Thrombosis Research 133 (2014) 334–339
Contents lists available at ScienceDirect
Thrombosis Research journal homepage: www.elsevier.com/locate/thromres
Regular Article
Red cell distribution width and risk for venous thromboembolism: A population-based cohort study Bengt Zöller a,⁎, Olle Melander b, Peter Svensson b,c, Gunnar Engström b a b c
Center for Primary Care Research, Lund University/Region Skåne, Malmö, Sweden Department of Clinical Sciences, Lund University, Clinical Research Center (CRC), Entrance 72, house 60, level 13, Skåne University Hospital, SE 205 02 Malmö, Sweden Department of Coagulation Disorders, Skåne University Hospital, Malmö, Sweden
a r t i c l e
i n f o
Article history: Received 18 September 2013 Received in revised form 10 December 2013 Accepted 11 December 2013 Available online 15 December 2013 Keywords: venous thromboembolism venous thrombosis pulmonary embolism epidemiology predictors
a b s t r a c t Introduction: Red cell distribution width (RDW) has been associated with venous thromboembolism (VTE), but whether RDW is a predictor of first event of VTE is unknown. We investigated the association between RDW and incidence of first event of VTE in a population-based cohort. Materials and Methods: RDW was measured in 27 042 subjects (aged 45–73 years, 60.6% women), without previous history of VTE or cancer within 5 years before follow-up, who participated in the Malmö Diet and Cancer study during 1991–1996. Incidence of VTE was identified from the patient register and the cause of death register during a mean follow-up of 13.8 years and studied in relation to RDW. Results: During follow-up, 991 subjects (57.5% women) were affected by VTE (pulmonary embolism or deep venous thrombosis of the lower limbs). After adjustment for potential confounding factors the hazard ratios (HR) for VTE for the second, third and fourth RDW quartiles 1.15 (95% confidence interval 0.94–1.41), 1.41 (1.14–1.73), 1.74 (1.38–2.21), respectively, were compared with the bottom quartile of RDW. In the multivariate model subjects with the top 5% of RDW values compared with the bottom quartile had an even higher risk (HR = 2.51, 1.78–2.54). In receiver operating characteristic (ROC) analysis, the male specific area under the ROC curve (AUC) for RDW was 0.57 (95% CI 0.54–0.59). The female specific AUC was 0.56 (95% CI 0.53–0.58). Conclusions: RDW was found to be associated with long-term incidence of first event of VTE among middle-aged subjects. © 2013 Elsevier Ltd. All rights reserved.
Introduction Venous thromboembolism (VTE) is the third most common cause of death from cardiovascular disease after coronary heart disease and stroke [1]. Although a number of acquired and genetic risk factors for VTE have been identified, in 30–50% of cases VTE is idiopathic [1]. During the last few decades research has therefore concentrated on identifying new VTE risk factors and creating predictive models for VTE [2]. Erythrocytes, or red blood cells (RBCs), are constituents in clots and thrombi formed in vivo [3]. RBCs play a prothrombotic role in blood coagulation by increasing blood viscosity and forcing platelets towards the vessel wall [3]. Incorporation of RBCs into a fibrin clot affects clot structure and mechanical properties [3]. Even small structural differences of RBCs may have a large influence on pathophysiology [3]. Moreover, RBCs actively participate in thrombin generation [4]. A subAbbreviations: CI, confidence interval; HR, Hazard ratio; RDW, Red cell distribution width; VTE, venous thromboembolism; DVT, deep venous thrombosis; PE, pulmonary embolism; MDC, Malmö Diet and Cancer study; CHD, coronary heart disease. ⁎ Corresponding author at: Center for Primary Health Care Research, CRC, building 28, floor 11, Jan Waldenströms gata 35, Skåne University Hospital, S-205 02 Malmö, Sweden. Tel.: +46 40 391954; fax: +46 40 391370. E-mail address: [email protected] (B. Zöller). 0049-3848/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.thromres.2013.12.013
fraction of red blood cells expresses phosphatidylserine on their surface. Unlike platelets, RBCs produce thrombin through the meizothrombin pathway, which has consequences in the context of clot formation and stabilization [4]. Polycythemia Vera has been associated with VTE [5], and hematocrit variations in the general population have been associated with VTE [6]. An increased focus on RBCs may therefore be justified, and may reveal novel mechanisms and risk factors for VTE. Red cell distribution width (RDW) is a measure of the size variation as well as an index of the heterogeneity of the erythrocytes (i.e. anisocytosis) [7–9]. RDW is part of routine hematology laboratory tests and is used for classification of anemia [7–9]. Recent studies have shown that RDW is associated with increased mortality in a number of cardiovascular disorders such as coronary artery disease, stroke, peripheral artery disease, heart failure, pulmonary embolism, and pulmonary arterial hypertension, reviewed by Montagnana et al. [10]. It is however unclear whether anisocytosis might be the cause, or a simple epiphenomenon due to conditions such as inflammation, impaired kidney function, malnutrition, or oxidative damage [10]. Two studies have determined RDW in patients with pulmonary embolism (PE) [11,12]. Both studies found that a high RDW level, i.e. anisocytosis, was an independent predictor of early PE-related mortality [11,12]. Recently, RDW has been associated with venous thromboembolism (VTE) in two case– control studies [13,14]. These case–control studies cannot exclude the
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339
335
possibility that anisocytosis was the result of the thrombotic event itself. To the best of our knowledge, there are no prospective studies on the association between RDW and risk of first event of VTE, i.e. deep venous thrombosis (DVT) and PE. The purpose of the present study was to determine whether RDW is an independent predictor of first event of VTE in asymptomatic middleaged subjects in the Malmö Diet and Cancer Study (MDC), which is a prospective cohort study in Malmö.
consecutively in fresh heparinized blood. Erythrocyte diameter was measured using a fully automated assay (SYSMEX K1000). Red cell distribution width was calculated as the width of the erythrocyte distribution curve at a relative height of 20% above the baseline. The intra-assay CV was 2.2%. Reference values were 36.4–46.3 fL in women and 35.1–43.9 fL in men [24].
Materials and Methods
All subjects were followed from the baseline examination until a first event of VTE, emigration, death or end of follow-up (December 31, 2008), whichever came first. VTE was defined as International Classification of Diseases – 8th revision (ICD-8 used before 1987), code 450 (PE), and 451 (DVT of the lower limbs) (ICD-9 used 1987–1996), code 415B (PE), and 451 (DVT of the lower limbs); and (ICD-10 used 1997–2008) I26 (PE), and I80 (DVT of the lower limbs) as the primary diagnosis. Patients with superficial thrombophlebitis were excluded (ICD-9 code 451A and ICD-10 code I80.0). The inpatient register, the outpatient register and the cause of death register were used to identify cases with VTE. All these registers have 100% coverage in Malmö during the whole follow-up time. A validation study has shown that a diagnosis of VTE in the SNHDR has a validity of 95% [25]. Such high validity has also been found for other cardiovascular disorders such as myocardial infarction (94%), stroke, heart failure, and atrial fibrillation [21,23,26–28]. The overall validity of the SNHDR is 87% [29]. A previous population-based prospective study was performed at the Malmö University Hospital in 1998–2006, of all in- and outpatients diagnosed with VTE in Malmö [30], i.e. at the same study center as the present study and with partially overlapping study periods. Among 1140 VTE patients, virtually all were diagnosed with an objective method such as phlebography, ultrasound or computer tomography [30]. Similar results were reported in another Swedish study that found that most Swedish patients with VTE are objectively diagnosed [31].
Study Population The Malmö Diet and Cancer study (MDC) is a prospective cohort study from the city of Malmö in the south of Sweden. Sample characteristics, data collection, and clinical definitions for MDC have been described previously [15–22]. Briefly, 28 449 men (n = 11 246, born 1923–1945) and women (n = 17 203, born 1923–1950) attended a baseline examination between March 1991 and September 1996. Participants underwent sampling of peripheral venous blood, measurement of blood pressure and anthropometric measures and filled out a self-administered questionnaire. Information on RDW was available in 28 363 subjects [19]. Subjects with previous VTE (n = 360 subjects) at the baseline examination were excluded. Patients with diagnosis of cancer within 5 years before baseline were excluded (n = 448). Of the remaining 27 563 subjects (8 patients were diagnosed with both VTE and cancer before baseline), 521 subjects were excluded due to missing information on blood pressure, smoking habits, alcohol consumption, education level, civil status, hemoglobin value, and body mass index (BMI). Thus, the final study population in the analysis consisted of 27 042 (10 660 [39.4%] men and 16 382 [60.6%] women) subjects, aged 45–73 years.
