- Email: [email protected]
The relation of C-reactive protein to vasoocclusive crisis in children with sickle cell disease
Blood Cells, Molecules, and Diseases 45 (2010) 293–296
Contents lists available at ScienceDirect
Blood Cells, Molecules, and Diseases j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y b c m d
The relation of C-reactive protein to vasoocclusive crisis in children with sickle cell disease Fatima A. Mohammed a, Najat Mahdi a,b, Mai A. Sater a, Khadija Al-Ola b, Wassim Y. Almawi a,⁎ a b
Department of Medical Biochemistry, College of Medicine and Medical Sciences, Arabian Gulf University, Manama, Bahrain Department of Pediatrics, Salmaniya Medical Complex, Manama, Bahrain
a r t i c l e
i n f o
Article history: Submitted 16 July 2010 Available online 1 September 2010 (Communicated by M. Lichtman, M.D., 28 July 2010) Keywords: C-reactive protein Sickle aell anemia Vasoocclusive crisis
a b s t r a c t In view of evidence linking sickle cell anemia (SCA) with chronic inflammation, and given the role of high sensitivity C-reactive protein (hs-CRP) as inflammatory mediator, we hypothesized that SCA vasoocclusive crisis (VOC) is associated with heightened hs-CRP levels. Study subjects comprised 104 SCA patients who experienced VOC event during the study period (VOC group), and 40 SCA patients who did not develop VOC for at least 9 months prior to blood collection (Steady-state group). hs-CRP determination was done by latexenhanced nephelometry. Higher hs-CRP levels were seen in VOC [median(range) = 31.3(1.14–363.0)] than steady-state [median(range) = 5(0.16–185.0)] groups (P b 0.001), with enrichment in high hs-CRP percentiles in VOC cases, which translated into step-wise increased VOC risk. Receiver-operating characteristic (ROC) analysis was employed in assessing the usefulness of hs-CRP as predictor of the frequency and severity of VOC. Spearman's correlation coefficient between hs-CRP and VOC was 0.65 (P b 0.001) among unselected patients (0.71 in males and 0.59 in females). hs-CRP area under ROC curves was 0.90 (95% CI = 0.85–0.94) among unselected patients, 0.94 (95% CI = 0.89–0.98) for males, and 0.85 (95% CI = 0.77–0.93) for females. Logistic regression analysis confirmed the positive association of increased hs-CRP levels with VOC, which correlated positively with VOC frequency (P b 0.001), type (P b 0.001), pain (P b 0.001), and need for hospitalization (P = 0.024). These data support strong association of increased hsCRP levels with VOC, which impacts VOC-related parameters, and support a role for hs-CRP in VOC follow-up. © 2010 Elsevier Inc. All rights reserved.
Introduction Sickle cell anemia (SCA) is an inherited hematological disorder, which arises from a single point mutation in codon 6 of the β-globin gene, resulting in glutamic acid-to-valine substitution [1,2]. Despite its monogenic origin, SCA has a heterogeneous phenotype of variable intensity, which is influenced by modifiable (environmental), and non-modifiable (ethnicity and sickle cell haplotype) factors [1,3]. SCA is associated with a number of acute and chronic complications, of which microvessel occlusion, also termed vasoocclusive crisis (VOC), is clinically the most significant pathological process [1,4]. VOC is complex in nature, and involves interactions between sickle-shaped red blood cells (ssRBCs), white blood cells, endothelium, plasma proteins, and additional factors [5]. This results in the occlusion of small capillaries by ssRBC, which in turn restricts blood flow to the affected organs, resulting in ischemia, pain, and organ damage [6,7]. Recent evidence implicate a state of chronic inflammation in the pathogenesis of SCA [8–10]. Increased levels of pro-inflammatory markers were detected in the sera of SCA patients, including high
sensitivity C-reactive protein (hs-CRP) [9,11]. CRP is a stable plasma acute phase reactant produced in the liver in response to inflammatory signals [12,13], and as inflammatory marker for predicting the risk of cardiovascular disease [13,14]. Circulating CRP levels are generally stable, are determined by its rate of synthesis, and fall within a characteristic range for each individual [15,16]. The association of CRP with VOC was suggested by the findings that heightened hs-CRP levels were detected in the sera of steadystate SCA patients [17,18], which increased further during VOC episodes [18,19]. In addition, the severity of VOC was attenuated by non-steroidal anti-inflammatory drugs (NSAIDs), and correlated with clinical improvement [20,21]. Despite the strong inflammatory facet of VOC, limited studies investigated the contribution of hs-CRP to VOC pathogenesis, often with inconclusive results. Here we examined the association between changes in hs-CRP levels on the development of VOC and VOC-related parameters in Bahraini SCA children. Materials and methods Study subjects
⁎ Corresponding author. Department of Medical Biochemistry, College of Medicine and Medical Sciences, Arabian Gulf University, PO Box 22979, Manama, Bahrain. Fax: + 973 17271090. E-mail address: [email protected] (W.Y. Almawi). 1079-9796/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.bcmd.2010.08.003
This was a case–control study, conducted between Sept. 2009 and March 2010 in the Department of Pediatrics, Salmaniya Medical Complex (Manama, Bahrain). Study subjects comprised 144 SCA
294
F.A. Mohammed et al. / Blood Cells, Molecules, and Diseases 45 (2010) 293–296
