol 2010; 92: 161–167.
Artz AS, Wickrema A, Dinner S, Godley LA, Kocherginsky M, Odenike O et al. Pretreatment C-reactive protein is a predictor for outcomes after reduced-intensity allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant 2008; 14: 1209–1216.
Kanda J, Mizumoto C, Ichinohe T, Kawabata H, Saito T, Yamashita K et al. Pretransplant serum ferritin and C-reactive protein as predictive factors for early bacterial infection after allogeneic hematopoietic cell transplantation. Bone Marrow Transplant 2010; 46: 208–216.
Ferrara JL, Levine JE, Reddy P, Holler E . Graft-versus-host disease. Lancet 2009; 373: 1550–1561.
Drahovsky D, Dunzendorfer U, Ziegenhagen G, Drahovsky M, Kellen JA . Reevaluation of C-reactive protein in cancer sera by radioimmunoassay and radial immunodiffusion. I. Diagnostic value and use in battery of conventional tumor markers. Oncology 1981; 38: 286–291.
Limper M, de Kruif MD, Duits AJ, Brandjes DP, van Gorp EC . The diagnostic role of procalcitonin and other biomarkers in discriminating infectious from non-infectious fever. J Infect 2010; 60: 409–416.
Koya J, Nannya Y, Ichikawa M, Kurokawa M . The clinical role of procalcitonin in hematopoietic SCT. Bone Marrow Transplant 2012; 47: 1326–1331.
The authors declare no conflict of interest.
About this article
Cite this article
Sato, M., Nakasone, H., Oshima, K. et al. Prediction of transplant-related complications by C-reactive protein levels before hematopoietic SCT. Bone Marrow Transplant 48, 698–702 (2013). https://doi.org/10.1038/bmt.2012.193
- transplant-related complications
- infectious events
- C-reactive protein
- hematopoietic SCT
Assessment of older adult candidates for allogeneic hematopoietic cell transplantation: updates and remaining questions
Expert Review of Hematology (2019)
Net reclassification improvement with serial biomarkers and bed-sided spirometry to early predict the need of organ support during the early post-transplantation in-hospital stay in allogeneic HCT recipients
Bone Marrow Transplantation (2019)
The Prognostic Impact of Pretransplantation Inflammatory and Nutritional Status in Adult Patients after Myeloablative Single Cord Blood Transplantation
Biology of Blood and Marrow Transplantation (2019)
Pretransplant serum FT3 levels in recipients predict early non-relapse mortality after myeloablative allogeneic haematopoietic cell transplantation from matched sibling donors