Ascertainment of Venous Thromboembolic Events
Measurements and Definitions Statistical Analysis Information on current use of statins, blood-pressure-lowering and anti-diabetic medications, physician-treated cancer, smoking habits, alcohol consumption, leisure-time physical activity, education level, and civil status were obtained from a self-administered questionnaire [15–22]. Weight and height were measured to the nearest 0.1 kg and 0.5 cm, respectively, with subjects wearing light clothing and no shoes. Current BMI was calculated as weight in kilograms divided by height in meters squared (kg/m2). Blood pressure was measured using a mercury-column sphygmomanometer after 10 min of rest in the supine position. Hyperglycemia was defined as fasting whole blood glucose level greater than 109 mg/dL (i.e. 6.0 mmol/L), self-reported physician’s diagnosis of diabetes or use of anti-diabetic medications. History of atrial fibrillation and also history of cardiovascular disease (CVD) at baseline was defined as diagnosis of atrial fibrillation and coronary heart disease (CHD) or stroke, respectively, in the Swedish National Hospital Discharge Register (SNHDR) [15,21,23]. Cancer during follow-up among individuals with first event of VTE (during followup) was obtained from Swedish Cancer Register. Subjects were categorized as current smokers (i.e. those who smoked regularly or occasionally) or non-smokers (i.e. former smokers and never smokers). High alcohol consumption was defined as N 40 g alcohol per day for men and N30 g per day for women [15,19]. High education level was defined as completed secondary school and at least one year of education from college, university or higher (e.g. N12 years education), modified from reference [19] and [21]. Civil status was categorized into married or not [20]. Low level of physical activity was defined as the lowest tertile of a score revealed through 18 questions covering a range of activities in the four seasons. The evaluation of the questionnaire has been previously reported [19,22]. Red cell distribution width, mean corpuscular volume (MCV), hemoglobin, platelet and leucocyte concentrations were analyzed
Detailed information on the statistical analysis is found in the supplementary material. P-values were calculated with two-sided Student’s ttest for continuous variables and with two-sided Fischer’s exact test for dichotomized variables. Confidence interval for incidence rates were calculated with the OpenEpi version 2.3.1 according to Rothman/Greenland as described [32,33]. Cox proportional hazards regression was used to examine the association between RDW and incidence of VTE. Hazard ratios (HR) with 95% confidence interval (CI) were calculated [32]. The area under the ROC (receiver operating characteristic) curve (AUC) was used as a measure of the overall performance of the ROC curve because it reflects the probability that the diagnostic test will classify correctly [34]. Positive [sensitivity/(1-specificity)] and negative [(1-sensitivity)/ specificity] likelihood ratios were also calculated to express the odds that a given value of a screening test outcome would be expected in a subject with or without VTE, respectively. Analyses were performed using IBM SPSS 21 (IBM, Armonk, New York, USA). Ethical Approval All participants provided written informed consent, and the study was approved by the ethics committee at Lund University, Lund, Sweden (LU 51/90). The MDC is registered in the US Library of Medicine as trial number NCT 01216228 (http://www.clinicaltrials.gov). Results Baseline Characteristics Cardiovascular risk factors at the baseline examination for the middle-aged MDC cohort in relation to the sex-specific quartiles of
336
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339
RDW have previously been published (not shown) [19]. Instead, cardiovascular risk factors at the baseline examination in relation to incident VTE during follow-up and sex between 1991 and 2008 are presented in Table 1 and Supplementary Table S1, respectively. Mean baseline RDW (± standard deviation, SD) was 40.5 (± 3.6) in men and 40.9 (±3.3) in women (Supplementary Table S1). Totally 991 (3.7%) of the included 27 042 individuals were affected by VTE during follow-up. Of hematological parameters at baseline, VTE during follow-up was associated with increased RDW, hemoglobin, and mean corpuscular volume (MCV) (Table 1). Sex, age, BMI, systolic blood pressure, diastolic blood pressure, hyperglycemia, use of blood pressure medication, low educational level, history of coronary heart disease (CHD) or stroke, and history of atrial fibrillation were associated with VTE during follow-up (Table 1). Incidence of First Event of VTE in Relation to Red Cell Distribution width and Known Cardiovascular Risk Factors During a median follow-up of 14.2 years (interquartile range 12.9–15.7 years), a total of 991 (3.7%) individuals were affected by VTE. The sum of follow-up time was 370 468 years corresponding to a VTE incidence rate of 2.7 (95% CI 2.5–2.8) per 1000 person years. Among the 991 VTE patients during follow-up 571 (57.6%) VTE events were coded as DVT, 382 (38.5%) as PE and 38 (3.8%) as both DVT and PE. The mean time between baseline and first event of VTE was 8.7 years (±4.2 SD). The RDW variable was divided into sex-specific quartiles as previously described (Table 2) [19]. The number of VTE cases in each quartile was 186 (2.8%) for the lowest quartile (Q1), 222 (3.3%) for the second quartile (Q2), 270 (3.9%) for the third quartile (Q3), and 313 (4.7%) for the highest quartile (Q4). The corresponding VTE incidence rates per 1000 person years were 2.0 (95% CI 1.7–2.3), 2.4 (95% CI 2.1–2.7), 2.9 (95% CI
Table 1 Baseline characteristics of subjects enrolled in the MDC study 1991–96 according to incident VTE during follow-up 1991–2008.
Sex (men/women) Age (years) Smoking (current, yes/no) Body mass index (kg/m2) Hemoglobin (g/L) Mean corpuscular volume (fL) Red cell distribution width Leucocyte count (109/L) Platelet count (109/L) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Hyperglycemia (yes/no) Blood pressure medication (yes/no) Statin medication (yes/no) Low educational level (yes/no) Unmarried (yes/no) High alcohol consumption (yes/no) Stroke or CHD (yes/no) Atrial fibrillation (yes/no)
No VTE (n = 26 051)
Incident VTE (n = 991)
10239/15812 (39%/61%) 58.0 ± 7.6 7390/18661 (28%/72%) 25.7 ± 3.9 141.7 ± 12.0 89.4 ± 4.2 40.7 ± 3.4 6.4 ± 2.3 230.4 ± 59,0 141.03 ± 19.98 85.5 ± 9.9 766/25285 (3%/97%) 4588/21463 (18%/82%) 558/25493 (2%/98%) 19982/6069 (77%/23%) 9019/17032 (35%/65%) 1134/24917 (4%/96%) 750/25301 (3%/97%) 242/25809 (1%/99%)
421/570 (42%/58%) 61.1 ± 7.4 277/714 (28%/72%) 27.0 ± 4.3 143.3 ± 12.2 89.8 ± 4.2 41.4 ± 3.3 6.6 ± 2.4 230.42 ± 78.99 145.2 ± 19.8 86.8 ± 9.9 46/945 (5%/95%) 231/760 (23%/77%) 20/971 (2%/98%) 816/175 (82%/18%) 364/627 (37%/63%) 33/958 (3%/97%) 44/947 (4%/96%) 19/972 (2%/98%)
p-value* 0.047 b0.001 0.800 b0.001 b0.001 0.003 b0.001 0.063 0.989 b0.001 b0.001 0.003 b0.001 0.898 b0.001 0.174
2.6–3.3), and 3.6 (95% CI 3.2–4.0), respectively. In the age- and sexadjusted model subjects in the two top quartiles (Q3 and Q4) compared to the bottom quartile (Q1) had significantly higher risk of VTE: HR 1.35 (95% CI 1.12–1.62) and HR 1.61 (95% CI 1.34–1.94), respectively. The risk for subjects in the second quartile (Q2) was not significantly increased, HR 1.13 (95% CI 0.93–1.38). In the age- and sex-adjusted model RDW, age, smoking, BMI, hemoglobin, MCV, leucocyte count, diastolic blood pressure, hyperglycemia, use of blood pressure medication, low educational level, being unmarried, history of CHD or stroke, and history of atrial fibrillation were all associated with an increased risk of VTE (Table 2). In the multivariate model subjects in the top (Q4) compared to the bottom quartile (Q1) of RDW had a significantly higher risk of VTE, HR 1.74 (95% CI 1.38–2.21) (Table 2 and Fig. 1). Subjects in Quartile 3 (Q3) had also a significantly higher risk of VTE, HR 1.41 (95% CI 1.14–1.73), while the risk for subjects in the second quartile (Q2) was not significantly increased HR 1.15 (95% CI 0.94–1.41). No other hematological variable than RDW was associated with an increased risk in the multivariable model. In the final multivariable model RDW, age, BMI, hyperglycemia, and being unmarried were independently associated with an increased risk of VTE (Table 2). In the multivariate model subjects with the top 5% RDW values (≥46.6 fL) compared with the bottom quartile had an even higher risk of VTE (HR 2.51, 95% CI 1.78–2.54, P b 0.001). If RDW was entered as a continuous variable into the final multivariable model the HR per fL was 1.05 (1.03–1.07, P b 0.001). No significant interaction was observed between RDW and other risk factors (age, BMI, hyperglycemia, unmarried) or hematological parameters for incidence of VTE (data not shown in Table). In a subanalysis we also determined the risk of DVT and PE separately. In the multivariable model the HR for PE for the quartiles Q2, Q3, and Q4 compared to Q1 was 1.29 (95% CI 0.95–1.75), 1.40 (95% CI 1.02–1.93) and 1.61 (95% CI 1.11–2.32), respectively. In the multivariable mode the HR for DVT for the quartiles Q2, Q3, and Q4 compared to Q1 was 1.05 (95% CI 0.81–1.36), 1.40 (95% CI 1.08–1.81) and 1.87 (95% CI 1.39–2.51), respectively. RWD, Sex and ROC Analysis No interaction was observed between RDW and sex. We nevertheless performed stratified analysis according to sex. In the multivariate model male subjects in the two top quartiles (Q3 and Q4) compared to the bottom quartile (Q1) had significantly higher risk of VTE: HR 1.44 (95% CI 1.04–1.99) and HR 1.70 (95% CI 1.16–2.47), respectively. The risk for male subjects in the second quartile (Q2) was not significantly increased, HR 1.13 (95% CI 0.87–1.67). In the multivariate model female subjects in the two top quartiles (Q3 and Q4) compared to the bottom quartile (Q1) had significantly higher risk for VTE: HR 1.40 (95% CI 1.07–1.83) and HR 1.79 (95% CI 1.32–2.42), respectively. The risk for female subjects in the second quartile (Q2) was not significantly increased, HR 1.13 (95% CI 0.87–1.47). The calculated male specific AUC for RDW was 0.57 (95% CI 0.54–0.59, p b 0.001). The female specific AUC for RDW was 0.56 (95% CI 0.53–0.58, p b 0.001). Cut-offs were derived from ROC analysis to maximize the sum of sensitivity and specificity. In males, a RDW cut-off point of 40.15 fL had sensitivity of 0.60 and specificity of 0.50, whereas in females a cut-off value of 40.55 fL had sensitivity of 0.58 and specificity of 0.50. The corresponding positive and negative likelihood ratios for both males and females were 1.2 and 0.8, respectively.
0.134
Sensitivity and Additional Analysis
0.005
In sensitivity analysis, inclusion of the 521 subjects with missing values in a sex- and age-adjusted model did not change the result: the HR for VTE for quartiles Q2, Q3, and Q4 compared to Q1 was 1.13 (95% CI 0.93–1.37), 1.35 (95% CI 1.12–1.62) and 1.60 (95% CI 1.34–1.92), respectively. We also excluded patients with personal history of atrial fibrillation, coronary heart disease, stroke, hyperglycemia, statin
0.004
Values are means with one standard deviation, or percentages with numbers in brackets. *P-values calculated with two-sided Student’s t-test for continuous variables and with two-sided Fischer’s exact test for dichotomized variables.