patients (mean age, 10.8 ± 7.7 years) diagnosed according to hemoglobin profile (Hb A, Hb S, Hb A2, and Hb F). Patients were assigned to 1 of 2 groups: SCA patients who had VOC event based on need for hospitalization and blood transfusion, and who presented with painful episodes (VOC group; n = 104), and SCA patient who reported no such events for the previous 9 months prior to date of blood draw (Steady-state group; n = 40). VOC patients presented with these complications: acute chest syndrome (n = 32), osteomyelitis (n = 28), priapism (n = 6), avascular necrosis (AVN) of the femoral head (n = 7), ischemic stroke (n = 4), osteomyeltits + acute chest syndrome (n = 8), osteomyeltits + stroke (n = 1), and undefined causes (n = 18). Inclusion criteria for steady-state SCA patients included asymptomatic with SCA, afebrile state, no VOC episode, hospitalization or transfusion for at least 9 months, and had not had any infection for at least 10 days prior to specimen collection, together with ethnic origin (only Bahraini Arabs). Study subjects' historical and clinical information were extracted and verified by qualified personnel. Peripheral venous blood was drawn into EDTA-anticoagulant tubes by a trained phlebotomist, and the platelet-poor plasma (PPP) was prepared by centrifugation for 10 min at 2000 g; PPP aliquots were stored at −70 °C pending analysis. The Arabian Gulf University Research and Ethics Committee approved the study protocol, and all participants (or guardians in the pediatric cases) gave written informed consent. Assessment of VOC pain VOC episode was defined as pain related to SCA complication; patients presenting with SCA-unrelated cause, including trauma and cancer were excluded. VOC painful episodes information were obtained by personal interview from patients or guardians by trained and qualified nurses, and/or pediatric residents. VOC pain assessment included age of onset, duration (days), frequency (number of episodes/year), type (localized, generalized), site (chest, back, abdomen, upper and lower limbs), severity (scored on a scale of 1 to 10), and management. VOC pain treatment consisted of NSAIDs [47 patients (45.2%)], narcotics [18 patients (17.3%)], with 16 patients (15.4%) treated with both NSAIDs and narcotics. Comparable frequencies of hydroxyurea- (P = 0.196) and folic acid- (P = 0.442) treated patients were seen in both SCA patient groups. hs-CRP measurement Measurement of hs-CRP in plasma samples was done by latexenhanced nephelometry on a BN II Nephelometer (Dade Behring, Milan, Italy). Samples were assayed in duplicate in each analytical run; the lower limit of detection was 0.15 mg/L, and the assay range
was 0.175–11.0 mg/L (initial dilution). Serial serum dilutions were made in measuring high hs-CRP (N10 mg/L) levels. Percentile hs-CRP values were estimated for comparison purposes. Statistical analyses Statistical analyses were performed on SPSS software (Statistical Package for the Social Sciences, version 17.0; SSPS Inc, Chicago, IL, USA). Since their values showed a nonparametric distribution, median hs-CRP plasma concentrations were computed and the differences in the distributions and in median values between VOC group and Steady-state control group were assessed by the use of Mann– Whitney comparison by ranks. Utility of hs-CRP as predictor of VOC was examined using receiver-operating characteristic (ROC) curves. In risk prediction models, hs-CRP concentrations were divided into percentiles. Using steady-state SCA control patients as the reference group, adjusted estimates of risk were obtained using conditional logistic regression models, done first at the univariate and later the multivariate levels; hs-CRP was used as continuous, and then as categorized variables. Results The baseline characteristics of the study participants are presented in Table 1. The VOC group consisted of 104 SCA patients who developed VOC episodes, whereas the steady-state control group consisted of 40 age- (P = 0.118) and gender- (P = 0.268) matched subjects who did not experience any VOC episodes 9 months prior to the collection of blood specimens. Statistically significant difference in HbF (P = 0.008), but not HbS (P = 0.138) or total hemoglobin (P = 0.252) was noted between the VOC and steady-state control groups. Apart from platelet count (P = 0.032), all hematological indices were comparable between the two groups. The median hsCRP concentration of the VOC group was significantly higher than that of steady-state controls (31.3 vs 5.0 mg/L; P b 0.001). Median hs-CRP value in steady-state SCA patients was 5.0 ng/mL (range, 0.16–185.0 mg/L), which was higher than that of healthy nonSCA individuals (b3.0 mg/L), thereby indicating that a state of lowgrade chronic inflammation accompanies SCA, even in asymptomatic cases. The median hs-CRP level in VOC cases of 31.3 mg/L (range 1.14– 363.0 mg/L) was significantly higher than that of steady-state control patients (P b 0.001) (Table 1). A systematic shift in hs-CRP distributions toward higher values was seen in VOC cases, which was apparent at the 50th percentile, and even more at higher percentiles (Table 2). ROC analyses were performed in order to determine the predictive value of hs-CRP levels in the assessment of the frequency and severity
Table 1 Characteristics of study participants. VOC group Age Gender (Male:Female) Hemoglobin Profile HbS (%) HbF (%) Total hemoglobin (Hb) (g/dL) Hematological Indices WBC (× 109/L) Platelets (×109/L) Hematocrit (%) Mean corpuscular volume (fL) Mean corpuscular Hb (pg) Mean corpuscular Hb Conc. (g/dL) Reticulocytes (%) CRP median (range) a b
a
Steady-state group
a
P
b
10.1 ± 5.7 70: 34
12.5 ± 12.8 20: 20
0.118 0.268
71.7 ± 9.6 19.1 ± 7.3 9.2 ± 1.4
69.0 ± 8.3 23.1 ± 8.1 9.5 ± 1.3
0.138 0.008 0.252
9.7 ± 5.0 381.6 ± 233.4 29.6 ± 4.6 76.8 ± 10.2 26.1 ± 4.2 33.6 ± 2.7 5.9 ± 3.8 31.3 (1.14–363.0)
9.1 ± 4.0 274.5 ± 168.6 27.7 ± 6.4 75.3 ± 8.3 25.6 ± 3.4 34.0 ± 1.6 5.8 ± 4.0 5 (0.16–185.0)
0.590 0.032 0.268 0.476 0.596 0.451 0.937 3.7 × 10− 6
Study subjects comprised 104 SCA patients who had any VOC event during study (VOC group) and 40 SCA patients who had no VOC events (Steady-state group). Pearson's chi square test (categorical variables), 2-tailed t-test (continuous variables).
F.A. Mohammed et al. / Blood Cells, Molecules, and Diseases 45 (2010) 293–296 Table 2 Influence of CRP cut-off Levels on VOC Risk. Cut-off (percentile)
Patients
P25 P50 P75 P90 P95
8 29 38 18 11
a
(7.7)a (27.9) (36.5) (17.3) (10.6)
295
Table 3 Correlation between CRP Levels and VOC outcomes.
Controls
P
OR
95% CI
Outcome
r2a
P
25 (53.2) 9 (19.1) 10 (21.3) 2 (4.3) 1 (2.1)
b 0.001 b 0.001 b 0.001 b 0.001 0.002
1.00 10.1 11.9 28.1 34.4
(Reference) 3.4–30.0 4.1–34.2 5.3–148.5 3.8–309.1
Frequency (episodes/month) VOC type (generalized vs. local) Pain scale (1–10) Age at VOC onset (years) Need for hospitalization VOC treatment (NSAIDs and/or narcotics) Hydroxyurea treatment HbS (%) HbF (%)
0.422 0.451 0.317 − 0.043 0.191 0.151 − 0.055 0.017 − 0.022
2.9 × 10− 6 1.0 × 10− 6 1.6 × 10− 4 0.622 0.024 0.071 0.511 0.854 0.811
Number of subjects (percent of total).
