- Email: [email protected]
The association between total leukocyte count and longevity: Evidence from longitudinal and cross-sectional data
Annals of Anatomy 204 (2016) 1–10
Contents lists available at ScienceDirect
Annals of Anatomy journal homepage: www.elsevier.com/locate/aanat
Research article
The association between total leukocyte count and longevity: Evidence from longitudinal and cross-sectional data Piotr Paweł Chmielewski a,∗ , Krzysztof Borysławski b , Krzysztof Chmielowiec c , Jolanta Chmielowiec d , Bartłomiej Strzelec a a
Department of Anatomy, Faculty of Medicine, Wroclaw Medical University, ul. Tytusa Chałubi´ nskiego 6a, 50-368 Wrocław, Poland ˙ Department of Anthropology, Institute of Biology, Wroclaw University of Environmental and Life Sciences, ul. Kozuchowska 5, 51-631 Wrocław, Poland c ˛ Regional Psychiatric Hospital for People with Mental Disorders, Cibórz 5, 66-213 Skape, Poland d Faculty of Education, Sociology and Health Sciences, University of Zielona Góra, Al. Wojska Polskiego 69, 65-762 Zielona Góra, Poland b
a r t i c l e
i n f o
Article history: Received 14 March 2015 Received in revised form 9 September 2015 Accepted 9 September 2015 Keywords: Leukocyte count Inflammation Aging Senescence Longevity Longitudinal studies
a b s t r a c t The aim of the study was to evaluate the relationship between age-dependent changes in total leukocyte count (TLC) and certain selected differential counts expressed as frequencies (granulocytes, band cells, eosinophils, lymphocytes, and monocytes) and longevity in physically healthy men and women aged 45+. Longitudinal data on cell counts from 142 subjects (68 men and 74 women; all aged 45–70 and examined for 25 years) were compared with cross-sectional data from 225 subjects (113 men and 112 women; this group was divided into four categories of average lifespan; i.e.: 53, 63, 68, and 76+ years of age). ANOVA, t-test, and regression analysis were employed. Secular changes in leukocyte count were controlled. Men had continuously higher TLC compared with women. Moreover, sex differences in patterns of changes with age were found. The longitudinal assessment revealed a U-shaped pattern of changes in TLC in men (y = 0.0026x2 − 0.2866x + 14.4374; R2 = 0.852) and women (y = 0.0048x2 − 0.5386x + 20.922; R2 = 0.938), whereas the cross-sectional comparison showed an inverted U-shaped pattern in men (y = −0.0021x2 + 0.2421x; R2 = 0.417) and women (y = −0.0017x2 + 0.2061x; R2 = 0.888). In general, the comparison of longitudinal and cross-sectional data on changes with age in TLC indicates that longevity favors individuals with lower yet normal TLC and this correlation is more pronounced in men. In conclusion, our findings are in line with previous longitudinal studies of aging and suggest that lower TLC within the normal range (4.0–10.0 × 103 L−1 ) can be a useful predictor of longevity in physically healthy individuals. © 2015 Elsevier GmbH. All rights reserved.
1. Introduction Like erythrocytes and platelets, leukocytes are derived from hematopoietic stem cells (HSCs) in the red bone marrow, which is a mesenchymal-derived complex structure consisting of hematopoietic precursors and a complex microenvironment that facilitates the maintenance of HSCs and supports the differentiation and maturation of the progenitors (Naeim et al., 2013). Unlike erythrocytes and platelets, leukocytes are large, nucleated, translucent, and morphologically heterogeneous cells which fulfill different defensive functions. There are two major categories of leukocytes: granulocytes (neutrophils, eosinophils, and basophils) and agranulocytes (monocytes and lymphocytes). Most of them have
∗ Corresponding author. Tel.: +48 71 784 13 45; fax: +48 71 784 00 79. E-mail address: [email protected] (P.P. Chmielewski). http://dx.doi.org/10.1016/j.aanat.2015.09.002 0940-9602/© 2015 Elsevier GmbH. All rights reserved.
a short lifespan, ranging from a few hours or several days (granulocytes) to a few months (monocytes) or even several years (lymphocytes). Leukocytes are constitutively produced throughout ontogeny and removed from the blood by the liver and spleen. Normally, they constitute less than 1% of whole blood (in adults, range, 4.0–10.0 × 103 L−1 ), but their number can double or can even increase tenfold within hours since there are reserve pools of these cells in bone marrow, spleen, and lymph nodes. Leukocytes can leave the blood by squeezing and migrating between the endothelial cells through pores in the capillary walls. The process of diapedesis is accelerated during inflammation. The functions of leukocytes are not confined to producing and distributing antibodies during the immune response but include removing toxins or wastes and destroying damaged or abnormal cells through phagocytosis. Phagocytic cells located in the reticular connective tissue are part of the mononuclear phagocyte system (MPS).
2
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Total leukocyte count (TLC) increases in response to infection, trauma, inflammation, and certain diseases. For example, sepsis is accompanied by a surge in the number of leukocytes (Aminzadeh and Parsa, 2011). There are, however, many factors that can affect leukocyte count in healthy subjects such as sex, genetic factors, stress level, diet, nutrition, and lifestyle, e.g. tobacco-induced inflammatory changes. First of all, leukocyte count changes with age: in newborns it is two to three times higher than in adults and in elderly people it diminishes gradually. Although leukopenia, neutropenia, and lymphocytopenia have long been recognized as indicators of poor health and severely impaired immunity, a growing body of evidence suggests now that higher TLC within the normal range is related to increased cumulative all-cause mortality and lower chances of long-term survival. Moreover, some recent studies have revealed that increased TLC is significantly associated with cardiovascular mortality in both sexes and with noncardiovascular mortality in women. Interestingly, hazard ratios were essentially unchanged by adjustment for risk factors such as smoking, hypertension, myocardial infarction, diabetes mellitus, total cholesterol, high-density lipoprotein cholesterol, and body mass index (Nilsson et al., 2014). The discovery of an increased risk of mortality in older women with elevated yet normal level of baseline leukocyte counts stimulated the ongoing debate about the relationship between TLC or differential cell counts and survival probability in adults and older people (Horne et al., 2005; Leng et al., 2005a,b, 2009). Higher level and upward changes in TLC can be used as an important clinical marker of chronic systemic inflammation and a negative prognostic in patients with cardiovascular disease (CVD), coronary heart disease (CHD), stroke, and cancer (Danesh et al., 1998; Erlinger et al., 2004; Wheeler et al., 2004; Leng and Fried, 2009). Ruggiero et al. (2007) demonstrated a nonlinear relationship between TLC and all-cause mortality and cancer. There was, however, a linear relationship between leukocyte count and mortality from CVD. Nevertheless, the role of TLC as an independent predictor of the first cardiovascular incident remains uncertain. Likewise, there is no evidence that leukocyte count is related to longevity in physically healthy people who are not at higher risk of CVD, CHD, stroke, and cancer. Although the results of several studies confirmed the positive relationship between TLC and mortality in older adults and elderly people, little work has been devoted to changes with age in leukocyte count in short and long-lived subjects on the basis of longitudinal studies of aging. The purpose of our study was to determine rates and patterns of such longitudinal as well as crosssectional age-related changes in TLC, including aging-associated alterations in differential cell counts expressed as frequencies (granulocytes, band cells, eosinophils, lymphocytes, and monocytes) in physically healthy men and women aged 45+, with special reference to the correlation with average lifespan (ALS) in the compared groups of subjects; i.e. short versus long-lived individuals.