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339
337
Table 2 Age- and sex-adjusted and multivariable hazard ratios (HR) with 95% confidence interval (CI) and p-values for first event of VTE during follow-up in the MDC cohort 1991–2008. Age- and sex- adjusted HR (95% CI) RDW Q1 (men b38.2 fL, women b38.6 fL, n = 6689) RDW Q2 (men 38.2–40.0 fL, women 38.6–40.5 fL, n = 6782) RDW Q3 (men 40.1–42.4 fL, women 40.6–42.7 fL, n = 6868) RDW Q4 (men N42.4 fL, women N42.7 fL, n = 6703) Male sex Age (per 1 year) Smoking (current) Body mass index (per kg/m2) Hemoglobin (per 1 g/L) MCV (per fL) Leucocyte count (per 109/L) Platelet count (per 109/L) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Hyperglycemia (yes/no) Blood pressure medication (yes/no) Statin medication (yes/no) Low educational level (yes/no) Unmarried (yes/no) High alcohol consumption (yes/no) Stroke or CHD (yes/no) Atrial fibrillation (yes/no)
reference 1.13 (0.93–1.38) 1.35 (1.12–1.62) 1.61 (1.34–1.94) 1.10 (0.97–1.25) 1.06 (1.06–1.07) 1.21 (1.05–1.39) 1.07 (1.06–1.09) 1.01 (1.00–1.01) 1.02 (1.01–1.04) 1.02 (1.01–1.03) 1.00 (1.00–1.00) 1.00 (1.00–1.01) 1.01 (1.00–1.01) 1.54 (1.15–2.08) 1.23 (1.06–1.43) 0.81 (0.52–1.26) 1.19 (1.01–1.40) 1.19 (1.04–1.36) 0.88 (0.62–1.25) 1.41 (1.04–1.92) 1.79 (1.13–2.82)
medication, and hypertension medication at baseline. The results obtained in a multivariate model were not changed to any major degree: the HR for VTE for the quartiles Q2, Q3, and Q4 compared to Q1 was 1.16 (95% CI 0.92–1.47), 1.39 (95% CI 1.09–1.76) and 1.77 (95% CI 1.34–2.33), respectively. We also performed stratified analysis with regard to follow-up time. In a multivariate model for the first 8 years of follow-up the HR for VTE for the RDW quartiles Q2, Q3, and Q4 compared to Q1 was 1.22 (95% CI 0.89–1.66), 1.44 (95% CI 1.05–1.99) and 1.60 (95% CI 1.11–2.31), respectively. In the multivariate model with follow-up between 8 and 18 years the HR for VTE for the RDW quartiles Q2, Q3, and Q4 compared to Q1 was 1.11 (95% CI 0.86–1.45), 1.39 (95% CI 1.06–1.81) and HR 1.86 (95% CI 1.37–2.53), respectively. Since cancer is a major risk factor for VTE, we performed a subanalysis in which all VTE cases with cancer prior to, or up to 180 days after the VTE event, were excluded. A total of 648 VTE cases remained in the analysis. In the multivariable model, the HR for VTE for quartiles Q2, Q3, and Q4 compared to Q1 was 1.14 (95% CI 0.89–1.45), 1.44 (95% CI 1.12–1.86) and 1.79 (95% CI 1.34–2.38), respectively. Physical activity data were available for 26 106 of the included individuals. Low physical activity was not associated with VTE in a sex- and age-adjusted model 1.02 (95% CI 0.88–1.18, p = 0.82), and when entered into the multivariate
Fig. 1. First VTE event free survival in relation to sex-specific quartiles (Q1-Q4) of RDW.
p-value
Multivariable HR (95% CI)
p-value
0.210 0.002 b0.001 0.132 b0.001 0.009 b0.001 0.025 0.009 0.002 0.126 0.062 0.041 0.004 0.007 0.341 0.042 0.009 0.485 0.028 0.012
reference 1.15 (0.94–1.41) 1.41 (1.14–1.73) 1.74 (1.38–2.21) 1.09 (0.93–1.27) 1.06 (1.04–1.07) 1.13 (0.97–1.31) 1.08 (1.06–1.09) 1.00 (0.99–1.01) 1.00 (0.98–1.02) 1.01 (0.99–1.03) 1.00 (1.00–1.00) 1.00 (1.00–1.01) 1.00 (0.99–1.01) 1.41 (1.04–1.91) 1.06 (0.90–1.24) 0.69 (0.44–1.09) 1.10 (0.94–1.31) 1.16 (1.02–1.32) 0.80 (0.56–1.14) 1.30 (0.95–1.79) 1.57 (0.99–2.49)
0.161 0.001 b0.001 0.277 b0.001 0.128 b0.001 0.837 0.688 0.520 0.330 0.668 0.805 0.026 0.513 0.110 0.220 0.027 0.220 0.103 0.056
model it did not affect the association between RDW and VTE (results not shown in Table).
Discussion Previous studies have found an association of RDW with early PErelated mortality in patients with acute pulmonary embolism [11,12]. RDW has also been associated with VTE in two recent case–control studies [13,14]. The present study is the first prospective cohort study to determine whether RDW is a predictor of first event of VTE. The present large, population-based cohort study shows a graded independent association between RDW and risk of first event of VTE among middle-aged subjects. This relationship was independent of potential confounders, including hematological parameters, multiple biological, lifestyle, and socioeconomic factors. The association was robust and remained significant in all different analysis performed. The results suggest that increased RDW is a novel independent risk factor for first event of VTE. The ROC analysis indicates that RDW as a sole marker is not useful for clinical prediction of VTE in the present study population. It is possible that RDW performs better as a prognostic marker in high-risk individuals such as patients with PE [11,12]. The mechanism underlying the relationship between RDW and VTE is unclear. Red cell distribution width has been associated with lowgrade systemic inflammation [35]. In a previous study, RDW was associated with leucocyte count and hsCRP, e.g. classic markers of inflammation [36]. In the present study the relationship between RDW and risk of VTE remained unchanged when leucocyte count was taken into account (Table 2). This suggests that inflammation is not the major mechanism for the increased incidence of VTE in the present study. However, adjustment for leukocyte count is not sufficient to draw firm conclusions on inflammation, but CRP data were not available. Red cell distribution width has been associated with several other risk factors, e.g. age, waist circumference, BMI, smoking, high alcohol consumption, diabetes, blood pressure, history of CHD, leucocyte count, and being unmarried [19,37]. RDW remained significant after adjustment for such risk factors, and there were no interactions between RDW and other risk factors or hematological parameters for the incidence of VTE. Malnutrition and deficiency of vitamin B12 and folic acid are other factors that have been associated with high RDW, because of their role in erythropoiesis [10]. We did not have information on plasma levels of these vitamins, and it therefore remains unclear whether these factors contributed to the relationship between RDW and incident VTE in our study.
338
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339
What is the relevance of an association between RDW and VTE in the present study and in the four previously published studies [11–14]? If RDW by itself is not the cause of the association with VTE, then RDW will just be another biomarker for VTE reflecting a moderate hypercoagulability. As RBCs are constituents of in vivo clots and thrombi it is possible that anisocytosis could increase the prothrombotic role of the RBCs [3]. Recently, a genome-wide association study among African Americans has identified two variants (rs1050828 and rs10493739) on chromosomes Xq28 (G6pD gene) and 1p31.1, respectively, that were associated with RDW [38]. Mendelian randomization studies may therefore be an important possibility in order to generate estimates for a causal effect of RDW on VTE [39]. It is worth pointing out the association between civil status and VTE risk confirming a recent study [40]. This is consistent with the growing literature about socioeconomic factors and VTE risk [25,40–42]. Moreover, an association between hyperglycemia was also observed in the present study, confirming previous studies [31,43–45].
Strengths and Limitations The study cohort included a large numbers of subjects and events during a long follow-up period [15–22]. The cardiovascular endpoints were obtained from national registers covering all Sweden [21,23,25–29]. Validation studies of cases obtained from the Swedish hospital discharge register have shown a 95% validity of VTE [24]. A previous study from Malmö has shown that virtually all VTE patients are diagnosed with an objective method such as phlebography, ultrasound or computer tomography [30], confirming that the majority of cases (91%) with VTE in Sweden are diagnosed with an objective method [31]. The present study focuses only on VTE coming into medical attention. Cases with subclinical VTE with minor symptoms not leading to a medical contact were not included. The registry does not allow for identification of proximal and distal deep venous thrombosis, which is a limitation of the study. However, only 28% of DVT of the lower limbs in Malmö are distal DVT [30]. Moreover, RDW was not only a risk factor for DVT but also for PE. The MDC is a population-based study with an attendance rate of 40%. Similar to most population-based studies, the incidence of disease in the MDC was higher in non-participants. The incidence of first VTE event in the present study (subjects aged 45–73 years and 60.6% women) was 2.7 per 1000 person-years, which is within limits reported from other studies [46–48]. It is well established that high-risk individuals are less likely to participate in population-based studies [49,50]. A potential underestimation of first VTE due to the “healthy cohort effect” is therefore not evident in the present study as the incidence of VTE was high. It seems likely that the “healthy cohort effect” is non-differential regarding RDW-associated VTE risk. Another major limitation of the present study is the lack of information on genetic risk factors for VTE and VTE-precipitating risk factors such as hormone replacement therapy, surgery or trauma. Though inherited factors are significant risk factors for VTE in the present age group, inherited factors are relatively more important at younger ages [48]. Thus, it will be important to verify the present results in studies with information about major genetic and circumstantial risk factors. Previously none of the established genetic risk factors for VTE have been reported to be associated with RDW [11–14], and it is not obvious how genetic variants like factor V Leiden Gln506 (rs6025) and prothrombin G20210A (rs1799963) could affect RDW [2]. All patients with self-reported physician-treated cancer during the last 5 years were excluded. We also excluded all cases with a cancer diagnosis prior to, or within 180 days after the VTE event, with essentially the same results. The association between RDW and VTE was observed among both men and women, and also after exclusion of patients with different metabolic and cardiovascular comorbidities. Still residual confounding is possible by various comorbidities or concomitant drugs not adjusted for in the analysis.