of VOC episodes (Fig. 1). Area-under-ROC curves provided good discriminatory power for hs-CRP and VOC episodes, and showed that varied hs-CRP levels revealed similar sensitivities and specificities for predicting the frequency and severity of VOC. The Spearman's correlation coefficient between hs-CRP and VOC was 0.65 (P b 0.001) among unselected patients, 0.71 (P b 0.001) in males and 0.59 (P b 0.001) in females. The area under ROC curves of hs-CRP was 0.90 (95% CI = 0.85–0.94) among unselected patient, 0.94 (95% CI = 0.89–0.98) for males, and 0.85 (95% CI = 0.77–0.93) for females. hs-CRP levels were then categorized into 6 strata: percentiles 1–25 (P25), 26–50 (P50), 51–75 (P75), 76–90 (P90) and N90 (P95), according to concentrations present in the control group, and analyzed in regression models, first at the univariate and later at the multivariate levels. Univariate regression analysis demonstrated a positive dose–effect relationship for hs-CRP with VOC, with increased VOC risk seen with increased hs-CRP levels (Table 2). The strongest OR was for P90 and P95 hs-CRP percentiles, in which they were associated with a 28.1-fold and 34.4-fold higher risk than P25, respectively (Table 2). Spearman correlation calculation demonstrated positive correlation between increased hs-CRP levels and VOC frequency (number of episodes/year; r2 = 0.422; P b 0.001), VOC type (localized vs. generalized; r2 = 0.451; P b 0.001), pain scale (r2 = 0.317; P b 0.001), and with the need for hospitalization (r2 = 0.191; P = 0.024). However, hs-CRP levels did not correlate with either HbS (P = 0.854) and HbF (P = 0.811) levels, or with with the age at VOC onset (P = 0.622), or with pain treatment regimen (P = 0.071) and hydroxyurea treatment (P = 0.511) (Table 3). On multivariate logistic regression analyses with SCA steady-state patients as the reference group, increased hs-CRP levels found to be independent risk factors for VOC, which was apparent at P50 (aOR = 9.5; 95% CI = 2.2–41.6), and increased at P75 (aOR = 12.8; 95% CI = 1.2–135.2), but later stabilized at P90 (aOR = 15.4; 95% CI = 2.5–94.7) and P95 (aOR = 15.5; 95% CI = 3.7–64.7) (Table 4).
a
Spearman correlation coefficients.
None of the other variables entered in the model (HbS, HbF, paletelet and WBC count) was found to be associated with VOC (Table 4). Discussion This study of Bahraini children with VOC demonstrated positive association between increased hs-CRP levels and the development of VOC and its severity, and thus the extent of microvessel occlusion in SCA. This was highlighted by the enrichment of high hs-CRP percentiles in VOC than in control SCA patients, and with the positive correlation between hs-CRP levels and the severity, frequency, and type of VOC, as well as the need for hospitalization. Logistic regression analysis confirmed the association of hs-CRP with VOC after controlling for potential confounders, thereby demonstrating significant contribution of hs-CRP, a well-established marker of inflammation [13,16], to VOC pathogenesis. In addition to the prevailing notion that VOC painful crisis is attributed to the occlusion of the small blood vessels by ssRBC leading to development of ischemia, increasing number of studies suggests a role for inflammation in the pathophysiology of VOC. This was exemplified by the heightened pro-inflammatory cytokine expression detected in the sera of SCA patients, both during steady-state and VOC episodes [18,19], which in turn elicits the inflammatory response accompanying VOC highlighted by elevation in hs-CRP levels [11,22]. In this study, hs-CRP levels in steady-state SCA patients (median = 5.0 mg/L) were higher than healthy non-SCA controls (normal range: 0–3 mg/L), suggesting a covert inflammatory response in SCA, even in the absence of crisis. This was in agreement with studies on pediatric [11] and adult [23,24] SCA patients, in which hs-CRP levels were elevated under steady-state conditions, supporting the contribution of a state of chronic inflammation to SCA [25]. It was proposed that intracellular hemoglobin polymerization, a hallmark of SCA, induces oxidative damage in the cell membrane membrane in SCA [25,26], which induces elevation in CRP due to blocked vascular endothelium by ssRBC, as was suggested [17]. Given the influence of age on the perception of and reaction to painful episodes, we focused on pediatric SCA patients, age 1.5– 16 years, with age- and ethnically-matched SCA patients without controls. Insofar as SCA phenotype influences the course of the Table 4 Logistic regression analysis for association of CRP with VOC. Cut-off (percentile) hs-CRP
Fig. 1. The ROC curve of serum levels of hs-CRP for evaluation of VOC. The Spearman's correlation coefficient between hs-CRP and VOC was 0.65 (P b 0.001) among unselected patients, and the area under ROC curves of hs-CRP was 0.90 (95% CI = 0.85–0.94) among unselected patients.
b
P50 P75 P90 P95
HbS (%) HbF (%) Platelets WBC a b
aOR = adjusted odds ratio. Adjusted for HbF levels and platelet count.
aORa
95% CI
9.5 12.8 15.4 15.5 0.98 0.95 1.00 1.05
2.2–41.6 1.2–135.2 2.5–94.7 3.7–64.7 0.90–1.08 0.88–1.03 0.99–1.01 0.90–1.23
296
F.A. Mohammed et al. / Blood Cells, Molecules, and Diseases 45 (2010) 293–296
disease, all patients and controls were homozygous HbSS/HbSβ0Thal, which may explain the higher baseline hs-CRP levels than healthy non-HbSS individuals, as was suggested [11,25], further supporting the notion of ongoing subclinical inflammation in HbSS/HbSβ0Thal patients, which predisposes them at higher risk for future VOC episodes [27]. More significantly, hs-CRP levels showed a strong statistical association by appropriate regression analysis models with significant VOC endpoints, namely VOC frequency, pain scale, VOC site, and increased hospitalizations for pain, in agreement with the study of Krishnan in which increased hs-CRP was also linked with increased need for hospitalization [11,28]. In the VOC group, hs-CRP showed excellent correlation with abnormality in the liver function tests ALT (r2 = 0.283; P = 0.014), LDH (r2 = 0.446; P = 0.025), alkaline phosphatase (r2 = 0.240; P = 0.041), and gamma-GT (r2 = 0.257; P = 0.034); no similar association was recorded for the steady-state control group. This underscores the intersection of altered liver function activity and inflammation with pathological manifestations of SCA, and makes hsCRP a reliable marker of SCA-related complications, including VOC, since it reflects not only the inflammation, but also other aspects of endothelial and coagulation dysfunctions linked with SCA [18,28]. While NSAIDs are the mainstay for VOC management [20], hs-CRP did not correlate with VOC treatment. This was in apparent disagreement with an earlier report suggesting that serial determinations of CRP may be helpful in monitoring the response to treatment in SCA [20]. In summary, our findings suggest that low-grade inflammation, highlighted by increased levels of the acute phase protein hs-CRP, accompanies steady-state SCA. Further increases in hs-CRP levels are associated with VOC episodes, thereby underscoring the clinical significance of hs-CRP as an important predictor of VOC. A predictive value of hs-CRP for VOC has previously been reported also for thalassemias [29,30]. It should be noted that elevation in hs-CRP is an indirect measurement of inflammation, and thus our results must be interpreted within the limitation of the experimental design. Another limitation lies in the difficulty of identifying the exact mechanism or pathway that triggers of elevated hs-CRP. Despite these shortcomings, our results support for the clinical notion for the role of a covert inflammatory response in SCA, and hence will be instrumental in the future management strategies of VOC in SCA patients. Conflict of interest disclosure The authors declare no competing financial interests. References [1] N. Conran, C.F. Franco-Penteado, F.F. Costa, Newer aspects of the pathophysiology of sickle cell disease vaso-occlusion, Hemoglobin 33 (2009) 1–16. [2] A.S. Mehanna, Sickle cell anemia and antisickling agents then and now, Curr. Med. Chem. 8 (2001) 79–88. [3] G.R. Buchanan, M.R. DeBaun, C.T. Quinn, M.H. Steinberg, Sickle cell disease, hematology, Am. Soc. Hematol. Educ. Program (2004) 35–47. [4] M.J. Stuart, R.L. Nagel, Sickle-cell disease, Lancet 364 (2004) 1343–1360. [5] P.S. Frenette, G.F. Atweh, Sickle cell disease: old discoveries, new concepts, and future promise, J. Clin. Invest. 117 (2007) 850–858.