2. Material and methods 2.1. Study population Out of the total number of patients (n = 3500) who lived at the Regional Psychiatric Hospital for People with Mental Disorders in Ciborz, Lubuskie Province, Poland, in the years 1960–2000, we have carefully selected longitudinal data on hematological parameters from 142 physically healthy individuals (including 68 men and 74 women) and cross-sectional data from 225 physically healthy individuals (including 113 men and 112 women). All subjects were Caucasian, born in the years 1911–1933. Their health was evaluated during regular physicals at the hospital on its premises by hospital
staff. In the years 1960–1989, the asylum provided custody for helpless and resourceless people from the underclass, thereby confining not only individuals with severe mental disorders but also persons who needed special care due to mild mental problems and poverty. The reason for keeping these people there for many years, was to separate them from the rest of the socialistic society, which ´ 2000; was a socially and politically motivated decision (Krzyminski, Borysławski et al., 2015; Chmielewski et al., 2015). During their lengthy stay at the hospital, the patients would take some powerful psychoactive drugs. Therefore, we selected solely data from patients who hardly ever had been treated with strong medicines or who had been treated so every once in a while. It is noteworthy that the patients lived for many years under very similar and relatively prosperous environmental conditions, maintained virtually the same lifestyle and had practically the same diet prescribed by a dietitian at the hospital. This fact undoubtedly boosts the value of the study sample and makes it quite unique. Nevertheless, detailed written data on diet, calorie intake, physical activity, and so forth were not available and were not controlled in the research. The patients from the longitudinal sample (n = 142), aged 45 at the beginning and 70 years at the end of the analyzed period, stayed continuously at the hospital and all lived to be at least 70 years of age. No further information on lifespan of these patients was available. In the years 1999–2000, there were sweeping reforms in the functioning of the medical institutions in Poland and those from the longitudinal sample who survived to be older than 70 years of age, were transported to other hospitals or medical institutions in the country. The patients from the cross-sectional sample (n = 225) were confined at the same hospital but differed in lifespan. Leading causes of death were predominantly aging-associated diseases such as ‘cardiorespiratory failure’ (CRF), cardiovascular disease (CVD), stroke, and cancer. The cross-sectional sample was divided into four categories of average lifespan (ALS), i.e.: (1) 53 years of age (the limit of individual lifespan for this group was 57.5 years of age; the group consisted of 22 men and 12 women, ALS was 53 and 52 years of age, respectively), (2) 63 years of age (the limit of individual lifespan for this group was 65 years; the group was comprised of 27 men and 30 women, ALS was 63 years for both sexes), (3) 68 years of age (the limit of individual lifespan for this group was 72.5 years; the group was made up of 49 men and 40 women, ALS was 67.6 years for men and 68 years for women), and (4) 76+ (there were no limits since this group comprised living subjects, 15 men and 30 women, their average age was 76 for both sexes). 2.2. Data collection and statistical analysis Data on hematological parameters were obtained from the hospital archives. All information used in the present study was derived from physical examinations of the inmates who were there in the years 1960–2000. The process of data collection was performed with permission and consent of hospital authorities. The study was carried out in accordance with the Declaration of Helsinki and consisted in observation only. No experiments were conducted. The medical files were anonymized so as not to divulge any personal or confidential information. On the basis of medical records and files that had been stored at the archives, we created a large computerized database in the years 2005–2007. Two types of materials were obtained: longitudinal and cross-sectional (see Section 2.1). All measurements were taken in accordance with internationally accepted standards and requirements by hospital staff. Blood samples from the median cubital vein were drawn by a nurse at the hospital. During the 25-year study period, complete blood tests were performed from 10 up to 18 times within each five-year period for a very long time, that is for 25 years in the case of the longitudinal sample. On the basis of such frequently repeated measurements for
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
3
Table 1 Characteristics of the patients (n = 142) from the longitudinal assessment at the onset of the study (all subjects aged 45) and in the consecutive age categories; p-values <0.05 are given in bold type. Age
Trait (unit)
Men (n = 68)
Women (n = 74)
t–test
Mean
SD
Mean
SD
p–value
45
TLC, 103 L−1 Granulocytes (%) Band cells (%) Eosinophils (%) Lymphocytes (%) Monocytes (%)
6.8 68.0 3.5 3.3 30.6 2.0
1.5 8.1 2.2 2.5 8.1 1.7
6.3 66.5 3.5 3.3 31.8 2.3
2.0 7.0 1.9 2.5 6.8 2.4
1.50 1.21 0.20 0.06 1.00 0.68
0.135 0.228 0.843 0.953 0.317 0.500
50
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.4 67.3 3.1 2.8 31.0 1.8
1.6 6.4 1.6 2.0 6.1 1.3
5.8 66.6 3.4 2.4 31.8 2.4
1.3 5.5 1.9 1.2 5.6 3.2
2.52 0.76 1.01 1.26 –0.83 –1.48
0.013 0.450 0.317 0.210 0.408 0.141
55
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.5 66.5 2.8 3.3 32.2 2.3
1.7 6.6 2.2 2.5 6.0 1.7
5.8 66.2 3.2 2.6 32.0 2.5
1.3 5.6 1.7 1.3 5.6 1.4
2.94 0.29 –0.46 0.51 0.20 –0.90
0.004 0.771 0.646 0.608 0.840 0.370
60
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.4 65.9 3.2 3.3 32.5 2.4
1.5 7.2 3.4 2.5 5.8 1.5
5.8 64.9 3.0 2.7 32.7 2.6
1.7 7.5 1.5 1.8 7.3 1.7
2.32 0.80 1.22 0.95 –0.23 –0.69
0.022 0.426 0.225 0.344 0.822 0.493
65
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.8 66.1 3.4 2.6 31.4 3.2
2.1 5.9 1.5 1.5 5.7 3.1
5.9 66.5 3.0 2.7 31.1 3.1
1.4 7.2 2.4 2.1 7.2 2.4
2.93 –0.38 1.32 –0.20 0.29 0.23
0.004 0.708 0.189 0.845 0.771 0.816
70
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.9 67.5 3.3 3.0 29.6 3.3
2.5 7.3 1.9 2.9 7.2 2.1
6.6 68.2 3.1 2.6 28.3 3.8
2.2 8.6 2.9 1.6 7.8 3.6
0.84 –0.47 0.62 1.13 0.99 –1.07
0.405 0.639 0.537 0.260 0.324 0.288
Table 2 Characteristics of the patients (n = 225) from the cross-sectional assessment in the consecutive age categories; p-values <0.05 are given in bold type. Age
Trait (unit)
Men (n = 113)
Women (n = 112)
t-test
p-value
Mean
SD
Mean
SD
53
TLC, 103 L−1 Granulocytes (%) Band cells (%) Eosinophils (%) Lymphocytes (%) Monocytes (%)
6.4 70.6 2.4 2.1 26.9 2.2
1.8 6.9 1.7 1.3 9.8 1.3
6.1 67.3 2.2 2.2 31.0 3.1
1.4 8.2 0.9 1.4 8.9 2.7
0.51 1.14 0.36 –0.34 –1.12 –1.16
0.611 0.263 0.719 0.737 0.272 0.258
63
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
7.6 64.2 2.0 2.4 34.9 1.7
2.0 6.3 1.5 1.9 6.5 1.6
6.3 65.5 2.2 2.4 33.0 1.8
1.5 7.6 1.3 1.8 8.3 1.5
2.74 –0.71 –0.58 0.13 0.95 –0.29
0.008 0.480 0.564 0.894 0.347 0.776
68
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.7 67.7 2.6 1.9 32.1 1.3
1.2 7.3 1.3 1.2 7.5 1.3
6.1 67.5 3.2 2.6 32.0 1.8
3.5 5.9 1.6 2.3 5.9 1.2
1.74 0.17 –2.03 –1.81 0.05 –2.10
0.086 0.869 0.045 0.074 0.960 0.039
76+
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
5.9 67.3 2.4 2.0 32.8 2.2
1.0 6.6 1.2 1.3 7.6 2.9
5.8 66.0 2.7 2.7 34.3 1.3
1.7 6.3 1.5 1.3 6.3 1.1
0.20 0.59 –0.48 –1.73 –0.65 1.39
0.843 0.558 0.636 0.092 0.519 0.174
4
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Fig. 1. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in total leukocyte count (TLC) in the consecutive age categories of men and women: arithmetic means ± standard deviations, differences between sexes tested by t-test (* p < 0.05, ** p < 0.01, *** p < 0.001), and models of regression with coefficients of determination (R2 ) are shown.
each subject, we calculated arithmetic means, standard deviations (SD), standard errors, and other statistics for each five-year interval. Thus an arithmetic mean for a given age category in both longitudinal and cross-sectional sample was always calculated on the basis of frequently repeated measurements in many individuals. Blood cell counting was performed by medical laboratory scientists using an optical microscope and Bürker chamber holding a specified volume of Giemsa stained diluted blood. In the years under study, automatic cell counters were not widely available in Poland and therefore only the traditional manual method was employed throughout the period of 25 years. It is worth mentioning that the manual method can be perceived as good and reliable since it is used as reference when different methods of cell counting are compared in terms of accuracy (Meintker et al., 2013). The traditional manual method is often perceived as more accurate and reliable than different automatic methods in the case of some anomalies of blood cells. To determine the rate and patterns of changes with age in the analyzed characteristics of the blood, one-way ANOVA, t-test, and regression analysis were performed. The method of least squares was used. The goodness of fit of a given regression model was confirmed only when a coefficient of determination (R2 ) reached the highest value and an unknown parameter (ˇ0 ) as well as a coefficient of
regression (ˇ1 ) were statistically significant (p ≤ 0.05). Five types of functions were tested: (I) linear function, y = ˇ1 age + ˇ0 , (II) logarithmic function, y = ˇ1 ln age + ˇ0 , (III) polynomial function, y = ˇ1 age2 + ˇ2 age + ˇ0 , (IV) exponential type I, y = ˇ1 agea , and (V) exponential type II, y = ˇ1 ea(age) , where (y) represents a value of an analyzed characteristic changing through aging, (ˇ2 ) denotes the second coefficient of regression, (a) stands for the exponent, and (e) is the base of the natural logarithm. 3. Results 3.1. Longitudinal study In the two samples, TLC and differential counts were normally distributed (K–S test, p > 0.2). The baseline characteristics of the study population in terms of total and differential cell counts are presented in Table 1. The regression analysis has revealed significant aging-associated changes in TLC in both sexes. In men and women, a U-shaped pattern was observed and an appropriate model of regression was polynomial (Fig. 1A). In the first five-year periods of study, leukocyte count decreased slightly but it increased in the later phases of aging. Interestingly, the greatest decrease in the number of leukocytes occurred at similar stages of ontogeny in men and women. There was a steady tendency for TLC to be