In conclusion, RDW among middle-aged subjects was associated with long-term incidence of first event of VTE. The possible mechanism underlying the association between RDW and incident VTE needs further investigation. Conflict of Interest Statement None. Acknowledgements Sources of Funding This work and the Malmö Diet and Cancer study were supported by grants from the Swedish Cancer Society, the Swedish Medical Research Council (2011-3891), the Swedish Heart and Lung Foundation (20130249, 20120352, 20110816), and the Malmö City Council and by funds from the Region Skåne, medical research and development of health care (ALF) funding from Region Skåne, Skåne University Hospital, Malmö and Lundström’s Foundation. The study sponsors had no role in the study design, in the collection, analysis and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication. Appendix A. Supplementary Data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.thromres.2013.12.013. References [1] Goldhaber SZ, Bounameaux H. Pulmonary embolism and deep vein thrombosis. Lancet 2012;379:1835–46. [2] Rosendaal FR. Venous thrombosis: the role of genes, environment, and behavior. Hematology Am Soc Hematol Educ Program 2005:1–12. [3] Gersh KC, Nagaswami C, Weisel JW. Fibrin network structure and clot mechanical properties are altered by incorporation of erythrocytes. Thromb Haemost 2009;102:1169–75. [4] Whelihan MF, Mann KG. The role of the red cell membrane in thrombin generation. Thromb Res 2013;131:377–82. [5] Panova-Noeva M, Marchetti M, Russo L, Tartari CJ, Leuzzi A, Finazzi G, et al. ADPinduced platelet aggregation and thrombin generation are increased in Essential Thrombocythemia and Polycythemia Vera. Thromb Res 2013;132:88–93. [6] Braekkan SK, Mathiesen EB, Njølstad I, Wilsgaard T, Hansen JB. Hematocrit and risk of venous thromboembolism in a general population. The Tromso study. Haematologica 2010;95:270–5. [7] Bessman JD, Gilmer Jr PR, Gardner FH. Improved classification of anemias by MCV and RDW. Am J Clin Pathol 1983;80:322–6. [8] Demir A, Yarali N, Fisgin T, Duru F, Kara A. Most reliable indices in differentiation between thalassemia trait and iron deficiency anemia. Pediatr Int 2002;44:612–6. [9] Lin CK, Lin JS, Chen SY, Jiang ML, Chiu CF. Comparison of hemoglobin and red blood cell distribution width in the differential diagnosis of microcytic anemia. Arch Pathol Lab Med 1992;116:1030–2. [10] Montagnana M, Cervellin G, Meschi T, Lippi G. The role of red blood cell distribution width in cardiovascular and thrombotic disorders. Clin Chem Lab Med 2011;50:635–41. [11] Zorlu A, Bektasoglu G, Guven FM, Dogan OT, Gucuk E, Ege MR, et al. Usefulness of admission red cell distribution width as a predictor of early mortality in patients with acute pulmonary embolism. Am J Cardiol 2012;109:128–34. [12] Ozsu S, Abul Y, Gunaydin S, Orem A, Ozlu T. Prognostic Value of Red Cell Distribution Width in Patients With Pulmonary Embolism. Clin Appl Thromb Hemost Nov 8 2012 [Epub ahead of print]. [13] Cay N, Unal O, Kartal MG, Ozdemir M, Tola M. Increased level of red blood cell distribution width is associated with deep venous thrombosis. Blood Coagul Fibrinolysis 2013;24:727–31. [14] Rezende SM, Lijfering WM, Rosendaal FR, Cannegieter S. Hematological variables and venous thrombosis: red cell distribution width and blood monocytes are associated with an increased risk. Haematologica Jul 26 2013 [Epub ahead of print]. [15] Zia E, Hedblad B, Pessah-Rasmussen H, Berglund G, Janzon L, Engström G. Blood pressure in relation to the incidence of cerebral infarction and intracerebral hemorrhage. Hypertensive hemorrhage: debated nomenclature is still relevant. Stroke 2007;38:2681–5. [16] Berglund G, Elmstahl S, Janzon L, Larsson SA. The Malmo Diet and Cancer Study. Design and feasibility. J Intern Med 1993;233:45–51. [17] Melander O, Newton-Cheh C, Almgren P, Hedblad B, Berglund G, Engström G, et al. Novel and conventional biomarkers for prediction of incident cardiovascular events in the community. JAMA 2009;302:49–57.
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339 [18] Smith JG, Newton-Cheh C, Almgren P, Struck J, Morgenthaler NG, Bergmann A, et al. Assessment of conventional cardiovascular risk factors and multiple biomarkers for the prediction of incident heart failure and atrial fibrillation. J Am Coll Cardiol 2010;56:1712–9. [19] Borné Y, Smith JG, Melander O, Hedblad B, Engström G. Red cell distribution width and risk for first hospitalization due to heart failure: a population-based cohort study. Eur J Heart Fail 2011;13:1355–61. [20] Rosvall M, Ostergren PO, Hedblad B, Isacsson SO, Janzon L, Berglund G. Occupational status, educational level, and the prevalence of carotid atherosclerosis in a general population sample of middle-aged Swedish men and women: results from the Malmö Diet and Cancer Study. Am J Epidemiol 2000;152:334–46. [21] Smith JG, Platonov PG, Hedblad B, Engström G, Melander O. Atrial fibrillation in the Malmö Diet and Cancer study: a study of occurrence, risk factors and diagnostic validity. Eur J Epidemiol 2010;25:95–102. [22] Li C, Aronsson CA, Hedblad B, Gullberg B, Wirfalt E, Berglund G. Ability of physical activity measurements to assess health-related risks. Eur J Clin Nutr 2009;63:1448–51. [23] Hammar N, Alfredsson L, Rosén M, Spetz CL, Kahan T, Ysberg AS. A national record linkage to study acute myocardial infarction incidence and case fatality in Sweden. Int J Epidemiol 2001;30(Suppl. 1):S30–4. [24] SYSMEX Lab. Referenzwertbereiche fur die Hämatologie: SYSMEX X-FAMILY; 03–2006. [25] Rosengren A, Freden M, Hansson P-O, Wilhelmsen L, Wedel H, Eriksson H. Psychosocial factors and venous thromboembolism: a long-term follow-up study of Swedish men. J Thromb Haemost 2008;6:558–64. [26] Ingelsson E, Arnlöv J, Sundström J, Lind L. The validity of a diagnosis of heart failure in a hospital discharge register. Eur J Heart Fail 2005;7:787–91. [27] Lindblad U, Råstam L, Ranstam J, Peterson M. Validity of register data on acute myocardial infarction and acute stroke: the Skaraborg Hypertension Project. Scand J Soc Med 1993;21:3–9. [28] Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall C, et al. External review and validation of the Swedish national inpatient register. BMC Public Health 2011;11:450. [29] National Board of Health and Welfare. Validity of the Diagnoses from the Swedish In-Care Register 1987 and 1995 [in Swedish]. Stockholm: Epidemiologiskt Centrum, Socialstyrelsen; 2000. [30] Isma N, Svensson PJ, Gottsäter A, Lindblad B. Prospective analysis of risk factors and distribution of venous thromboembolism in the population-based Malmö Thrombophilia Study (MATS). Thromb Res 2009;124:663–6. [31] Petrauskiene V, Falk M, Waernbaum I, Norberg M, Eriksson JW. The risk of venous thromboembolism is markedly elevated in patients with diabetes. Diabetologia 2005;48:1017–21. [32] Rothman KJ, Greenland S, editors. Modern epidemiology. 2nd ed. Philadelphia, PA: Lippincott-Raven; 1998. [33] Dean AG, Sullivan KM, Soe MM. OpenEpi: Open Source Epidemiologic Statistics for Public Health, Version. www.OpenEpi.com . [updated 2013/04/06, accessed 2013/06/26].
339
[34] Hanley JA. The robustness of the “binormal” assumptions used in fitting ROC curves. Med Decis Making 1988;8:197–203. [35] Perlstein TS, Weuve J, Pfeffer MA, Beckman JA. Red blood cell distribution width and mortality risk in a community-based prospective cohort. Arch Intern Med 2009;169:588–94. [36] Lippi G, Targher G, Montagnana M, Salvagno GL, Zoppini G, Guidi GC. Relation between red blood cell distribution width and inflammatory biomarkers in a large cohort of unselected outpatients. Arch Pathol Lab Med 2009;133:628–32. [37] Fujita B, Strodthoff D, Fritzenwanger M, Pfeil A, Ferrari M, Goebel B, et al. Altered red blood cell distribution width in overweight adolescents and its association with markers of inflammation. Pediatr Obes 2013;8:385–91. [38] Chen Z, Tang H, Qayyum R, Schick UM, Nalls MA, Handsaker R, et al. Genome-wide association analysis of red blood cell traits in African Americans: the COGENT Network. Hum Mol Genet 2013;22:2529–38. [39] Timpson NJ, Wade KH, Smith GD. Mendelian randomization: application to cardiovascular disease. Curr Hypertens Rep 2012;14:29–37. [40] Zöller B, Li X, Sundquist J, Sundquist K. Neighborhood deprivation and hospitalization for venous thromboembolism in Sweden. J Thromb Thrombolysis 2012;34:374–82. [41] Zöller B, Li X, Sundquist J, Sundquist K. Socioeconomic and occupational risk factors for venous thromboembolism in Sweden: a nationwide epidemiological study. Thromb Res 2012;129:577–82. [42] Isma N, Merlo J, Ohlsson H, Svensson PJ, Lindblad B, Gottsäter A. Socioeconomic factors and concomitant diseases are related to the risk for venous thromboembolism during long time follow-up. J Thromb Thrombolysis 2013;36:58–64. [43] Hermanides J, Cohn DM, Devries JH, Kamphuisen PW, Huijgen R, Meijers JC, et al. Venous thrombosis is associated with hyperglycemia at diagnosis: a case–control study. J Thromb Haemost 2009;7:945–9. [44] Stein PD, Goldman J, Matta F, Yaekoub AY. Diabetes mellitus and risk of venous thromboembolism. Am J Med Sci 2009;337:259–64. [45] Piazza G, Goldhaber SZ, Kroll A, Goldberg RJ, Emery C, Spencer FA. Venous thromboembolism in patients with diabetes mellitus. Am J Med 2012;125:709–16. [46] Silverstein MD, Heit JA, Mohr DN, Petterson TM, O’Fallon WM, Melton LJ. Trends in the incidence of deep vein thrombosis and pulmonary embolism: a 25-year population-based cohort study. Arch Intern Med 1998;158:585–93. [47] White R. The epidemiology of venous thromboembolism. Circulation 2003;107:I4–8. [48] Zöller B, Li X, Sundquist J, Sundquist K. Age- and gender-specific familial risks for venous thromboembolism: a nationwide epidemiological study based on hospitalizations in Sweden. Circulation 2011;124:1012–20. [49] Manjer J, Carlsson S, Elmstahl S, Gullberg B, Janzon L, Lindstrom M, et al. The Malmo Diet and Cancer Study: representativity, cancer incidence and mortality in participants and non-participants. Eur J Cancer Prev 2001;10:489–99. [50] Manjer J, Elmstahl S, Janzon L, Berglund G. Invitation to a population-based cohort study: differences between subjects recruited using various strategies. Scand J Public Health 2002;30:103–12.