[6] D.K. Kaul, E. Finnegan, G.A. Barabino, Sickle red cell–endothelium interactions, Microcirculation 16 (2009) 97–111. [7] R. Zennadi, A. Chien, K. Xu, M. Batchvarova, M.J. Telen, Sickle red cells induce adhesion of lymphocytes and monocytes to endothelium, Blood 112 (2008) 3474–3483. [8] J.D. Belcher, C.J. Bryant, J. Nguyen, P.R. Bowlin, M.C. Kielbik, J.C. Bischof, R.P. Hebbel, G.M. Vercellotti, Transgenic sickle mice have vascular inflammation, Blood 101 (2003) 3953–3959. [9] C. Lanaro, C.F. Franco-Penteado, D.M. Albuqueque, S.T. Saad, N. Conran, F.F. Costa, Altered levels of cytokines and inflammatory mediators in plasma and leukocytes of sickle cell anemia patients and effects of hydroxyurea therapy, J. Leukoc. Biol. 85 (2009) 235–242. [10] O.S. Platt, Sickle cell anemia as an inflammatory disease, J. Clin. Invest. 106 (2000) 337–338. [11] S. Krishnan, Y. Setty, S.G. Betal, V. Vijender, K. Rao, C. Dampier, M. Stuart, Increased levels of the inflammatory biomarker C-reactive protein at baseline are associated with childhood sickle cell vasocclusive crises, Br. J. Haematol. 148 (2010) 797–804. [12] A.M. Dupuy, N. Terrier, L. Sénécal, M. Morena, H. Leray, B. Canaud, J.P. Cristol, Is Creactive protein a marker of inflammation? Nephrologie 24 (2003) 337–341. [13] J. Genest, C-reactive protein: risk factor, biomarker and/or therapeutic target? Can. J. Cardiol. 26 (Suppl A) (2010) 41A–44A. [14] V.B. Martínez, J.R. González-Juanatey, Markers of inflammation and cardiovascular disease: clinical applications of C-reactive protein determination, Am. J. Cardiovasc. Drugs 9 (Suppl 1) (2009) 3–7. [15] E.M. Macy, T.E. Hayes, R.P. Tracy, Variability in the measurement of C-reactive protein in healthy subjects: implications for reference intervals and epidemiological applications, Clin. Chem. 43 (1997) 52–58. [16] M.B. Pepys, G.M. Hirschfield, C-reactive protein: a critical update, J. Clin. Invest. 111 (2003) 1805–1812. [17] S.A. Akohoue, S. Shankar, G.L. Milne, J. Morrow, K.Y. Chen, W.U. Ajayi, M.S. Buchowski, Energy expenditure, inflammation, and oxidative stress in steadystate adolescents with sickle cell anemia, Pediatr. Res. 61 (2007) 233–238. [18] K.L. Bourantas, G.N. Dalekos, A. Makis, A. Chaidos, S. Tsiara, A. Mavridis, Acute phase proteins and interleukins in steady state sickle cell disease, Eur. J. Haematol. 61 (1998) 49–54. [19] A. Pathare, S.A. Kindi, S. Daar, D. Dennison, Cytokines in sickle cell disease, Hematology 8 (2003) 329–337. [20] D. Monnet, I. Diallo, A. Sangare, A.E. Yapo, Clinical value of C-reactive protein, alpha 1-glycoprotein acid and transferrin assay in homozygous sickle cell disease, Bull. Soc. Pathol. Exot. 86 (1993) 282–285. [21] E. Perlin, H. Finke, O. Castro, S. Rana, J. Pittman, R. Burt, C. Ruff, D. McHugh, Enhancement of pain control with ketorolac tromethamine in patients with sickle cell vaso-occlusive crisis, Am. J. Hematol. 46 (1994) 43–47. [22] A. Singhal, J.F. Doherty, J.G. Raynes, K.P. McAdam, P.W. Thomas, B.E. Serjeant, G.R. Serjeant, Is there an acute-phase response in steady-state sickle cell disease? Lancet 341 (1993) 651–653. [23] A.C. Makis, E.C. Hatzimichael, J. Stebbing, K.L. Bourantas, C-reactive protein and vascular cell adhesion molecule-1 as markers of severity in sickle cell disease, Arch. Int. Med. 166 (2006) 366–368. [24] J.S. Mohan, G.Y. Lip, J. Wright, D. Bareford, A.D. Blann, Plasma levels of tissue factor and soluble E-selectin in sickle cell disease: relationship to genotype and to inflammation, Blood Coagul Fibrinolysis 16 (2005) 209–214. [25] J. Stuart, P.C.W. Stone, N.O. Akinola, J.R. Gallimore, M.B. Pepys, Monitoring the acute phase response to vaso-occlusive crisis in sickle cell disease, J. Clin. Pathol. 47 (1994) 166–169. [26] M. Aslan, D. Thornley-Brown, B.A. Freeman, Reactive species in sickle cell disease, Ann. N.Y. Acad. Sci. 899 (2000) 375–391. [27] N.O. Akinola, S.M. Stevens, I.M. Franklin, G.B. Nash, J. Stuart, Subclinical ischaemic episodes during the steady state of sickle cell anaemia, J. Clin. Pathol. 45 (1992) 902–906. [28] C.C. Hedo, Y.A. Aken'ova, I.E. Okpala, A.O. Durojaiye, L.S. Salimonu, Acute phase reactants and severity of homozygous sickle cell disease, J. Intern. Med. 233 (1993) 467–470. [29] N. Archararit, S. Chuncharunee, A. Pornvoranunt, K. Atamasirikul, B. Rachakom, V. Atichartakarn, Serum C-reactive protein level in postsplenectomized thalassemic patients, J. Med. Assoc. Thai. 83 (Suppl 1) (2000) S63–S69. [30] I. Kanavaki, P. Makrythanasis, C. Lazaropoulou, M. Tsironi, A. Kattamis, I. Rombos, I. Papassotiriou, Soluble endothelial adhesion molecules and inflammation markers in patients with beta-thalassemia intermedia, Blood Cells Mol. Dis. 43 (2009) 230–234.