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
5
Fig. 2. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in granulocyte count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
higher in men throughout the 25-year study period. Significant differences between sexes were found in the age category of 50 (t = 2.52, p = 0.013), 55 (t = 2.94, p = 0.004), 60 (t = 2.32, p = 0.022), and 65 (t = 2.93, p = 0.004). The curve of regression for the frequency of granulocytes assumed a U-shaped pattern of alterations in both sexes (Fig. 2A). The best fitting model was polynomial in men (y = 0.0104x2 − 1.2383x + 102.797, R2 = 0.878) and women (y = 0.0111x2 − 1.24x + 100.135, R2 = 0.660). The lowest level of these cells occurred at the age of 60 in men (65.9%) and women (64.9%), whereas the highest level occurred at the age of 45 in men (68.0%) and at the age of 70 in women (68.2%). No significant sex differences in granulocyte count were observed over the whole period under study (ANOVA, p > 0.05 for each age category). Band cells count (BCC) in men decreased in the first five-year period and then it rose (y = 0.0005x2 − 0.0608x + 4.9975, R2 = 0.089). Therefore, the lowest level occurred at the age of 50 (3.1%). In both sexes, the highest BCC was at the age of 45 (Table 1, Fig. 3A). By contrast, there was a steady decline in BCC during the study period in women (y = 0.0013x2 − 0.1729x + 8.693, R2 = 0.966). There were no significant changes with age in the frequency of eosinophil granulocytes in men and women (Fig. 4A). The models of regression were polynomial (for men: y = 0.0018x2 − 0.2142x + 9.2638, R2 = 0.329; for women: y = 0.002x2 − 0.2462x + 10.1298, R2 = 0.444). The frequency of lymphocytes in men and women showed an inverted U-shaped pattern and the highest level of these cells occurred at the age of 60
(32.5% and 32.7%, respectively, t = 0.225, p = 0.057). No significant sex differences were found (Fig. 5A). The patterns of changes with age were similar and commensurate in both sexes. The model was polynomial in men (y =−0.0144x2 + 1.6325x − 14.1607, R2 = 0.887) and women (y =−0.0148x2 + 1.599x − 10.5286, R2 = 0.870). A gradual increase in the frequency of monocytes was found in men and women (Fig. 6A). The model of regression was of the exponential type II in men (y = 0.5739e0.0251x , R2 = 0.874) but polynomial in women (y = 0.0032x2 − 0.3097x + 9.8247, R2 = 0.974). 3.2. Cross-sectional study Table 2 shows the baseline characteristics of the subjects from the cross-sectional assessment in terms of total and differential cell counts. In patients who differed in lifespan, the curve of regression assumed an inverted U-shaped pattern of changes and the models tuned out to be polynomial (for men: y = −0.0021x2 + 0.2421x, R2 = 0.417; for women: y = −0.0017x2 + 0.2061x, R2 = 0.888; Fig. 1B). Men had significantly higher TLC than women at the category of age 63 (t = 2.74, p = 0.008). In women, TLC increased in the first period and then diminished continuously. In general, after the age of 60, the longer the patients lived, the lower TLC they had. Thus individuals with low yet normal TLC lived significantly longer and were more likely to survive to the age category of 76+ compared with those who had high but normal TLC (for men: n = 15, mean ± SD: 5.9 ± 1.0 × 103 L−1 , range 4.6–7.6 × 103 L−1 ; for women: n = 24,
6
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Fig. 3. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in band cell count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
5.8 ± 1.7 × 103 L−1 , range 3.6–10.2 × 103 L−1 ). The relationship was more pronounced in men. With respect to the frequency of granulocytes, the model of regression was polynomial in both sexes Fig. 2B). The function was U-shaped in men (y = 0.0238x2 − 3.1766x + 171.8592, R2 = 0.616), whereas a slow decline occurred in women (y = 0.0015x2 − 0.2228x + 74.654, R2 = 0.138). There were no significant changes with age in BCC (%) in both sexes (Fig. 3B) and the function of regression was polynomial in men (y = 0.0013x2 − 0.1683x + 7.5496, R2 = 0.168) and women (y = 0.0014x2 + 0.2002x − 4.5475, R2 = 0.349). The frequency of eosinophil granulocytes showed an inverted U-shaped pattern in men and the model of regression was polynomial (y = −0.0006x2 + 0.075x, R2 = 0.215). Although there was a slow and gradual increase in the frequency of eosinophil granulocytes in women (exponential type II function, y = 1.4155e0.0086x , R2 = 0.929), no significant sex differences in terms of the frequency were observed in any of the four tested age categories (Fig. 4B). The changes with age in the frequency of lymphocytes followed an inverted U-shaped pattern in men (y = −0.0283x2 + 3.8734x − 98.542, R2 = 0.798), while there was a slow and irregular increase in the frequency of these agranulocytes in women (y = −0.0062x2 + 0.9062x, R2 = 0.565). No significant sex differences in the frequency of lymphocytes were observed (Fig. 5B). The U-shaped pattern of changes with age in the frequency of monocytes (Fig. 6B) was found in men (polynomial function, y = 0.0057x2 − 0.7375x + 25.411, R2 = 0.855; the goodness of fit of
this model was statistically nonsignificant, p > 0.05), whereas a significant decrease was observed in women (logarithmic function, y = −4.8112 ln(x) + 22.0117, R2 = 0.935). 4. Discussion It has been previously established that both morbidity and mortality are associated with changes in leukocyte count in adults and older people. Therefore, it was necessary to explore the association between leukocyte count and longevity in physically healthy individuals. Consistent with the findings of earlier research on the relationship between TLC and all-cause mortality (Leng et al., 2005a,b), our analysis showed that the individuals with lower yet normal TLC lived on average longer compared with those who had higher yet normal TLC (ANOVA and regression analysis, p < 0.05). It appears that the prognostic value of these indicators is stronger in women because the models of regression of age-related changes are better adjusted. The longitudinal data (Figs. 1A–6A; Table 1) show what changes in TLC and in the frequency of different types of leukocytes occur with age. For example, there is a gradual increase in the frequency of monocytes in both sexes (Fig. 6A). The cross-sectional data (Figs. 1B–6B; Table 2) show that there can be differences in the mean values of the compared parameters between those who had shorter and longer ALS. For example, the longer the women lived, the lower frequency of monocytes they had (Fig. 6B). The model of regression was well adjusted in women but not in men. Furthermore, it seems
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
7
Fig. 4. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in eosinophil count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
that when TLC decreases with age (e.g. at the age of 55), there is an increase in the frequency of lymphocytes and when TLC increases with age, there is a decrease in the frequency of lymphocytes (the longitudinal data, cf. Figs. 1A and 5A). In the cross-sectional data, the frequency of granulocytes in women from the two last categories of average lifespan (i.e. 68 and 76+) is always greater than at the age of 63 (Fig. 2B), albeit these differences are on the border of statistical significance. The longitudinal data show that the frequency of lymphocytes in women decreases in the three last five-year periods; i.e. starting from the age of 60 onward (Fig. 5A), whereas the crosssectional data show an increase in the frequency of lymphocytes since the lowest value occurs at the age of 53 and the highest value is observed at the age of 76+ (Fig. 5B). In men, these differences and age-dependent changes in the frequency of these selected types of leukocytes were statistically nonsignificant. Therefore, the comparison of longitudinal and cross-sectional changes in differential counts points to higher number of granulocytes and lymphocytes as predictors of longer lifespan in women but not in men. Although TLC increases with advancing age in the longitudinal sample, starting from the age of 60 onward in men and starting from the age of 55 onward in women, it decreases with age in the cross-sectional sample starting from the age of 63 onward in both sexes. The comparison of longitudinal and cross-sectional data revealed that TLC within the normal range (4.0–10.0 × 103 L−1 ) is inversely correlated with longevity in both sexes, but especially in men. The observed pattern of differences in longitudinal and cross-sectional changes with age in TLC provides evidence that
individuals with lower but normal TLC tend to live longer compared with individuals with higher but normal TLC. The lower level of TLC in the cross-sectional sample at the age of 53 (Fig. 1B) can be explained by the fact that at this stage of ontogeny there was an age-related decrease in TLC, as revealed by the longitudinal data (Fig. 1A). Interestingly, around the five and six decade of life, there is excess cardiovascular mortality in men compared with women in the Polish population. It is possible that men who had higher average levels of TLC were more likely to die compared with those who had lower average levels of TLC. Thus the first group was lost to the population and the arithmetic mean decreased as a consequence of this phenomenon. Some authors reported the existence of downward secular changes in TLC in the years 1958–2002 (Ruggiero et al., 2007). It should be stressed that the discrepancies between patterns of age-associated alterations in TLC that were observed in the longitudinal and cross-sectional samples in the present study were not caused by such trends in TLC because all these changes were strictly controlled. At the stage of selecting subjects to the groups, we chose only inmates from known birth cohorts (1911–1933) and the secular trends in total and differential counts were under control (secular changes were statistically nonsignificant). Therefore, higher TLC seems to be a risk factor whose predictive value remains comparatively strong and relevant to the assessment of long-term survival in the consecutive birth cohorts. As to sex differences, men had constantly higher TLC than women at each consecutive stage of aging. The difference between
8
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Fig. 5. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in lymphocyte count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