Contents lists available at ScienceDirect
Thrombosis Research journal homepage: www.elsevier.com/locate/thromres
Regular Article
Red cell distribution width and risk for venous thromboembolism: A population-based cohort study Bengt Zöller a,⁎, Olle Melander b, Peter Svensson b,c, Gunnar Engström b a b c
Center for Primary Care Research, Lund University/Region Skåne, Malmö, Sweden Department of Clinical Sciences, Lund University, Clinical Research Center (CRC), Entrance 72, house 60, level 13, Skåne University Hospital, SE 205 02 Malmö, Sweden Department of Coagulation Disorders, Skåne University Hospital, Malmö, Sweden
a r t i c l e
i n f o
Article history: Received 18 September 2013 Received in revised form 10 December 2013 Accepted 11 December 2013 Available online 15 December 2013 Keywords: venous thromboembolism venous thrombosis pulmonary embolism epidemiology predictors
a b s t r a c t Introduction: Red cell distribution width (RDW) has been associated with venous thromboembolism (VTE), but whether RDW is a predictor of first event of VTE is unknown. We investigated the association between RDW and incidence of first event of VTE in a population-based cohort. Materials and Methods: RDW was measured in 27 042 subjects (aged 45–73 years, 60.6% women), without previous history of VTE or cancer within 5 years before follow-up, who participated in the Malmö Diet and Cancer study during 1991–1996. Incidence of VTE was identified from the patient register and the cause of death register during a mean follow-up of 13.8 years and studied in relation to RDW. Results: During follow-up, 991 subjects (57.5% women) were affected by VTE (pulmonary embolism or deep venous thrombosis of the lower limbs). After adjustment for potential confounding factors the hazard ratios (HR) for VTE for the second, third and fourth RDW quartiles 1.15 (95% confidence interval 0.94–1.41), 1.41 (1.14–1.73), 1.74 (1.38–2.21), respectively, were compared with the bottom quartile of RDW. In the multivariate model subjects with the top 5% of RDW values compared with the bottom quartile had an even higher risk (HR = 2.51, 1.78–2.54). In receiver operating characteristic (ROC) analysis, the male specific area under the ROC curve (AUC) for RDW was 0.57 (95% CI 0.54–0.59). The female specific AUC was 0.56 (95% CI 0.53–0.58). Conclusions: RDW was found to be associated with long-term incidence of first event of VTE among middle-aged subjects. © 2013 Elsevier Ltd. All rights reserved.
Introduction Venous thromboembolism (VTE) is the third most common cause of death from cardiovascular disease after coronary heart disease and stroke [1]. Although a number of acquired and genetic risk factors for VTE have been identified, in 30–50% of cases VTE is idiopathic [1]. During the last few decades research has therefore concentrated on identifying new VTE risk factors and creating predictive models for VTE [2]. Erythrocytes, or red blood cells (RBCs), are constituents in clots and thrombi formed in vivo [3]. RBCs play a prothrombotic role in blood coagulation by increasing blood viscosity and forcing platelets towards the vessel wall [3]. Incorporation of RBCs into a fibrin clot affects clot structure and mechanical properties [3]. Even small structural differences of RBCs may have a large influence on pathophysiology [3]. Moreover, RBCs actively participate in thrombin generation [4]. A subAbbreviations: CI, confidence interval; HR, Hazard ratio; RDW, Red cell distribution width; VTE, venous thromboembolism; DVT, deep venous thrombosis; PE, pulmonary embolism; MDC, Malmö Diet and Cancer study; CHD, coronary heart disease. ⁎ Corresponding author at: Center for Primary Health Care Research, CRC, building 28, floor 11, Jan Waldenströms gata 35, Skåne University Hospital, S-205 02 Malmö, Sweden. Tel.: +46 40 391954; fax: +46 40 391370. E-mail address: [email protected] (B. Zöller). 0049-3848/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.thromres.2013.12.013
fraction of red blood cells expresses phosphatidylserine on their surface. Unlike platelets, RBCs produce thrombin through the meizothrombin pathway, which has consequences in the context of clot formation and stabilization [4]. Polycythemia Vera has been associated with VTE [5], and hematocrit variations in the general population have been associated with VTE [6]. An increased focus on RBCs may therefore be justified, and may reveal novel mechanisms and risk factors for VTE. Red cell distribution width (RDW) is a measure of the size variation as well as an index of the heterogeneity of the erythrocytes (i.e. anisocytosis) [7–9]. RDW is part of routine hematology laboratory tests and is used for classification of anemia [7–9]. Recent studies have shown that RDW is associated with increased mortality in a number of cardiovascular disorders such as coronary artery disease, stroke, peripheral artery disease, heart failure, pulmonary embolism, and pulmonary arterial hypertension, reviewed by Montagnana et al. [10]. It is however unclear whether anisocytosis might be the cause, or a simple epiphenomenon due to conditions such as inflammation, impaired kidney function, malnutrition, or oxidative damage [10]. Two studies have determined RDW in patients with pulmonary embolism (PE) [11,12]. Both studies found that a high RDW level, i.e. anisocytosis, was an independent predictor of early PE-related mortality [11,12]. Recently, RDW has been associated with venous thromboembolism (VTE) in two case– control studies [13,14]. These case–control studies cannot exclude the
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339
335
possibility that anisocytosis was the result of the thrombotic event itself. To the best of our knowledge, there are no prospective studies on the association between RDW and risk of first event of VTE, i.e. deep venous thrombosis (DVT) and PE. The purpose of the present study was to determine whether RDW is an independent predictor of first event of VTE in asymptomatic middleaged subjects in the Malmö Diet and Cancer Study (MDC), which is a prospective cohort study in Malmö.
consecutively in fresh heparinized blood. Erythrocyte diameter was measured using a fully automated assay (SYSMEX K1000). Red cell distribution width was calculated as the width of the erythrocyte distribution curve at a relative height of 20% above the baseline. The intra-assay CV was 2.2%. Reference values were 36.4–46.3 fL in women and 35.1–43.9 fL in men [24].
Materials and Methods
All subjects were followed from the baseline examination until a first event of VTE, emigration, death or end of follow-up (December 31, 2008), whichever came first. VTE was defined as International Classification of Diseases – 8th revision (ICD-8 used before 1987), code 450 (PE), and 451 (DVT of the lower limbs) (ICD-9 used 1987–1996), code 415B (PE), and 451 (DVT of the lower limbs); and (ICD-10 used 1997–2008) I26 (PE), and I80 (DVT of the lower limbs) as the primary diagnosis. Patients with superficial thrombophlebitis were excluded (ICD-9 code 451A and ICD-10 code I80.0). The inpatient register, the outpatient register and the cause of death register were used to identify cases with VTE. All these registers have 100% coverage in Malmö during the whole follow-up time. A validation study has shown that a diagnosis of VTE in the SNHDR has a validity of 95% [25]. Such high validity has also been found for other cardiovascular disorders such as myocardial infarction (94%), stroke, heart failure, and atrial fibrillation [21,23,26–28]. The overall validity of the SNHDR is 87% [29]. A previous population-based prospective study was performed at the Malmö University Hospital in 1998–2006, of all in- and outpatients diagnosed with VTE in Malmö [30], i.e. at the same study center as the present study and with partially overlapping study periods. Among 1140 VTE patients, virtually all were diagnosed with an objective method such as phlebography, ultrasound or computer tomography [30]. Similar results were reported in another Swedish study that found that most Swedish patients with VTE are objectively diagnosed [31].
Study Population The Malmö Diet and Cancer study (MDC) is a prospective cohort study from the city of Malmö in the south of Sweden. Sample characteristics, data collection, and clinical definitions for MDC have been described previously [15–22]. Briefly, 28 449 men (n = 11 246, born 1923–1945) and women (n = 17 203, born 1923–1950) attended a baseline examination between March 1991 and September 1996. Participants underwent sampling of peripheral venous blood, measurement of blood pressure and anthropometric measures and filled out a self-administered questionnaire. Information on RDW was available in 28 363 subjects [19]. Subjects with previous VTE (n = 360 subjects) at the baseline examination were excluded. Patients with diagnosis of cancer within 5 years before baseline were excluded (n = 448). Of the remaining 27 563 subjects (8 patients were diagnosed with both VTE and cancer before baseline), 521 subjects were excluded due to missing information on blood pressure, smoking habits, alcohol consumption, education level, civil status, hemoglobin value, and body mass index (BMI). Thus, the final study population in the analysis consisted of 27 042 (10 660 [39.4%] men and 16 382 [60.6%] women) subjects, aged 45–73 years.