Contents lists available at ScienceDirect
Blood Cells, Molecules, and Diseases j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y b c m d
The relation of C-reactive protein to vasoocclusive crisis in children with sickle cell disease Fatima A. Mohammed a, Najat Mahdi a,b, Mai A. Sater a, Khadija Al-Ola b, Wassim Y. Almawi a,⁎ a b
Department of Medical Biochemistry, College of Medicine and Medical Sciences, Arabian Gulf University, Manama, Bahrain Department of Pediatrics, Salmaniya Medical Complex, Manama, Bahrain
a r t i c l e
i n f o
Article history: Submitted 16 July 2010 Available online 1 September 2010 (Communicated by M. Lichtman, M.D., 28 July 2010) Keywords: C-reactive protein Sickle aell anemia Vasoocclusive crisis
a b s t r a c t In view of evidence linking sickle cell anemia (SCA) with chronic inflammation, and given the role of high sensitivity C-reactive protein (hs-CRP) as inflammatory mediator, we hypothesized that SCA vasoocclusive crisis (VOC) is associated with heightened hs-CRP levels. Study subjects comprised 104 SCA patients who experienced VOC event during the study period (VOC group), and 40 SCA patients who did not develop VOC for at least 9 months prior to blood collection (Steady-state group). hs-CRP determination was done by latexenhanced nephelometry. Higher hs-CRP levels were seen in VOC [median(range) = 31.3(1.14–363.0)] than steady-state [median(range) = 5(0.16–185.0)] groups (P b 0.001), with enrichment in high hs-CRP percentiles in VOC cases, which translated into step-wise increased VOC risk. Receiver-operating characteristic (ROC) analysis was employed in assessing the usefulness of hs-CRP as predictor of the frequency and severity of VOC. Spearman's correlation coefficient between hs-CRP and VOC was 0.65 (P b 0.001) among unselected patients (0.71 in males and 0.59 in females). hs-CRP area under ROC curves was 0.90 (95% CI = 0.85–0.94) among unselected patients, 0.94 (95% CI = 0.89–0.98) for males, and 0.85 (95% CI = 0.77–0.93) for females. Logistic regression analysis confirmed the positive association of increased hs-CRP levels with VOC, which correlated positively with VOC frequency (P b 0.001), type (P b 0.001), pain (P b 0.001), and need for hospitalization (P = 0.024). These data support strong association of increased hsCRP levels with VOC, which impacts VOC-related parameters, and support a role for hs-CRP in VOC follow-up. © 2010 Elsevier Inc. All rights reserved.
Introduction Sickle cell anemia (SCA) is an inherited hematological disorder, which arises from a single point mutation in codon 6 of the β-globin gene, resulting in glutamic acid-to-valine substitution [1,2]. Despite its monogenic origin, SCA has a heterogeneous phenotype of variable intensity, which is influenced by modifiable (environmental), and non-modifiable (ethnicity and sickle cell haplotype) factors [1,3]. SCA is associated with a number of acute and chronic complications, of which microvessel occlusion, also termed vasoocclusive crisis (VOC), is clinically the most significant pathological process [1,4]. VOC is complex in nature, and involves interactions between sickle-shaped red blood cells (ssRBCs), white blood cells, endothelium, plasma proteins, and additional factors [5]. This results in the occlusion of small capillaries by ssRBC, which in turn restricts blood flow to the affected organs, resulting in ischemia, pain, and organ damage [6,7]. Recent evidence implicate a state of chronic inflammation in the pathogenesis of SCA [8–10]. Increased levels of pro-inflammatory markers were detected in the sera of SCA patients, including high
sensitivity C-reactive protein (hs-CRP) [9,11]. CRP is a stable plasma acute phase reactant produced in the liver in response to inflammatory signals [12,13], and as inflammatory marker for predicting the risk of cardiovascular disease [13,14]. Circulating CRP levels are generally stable, are determined by its rate of synthesis, and fall within a characteristic range for each individual [15,16]. The association of CRP with VOC was suggested by the findings that heightened hs-CRP levels were detected in the sera of steadystate SCA patients [17,18], which increased further during VOC episodes [18,19]. In addition, the severity of VOC was attenuated by non-steroidal anti-inflammatory drugs (NSAIDs), and correlated with clinical improvement [20,21]. Despite the strong inflammatory facet of VOC, limited studies investigated the contribution of hs-CRP to VOC pathogenesis, often with inconclusive results. Here we examined the association between changes in hs-CRP levels on the development of VOC and VOC-related parameters in Bahraini SCA children. Materials and methods Study subjects
⁎ Corresponding author. Department of Medical Biochemistry, College of Medicine and Medical Sciences, Arabian Gulf University, PO Box 22979, Manama, Bahrain. Fax: + 973 17271090. E-mail address: [email protected] (W.Y. Almawi). 1079-9796/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.bcmd.2010.08.003
This was a case–control study, conducted between Sept. 2009 and March 2010 in the Department of Pediatrics, Salmaniya Medical Complex (Manama, Bahrain). Study subjects comprised 144 SCA
294
F.A. Mohammed et al. / Blood Cells, Molecules, and Diseases 45 (2010) 293–296