sexes in TLC ranged from 0.6 to 0.9 × 103 L−1 . The age-dependent decline in leukocyte count remained larger in women until the age of 55, which was probably the main cause of this divergence. The observed variances, however, can be interpreted in the context of sex-related patterns of immunosenescence (Gruver et al., 2007; Leng and Fried, 2009). In older ages, the immune system experiences an age-dependent deterioration in function, which is a consequence of a continuous attrition caused by chronic antigenic overload and consists in a gradual decrease in immunological competence (Franceschi et al., 1999). More recent studies revealed specific differences between sexes in the function of the immune system that proceed from faster immunosenescence in men and consist in different patterns of age-associated changes in subpopulations of T cells and B cells (Caruso et al., 2013; Hirokawa et al., 2013). Moller et al. (2009) reported that the rate of telomere attrition is slower in women. Epel et al. (2009) showed that the rate of leukocyte telomere shortening may serve as a reliable predictor of mortality from CVD in older men. Furthermore, it is well established that systemic inflammation and related factors play an important role in human senescence (Gruver et al., 2007; Leng and Fried, 2009). Interestingly, the gender gap in life expectancy is also partly attributable to specific differences in function of the immune system in older men and women. The results of previous studies revealed that patients with higher number of leukocytes and neutrophil granulocytes and lower
lymphocyte count were at a higher risk of death, after allowing for age, gender, race, education level, smoking, or body mass index (BMI). In postmenopausal women, TLC at the level of >6.7 × 103 L−1 successfully identified high-risk individuals who were not identified by other traditional CVD risk factors (Margolis et al., 2005). Subsequently, it was demonstrated that in patients with >6.0 × 103 L−1 , lower age at death can be forecast (Ruggiero et al., 2007). Therefore, under some conditions, higher level of TLC may serve as a risk factor in respect of morbidity, CVD incidents, and allcause mortality (Weiss et al., 1995). Some researchers have come to the conclusion that TLC can be used as a proxy of systemic inflammation and a reliable marker of subclinical disease in older adults and elderly people. Other authors suggest that TLC deserves attention as a clinically useful predictor of survival in the 75-year-olds, especially in women (Nilsson et al., 2014). Certain limitations of the present study should be acknowledged. First, we analyzed long-term trends in rapidly reacting parameters in which severe changes may occur within hours due to recruitment from vascular or bone marrow reserve pools, or even within minutes due to changes in endothelial adhesion or transmigration. However, these effects seem to be relatively not so important to the final results of our analysis because only averaged data obtained from frequently repeated measurements in many individuals were used. Extreme values, i.e. both minimum
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
9
Fig. 6. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in monocyte count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
and maximum from each subject, were routinely removed during the process of statistical analysis. The arithmetic means were calculated for each five-year interval on the basis of repeated measurements (from 10 up to 18 times within a five-year period) for each subject and then for the compared groups of subjects. For example, the mean value ± standard deviation of TLC for the age of 45 in the longitudinal sample in men is 6.797 ± 1.546 × 103 L−1 , which was established on the basis of numerous measurements in 68 individuals. Therefore, there is a low probability that the final results of such a long-lasting population study, which consisted in comparison of longitudinal and cross-sectional data, could have been encumbered with these effects. However, we have no exact data on the reliability of repeated measurements and therefore we cannot provide a mathematical proof or other evidence that such otherwise important effects did not interfere with the observed long-term trends in TLC, a limitation of our study that should be avoided in the future investigations. Second, the patients constituted a fairly specific group of subjects because they lived for many years in the hospital and some of them were treated for psychiatric disorders with psychoactive drugs. However, no statistically significant differences were found between the group of subjects who underwent such medical treatment and those who took these drugs every now and then (ANOVA and test, t-p > 0.05), except for higher number of eosinophils in the treated inmates (at the age of 45, n = 16, 4.5 ± 3.1 × 103 vs. n = 52, 3.0 ± 2.2 × 103 , t = 2.13, p = 0.037, F = 1.88, p = 0.096). Third, the baseline TLC was comparatively high at the commencement of the study and averaged >6.0 × 103 , which may be
attributed to poor health status of the inmates. Similarly, TLC was higher in older ages as the U-shaped pattern was observed instead of a gradual decline in leukocyte count. The latter is a typical pattern of changes with age in TLC in healthy older individuals. Acknowledgments The authors would like to thank the anonymous reviewers for their helpful comments and constructive suggestions. References Aminzadeh, Z., Parsa, E., 2011. Relationship between age and peripheral white blood cell count in patients with sepsis. Int. J. Prev. Med. 2, 238–242. Borysławski, K., Chmielowiec, K., Chmielewski, P., Chmielowiec, J., 2015. Changes with age in selected anthropometric, physiological and biochemical traits and their association with lifespan of men and women. Monographs of Physical Anthropology, vol. 2. DN Publisher, Wrocław (In Polish). Caruso, C., Accardi, G., Virruso, C., Candore, G., 2013. Sex, gender and immunosenescence: a key to understand the different lifespan between men and women? Immun. Ageing 10, 20. Chmielewski, P., Borysławski, K., Chmielowiec, K., Chmielowiec, J., 2015. Height loss with advancing age in a hospitalized population of Polish men and women: magnitude, pattern and associations with mortality. Anthropol. Rev. 78 (2), 157–168. Danesh, J., Collins, R., Appleby, P., Peto, R., 1998. Association of fibrinogen, C-reactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies. JAMA 279, 1477–1482. Epel, E.S., Merkin, S.S., Cawthon, R., Blackburn, E.H., Adler, N.E., Pletcher, M.J., Seeman, T.E., 2009. The rate of leukocyte telomere shortening predicts mortality from cardiovascular disease in elderly men. Aging 1, 81–88.
10
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Erlinger, T.P., Muntner, P., Helzlsouer, K.J., 2004. WBC count and the risk of cancer mortality in a national sample of U.S. adults: results from the Second National Health and Nutrition Examination Survey mortality study. Cancer Epidemiol. Biomarkers Prev. 13, 1052–1056. Franceschi, C., Valensin, S., Fagnoni, F., Barbi, C., Bonafe, M., 1999. Biomarkers of immunosenescence within an evolutionary perspective: the challenge of heterogeneity and the role of antigenic load. Exp. Gerontol. 34 (8), 911–921. Gruver, A.L., Hudson, L.L., Sempowski, G.D., 2007. Immunosenescence of ageing. J. Pathol. 211, 144–156. Hirokawa, K., Utsuyama, M., Hayashi, Y., Kitagawa, M., Makinodan, T., Fulop, T., 2013. Slower immune system aging in women versus men in the Japanese population. Immun. Ageing 10, 19. Horne, B.D., Anderson, J.L., John, J.M., Weaver, A., Bair, T.L., Jensen, K.R., Renlund, D.G., Muhlestein, J.B., Intermountain Heart Collaborative (IHC) Study Group, 2005. Which white blood cell subtypes predict increased cardiovascular risk? J. Am. Coll. Cardiol. 45, 1638–1643. ´ Krzyminski, S., 2000. Kilka refleksji psychiatry na przełomie stuleci (Some reflections ˛ Psychiatrii i Neurologii 9, 507–508 (In Polish). at the turn of the century). Postepy Leng, S.X., Xue, Q.L., Huang, Y., Ferrucci, L., Fried, L.P., Walston, J.D., 2005a. Baseline total and specific differential white blood cell counts and 5-year all-cause mortality in community-dwelling older women. Exp. Gerontol. 40, 982–987. Leng, S.X., Xue, Q.L., Huang, Y., Semba, R., Chaves, P., Bandeen-Roche, K., Fried, L., Walston, J., 2005b. Total and differential white blood cell counts and their associations with circulating interleukin-6 levels in community-dwelling older women. J. Gerontol., A: Biol. Sci. Med. 60, 195–199. Leng, S.X., Fried, L.P., 2009. Inflammatory markers and frailty. In: Fulop, T., Franceschi, C., Hirokawa, K., Pawelec, G. (Eds.), Handbook on Immunosenescence. Basic Understanding and Clinical Applications. Springer, New York, NY. Leng, S.X., Xue, Q.L., Tian, J., Huang, Y., Yeh, S.H., Fried, L.P., 2009. Associations of neutrophil and monocyte counts with frailty in community-dwelling disabled older
women: results from the Women’s Health and Aging Studies I. Exp. Gerontol. 44, 511–516. Margolis, K.L., Manson, J.E., Greenland, P., Rodabough, R.J., Bray, P.F., Safford, M., Grimm Jr., R.H., Howard, B.V., Assaf, A.R., Prentice, R., Women’s Health Initiative Research Group, 2005. Leukocyte count as a predictor of cardiovascular events and mortality in postmenopausal women: the Women’s Health Initiative Observational Study. Arch. Intern. Med. 165, 500–508. Meintker, L., Ringwald, J., Rauh, M., Krause, S.W., 2013. Comparison of automated differential blood cell counts from Abbott Sapphire, Simens Advia 120, Beckman Coulter DxH 800, and Sysmex XE-2100 in normal and pathologic samples. Am. J. Clin. Pathol. 139, 641–650. Moller, P., Mayer, S., Mattfeldt, T., Muller, K., Wiegand, P., Bruderlein, S., 2009. Sexrelated differences in length and erosion dynamics of human telomeres favor females. Aging 1, 733–739. Naeim, F., Rao, P.N., Song, S.X., Grody, W.W., 2013. Atlas of hematopathology. Morphology, immunophenotype, cytogenetics, and molecular approaches. Academic Press, New York, NY. Nilsson, G., Hedberg, P., Öhrvik, J., 2014. White blood cell count in elderly is clinically useful in predicting long-term survival. J. Aging Res. 2014, 475093. Ruggiero, C., Metter, E.J., Cherubini, A., Maggio, M., Sen, R., Najjar, S.S., Windham, G.B., Ble, A., Senin, U., Ferrucci, L., 2007. White blood cell count and mortality in the Baltimore Longitudinal Study of Aging. J. Am. Coll. Cardiol. 49, 1841–1850. Weiss, S.T., Segal, M.R., Sparrow, D., Wager, C., 1995. Relation of FEV1 and peripheral blood leukocyte count to total mortality, The Normative Aging Study. Am. J. Epidemiol. 142, 493–498. Wheeler, J.G., Mussolino, M.E., Gillum, R.F., Danesh, J., 2004. Associations between differential leucocyte count and incident coronary heart disease: 1764 incident cases from seven prospective studies of 30,374 individuals. Eur. Heart J. 25, 1287–1292.