Ascertainment of Venous Thromboembolic Events
Measurements and Definitions Statistical Analysis Information on current use of statins, blood-pressure-lowering and anti-diabetic medications, physician-treated cancer, smoking habits, alcohol consumption, leisure-time physical activity, education level, and civil status were obtained from a self-administered questionnaire [15–22]. Weight and height were measured to the nearest 0.1 kg and 0.5 cm, respectively, with subjects wearing light clothing and no shoes. Current BMI was calculated as weight in kilograms divided by height in meters squared (kg/m2). Blood pressure was measured using a mercury-column sphygmomanometer after 10 min of rest in the supine position. Hyperglycemia was defined as fasting whole blood glucose level greater than 109 mg/dL (i.e. 6.0 mmol/L), self-reported physician’s diagnosis of diabetes or use of anti-diabetic medications. History of atrial fibrillation and also history of cardiovascular disease (CVD) at baseline was defined as diagnosis of atrial fibrillation and coronary heart disease (CHD) or stroke, respectively, in the Swedish National Hospital Discharge Register (SNHDR) [15,21,23]. Cancer during follow-up among individuals with first event of VTE (during followup) was obtained from Swedish Cancer Register. Subjects were categorized as current smokers (i.e. those who smoked regularly or occasionally) or non-smokers (i.e. former smokers and never smokers). High alcohol consumption was defined as N 40 g alcohol per day for men and N30 g per day for women [15,19]. High education level was defined as completed secondary school and at least one year of education from college, university or higher (e.g. N12 years education), modified from reference [19] and [21]. Civil status was categorized into married or not [20]. Low level of physical activity was defined as the lowest tertile of a score revealed through 18 questions covering a range of activities in the four seasons. The evaluation of the questionnaire has been previously reported [19,22]. Red cell distribution width, mean corpuscular volume (MCV), hemoglobin, platelet and leucocyte concentrations were analyzed
Detailed information on the statistical analysis is found in the supplementary material. P-values were calculated with two-sided Student’s ttest for continuous variables and with two-sided Fischer’s exact test for dichotomized variables. Confidence interval for incidence rates were calculated with the OpenEpi version 2.3.1 according to Rothman/Greenland as described [32,33]. Cox proportional hazards regression was used to examine the association between RDW and incidence of VTE. Hazard ratios (HR) with 95% confidence interval (CI) were calculated [32]. The area under the ROC (receiver operating characteristic) curve (AUC) was used as a measure of the overall performance of the ROC curve because it reflects the probability that the diagnostic test will classify correctly [34]. Positive [sensitivity/(1-specificity)] and negative [(1-sensitivity)/ specificity] likelihood ratios were also calculated to express the odds that a given value of a screening test outcome would be expected in a subject with or without VTE, respectively. Analyses were performed using IBM SPSS 21 (IBM, Armonk, New York, USA). Ethical Approval All participants provided written informed consent, and the study was approved by the ethics committee at Lund University, Lund, Sweden (LU 51/90). The MDC is registered in the US Library of Medicine as trial number NCT 01216228 (http://www.clinicaltrials.gov). Results Baseline Characteristics Cardiovascular risk factors at the baseline examination for the middle-aged MDC cohort in relation to the sex-specific quartiles of
336
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339
RDW have previously been published (not shown) [19]. Instead, cardiovascular risk factors at the baseline examination in relation to incident VTE during follow-up and sex between 1991 and 2008 are presented in Table 1 and Supplementary Table S1, respectively. Mean baseline RDW (± standard deviation, SD) was 40.5 (± 3.6) in men and 40.9 (±3.3) in women (Supplementary Table S1). Totally 991 (3.7%) of the included 27 042 individuals were affected by VTE during follow-up. Of hematological parameters at baseline, VTE during follow-up was associated with increased RDW, hemoglobin, and mean corpuscular volume (MCV) (Table 1). Sex, age, BMI, systolic blood pressure, diastolic blood pressure, hyperglycemia, use of blood pressure medication, low educational level, history of coronary heart disease (CHD) or stroke, and history of atrial fibrillation were associated with VTE during follow-up (Table 1). Incidence of First Event of VTE in Relation to Red Cell Distribution width and Known Cardiovascular Risk Factors During a median follow-up of 14.2 years (interquartile range 12.9–15.7 years), a total of 991 (3.7%) individuals were affected by VTE. The sum of follow-up time was 370 468 years corresponding to a VTE incidence rate of 2.7 (95% CI 2.5–2.8) per 1000 person years. Among the 991 VTE patients during follow-up 571 (57.6%) VTE events were coded as DVT, 382 (38.5%) as PE and 38 (3.8%) as both DVT and PE. The mean time between baseline and first event of VTE was 8.7 years (±4.2 SD). The RDW variable was divided into sex-specific quartiles as previously described (Table 2) [19]. The number of VTE cases in each quartile was 186 (2.8%) for the lowest quartile (Q1), 222 (3.3%) for the second quartile (Q2), 270 (3.9%) for the third quartile (Q3), and 313 (4.7%) for the highest quartile (Q4). The corresponding VTE incidence rates per 1000 person years were 2.0 (95% CI 1.7–2.3), 2.4 (95% CI 2.1–2.7), 2.9 (95% CI
Table 1 Baseline characteristics of subjects enrolled in the MDC study 1991–96 according to incident VTE during follow-up 1991–2008.
Sex (men/women) Age (years) Smoking (current, yes/no) Body mass index (kg/m2) Hemoglobin (g/L) Mean corpuscular volume (fL) Red cell distribution width Leucocyte count (109/L) Platelet count (109/L) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Hyperglycemia (yes/no) Blood pressure medication (yes/no) Statin medication (yes/no) Low educational level (yes/no) Unmarried (yes/no) High alcohol consumption (yes/no) Stroke or CHD (yes/no) Atrial fibrillation (yes/no)
No VTE (n = 26 051)
Incident VTE (n = 991)
10239/15812 (39%/61%) 58.0 ± 7.6 7390/18661 (28%/72%) 25.7 ± 3.9 141.7 ± 12.0 89.4 ± 4.2 40.7 ± 3.4 6.4 ± 2.3 230.4 ± 59,0 141.03 ± 19.98 85.5 ± 9.9 766/25285 (3%/97%) 4588/21463 (18%/82%) 558/25493 (2%/98%) 19982/6069 (77%/23%) 9019/17032 (35%/65%) 1134/24917 (4%/96%) 750/25301 (3%/97%) 242/25809 (1%/99%)
421/570 (42%/58%) 61.1 ± 7.4 277/714 (28%/72%) 27.0 ± 4.3 143.3 ± 12.2 89.8 ± 4.2 41.4 ± 3.3 6.6 ± 2.4 230.42 ± 78.99 145.2 ± 19.8 86.8 ± 9.9 46/945 (5%/95%) 231/760 (23%/77%) 20/971 (2%/98%) 816/175 (82%/18%) 364/627 (37%/63%) 33/958 (3%/97%) 44/947 (4%/96%) 19/972 (2%/98%)
p-value* 0.047 b0.001 0.800 b0.001 b0.001 0.003 b0.001 0.063 0.989 b0.001 b0.001 0.003 b0.001 0.898 b0.001 0.174
2.6–3.3), and 3.6 (95% CI 3.2–4.0), respectively. In the age- and sexadjusted model subjects in the two top quartiles (Q3 and Q4) compared to the bottom quartile (Q1) had significantly higher risk of VTE: HR 1.35 (95% CI 1.12–1.62) and HR 1.61 (95% CI 1.34–1.94), respectively. The risk for subjects in the second quartile (Q2) was not significantly increased, HR 1.13 (95% CI 0.93–1.38). In the age- and sex-adjusted model RDW, age, smoking, BMI, hemoglobin, MCV, leucocyte count, diastolic blood pressure, hyperglycemia, use of blood pressure medication, low educational level, being unmarried, history of CHD or stroke, and history of atrial fibrillation were all associated with an increased risk of VTE (Table 2). In the multivariate model subjects in the top (Q4) compared to the bottom quartile (Q1) of RDW had a significantly higher risk of VTE, HR 1.74 (95% CI 1.38–2.21) (Table 2 and Fig. 1). Subjects in Quartile 3 (Q3) had also a significantly higher risk of VTE, HR 1.41 (95% CI 1.14–1.73), while the risk for subjects in the second quartile (Q2) was not significantly increased HR 1.15 (95% CI 0.94–1.41). No other hematological variable than RDW was associated with an increased risk in the multivariable model. In the final multivariable model RDW, age, BMI, hyperglycemia, and being unmarried were independently associated with an increased risk of VTE (Table 2). In the multivariate model subjects with the top 5% RDW values (≥46.6 fL) compared with the bottom quartile had an even higher risk of VTE (HR 2.51, 95% CI 1.78–2.54, P b 0.001). If RDW was entered as a continuous variable into the final multivariable model the HR per fL was 1.05 (1.03–1.07, P b 0.001). No significant interaction was observed between RDW and other risk factors (age, BMI, hyperglycemia, unmarried) or hematological parameters for incidence of VTE (data not shown in Table). In a subanalysis we also determined the risk of DVT and PE separately. In the multivariable model the HR for PE for the quartiles Q2, Q3, and Q4 compared to Q1 was 1.29 (95% CI 0.95–1.75), 1.40 (95% CI 1.02–1.93) and 1.61 (95% CI 1.11–2.32), respectively. In the multivariable mode the HR for DVT for the quartiles Q2, Q3, and Q4 compared to Q1 was 1.05 (95% CI 0.81–1.36), 1.40 (95% CI 1.08–1.81) and 1.87 (95% CI 1.39–2.51), respectively. RWD, Sex and ROC Analysis No interaction was observed between RDW and sex. We nevertheless performed stratified analysis according to sex. In the multivariate model male subjects in the two top quartiles (Q3 and Q4) compared to the bottom quartile (Q1) had significantly higher risk of VTE: HR 1.44 (95% CI 1.04–1.99) and HR 1.70 (95% CI 1.16–2.47), respectively. The risk for male subjects in the second quartile (Q2) was not significantly increased, HR 1.13 (95% CI 0.87–1.67). In the multivariate model female subjects in the two top quartiles (Q3 and Q4) compared to the bottom quartile (Q1) had significantly higher risk for VTE: HR 1.40 (95% CI 1.07–1.83) and HR 1.79 (95% CI 1.32–2.42), respectively. The risk for female subjects in the second quartile (Q2) was not significantly increased, HR 1.13 (95% CI 0.87–1.47). The calculated male specific AUC for RDW was 0.57 (95% CI 0.54–0.59, p b 0.001). The female specific AUC for RDW was 0.56 (95% CI 0.53–0.58, p b 0.001). Cut-offs were derived from ROC analysis to maximize the sum of sensitivity and specificity. In males, a RDW cut-off point of 40.15 fL had sensitivity of 0.60 and specificity of 0.50, whereas in females a cut-off value of 40.55 fL had sensitivity of 0.58 and specificity of 0.50. The corresponding positive and negative likelihood ratios for both males and females were 1.2 and 0.8, respectively.
0.134
Sensitivity and Additional Analysis
0.005
In sensitivity analysis, inclusion of the 521 subjects with missing values in a sex- and age-adjusted model did not change the result: the HR for VTE for quartiles Q2, Q3, and Q4 compared to Q1 was 1.13 (95% CI 0.93–1.37), 1.35 (95% CI 1.12–1.62) and 1.60 (95% CI 1.34–1.92), respectively. We also excluded patients with personal history of atrial fibrillation, coronary heart disease, stroke, hyperglycemia, statin
0.004
Values are means with one standard deviation, or percentages with numbers in brackets. *P-values calculated with two-sided Student’s t-test for continuous variables and with two-sided Fischer’s exact test for dichotomized variables.