patients (mean age, 10.8 ± 7.7 years) diagnosed according to hemoglobin profile (Hb A, Hb S, Hb A2, and Hb F). Patients were assigned to 1 of 2 groups: SCA patients who had VOC event based on need for hospitalization and blood transfusion, and who presented with painful episodes (VOC group; n = 104), and SCA patient who reported no such events for the previous 9 months prior to date of blood draw (Steady-state group; n = 40). VOC patients presented with these complications: acute chest syndrome (n = 32), osteomyelitis (n = 28), priapism (n = 6), avascular necrosis (AVN) of the femoral head (n = 7), ischemic stroke (n = 4), osteomyeltits + acute chest syndrome (n = 8), osteomyeltits + stroke (n = 1), and undefined causes (n = 18). Inclusion criteria for steady-state SCA patients included asymptomatic with SCA, afebrile state, no VOC episode, hospitalization or transfusion for at least 9 months, and had not had any infection for at least 10 days prior to specimen collection, together with ethnic origin (only Bahraini Arabs). Study subjects' historical and clinical information were extracted and verified by qualified personnel. Peripheral venous blood was drawn into EDTA-anticoagulant tubes by a trained phlebotomist, and the platelet-poor plasma (PPP) was prepared by centrifugation for 10 min at 2000 g; PPP aliquots were stored at −70 °C pending analysis. The Arabian Gulf University Research and Ethics Committee approved the study protocol, and all participants (or guardians in the pediatric cases) gave written informed consent. Assessment of VOC pain VOC episode was defined as pain related to SCA complication; patients presenting with SCA-unrelated cause, including trauma and cancer were excluded. VOC painful episodes information were obtained by personal interview from patients or guardians by trained and qualified nurses, and/or pediatric residents. VOC pain assessment included age of onset, duration (days), frequency (number of episodes/year), type (localized, generalized), site (chest, back, abdomen, upper and lower limbs), severity (scored on a scale of 1 to 10), and management. VOC pain treatment consisted of NSAIDs [47 patients (45.2%)], narcotics [18 patients (17.3%)], with 16 patients (15.4%) treated with both NSAIDs and narcotics. Comparable frequencies of hydroxyurea- (P = 0.196) and folic acid- (P = 0.442) treated patients were seen in both SCA patient groups. hs-CRP measurement Measurement of hs-CRP in plasma samples was done by latexenhanced nephelometry on a BN II Nephelometer (Dade Behring, Milan, Italy). Samples were assayed in duplicate in each analytical run; the lower limit of detection was 0.15 mg/L, and the assay range
was 0.175–11.0 mg/L (initial dilution). Serial serum dilutions were made in measuring high hs-CRP (N10 mg/L) levels. Percentile hs-CRP values were estimated for comparison purposes. Statistical analyses Statistical analyses were performed on SPSS software (Statistical Package for the Social Sciences, version 17.0; SSPS Inc, Chicago, IL, USA). Since their values showed a nonparametric distribution, median hs-CRP plasma concentrations were computed and the differences in the distributions and in median values between VOC group and Steady-state control group were assessed by the use of Mann– Whitney comparison by ranks. Utility of hs-CRP as predictor of VOC was examined using receiver-operating characteristic (ROC) curves. In risk prediction models, hs-CRP concentrations were divided into percentiles. Using steady-state SCA control patients as the reference group, adjusted estimates of risk were obtained using conditional logistic regression models, done first at the univariate and later the multivariate levels; hs-CRP was used as continuous, and then as categorized variables. Results The baseline characteristics of the study participants are presented in Table 1. The VOC group consisted of 104 SCA patients who developed VOC episodes, whereas the steady-state control group consisted of 40 age- (P = 0.118) and gender- (P = 0.268) matched subjects who did not experience any VOC episodes 9 months prior to the collection of blood specimens. Statistically significant difference in HbF (P = 0.008), but not HbS (P = 0.138) or total hemoglobin (P = 0.252) was noted between the VOC and steady-state control groups. Apart from platelet count (P = 0.032), all hematological indices were comparable between the two groups. The median hsCRP concentration of the VOC group was significantly higher than that of steady-state controls (31.3 vs 5.0 mg/L; P b 0.001). Median hs-CRP value in steady-state SCA patients was 5.0 ng/mL (range, 0.16–185.0 mg/L), which was higher than that of healthy nonSCA individuals (b3.0 mg/L), thereby indicating that a state of lowgrade chronic inflammation accompanies SCA, even in asymptomatic cases. The median hs-CRP level in VOC cases of 31.3 mg/L (range 1.14– 363.0 mg/L) was significantly higher than that of steady-state control patients (P b 0.001) (Table 1). A systematic shift in hs-CRP distributions toward higher values was seen in VOC cases, which was apparent at the 50th percentile, and even more at higher percentiles (Table 2). ROC analyses were performed in order to determine the predictive value of hs-CRP levels in the assessment of the frequency and severity
Table 1 Characteristics of study participants. VOC group Age Gender (Male:Female) Hemoglobin Profile HbS (%) HbF (%) Total hemoglobin (Hb) (g/dL) Hematological Indices WBC (× 109/L) Platelets (×109/L) Hematocrit (%) Mean corpuscular volume (fL) Mean corpuscular Hb (pg) Mean corpuscular Hb Conc. (g/dL) Reticulocytes (%) CRP median (range) a b
a
Steady-state group
a
P
b
10.1 ± 5.7 70: 34
12.5 ± 12.8 20: 20
0.118 0.268
71.7 ± 9.6 19.1 ± 7.3 9.2 ± 1.4
69.0 ± 8.3 23.1 ± 8.1 9.5 ± 1.3
0.138 0.008 0.252
9.7 ± 5.0 381.6 ± 233.4 29.6 ± 4.6 76.8 ± 10.2 26.1 ± 4.2 33.6 ± 2.7 5.9 ± 3.8 31.3 (1.14–363.0)
9.1 ± 4.0 274.5 ± 168.6 27.7 ± 6.4 75.3 ± 8.3 25.6 ± 3.4 34.0 ± 1.6 5.8 ± 4.0 5 (0.16–185.0)
0.590 0.032 0.268 0.476 0.596 0.451 0.937 3.7 × 10− 6
Study subjects comprised 104 SCA patients who had any VOC event during study (VOC group) and 40 SCA patients who had no VOC events (Steady-state group). Pearson's chi square test (categorical variables), 2-tailed t-test (continuous variables).
F.A. Mohammed et al. / Blood Cells, Molecules, and Diseases 45 (2010) 293–296 Table 2 Influence of CRP cut-off Levels on VOC Risk. Cut-off (percentile)
Patients
P25 P50 P75 P90 P95
8 29 38 18 11
a
(7.7)a (27.9) (36.5) (17.3) (10.6)
295
Table 3 Correlation between CRP Levels and VOC outcomes.
Controls
P
OR
95% CI
Outcome
r2a
P
25 (53.2) 9 (19.1) 10 (21.3) 2 (4.3) 1 (2.1)
b 0.001 b 0.001 b 0.001 b 0.001 0.002
1.00 10.1 11.9 28.1 34.4
(Reference) 3.4–30.0 4.1–34.2 5.3–148.5 3.8–309.1
Frequency (episodes/month) VOC type (generalized vs. local) Pain scale (1–10) Age at VOC onset (years) Need for hospitalization VOC treatment (NSAIDs and/or narcotics) Hydroxyurea treatment HbS (%) HbF (%)
0.422 0.451 0.317 − 0.043 0.191 0.151 − 0.055 0.017 − 0.022
2.9 × 10− 6 1.0 × 10− 6 1.6 × 10− 4 0.622 0.024 0.071 0.511 0.854 0.811
Number of subjects (percent of total).