Contents lists available at ScienceDirect
Annals of Anatomy journal homepage: www.elsevier.com/locate/aanat
Research article
The association between total leukocyte count and longevity: Evidence from longitudinal and cross-sectional data Piotr Paweł Chmielewski a,∗ , Krzysztof Borysławski b , Krzysztof Chmielowiec c , Jolanta Chmielowiec d , Bartłomiej Strzelec a a
Department of Anatomy, Faculty of Medicine, Wroclaw Medical University, ul. Tytusa Chałubi´ nskiego 6a, 50-368 Wrocław, Poland ˙ Department of Anthropology, Institute of Biology, Wroclaw University of Environmental and Life Sciences, ul. Kozuchowska 5, 51-631 Wrocław, Poland c ˛ Regional Psychiatric Hospital for People with Mental Disorders, Cibórz 5, 66-213 Skape, Poland d Faculty of Education, Sociology and Health Sciences, University of Zielona Góra, Al. Wojska Polskiego 69, 65-762 Zielona Góra, Poland b
a r t i c l e
i n f o
Article history: Received 14 March 2015 Received in revised form 9 September 2015 Accepted 9 September 2015 Keywords: Leukocyte count Inflammation Aging Senescence Longevity Longitudinal studies
a b s t r a c t The aim of the study was to evaluate the relationship between age-dependent changes in total leukocyte count (TLC) and certain selected differential counts expressed as frequencies (granulocytes, band cells, eosinophils, lymphocytes, and monocytes) and longevity in physically healthy men and women aged 45+. Longitudinal data on cell counts from 142 subjects (68 men and 74 women; all aged 45–70 and examined for 25 years) were compared with cross-sectional data from 225 subjects (113 men and 112 women; this group was divided into four categories of average lifespan; i.e.: 53, 63, 68, and 76+ years of age). ANOVA, t-test, and regression analysis were employed. Secular changes in leukocyte count were controlled. Men had continuously higher TLC compared with women. Moreover, sex differences in patterns of changes with age were found. The longitudinal assessment revealed a U-shaped pattern of changes in TLC in men (y = 0.0026x2 − 0.2866x + 14.4374; R2 = 0.852) and women (y = 0.0048x2 − 0.5386x + 20.922; R2 = 0.938), whereas the cross-sectional comparison showed an inverted U-shaped pattern in men (y = −0.0021x2 + 0.2421x; R2 = 0.417) and women (y = −0.0017x2 + 0.2061x; R2 = 0.888). In general, the comparison of longitudinal and cross-sectional data on changes with age in TLC indicates that longevity favors individuals with lower yet normal TLC and this correlation is more pronounced in men. In conclusion, our findings are in line with previous longitudinal studies of aging and suggest that lower TLC within the normal range (4.0–10.0 × 103 L−1 ) can be a useful predictor of longevity in physically healthy individuals. © 2015 Elsevier GmbH. All rights reserved.
1. Introduction Like erythrocytes and platelets, leukocytes are derived from hematopoietic stem cells (HSCs) in the red bone marrow, which is a mesenchymal-derived complex structure consisting of hematopoietic precursors and a complex microenvironment that facilitates the maintenance of HSCs and supports the differentiation and maturation of the progenitors (Naeim et al., 2013). Unlike erythrocytes and platelets, leukocytes are large, nucleated, translucent, and morphologically heterogeneous cells which fulfill different defensive functions. There are two major categories of leukocytes: granulocytes (neutrophils, eosinophils, and basophils) and agranulocytes (monocytes and lymphocytes). Most of them have
∗ Corresponding author. Tel.: +48 71 784 13 45; fax: +48 71 784 00 79. E-mail address: [email protected] (P.P. Chmielewski). http://dx.doi.org/10.1016/j.aanat.2015.09.002 0940-9602/© 2015 Elsevier GmbH. All rights reserved.
a short lifespan, ranging from a few hours or several days (granulocytes) to a few months (monocytes) or even several years (lymphocytes). Leukocytes are constitutively produced throughout ontogeny and removed from the blood by the liver and spleen. Normally, they constitute less than 1% of whole blood (in adults, range, 4.0–10.0 × 103 L−1 ), but their number can double or can even increase tenfold within hours since there are reserve pools of these cells in bone marrow, spleen, and lymph nodes. Leukocytes can leave the blood by squeezing and migrating between the endothelial cells through pores in the capillary walls. The process of diapedesis is accelerated during inflammation. The functions of leukocytes are not confined to producing and distributing antibodies during the immune response but include removing toxins or wastes and destroying damaged or abnormal cells through phagocytosis. Phagocytic cells located in the reticular connective tissue are part of the mononuclear phagocyte system (MPS).
2
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Total leukocyte count (TLC) increases in response to infection, trauma, inflammation, and certain diseases. For example, sepsis is accompanied by a surge in the number of leukocytes (Aminzadeh and Parsa, 2011). There are, however, many factors that can affect leukocyte count in healthy subjects such as sex, genetic factors, stress level, diet, nutrition, and lifestyle, e.g. tobacco-induced inflammatory changes. First of all, leukocyte count changes with age: in newborns it is two to three times higher than in adults and in elderly people it diminishes gradually. Although leukopenia, neutropenia, and lymphocytopenia have long been recognized as indicators of poor health and severely impaired immunity, a growing body of evidence suggests now that higher TLC within the normal range is related to increased cumulative all-cause mortality and lower chances of long-term survival. Moreover, some recent studies have revealed that increased TLC is significantly associated with cardiovascular mortality in both sexes and with noncardiovascular mortality in women. Interestingly, hazard ratios were essentially unchanged by adjustment for risk factors such as smoking, hypertension, myocardial infarction, diabetes mellitus, total cholesterol, high-density lipoprotein cholesterol, and body mass index (Nilsson et al., 2014). The discovery of an increased risk of mortality in older women with elevated yet normal level of baseline leukocyte counts stimulated the ongoing debate about the relationship between TLC or differential cell counts and survival probability in adults and older people (Horne et al., 2005; Leng et al., 2005a,b, 2009). Higher level and upward changes in TLC can be used as an important clinical marker of chronic systemic inflammation and a negative prognostic in patients with cardiovascular disease (CVD), coronary heart disease (CHD), stroke, and cancer (Danesh et al., 1998; Erlinger et al., 2004; Wheeler et al., 2004; Leng and Fried, 2009). Ruggiero et al. (2007) demonstrated a nonlinear relationship between TLC and all-cause mortality and cancer. There was, however, a linear relationship between leukocyte count and mortality from CVD. Nevertheless, the role of TLC as an independent predictor of the first cardiovascular incident remains uncertain. Likewise, there is no evidence that leukocyte count is related to longevity in physically healthy people who are not at higher risk of CVD, CHD, stroke, and cancer. Although the results of several studies confirmed the positive relationship between TLC and mortality in older adults and elderly people, little work has been devoted to changes with age in leukocyte count in short and long-lived subjects on the basis of longitudinal studies of aging. The purpose of our study was to determine rates and patterns of such longitudinal as well as crosssectional age-related changes in TLC, including aging-associated alterations in differential cell counts expressed as frequencies (granulocytes, band cells, eosinophils, lymphocytes, and monocytes) in physically healthy men and women aged 45+, with special reference to the correlation with average lifespan (ALS) in the compared groups of subjects; i.e. short versus long-lived individuals.