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339
337
Table 2 Age- and sex-adjusted and multivariable hazard ratios (HR) with 95% confidence interval (CI) and p-values for first event of VTE during follow-up in the MDC cohort 1991–2008. Age- and sex- adjusted HR (95% CI) RDW Q1 (men b38.2 fL, women b38.6 fL, n = 6689) RDW Q2 (men 38.2–40.0 fL, women 38.6–40.5 fL, n = 6782) RDW Q3 (men 40.1–42.4 fL, women 40.6–42.7 fL, n = 6868) RDW Q4 (men N42.4 fL, women N42.7 fL, n = 6703) Male sex Age (per 1 year) Smoking (current) Body mass index (per kg/m2) Hemoglobin (per 1 g/L) MCV (per fL) Leucocyte count (per 109/L) Platelet count (per 109/L) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Hyperglycemia (yes/no) Blood pressure medication (yes/no) Statin medication (yes/no) Low educational level (yes/no) Unmarried (yes/no) High alcohol consumption (yes/no) Stroke or CHD (yes/no) Atrial fibrillation (yes/no)
reference 1.13 (0.93–1.38) 1.35 (1.12–1.62) 1.61 (1.34–1.94) 1.10 (0.97–1.25) 1.06 (1.06–1.07) 1.21 (1.05–1.39) 1.07 (1.06–1.09) 1.01 (1.00–1.01) 1.02 (1.01–1.04) 1.02 (1.01–1.03) 1.00 (1.00–1.00) 1.00 (1.00–1.01) 1.01 (1.00–1.01) 1.54 (1.15–2.08) 1.23 (1.06–1.43) 0.81 (0.52–1.26) 1.19 (1.01–1.40) 1.19 (1.04–1.36) 0.88 (0.62–1.25) 1.41 (1.04–1.92) 1.79 (1.13–2.82)
medication, and hypertension medication at baseline. The results obtained in a multivariate model were not changed to any major degree: the HR for VTE for the quartiles Q2, Q3, and Q4 compared to Q1 was 1.16 (95% CI 0.92–1.47), 1.39 (95% CI 1.09–1.76) and 1.77 (95% CI 1.34–2.33), respectively. We also performed stratified analysis with regard to follow-up time. In a multivariate model for the first 8 years of follow-up the HR for VTE for the RDW quartiles Q2, Q3, and Q4 compared to Q1 was 1.22 (95% CI 0.89–1.66), 1.44 (95% CI 1.05–1.99) and 1.60 (95% CI 1.11–2.31), respectively. In the multivariate model with follow-up between 8 and 18 years the HR for VTE for the RDW quartiles Q2, Q3, and Q4 compared to Q1 was 1.11 (95% CI 0.86–1.45), 1.39 (95% CI 1.06–1.81) and HR 1.86 (95% CI 1.37–2.53), respectively. Since cancer is a major risk factor for VTE, we performed a subanalysis in which all VTE cases with cancer prior to, or up to 180 days after the VTE event, were excluded. A total of 648 VTE cases remained in the analysis. In the multivariable model, the HR for VTE for quartiles Q2, Q3, and Q4 compared to Q1 was 1.14 (95% CI 0.89–1.45), 1.44 (95% CI 1.12–1.86) and 1.79 (95% CI 1.34–2.38), respectively. Physical activity data were available for 26 106 of the included individuals. Low physical activity was not associated with VTE in a sex- and age-adjusted model 1.02 (95% CI 0.88–1.18, p = 0.82), and when entered into the multivariate
Fig. 1. First VTE event free survival in relation to sex-specific quartiles (Q1-Q4) of RDW.
p-value
Multivariable HR (95% CI)
p-value
0.210 0.002 b0.001 0.132 b0.001 0.009 b0.001 0.025 0.009 0.002 0.126 0.062 0.041 0.004 0.007 0.341 0.042 0.009 0.485 0.028 0.012
reference 1.15 (0.94–1.41) 1.41 (1.14–1.73) 1.74 (1.38–2.21) 1.09 (0.93–1.27) 1.06 (1.04–1.07) 1.13 (0.97–1.31) 1.08 (1.06–1.09) 1.00 (0.99–1.01) 1.00 (0.98–1.02) 1.01 (0.99–1.03) 1.00 (1.00–1.00) 1.00 (1.00–1.01) 1.00 (0.99–1.01) 1.41 (1.04–1.91) 1.06 (0.90–1.24) 0.69 (0.44–1.09) 1.10 (0.94–1.31) 1.16 (1.02–1.32) 0.80 (0.56–1.14) 1.30 (0.95–1.79) 1.57 (0.99–2.49)
0.161 0.001 b0.001 0.277 b0.001 0.128 b0.001 0.837 0.688 0.520 0.330 0.668 0.805 0.026 0.513 0.110 0.220 0.027 0.220 0.103 0.056
model it did not affect the association between RDW and VTE (results not shown in Table).
Discussion Previous studies have found an association of RDW with early PErelated mortality in patients with acute pulmonary embolism [11,12]. RDW has also been associated with VTE in two recent case–control studies [13,14]. The present study is the first prospective cohort study to determine whether RDW is a predictor of first event of VTE. The present large, population-based cohort study shows a graded independent association between RDW and risk of first event of VTE among middle-aged subjects. This relationship was independent of potential confounders, including hematological parameters, multiple biological, lifestyle, and socioeconomic factors. The association was robust and remained significant in all different analysis performed. The results suggest that increased RDW is a novel independent risk factor for first event of VTE. The ROC analysis indicates that RDW as a sole marker is not useful for clinical prediction of VTE in the present study population. It is possible that RDW performs better as a prognostic marker in high-risk individuals such as patients with PE [11,12]. The mechanism underlying the relationship between RDW and VTE is unclear. Red cell distribution width has been associated with lowgrade systemic inflammation [35]. In a previous study, RDW was associated with leucocyte count and hsCRP, e.g. classic markers of inflammation [36]. In the present study the relationship between RDW and risk of VTE remained unchanged when leucocyte count was taken into account (Table 2). This suggests that inflammation is not the major mechanism for the increased incidence of VTE in the present study. However, adjustment for leukocyte count is not sufficient to draw firm conclusions on inflammation, but CRP data were not available. Red cell distribution width has been associated with several other risk factors, e.g. age, waist circumference, BMI, smoking, high alcohol consumption, diabetes, blood pressure, history of CHD, leucocyte count, and being unmarried [19,37]. RDW remained significant after adjustment for such risk factors, and there were no interactions between RDW and other risk factors or hematological parameters for the incidence of VTE. Malnutrition and deficiency of vitamin B12 and folic acid are other factors that have been associated with high RDW, because of their role in erythropoiesis [10]. We did not have information on plasma levels of these vitamins, and it therefore remains unclear whether these factors contributed to the relationship between RDW and incident VTE in our study.
338
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339
What is the relevance of an association between RDW and VTE in the present study and in the four previously published studies [11–14]? If RDW by itself is not the cause of the association with VTE, then RDW will just be another biomarker for VTE reflecting a moderate hypercoagulability. As RBCs are constituents of in vivo clots and thrombi it is possible that anisocytosis could increase the prothrombotic role of the RBCs [3]. Recently, a genome-wide association study among African Americans has identified two variants (rs1050828 and rs10493739) on chromosomes Xq28 (G6pD gene) and 1p31.1, respectively, that were associated with RDW [38]. Mendelian randomization studies may therefore be an important possibility in order to generate estimates for a causal effect of RDW on VTE [39]. It is worth pointing out the association between civil status and VTE risk confirming a recent study [40]. This is consistent with the growing literature about socioeconomic factors and VTE risk [25,40–42]. Moreover, an association between hyperglycemia was also observed in the present study, confirming previous studies [31,43–45].
Strengths and Limitations The study cohort included a large numbers of subjects and events during a long follow-up period [15–22]. The cardiovascular endpoints were obtained from national registers covering all Sweden [21,23,25–29]. Validation studies of cases obtained from the Swedish hospital discharge register have shown a 95% validity of VTE [24]. A previous study from Malmö has shown that virtually all VTE patients are diagnosed with an objective method such as phlebography, ultrasound or computer tomography [30], confirming that the majority of cases (91%) with VTE in Sweden are diagnosed with an objective method [31]. The present study focuses only on VTE coming into medical attention. Cases with subclinical VTE with minor symptoms not leading to a medical contact were not included. The registry does not allow for identification of proximal and distal deep venous thrombosis, which is a limitation of the study. However, only 28% of DVT of the lower limbs in Malmö are distal DVT [30]. Moreover, RDW was not only a risk factor for DVT but also for PE. The MDC is a population-based study with an attendance rate of 40%. Similar to most population-based studies, the incidence of disease in the MDC was higher in non-participants. The incidence of first VTE event in the present study (subjects aged 45–73 years and 60.6% women) was 2.7 per 1000 person-years, which is within limits reported from other studies [46–48]. It is well established that high-risk individuals are less likely to participate in population-based studies [49,50]. A potential underestimation of first VTE due to the “healthy cohort effect” is therefore not evident in the present study as the incidence of VTE was high. It seems likely that the “healthy cohort effect” is non-differential regarding RDW-associated VTE risk. Another major limitation of the present study is the lack of information on genetic risk factors for VTE and VTE-precipitating risk factors such as hormone replacement therapy, surgery or trauma. Though inherited factors are significant risk factors for VTE in the present age group, inherited factors are relatively more important at younger ages [48]. Thus, it will be important to verify the present results in studies with information about major genetic and circumstantial risk factors. Previously none of the established genetic risk factors for VTE have been reported to be associated with RDW [11–14], and it is not obvious how genetic variants like factor V Leiden Gln506 (rs6025) and prothrombin G20210A (rs1799963) could affect RDW [2]. All patients with self-reported physician-treated cancer during the last 5 years were excluded. We also excluded all cases with a cancer diagnosis prior to, or within 180 days after the VTE event, with essentially the same results. The association between RDW and VTE was observed among both men and women, and also after exclusion of patients with different metabolic and cardiovascular comorbidities. Still residual confounding is possible by various comorbidities or concomitant drugs not adjusted for in the analysis.