of VOC episodes (Fig. 1). Area-under-ROC curves provided good discriminatory power for hs-CRP and VOC episodes, and showed that varied hs-CRP levels revealed similar sensitivities and specificities for predicting the frequency and severity of VOC. The Spearman's correlation coefficient between hs-CRP and VOC was 0.65 (P b 0.001) among unselected patients, 0.71 (P b 0.001) in males and 0.59 (P b 0.001) in females. The area under ROC curves of hs-CRP was 0.90 (95% CI = 0.85–0.94) among unselected patient, 0.94 (95% CI = 0.89–0.98) for males, and 0.85 (95% CI = 0.77–0.93) for females. hs-CRP levels were then categorized into 6 strata: percentiles 1–25 (P25), 26–50 (P50), 51–75 (P75), 76–90 (P90) and N90 (P95), according to concentrations present in the control group, and analyzed in regression models, first at the univariate and later at the multivariate levels. Univariate regression analysis demonstrated a positive dose–effect relationship for hs-CRP with VOC, with increased VOC risk seen with increased hs-CRP levels (Table 2). The strongest OR was for P90 and P95 hs-CRP percentiles, in which they were associated with a 28.1-fold and 34.4-fold higher risk than P25, respectively (Table 2). Spearman correlation calculation demonstrated positive correlation between increased hs-CRP levels and VOC frequency (number of episodes/year; r2 = 0.422; P b 0.001), VOC type (localized vs. generalized; r2 = 0.451; P b 0.001), pain scale (r2 = 0.317; P b 0.001), and with the need for hospitalization (r2 = 0.191; P = 0.024). However, hs-CRP levels did not correlate with either HbS (P = 0.854) and HbF (P = 0.811) levels, or with with the age at VOC onset (P = 0.622), or with pain treatment regimen (P = 0.071) and hydroxyurea treatment (P = 0.511) (Table 3). On multivariate logistic regression analyses with SCA steady-state patients as the reference group, increased hs-CRP levels found to be independent risk factors for VOC, which was apparent at P50 (aOR = 9.5; 95% CI = 2.2–41.6), and increased at P75 (aOR = 12.8; 95% CI = 1.2–135.2), but later stabilized at P90 (aOR = 15.4; 95% CI = 2.5–94.7) and P95 (aOR = 15.5; 95% CI = 3.7–64.7) (Table 4).
a
Spearman correlation coefficients.
None of the other variables entered in the model (HbS, HbF, paletelet and WBC count) was found to be associated with VOC (Table 4). Discussion This study of Bahraini children with VOC demonstrated positive association between increased hs-CRP levels and the development of VOC and its severity, and thus the extent of microvessel occlusion in SCA. This was highlighted by the enrichment of high hs-CRP percentiles in VOC than in control SCA patients, and with the positive correlation between hs-CRP levels and the severity, frequency, and type of VOC, as well as the need for hospitalization. Logistic regression analysis confirmed the association of hs-CRP with VOC after controlling for potential confounders, thereby demonstrating significant contribution of hs-CRP, a well-established marker of inflammation [13,16], to VOC pathogenesis. In addition to the prevailing notion that VOC painful crisis is attributed to the occlusion of the small blood vessels by ssRBC leading to development of ischemia, increasing number of studies suggests a role for inflammation in the pathophysiology of VOC. This was exemplified by the heightened pro-inflammatory cytokine expression detected in the sera of SCA patients, both during steady-state and VOC episodes [18,19], which in turn elicits the inflammatory response accompanying VOC highlighted by elevation in hs-CRP levels [11,22]. In this study, hs-CRP levels in steady-state SCA patients (median = 5.0 mg/L) were higher than healthy non-SCA controls (normal range: 0–3 mg/L), suggesting a covert inflammatory response in SCA, even in the absence of crisis. This was in agreement with studies on pediatric [11] and adult [23,24] SCA patients, in which hs-CRP levels were elevated under steady-state conditions, supporting the contribution of a state of chronic inflammation to SCA [25]. It was proposed that intracellular hemoglobin polymerization, a hallmark of SCA, induces oxidative damage in the cell membrane membrane in SCA [25,26], which induces elevation in CRP due to blocked vascular endothelium by ssRBC, as was suggested [17]. Given the influence of age on the perception of and reaction to painful episodes, we focused on pediatric SCA patients, age 1.5– 16 years, with age- and ethnically-matched SCA patients without controls. Insofar as SCA phenotype influences the course of the Table 4 Logistic regression analysis for association of CRP with VOC. Cut-off (percentile) hs-CRP
Fig. 1. The ROC curve of serum levels of hs-CRP for evaluation of VOC. The Spearman's correlation coefficient between hs-CRP and VOC was 0.65 (P b 0.001) among unselected patients, and the area under ROC curves of hs-CRP was 0.90 (95% CI = 0.85–0.94) among unselected patients.
b
P50 P75 P90 P95
HbS (%) HbF (%) Platelets WBC a b
aOR = adjusted odds ratio. Adjusted for HbF levels and platelet count.
aORa
95% CI
9.5 12.8 15.4 15.5 0.98 0.95 1.00 1.05
2.2–41.6 1.2–135.2 2.5–94.7 3.7–64.7 0.90–1.08 0.88–1.03 0.99–1.01 0.90–1.23
296
F.A. Mohammed et al. / Blood Cells, Molecules, and Diseases 45 (2010) 293–296
disease, all patients and controls were homozygous HbSS/HbSβ0Thal, which may explain the higher baseline hs-CRP levels than healthy non-HbSS individuals, as was suggested [11,25], further supporting the notion of ongoing subclinical inflammation in HbSS/HbSβ0Thal patients, which predisposes them at higher risk for future VOC episodes [27]. More significantly, hs-CRP levels showed a strong statistical association by appropriate regression analysis models with significant VOC endpoints, namely VOC frequency, pain scale, VOC site, and increased hospitalizations for pain, in agreement with the study of Krishnan in which increased hs-CRP was also linked with increased need for hospitalization [11,28]. In the VOC group, hs-CRP showed excellent correlation with abnormality in the liver function tests ALT (r2 = 0.283; P = 0.014), LDH (r2 = 0.446; P = 0.025), alkaline phosphatase (r2 = 0.240; P = 0.041), and gamma-GT (r2 = 0.257; P = 0.034); no similar association was recorded for the steady-state control group. This underscores the intersection of altered liver function activity and inflammation with pathological manifestations of SCA, and makes hsCRP a reliable marker of SCA-related complications, including VOC, since it reflects not only the inflammation, but also other aspects of endothelial and coagulation dysfunctions linked with SCA [18,28]. While NSAIDs are the mainstay for VOC management [20], hs-CRP did not correlate with VOC treatment. This was in apparent disagreement with an earlier report suggesting that serial determinations of CRP may be helpful in monitoring the response to treatment in SCA [20]. In summary, our findings suggest that low-grade inflammation, highlighted by increased levels of the acute phase protein hs-CRP, accompanies steady-state SCA. Further increases in hs-CRP levels are associated with VOC episodes, thereby underscoring the clinical significance of hs-CRP as an important predictor of VOC. A predictive value of hs-CRP for VOC has previously been reported also for thalassemias [29,30]. It should be noted that elevation in hs-CRP is an indirect measurement of inflammation, and thus our results must be interpreted within the limitation of the experimental design. Another limitation lies in the difficulty of identifying the exact mechanism or pathway that triggers of elevated hs-CRP. Despite these shortcomings, our results support for the clinical notion for the role of a covert inflammatory response in SCA, and hence will be instrumental in the future management strategies of VOC in SCA patients. Conflict of interest disclosure The authors declare no competing financial interests. References [1] N. Conran, C.F. Franco-Penteado, F.F. Costa, Newer aspects of the pathophysiology of sickle cell disease vaso-occlusion, Hemoglobin 33 (2009) 1–16. [2] A.S. Mehanna, Sickle cell anemia and antisickling agents then and now, Curr. Med. Chem. 8 (2001) 79–88. [3] G.R. Buchanan, M.R. DeBaun, C.T. Quinn, M.H. Steinberg, Sickle cell disease, hematology, Am. Soc. Hematol. Educ. Program (2004) 35–47. [4] M.J. Stuart, R.L. Nagel, Sickle-cell disease, Lancet 364 (2004) 1343–1360. [5] P.S. Frenette, G.F. Atweh, Sickle cell disease: old discoveries, new concepts, and future promise, J. Clin. Invest. 117 (2007) 850–858.