2. Material and methods 2.1. Study population Out of the total number of patients (n = 3500) who lived at the Regional Psychiatric Hospital for People with Mental Disorders in Ciborz, Lubuskie Province, Poland, in the years 1960–2000, we have carefully selected longitudinal data on hematological parameters from 142 physically healthy individuals (including 68 men and 74 women) and cross-sectional data from 225 physically healthy individuals (including 113 men and 112 women). All subjects were Caucasian, born in the years 1911–1933. Their health was evaluated during regular physicals at the hospital on its premises by hospital
staff. In the years 1960–1989, the asylum provided custody for helpless and resourceless people from the underclass, thereby confining not only individuals with severe mental disorders but also persons who needed special care due to mild mental problems and poverty. The reason for keeping these people there for many years, was to separate them from the rest of the socialistic society, which ´ 2000; was a socially and politically motivated decision (Krzyminski, Borysławski et al., 2015; Chmielewski et al., 2015). During their lengthy stay at the hospital, the patients would take some powerful psychoactive drugs. Therefore, we selected solely data from patients who hardly ever had been treated with strong medicines or who had been treated so every once in a while. It is noteworthy that the patients lived for many years under very similar and relatively prosperous environmental conditions, maintained virtually the same lifestyle and had practically the same diet prescribed by a dietitian at the hospital. This fact undoubtedly boosts the value of the study sample and makes it quite unique. Nevertheless, detailed written data on diet, calorie intake, physical activity, and so forth were not available and were not controlled in the research. The patients from the longitudinal sample (n = 142), aged 45 at the beginning and 70 years at the end of the analyzed period, stayed continuously at the hospital and all lived to be at least 70 years of age. No further information on lifespan of these patients was available. In the years 1999–2000, there were sweeping reforms in the functioning of the medical institutions in Poland and those from the longitudinal sample who survived to be older than 70 years of age, were transported to other hospitals or medical institutions in the country. The patients from the cross-sectional sample (n = 225) were confined at the same hospital but differed in lifespan. Leading causes of death were predominantly aging-associated diseases such as ‘cardiorespiratory failure’ (CRF), cardiovascular disease (CVD), stroke, and cancer. The cross-sectional sample was divided into four categories of average lifespan (ALS), i.e.: (1) 53 years of age (the limit of individual lifespan for this group was 57.5 years of age; the group consisted of 22 men and 12 women, ALS was 53 and 52 years of age, respectively), (2) 63 years of age (the limit of individual lifespan for this group was 65 years; the group was comprised of 27 men and 30 women, ALS was 63 years for both sexes), (3) 68 years of age (the limit of individual lifespan for this group was 72.5 years; the group was made up of 49 men and 40 women, ALS was 67.6 years for men and 68 years for women), and (4) 76+ (there were no limits since this group comprised living subjects, 15 men and 30 women, their average age was 76 for both sexes). 2.2. Data collection and statistical analysis Data on hematological parameters were obtained from the hospital archives. All information used in the present study was derived from physical examinations of the inmates who were there in the years 1960–2000. The process of data collection was performed with permission and consent of hospital authorities. The study was carried out in accordance with the Declaration of Helsinki and consisted in observation only. No experiments were conducted. The medical files were anonymized so as not to divulge any personal or confidential information. On the basis of medical records and files that had been stored at the archives, we created a large computerized database in the years 2005–2007. Two types of materials were obtained: longitudinal and cross-sectional (see Section 2.1). All measurements were taken in accordance with internationally accepted standards and requirements by hospital staff. Blood samples from the median cubital vein were drawn by a nurse at the hospital. During the 25-year study period, complete blood tests were performed from 10 up to 18 times within each five-year period for a very long time, that is for 25 years in the case of the longitudinal sample. On the basis of such frequently repeated measurements for
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
3
Table 1 Characteristics of the patients (n = 142) from the longitudinal assessment at the onset of the study (all subjects aged 45) and in the consecutive age categories; p-values <0.05 are given in bold type. Age
Trait (unit)
Men (n = 68)
Women (n = 74)
t–test
Mean
SD
Mean
SD
p–value
45
TLC, 103 L−1 Granulocytes (%) Band cells (%) Eosinophils (%) Lymphocytes (%) Monocytes (%)
6.8 68.0 3.5 3.3 30.6 2.0
1.5 8.1 2.2 2.5 8.1 1.7
6.3 66.5 3.5 3.3 31.8 2.3
2.0 7.0 1.9 2.5 6.8 2.4
1.50 1.21 0.20 0.06 1.00 0.68
0.135 0.228 0.843 0.953 0.317 0.500
50
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.4 67.3 3.1 2.8 31.0 1.8
1.6 6.4 1.6 2.0 6.1 1.3
5.8 66.6 3.4 2.4 31.8 2.4
1.3 5.5 1.9 1.2 5.6 3.2
2.52 0.76 1.01 1.26 –0.83 –1.48
0.013 0.450 0.317 0.210 0.408 0.141
55
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.5 66.5 2.8 3.3 32.2 2.3
1.7 6.6 2.2 2.5 6.0 1.7
5.8 66.2 3.2 2.6 32.0 2.5
1.3 5.6 1.7 1.3 5.6 1.4
2.94 0.29 –0.46 0.51 0.20 –0.90
0.004 0.771 0.646 0.608 0.840 0.370
60
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.4 65.9 3.2 3.3 32.5 2.4
1.5 7.2 3.4 2.5 5.8 1.5
5.8 64.9 3.0 2.7 32.7 2.6
1.7 7.5 1.5 1.8 7.3 1.7
2.32 0.80 1.22 0.95 –0.23 –0.69
0.022 0.426 0.225 0.344 0.822 0.493
65
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.8 66.1 3.4 2.6 31.4 3.2
2.1 5.9 1.5 1.5 5.7 3.1
5.9 66.5 3.0 2.7 31.1 3.1
1.4 7.2 2.4 2.1 7.2 2.4
2.93 –0.38 1.32 –0.20 0.29 0.23
0.004 0.708 0.189 0.845 0.771 0.816
70
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.9 67.5 3.3 3.0 29.6 3.3
2.5 7.3 1.9 2.9 7.2 2.1
6.6 68.2 3.1 2.6 28.3 3.8
2.2 8.6 2.9 1.6 7.8 3.6
0.84 –0.47 0.62 1.13 0.99 –1.07
0.405 0.639 0.537 0.260 0.324 0.288
Table 2 Characteristics of the patients (n = 225) from the cross-sectional assessment in the consecutive age categories; p-values <0.05 are given in bold type. Age
Trait (unit)
Men (n = 113)
Women (n = 112)
t-test
p-value
Mean
SD
Mean
SD
53
TLC, 103 L−1 Granulocytes (%) Band cells (%) Eosinophils (%) Lymphocytes (%) Monocytes (%)
6.4 70.6 2.4 2.1 26.9 2.2
1.8 6.9 1.7 1.3 9.8 1.3
6.1 67.3 2.2 2.2 31.0 3.1
1.4 8.2 0.9 1.4 8.9 2.7
0.51 1.14 0.36 –0.34 –1.12 –1.16
0.611 0.263 0.719 0.737 0.272 0.258
63
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
7.6 64.2 2.0 2.4 34.9 1.7
2.0 6.3 1.5 1.9 6.5 1.6
6.3 65.5 2.2 2.4 33.0 1.8
1.5 7.6 1.3 1.8 8.3 1.5
2.74 –0.71 –0.58 0.13 0.95 –0.29
0.008 0.480 0.564 0.894 0.347 0.776
68
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
6.7 67.7 2.6 1.9 32.1 1.3
1.2 7.3 1.3 1.2 7.5 1.3
6.1 67.5 3.2 2.6 32.0 1.8
3.5 5.9 1.6 2.3 5.9 1.2
1.74 0.17 –2.03 –1.81 0.05 –2.10
0.086 0.869 0.045 0.074 0.960 0.039
76+
TLC Granulocytes Band cells Eosinophils Lymphocytes Monocytes
5.9 67.3 2.4 2.0 32.8 2.2
1.0 6.6 1.2 1.3 7.6 2.9
5.8 66.0 2.7 2.7 34.3 1.3
1.7 6.3 1.5 1.3 6.3 1.1
0.20 0.59 –0.48 –1.73 –0.65 1.39
0.843 0.558 0.636 0.092 0.519 0.174
4
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Fig. 1. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in total leukocyte count (TLC) in the consecutive age categories of men and women: arithmetic means ± standard deviations, differences between sexes tested by t-test (* p < 0.05, ** p < 0.01, *** p < 0.001), and models of regression with coefficients of determination (R2 ) are shown.
each subject, we calculated arithmetic means, standard deviations (SD), standard errors, and other statistics for each five-year interval. Thus an arithmetic mean for a given age category in both longitudinal and cross-sectional sample was always calculated on the basis of frequently repeated measurements in many individuals. Blood cell counting was performed by medical laboratory scientists using an optical microscope and Bürker chamber holding a specified volume of Giemsa stained diluted blood. In the years under study, automatic cell counters were not widely available in Poland and therefore only the traditional manual method was employed throughout the period of 25 years. It is worth mentioning that the manual method can be perceived as good and reliable since it is used as reference when different methods of cell counting are compared in terms of accuracy (Meintker et al., 2013). The traditional manual method is often perceived as more accurate and reliable than different automatic methods in the case of some anomalies of blood cells. To determine the rate and patterns of changes with age in the analyzed characteristics of the blood, one-way ANOVA, t-test, and regression analysis were performed. The method of least squares was used. The goodness of fit of a given regression model was confirmed only when a coefficient of determination (R2 ) reached the highest value and an unknown parameter (ˇ0 ) as well as a coefficient of
regression (ˇ1 ) were statistically significant (p ≤ 0.05). Five types of functions were tested: (I) linear function, y = ˇ1 age + ˇ0 , (II) logarithmic function, y = ˇ1 ln age + ˇ0 , (III) polynomial function, y = ˇ1 age2 + ˇ2 age + ˇ0 , (IV) exponential type I, y = ˇ1 agea , and (V) exponential type II, y = ˇ1 ea(age) , where (y) represents a value of an analyzed characteristic changing through aging, (ˇ2 ) denotes the second coefficient of regression, (a) stands for the exponent, and (e) is the base of the natural logarithm. 3. Results 3.1. Longitudinal study In the two samples, TLC and differential counts were normally distributed (K–S test, p > 0.2). The baseline characteristics of the study population in terms of total and differential cell counts are presented in Table 1. The regression analysis has revealed significant aging-associated changes in TLC in both sexes. In men and women, a U-shaped pattern was observed and an appropriate model of regression was polynomial (Fig. 1A). In the first five-year periods of study, leukocyte count decreased slightly but it increased in the later phases of aging. Interestingly, the greatest decrease in the number of leukocytes occurred at similar stages of ontogeny in men and women. There was a steady tendency for TLC to be
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
5
Fig. 2. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in granulocyte count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
higher in men throughout the 25-year study period. Significant differences between sexes were found in the age category of 50 (t = 2.52, p = 0.013), 55 (t = 2.94, p = 0.004), 60 (t = 2.32, p = 0.022), and 65 (t = 2.93, p = 0.004). The curve of regression for the frequency of granulocytes assumed a U-shaped pattern of alterations in both sexes (Fig. 2A). The best fitting model was polynomial in men (y = 0.0104x2 − 1.2383x + 102.797, R2 = 0.878) and women (y = 0.0111x2 − 1.24x + 100.135, R2 = 0.660). The lowest level of these cells occurred at the age of 60 in men (65.9%) and women (64.9%), whereas the highest level occurred at the age of 45 in men (68.0%) and at the age of 70 in women (68.2%). No significant sex differences in granulocyte count were observed over the whole period under study (ANOVA, p > 0.05 for each age category). Band cells count (BCC) in men decreased in the first five-year period and then it rose (y = 0.0005x2 − 0.0608x + 4.9975, R2 = 0.089). Therefore, the lowest level occurred at the age of 50 (3.1%). In both sexes, the highest BCC was at the age of 45 (Table 1, Fig. 3A). By contrast, there was a steady decline in BCC during the study period in women (y = 0.0013x2 − 0.1729x + 8.693, R2 = 0.966). There were no significant changes with age in the frequency of eosinophil granulocytes in men and women (Fig. 4A). The models of regression were polynomial (for men: y = 0.0018x2 − 0.2142x + 9.2638, R2 = 0.329; for women: y = 0.002x2 − 0.2462x + 10.1298, R2 = 0.444). The frequency of lymphocytes in men and women showed an inverted U-shaped pattern and the highest level of these cells occurred at the age of 60