In conclusion, RDW among middle-aged subjects was associated with long-term incidence of first event of VTE. The possible mechanism underlying the association between RDW and incident VTE needs further investigation. Conflict of Interest Statement None. Acknowledgements Sources of Funding This work and the Malmö Diet and Cancer study were supported by grants from the Swedish Cancer Society, the Swedish Medical Research Council (2011-3891), the Swedish Heart and Lung Foundation (20130249, 20120352, 20110816), and the Malmö City Council and by funds from the Region Skåne, medical research and development of health care (ALF) funding from Region Skåne, Skåne University Hospital, Malmö and Lundström’s Foundation. The study sponsors had no role in the study design, in the collection, analysis and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication. Appendix A. Supplementary Data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.thromres.2013.12.013. References [1] Goldhaber SZ, Bounameaux H. Pulmonary embolism and deep vein thrombosis. Lancet 2012;379:1835–46. [2] Rosendaal FR. Venous thrombosis: the role of genes, environment, and behavior. Hematology Am Soc Hematol Educ Program 2005:1–12. [3] Gersh KC, Nagaswami C, Weisel JW. Fibrin network structure and clot mechanical properties are altered by incorporation of erythrocytes. Thromb Haemost 2009;102:1169–75. [4] Whelihan MF, Mann KG. The role of the red cell membrane in thrombin generation. Thromb Res 2013;131:377–82. [5] Panova-Noeva M, Marchetti M, Russo L, Tartari CJ, Leuzzi A, Finazzi G, et al. ADPinduced platelet aggregation and thrombin generation are increased in Essential Thrombocythemia and Polycythemia Vera. Thromb Res 2013;132:88–93. [6] Braekkan SK, Mathiesen EB, Njølstad I, Wilsgaard T, Hansen JB. Hematocrit and risk of venous thromboembolism in a general population. The Tromso study. Haematologica 2010;95:270–5. [7] Bessman JD, Gilmer Jr PR, Gardner FH. Improved classification of anemias by MCV and RDW. Am J Clin Pathol 1983;80:322–6. [8] Demir A, Yarali N, Fisgin T, Duru F, Kara A. Most reliable indices in differentiation between thalassemia trait and iron deficiency anemia. Pediatr Int 2002;44:612–6. [9] Lin CK, Lin JS, Chen SY, Jiang ML, Chiu CF. Comparison of hemoglobin and red blood cell distribution width in the differential diagnosis of microcytic anemia. Arch Pathol Lab Med 1992;116:1030–2. [10] Montagnana M, Cervellin G, Meschi T, Lippi G. The role of red blood cell distribution width in cardiovascular and thrombotic disorders. Clin Chem Lab Med 2011;50:635–41. [11] Zorlu A, Bektasoglu G, Guven FM, Dogan OT, Gucuk E, Ege MR, et al. Usefulness of admission red cell distribution width as a predictor of early mortality in patients with acute pulmonary embolism. Am J Cardiol 2012;109:128–34. [12] Ozsu S, Abul Y, Gunaydin S, Orem A, Ozlu T. Prognostic Value of Red Cell Distribution Width in Patients With Pulmonary Embolism. Clin Appl Thromb Hemost Nov 8 2012 [Epub ahead of print]. [13] Cay N, Unal O, Kartal MG, Ozdemir M, Tola M. Increased level of red blood cell distribution width is associated with deep venous thrombosis. Blood Coagul Fibrinolysis 2013;24:727–31. [14] Rezende SM, Lijfering WM, Rosendaal FR, Cannegieter S. Hematological variables and venous thrombosis: red cell distribution width and blood monocytes are associated with an increased risk. Haematologica Jul 26 2013 [Epub ahead of print]. [15] Zia E, Hedblad B, Pessah-Rasmussen H, Berglund G, Janzon L, Engström G. Blood pressure in relation to the incidence of cerebral infarction and intracerebral hemorrhage. Hypertensive hemorrhage: debated nomenclature is still relevant. Stroke 2007;38:2681–5. [16] Berglund G, Elmstahl S, Janzon L, Larsson SA. The Malmo Diet and Cancer Study. Design and feasibility. J Intern Med 1993;233:45–51. [17] Melander O, Newton-Cheh C, Almgren P, Hedblad B, Berglund G, Engström G, et al. Novel and conventional biomarkers for prediction of incident cardiovascular events in the community. JAMA 2009;302:49–57.
B. Zöller et al. / Thrombosis Research 133 (2014) 334–339 [18] Smith JG, Newton-Cheh C, Almgren P, Struck J, Morgenthaler NG, Bergmann A, et al. Assessment of conventional cardiovascular risk factors and multiple biomarkers for the prediction of incident heart failure and atrial fibrillation. J Am Coll Cardiol 2010;56:1712–9. [19] Borné Y, Smith JG, Melander O, Hedblad B, Engström G. Red cell distribution width and risk for first hospitalization due to heart failure: a population-based cohort study. Eur J Heart Fail 2011;13:1355–61. [20] Rosvall M, Ostergren PO, Hedblad B, Isacsson SO, Janzon L, Berglund G. Occupational status, educational level, and the prevalence of carotid atherosclerosis in a general population sample of middle-aged Swedish men and women: results from the Malmö Diet and Cancer Study. Am J Epidemiol 2000;152:334–46. [21] Smith JG, Platonov PG, Hedblad B, Engström G, Melander O. Atrial fibrillation in the Malmö Diet and Cancer study: a study of occurrence, risk factors and diagnostic validity. Eur J Epidemiol 2010;25:95–102. [22] Li C, Aronsson CA, Hedblad B, Gullberg B, Wirfalt E, Berglund G. Ability of physical activity measurements to assess health-related risks. Eur J Clin Nutr 2009;63:1448–51. [23] Hammar N, Alfredsson L, Rosén M, Spetz CL, Kahan T, Ysberg AS. A national record linkage to study acute myocardial infarction incidence and case fatality in Sweden. Int J Epidemiol 2001;30(Suppl. 1):S30–4. [24] SYSMEX Lab. Referenzwertbereiche fur die Hämatologie: SYSMEX X-FAMILY; 03–2006. [25] Rosengren A, Freden M, Hansson P-O, Wilhelmsen L, Wedel H, Eriksson H. Psychosocial factors and venous thromboembolism: a long-term follow-up study of Swedish men. J Thromb Haemost 2008;6:558–64. [26] Ingelsson E, Arnlöv J, Sundström J, Lind L. The validity of a diagnosis of heart failure in a hospital discharge register. Eur J Heart Fail 2005;7:787–91. [27] Lindblad U, Råstam L, Ranstam J, Peterson M. Validity of register data on acute myocardial infarction and acute stroke: the Skaraborg Hypertension Project. Scand J Soc Med 1993;21:3–9. [28] Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall C, et al. External review and validation of the Swedish national inpatient register. BMC Public Health 2011;11:450. [29] National Board of Health and Welfare. Validity of the Diagnoses from the Swedish In-Care Register 1987 and 1995 [in Swedish]. Stockholm: Epidemiologiskt Centrum, Socialstyrelsen; 2000. [30] Isma N, Svensson PJ, Gottsäter A, Lindblad B. Prospective analysis of risk factors and distribution of venous thromboembolism in the population-based Malmö Thrombophilia Study (MATS). Thromb Res 2009;124:663–6. [31] Petrauskiene V, Falk M, Waernbaum I, Norberg M, Eriksson JW. The risk of venous thromboembolism is markedly elevated in patients with diabetes. Diabetologia 2005;48:1017–21. [32] Rothman KJ, Greenland S, editors. Modern epidemiology. 2nd ed. Philadelphia, PA: Lippincott-Raven; 1998. [33] Dean AG, Sullivan KM, Soe MM. OpenEpi: Open Source Epidemiologic Statistics for Public Health, Version. www.OpenEpi.com . [updated 2013/04/06, accessed 2013/06/26].
339
[34] Hanley JA. The robustness of the “binormal” assumptions used in fitting ROC curves. Med Decis Making 1988;8:197–203. [35] Perlstein TS, Weuve J, Pfeffer MA, Beckman JA. Red blood cell distribution width and mortality risk in a community-based prospective cohort. Arch Intern Med 2009;169:588–94. [36] Lippi G, Targher G, Montagnana M, Salvagno GL, Zoppini G, Guidi GC. Relation between red blood cell distribution width and inflammatory biomarkers in a large cohort of unselected outpatients. Arch Pathol Lab Med 2009;133:628–32. [37] Fujita B, Strodthoff D, Fritzenwanger M, Pfeil A, Ferrari M, Goebel B, et al. Altered red blood cell distribution width in overweight adolescents and its association with markers of inflammation. Pediatr Obes 2013;8:385–91. [38] Chen Z, Tang H, Qayyum R, Schick UM, Nalls MA, Handsaker R, et al. Genome-wide association analysis of red blood cell traits in African Americans: the COGENT Network. Hum Mol Genet 2013;22:2529–38. [39] Timpson NJ, Wade KH, Smith GD. Mendelian randomization: application to cardiovascular disease. Curr Hypertens Rep 2012;14:29–37. [40] Zöller B, Li X, Sundquist J, Sundquist K. Neighborhood deprivation and hospitalization for venous thromboembolism in Sweden. J Thromb Thrombolysis 2012;34:374–82. [41] Zöller B, Li X, Sundquist J, Sundquist K. Socioeconomic and occupational risk factors for venous thromboembolism in Sweden: a nationwide epidemiological study. Thromb Res 2012;129:577–82. [42] Isma N, Merlo J, Ohlsson H, Svensson PJ, Lindblad B, Gottsäter A. Socioeconomic factors and concomitant diseases are related to the risk for venous thromboembolism during long time follow-up. J Thromb Thrombolysis 2013;36:58–64. [43] Hermanides J, Cohn DM, Devries JH, Kamphuisen PW, Huijgen R, Meijers JC, et al. Venous thrombosis is associated with hyperglycemia at diagnosis: a case–control study. J Thromb Haemost 2009;7:945–9. [44] Stein PD, Goldman J, Matta F, Yaekoub AY. Diabetes mellitus and risk of venous thromboembolism. Am J Med Sci 2009;337:259–64. [45] Piazza G, Goldhaber SZ, Kroll A, Goldberg RJ, Emery C, Spencer FA. Venous thromboembolism in patients with diabetes mellitus. Am J Med 2012;125:709–16. [46] Silverstein MD, Heit JA, Mohr DN, Petterson TM, O’Fallon WM, Melton LJ. Trends in the incidence of deep vein thrombosis and pulmonary embolism: a 25-year population-based cohort study. Arch Intern Med 1998;158:585–93. [47] White R. The epidemiology of venous thromboembolism. Circulation 2003;107:I4–8. [48] Zöller B, Li X, Sundquist J, Sundquist K. Age- and gender-specific familial risks for venous thromboembolism: a nationwide epidemiological study based on hospitalizations in Sweden. Circulation 2011;124:1012–20. [49] Manjer J, Carlsson S, Elmstahl S, Gullberg B, Janzon L, Lindstrom M, et al. The Malmo Diet and Cancer Study: representativity, cancer incidence and mortality in participants and non-participants. Eur J Cancer Prev 2001;10:489–99. [50] Manjer J, Elmstahl S, Janzon L, Berglund G. Invitation to a population-based cohort study: differences between subjects recruited using various strategies. Scand J Public Health 2002;30:103–12.