[6] D.K. Kaul, E. Finnegan, G.A. Barabino, Sickle red cell–endothelium interactions, Microcirculation 16 (2009) 97–111. [7] R. Zennadi, A. Chien, K. Xu, M. Batchvarova, M.J. Telen, Sickle red cells induce adhesion of lymphocytes and monocytes to endothelium, Blood 112 (2008) 3474–3483. [8] J.D. Belcher, C.J. Bryant, J. Nguyen, P.R. Bowlin, M.C. Kielbik, J.C. Bischof, R.P. Hebbel, G.M. Vercellotti, Transgenic sickle mice have vascular inflammation, Blood 101 (2003) 3953–3959. [9] C. Lanaro, C.F. Franco-Penteado, D.M. Albuqueque, S.T. Saad, N. Conran, F.F. Costa, Altered levels of cytokines and inflammatory mediators in plasma and leukocytes of sickle cell anemia patients and effects of hydroxyurea therapy, J. Leukoc. Biol. 85 (2009) 235–242. [10] O.S. Platt, Sickle cell anemia as an inflammatory disease, J. Clin. Invest. 106 (2000) 337–338. [11] S. Krishnan, Y. Setty, S.G. Betal, V. Vijender, K. Rao, C. Dampier, M. Stuart, Increased levels of the inflammatory biomarker C-reactive protein at baseline are associated with childhood sickle cell vasocclusive crises, Br. J. Haematol. 148 (2010) 797–804. [12] A.M. Dupuy, N. Terrier, L. Sénécal, M. Morena, H. Leray, B. Canaud, J.P. Cristol, Is Creactive protein a marker of inflammation? Nephrologie 24 (2003) 337–341. [13] J. Genest, C-reactive protein: risk factor, biomarker and/or therapeutic target? Can. J. Cardiol. 26 (Suppl A) (2010) 41A–44A. [14] V.B. Martínez, J.R. González-Juanatey, Markers of inflammation and cardiovascular disease: clinical applications of C-reactive protein determination, Am. J. Cardiovasc. Drugs 9 (Suppl 1) (2009) 3–7. [15] E.M. Macy, T.E. Hayes, R.P. Tracy, Variability in the measurement of C-reactive protein in healthy subjects: implications for reference intervals and epidemiological applications, Clin. Chem. 43 (1997) 52–58. [16] M.B. Pepys, G.M. Hirschfield, C-reactive protein: a critical update, J. Clin. Invest. 111 (2003) 1805–1812. [17] S.A. Akohoue, S. Shankar, G.L. Milne, J. Morrow, K.Y. Chen, W.U. Ajayi, M.S. Buchowski, Energy expenditure, inflammation, and oxidative stress in steadystate adolescents with sickle cell anemia, Pediatr. Res. 61 (2007) 233–238. [18] K.L. Bourantas, G.N. Dalekos, A. Makis, A. Chaidos, S. Tsiara, A. Mavridis, Acute phase proteins and interleukins in steady state sickle cell disease, Eur. J. Haematol. 61 (1998) 49–54. [19] A. Pathare, S.A. Kindi, S. Daar, D. Dennison, Cytokines in sickle cell disease, Hematology 8 (2003) 329–337. [20] D. Monnet, I. Diallo, A. Sangare, A.E. Yapo, Clinical value of C-reactive protein, alpha 1-glycoprotein acid and transferrin assay in homozygous sickle cell disease, Bull. Soc. Pathol. Exot. 86 (1993) 282–285. [21] E. Perlin, H. Finke, O. Castro, S. Rana, J. Pittman, R. Burt, C. Ruff, D. McHugh, Enhancement of pain control with ketorolac tromethamine in patients with sickle cell vaso-occlusive crisis, Am. J. Hematol. 46 (1994) 43–47. [22] A. Singhal, J.F. Doherty, J.G. Raynes, K.P. McAdam, P.W. Thomas, B.E. Serjeant, G.R. Serjeant, Is there an acute-phase response in steady-state sickle cell disease? Lancet 341 (1993) 651–653. [23] A.C. Makis, E.C. Hatzimichael, J. Stebbing, K.L. Bourantas, C-reactive protein and vascular cell adhesion molecule-1 as markers of severity in sickle cell disease, Arch. Int. Med. 166 (2006) 366–368. [24] J.S. Mohan, G.Y. Lip, J. Wright, D. Bareford, A.D. Blann, Plasma levels of tissue factor and soluble E-selectin in sickle cell disease: relationship to genotype and to inflammation, Blood Coagul Fibrinolysis 16 (2005) 209–214. [25] J. Stuart, P.C.W. Stone, N.O. Akinola, J.R. Gallimore, M.B. Pepys, Monitoring the acute phase response to vaso-occlusive crisis in sickle cell disease, J. Clin. Pathol. 47 (1994) 166–169. [26] M. Aslan, D. Thornley-Brown, B.A. Freeman, Reactive species in sickle cell disease, Ann. N.Y. Acad. Sci. 899 (2000) 375–391. [27] N.O. Akinola, S.M. Stevens, I.M. Franklin, G.B. Nash, J. Stuart, Subclinical ischaemic episodes during the steady state of sickle cell anaemia, J. Clin. Pathol. 45 (1992) 902–906. [28] C.C. Hedo, Y.A. Aken'ova, I.E. Okpala, A.O. Durojaiye, L.S. Salimonu, Acute phase reactants and severity of homozygous sickle cell disease, J. Intern. Med. 233 (1993) 467–470. [29] N. Archararit, S. Chuncharunee, A. Pornvoranunt, K. Atamasirikul, B. Rachakom, V. Atichartakarn, Serum C-reactive protein level in postsplenectomized thalassemic patients, J. Med. Assoc. Thai. 83 (Suppl 1) (2000) S63–S69. [30] I. Kanavaki, P. Makrythanasis, C. Lazaropoulou, M. Tsironi, A. Kattamis, I. Rombos, I. Papassotiriou, Soluble endothelial adhesion molecules and inflammation markers in patients with beta-thalassemia intermedia, Blood Cells Mol. Dis. 43 (2009) 230–234.