(32.5% and 32.7%, respectively, t = 0.225, p = 0.057). No significant sex differences were found (Fig. 5A). The patterns of changes with age were similar and commensurate in both sexes. The model was polynomial in men (y =−0.0144x2 + 1.6325x − 14.1607, R2 = 0.887) and women (y =−0.0148x2 + 1.599x − 10.5286, R2 = 0.870). A gradual increase in the frequency of monocytes was found in men and women (Fig. 6A). The model of regression was of the exponential type II in men (y = 0.5739e0.0251x , R2 = 0.874) but polynomial in women (y = 0.0032x2 − 0.3097x + 9.8247, R2 = 0.974). 3.2. Cross-sectional study Table 2 shows the baseline characteristics of the subjects from the cross-sectional assessment in terms of total and differential cell counts. In patients who differed in lifespan, the curve of regression assumed an inverted U-shaped pattern of changes and the models tuned out to be polynomial (for men: y = −0.0021x2 + 0.2421x, R2 = 0.417; for women: y = −0.0017x2 + 0.2061x, R2 = 0.888; Fig. 1B). Men had significantly higher TLC than women at the category of age 63 (t = 2.74, p = 0.008). In women, TLC increased in the first period and then diminished continuously. In general, after the age of 60, the longer the patients lived, the lower TLC they had. Thus individuals with low yet normal TLC lived significantly longer and were more likely to survive to the age category of 76+ compared with those who had high but normal TLC (for men: n = 15, mean ± SD: 5.9 ± 1.0 × 103 L−1 , range 4.6–7.6 × 103 L−1 ; for women: n = 24,
6
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Fig. 3. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in band cell count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
5.8 ± 1.7 × 103 L−1 , range 3.6–10.2 × 103 L−1 ). The relationship was more pronounced in men. With respect to the frequency of granulocytes, the model of regression was polynomial in both sexes Fig. 2B). The function was U-shaped in men (y = 0.0238x2 − 3.1766x + 171.8592, R2 = 0.616), whereas a slow decline occurred in women (y = 0.0015x2 − 0.2228x + 74.654, R2 = 0.138). There were no significant changes with age in BCC (%) in both sexes (Fig. 3B) and the function of regression was polynomial in men (y = 0.0013x2 − 0.1683x + 7.5496, R2 = 0.168) and women (y = 0.0014x2 + 0.2002x − 4.5475, R2 = 0.349). The frequency of eosinophil granulocytes showed an inverted U-shaped pattern in men and the model of regression was polynomial (y = −0.0006x2 + 0.075x, R2 = 0.215). Although there was a slow and gradual increase in the frequency of eosinophil granulocytes in women (exponential type II function, y = 1.4155e0.0086x , R2 = 0.929), no significant sex differences in terms of the frequency were observed in any of the four tested age categories (Fig. 4B). The changes with age in the frequency of lymphocytes followed an inverted U-shaped pattern in men (y = −0.0283x2 + 3.8734x − 98.542, R2 = 0.798), while there was a slow and irregular increase in the frequency of these agranulocytes in women (y = −0.0062x2 + 0.9062x, R2 = 0.565). No significant sex differences in the frequency of lymphocytes were observed (Fig. 5B). The U-shaped pattern of changes with age in the frequency of monocytes (Fig. 6B) was found in men (polynomial function, y = 0.0057x2 − 0.7375x + 25.411, R2 = 0.855; the goodness of fit of
this model was statistically nonsignificant, p > 0.05), whereas a significant decrease was observed in women (logarithmic function, y = −4.8112 ln(x) + 22.0117, R2 = 0.935). 4. Discussion It has been previously established that both morbidity and mortality are associated with changes in leukocyte count in adults and older people. Therefore, it was necessary to explore the association between leukocyte count and longevity in physically healthy individuals. Consistent with the findings of earlier research on the relationship between TLC and all-cause mortality (Leng et al., 2005a,b), our analysis showed that the individuals with lower yet normal TLC lived on average longer compared with those who had higher yet normal TLC (ANOVA and regression analysis, p < 0.05). It appears that the prognostic value of these indicators is stronger in women because the models of regression of age-related changes are better adjusted. The longitudinal data (Figs. 1A–6A; Table 1) show what changes in TLC and in the frequency of different types of leukocytes occur with age. For example, there is a gradual increase in the frequency of monocytes in both sexes (Fig. 6A). The cross-sectional data (Figs. 1B–6B; Table 2) show that there can be differences in the mean values of the compared parameters between those who had shorter and longer ALS. For example, the longer the women lived, the lower frequency of monocytes they had (Fig. 6B). The model of regression was well adjusted in women but not in men. Furthermore, it seems
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
7
Fig. 4. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in eosinophil count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
that when TLC decreases with age (e.g. at the age of 55), there is an increase in the frequency of lymphocytes and when TLC increases with age, there is a decrease in the frequency of lymphocytes (the longitudinal data, cf. Figs. 1A and 5A). In the cross-sectional data, the frequency of granulocytes in women from the two last categories of average lifespan (i.e. 68 and 76+) is always greater than at the age of 63 (Fig. 2B), albeit these differences are on the border of statistical significance. The longitudinal data show that the frequency of lymphocytes in women decreases in the three last five-year periods; i.e. starting from the age of 60 onward (Fig. 5A), whereas the crosssectional data show an increase in the frequency of lymphocytes since the lowest value occurs at the age of 53 and the highest value is observed at the age of 76+ (Fig. 5B). In men, these differences and age-dependent changes in the frequency of these selected types of leukocytes were statistically nonsignificant. Therefore, the comparison of longitudinal and cross-sectional changes in differential counts points to higher number of granulocytes and lymphocytes as predictors of longer lifespan in women but not in men. Although TLC increases with advancing age in the longitudinal sample, starting from the age of 60 onward in men and starting from the age of 55 onward in women, it decreases with age in the cross-sectional sample starting from the age of 63 onward in both sexes. The comparison of longitudinal and cross-sectional data revealed that TLC within the normal range (4.0–10.0 × 103 L−1 ) is inversely correlated with longevity in both sexes, but especially in men. The observed pattern of differences in longitudinal and cross-sectional changes with age in TLC provides evidence that
individuals with lower but normal TLC tend to live longer compared with individuals with higher but normal TLC. The lower level of TLC in the cross-sectional sample at the age of 53 (Fig. 1B) can be explained by the fact that at this stage of ontogeny there was an age-related decrease in TLC, as revealed by the longitudinal data (Fig. 1A). Interestingly, around the five and six decade of life, there is excess cardiovascular mortality in men compared with women in the Polish population. It is possible that men who had higher average levels of TLC were more likely to die compared with those who had lower average levels of TLC. Thus the first group was lost to the population and the arithmetic mean decreased as a consequence of this phenomenon. Some authors reported the existence of downward secular changes in TLC in the years 1958–2002 (Ruggiero et al., 2007). It should be stressed that the discrepancies between patterns of age-associated alterations in TLC that were observed in the longitudinal and cross-sectional samples in the present study were not caused by such trends in TLC because all these changes were strictly controlled. At the stage of selecting subjects to the groups, we chose only inmates from known birth cohorts (1911–1933) and the secular trends in total and differential counts were under control (secular changes were statistically nonsignificant). Therefore, higher TLC seems to be a risk factor whose predictive value remains comparatively strong and relevant to the assessment of long-term survival in the consecutive birth cohorts. As to sex differences, men had constantly higher TLC than women at each consecutive stage of aging. The difference between
8
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Fig. 5. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in lymphocyte count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
sexes in TLC ranged from 0.6 to 0.9 × 103 L−1 . The age-dependent decline in leukocyte count remained larger in women until the age of 55, which was probably the main cause of this divergence. The observed variances, however, can be interpreted in the context of sex-related patterns of immunosenescence (Gruver et al., 2007; Leng and Fried, 2009). In older ages, the immune system experiences an age-dependent deterioration in function, which is a consequence of a continuous attrition caused by chronic antigenic overload and consists in a gradual decrease in immunological competence (Franceschi et al., 1999). More recent studies revealed specific differences between sexes in the function of the immune system that proceed from faster immunosenescence in men and consist in different patterns of age-associated changes in subpopulations of T cells and B cells (Caruso et al., 2013; Hirokawa et al., 2013). Moller et al. (2009) reported that the rate of telomere attrition is slower in women. Epel et al. (2009) showed that the rate of leukocyte telomere shortening may serve as a reliable predictor of mortality from CVD in older men. Furthermore, it is well established that systemic inflammation and related factors play an important role in human senescence (Gruver et al., 2007; Leng and Fried, 2009). Interestingly, the gender gap in life expectancy is also partly attributable to specific differences in function of the immune system in older men and women. The results of previous studies revealed that patients with higher number of leukocytes and neutrophil granulocytes and lower
lymphocyte count were at a higher risk of death, after allowing for age, gender, race, education level, smoking, or body mass index (BMI). In postmenopausal women, TLC at the level of >6.7 × 103 L−1 successfully identified high-risk individuals who were not identified by other traditional CVD risk factors (Margolis et al., 2005). Subsequently, it was demonstrated that in patients with >6.0 × 103 L−1 , lower age at death can be forecast (Ruggiero et al., 2007). Therefore, under some conditions, higher level of TLC may serve as a risk factor in respect of morbidity, CVD incidents, and allcause mortality (Weiss et al., 1995). Some researchers have come to the conclusion that TLC can be used as a proxy of systemic inflammation and a reliable marker of subclinical disease in older adults and elderly people. Other authors suggest that TLC deserves attention as a clinically useful predictor of survival in the 75-year-olds, especially in women (Nilsson et al., 2014). Certain limitations of the present study should be acknowledged. First, we analyzed long-term trends in rapidly reacting parameters in which severe changes may occur within hours due to recruitment from vascular or bone marrow reserve pools, or even within minutes due to changes in endothelial adhesion or transmigration. However, these effects seem to be relatively not so important to the final results of our analysis because only averaged data obtained from frequently repeated measurements in many individuals were used. Extreme values, i.e. both minimum
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
9
Fig. 6. Longitudinal changes with aging (A), n = 142, and cross-sectional changes with age (B), n = 225, in monocyte count (frequency) in the consecutive age categories of men and women: description is given in the caption of Fig. 1.
and maximum from each subject, were routinely removed during the process of statistical analysis. The arithmetic means were calculated for each five-year interval on the basis of repeated measurements (from 10 up to 18 times within a five-year period) for each subject and then for the compared groups of subjects. For example, the mean value ± standard deviation of TLC for the age of 45 in the longitudinal sample in men is 6.797 ± 1.546 × 103 L−1 , which was established on the basis of numerous measurements in 68 individuals. Therefore, there is a low probability that the final results of such a long-lasting population study, which consisted in comparison of longitudinal and cross-sectional data, could have been encumbered with these effects. However, we have no exact data on the reliability of repeated measurements and therefore we cannot provide a mathematical proof or other evidence that such otherwise important effects did not interfere with the observed long-term trends in TLC, a limitation of our study that should be avoided in the future investigations. Second, the patients constituted a fairly specific group of subjects because they lived for many years in the hospital and some of them were treated for psychiatric disorders with psychoactive drugs. However, no statistically significant differences were found between the group of subjects who underwent such medical treatment and those who took these drugs every now and then (ANOVA and test, t-p > 0.05), except for higher number of eosinophils in the treated inmates (at the age of 45, n = 16, 4.5 ± 3.1 × 103 vs. n = 52, 3.0 ± 2.2 × 103 , t = 2.13, p = 0.037, F = 1.88, p = 0.096). Third, the baseline TLC was comparatively high at the commencement of the study and averaged >6.0 × 103 , which may be
attributed to poor health status of the inmates. Similarly, TLC was higher in older ages as the U-shaped pattern was observed instead of a gradual decline in leukocyte count. The latter is a typical pattern of changes with age in TLC in healthy older individuals. Acknowledgments The authors would like to thank the anonymous reviewers for their helpful comments and constructive suggestions. References Aminzadeh, Z., Parsa, E., 2011. Relationship between age and peripheral white blood cell count in patients with sepsis. Int. J. Prev. Med. 2, 238–242. Borysławski, K., Chmielowiec, K., Chmielewski, P., Chmielowiec, J., 2015. Changes with age in selected anthropometric, physiological and biochemical traits and their association with lifespan of men and women. Monographs of Physical Anthropology, vol. 2. DN Publisher, Wrocław (In Polish). Caruso, C., Accardi, G., Virruso, C., Candore, G., 2013. Sex, gender and immunosenescence: a key to understand the different lifespan between men and women? Immun. Ageing 10, 20. Chmielewski, P., Borysławski, K., Chmielowiec, K., Chmielowiec, J., 2015. Height loss with advancing age in a hospitalized population of Polish men and women: magnitude, pattern and associations with mortality. Anthropol. Rev. 78 (2), 157–168. Danesh, J., Collins, R., Appleby, P., Peto, R., 1998. Association of fibrinogen, C-reactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies. JAMA 279, 1477–1482. Epel, E.S., Merkin, S.S., Cawthon, R., Blackburn, E.H., Adler, N.E., Pletcher, M.J., Seeman, T.E., 2009. The rate of leukocyte telomere shortening predicts mortality from cardiovascular disease in elderly men. Aging 1, 81–88.
10
P.P. Chmielewski et al. / Annals of Anatomy 204 (2016) 1–10
Erlinger, T.P., Muntner, P., Helzlsouer, K.J., 2004. WBC count and the risk of cancer mortality in a national sample of U.S. adults: results from the Second National Health and Nutrition Examination Survey mortality study. Cancer Epidemiol. Biomarkers Prev. 13, 1052–1056. Franceschi, C., Valensin, S., Fagnoni, F., Barbi, C., Bonafe, M., 1999. Biomarkers of immunosenescence within an evolutionary perspective: the challenge of heterogeneity and the role of antigenic load. Exp. Gerontol. 34 (8), 911–921. Gruver, A.L., Hudson, L.L., Sempowski, G.D., 2007. Immunosenescence of ageing. J. Pathol. 211, 144–156. Hirokawa, K., Utsuyama, M., Hayashi, Y., Kitagawa, M., Makinodan, T., Fulop, T., 2013. Slower immune system aging in women versus men in the Japanese population. Immun. Ageing 10, 19. Horne, B.D., Anderson, J.L., John, J.M., Weaver, A., Bair, T.L., Jensen, K.R., Renlund, D.G., Muhlestein, J.B., Intermountain Heart Collaborative (IHC) Study Group, 2005. Which white blood cell subtypes predict increased cardiovascular risk? J. Am. Coll. Cardiol. 45, 1638–1643. ´ Krzyminski, S., 2000. Kilka refleksji psychiatry na przełomie stuleci (Some reflections ˛ Psychiatrii i Neurologii 9, 507–508 (In Polish). at the turn of the century). Postepy Leng, S.X., Xue, Q.L., Huang, Y., Ferrucci, L., Fried, L.P., Walston, J.D., 2005a. Baseline total and specific differential white blood cell counts and 5-year all-cause mortality in community-dwelling older women. Exp. Gerontol. 40, 982–987. Leng, S.X., Xue, Q.L., Huang, Y., Semba, R., Chaves, P., Bandeen-Roche, K., Fried, L., Walston, J., 2005b. Total and differential white blood cell counts and their associations with circulating interleukin-6 levels in community-dwelling older women. J. Gerontol., A: Biol. Sci. Med. 60, 195–199. Leng, S.X., Fried, L.P., 2009. Inflammatory markers and frailty. In: Fulop, T., Franceschi, C., Hirokawa, K., Pawelec, G. (Eds.), Handbook on Immunosenescence. Basic Understanding and Clinical Applications. Springer, New York, NY. Leng, S.X., Xue, Q.L., Tian, J., Huang, Y., Yeh, S.H., Fried, L.P., 2009. Associations of neutrophil and monocyte counts with frailty in community-dwelling disabled older
women: results from the Women’s Health and Aging Studies I. Exp. Gerontol. 44, 511–516. Margolis, K.L., Manson, J.E., Greenland, P., Rodabough, R.J., Bray, P.F., Safford, M., Grimm Jr., R.H., Howard, B.V., Assaf, A.R., Prentice, R., Women’s Health Initiative Research Group, 2005. Leukocyte count as a predictor of cardiovascular events and mortality in postmenopausal women: the Women’s Health Initiative Observational Study. Arch. Intern. Med. 165, 500–508. Meintker, L., Ringwald, J., Rauh, M., Krause, S.W., 2013. Comparison of automated differential blood cell counts from Abbott Sapphire, Simens Advia 120, Beckman Coulter DxH 800, and Sysmex XE-2100 in normal and pathologic samples. Am. J. Clin. Pathol. 139, 641–650. Moller, P., Mayer, S., Mattfeldt, T., Muller, K., Wiegand, P., Bruderlein, S., 2009. Sexrelated differences in length and erosion dynamics of human telomeres favor females. Aging 1, 733–739. Naeim, F., Rao, P.N., Song, S.X., Grody, W.W., 2013. Atlas of hematopathology. Morphology, immunophenotype, cytogenetics, and molecular approaches. Academic Press, New York, NY. Nilsson, G., Hedberg, P., Öhrvik, J., 2014. White blood cell count in elderly is clinically useful in predicting long-term survival. J. Aging Res. 2014, 475093. Ruggiero, C., Metter, E.J., Cherubini, A., Maggio, M., Sen, R., Najjar, S.S., Windham, G.B., Ble, A., Senin, U., Ferrucci, L., 2007. White blood cell count and mortality in the Baltimore Longitudinal Study of Aging. J. Am. Coll. Cardiol. 49, 1841–1850. Weiss, S.T., Segal, M.R., Sparrow, D., Wager, C., 1995. Relation of FEV1 and peripheral blood leukocyte count to total mortality, The Normative Aging Study. Am. J. Epidemiol. 142, 493–498. Wheeler, J.G., Mussolino, M.E., Gillum, R.F., Danesh, J., 2004. Associations between differential leucocyte count and incident coronary heart disease: 1764 incident cases from seven prospective studies of 30,374 individuals. Eur. Heart J. 25, 1287–1292.