Secondary Logo

Journal Logo

Original Article

No Association Between Telomere Length and Survival Among the Elderly and Oldest Old

Bischoff, Claus*; Petersen, Hans Christian†‡; Graakjaer, Jesper*; Andersen-Ranberg, Karen; Vaupel, James W.‡¶; Bohr, Vilhelm A.§; Kølvraa, Steen*∥; Christensen, Kaare

Author Information
doi: 10.1097/01.ede.0000199436.55248.10


Click on the links below to access all the ArticlePlus for this article.

Please note that ArticlePlus files may launch a viewer application outside of your web browser.

All higher organisms have their DNA organized as linear molecules which, together with a number of proteins, form the chromosomes. The 2 ends of linear DNA fragments would normally pose a problem for the cell, because such free ends are at risk for both degradation and end-to-end fusions. In organisms with linear chromosomes, this problem is solved by the formation of a special loop structure at both chromosome ends. The loop is made up of a chromosomal DNA part and a protein part. The DNA part consists of the telomere sequence TTAGGG in tandem repeats, and the protein part consists of a number of proteins that bind and interact with the telomeric DNA. This structure, termed the t-loop,1 is believed to be very stable, thereby protecting the chromosome ends.

Due to the nature of the replication process, the extreme tips of the linear chromosomes are not replicated during the S-phase of the cell cycle, causing slight shortening of the telomeric DNA at each cell division.2 The “telomere hypothesis” of cellular aging postulates that this telomere shortening acts as a counting mechanism, tracking the number of cell divisions undergone by a given cell line. When the telomere length falls below a critical threshold, the cell line has expended its proliferative potential and goes into senescence.3–5 It has been suggested that senescence is either dependent on mean telomere length reaching a critical lower limit or a critical length being reached by the telomere on a given chromosome.6,7

Although it is well established that telomere shortening has an essential role in in vitro aging of somatic cells, there is so far no conclusive evidence of an involvement of telomere erosion in aging in vivo. Indirect support for such an involvement is suggested by the fact that the telomeres of many organs shorten with age.8 An attempt to approach this problem more directly was carried out by Blasco et al,9 who constructed a telomerase knockout mouse. Surprisingly, neither the knockout mouse nor several of the following generations experienced any effects. The mice were healthy until generation 6, in which certain symptoms and signs similar to normal aging were observed.10,11 If these effects are due to telomere shortening, the late debut can likely be ascribed to the fact that the mouse strain used, Mus musculus, has extremely long telomeres.

The human disease dyskeratosis congenita seems to be a human analog to the knockout mouse.12 In the autosomal-dominant variant of this disorder, there are inactivating mutations in the RNA component of the telomerase, and in the X-linked variant, there are inactivating mutations in the gene coding for dyskerin, a protein implicated in telomerase function. Individuals with this disorder are normal at birth, develop affections of skin and mucosal membranes as well as bone marrow dysfunction in adolescence, and die in the third decade. The fact that dysfunction is seen mainly in organs with high cell turnover could indicate that cell renewal from stem cells is defective in this condition and consequently that a negative effect of excessive telomere shortening may be mediated through this stem cell function.

Telomere length has also recently been shown to be associated with cardiovascular diseases13,14 and dementia.15,16 A possible association between telomere length and mortality late in life in humans was recently suggested by Cawthon et al.17 These authors measured telomere length on blood samples drawn approximately 20 years earlier from healthy individuals, then 60 to 97 years of age, and they correlated this telomere length to subsequent survival. They found that individuals with shorter telomeres had a poorer survival than those with longer telomeres. Here we analyzed the telomere length in leukocytes from whole blood obtained from a cohort of 812 individuals, age 73 to 101 years.


The DNA on which this study is based was obtained from 3 different study populations.

The Longitudinal Study of Aging Danish Twins

In 1995, the Longitudinal Study of Aging Danish Twins (LSADT) began by assessing all cooperating Danish twins age 75 years and older (a sample of 2401 individuals) who were registered in the population-based nationwide Danish Twin Registry.18 The assessment was a home-based 2-hour multidimensional interview, including cognitive and physical performance tests. The tests were repeated in 1997, including all twins who participated in 1995 as well as a sample of previously unassessed twins who were between 73 and 76 years old in 1997. The 1997 survey comprised 2172 individuals corresponding to a participation rate of 79%. We subsequently asked individuals from participating twin pairs (n = 974) to allow a trained technician to visit them in their homes and draw a sample of blood. A total of 689 individuals provided blood samples, from which telomere length could be determined for 652 individuals aged 73 to 94 years.19 For the remaining 37 samples, the amount or the quality of the DNA was not sufficient for analysis.

The Danish 1905 Cohort Study

To study nonagenarians, all Danes born in 1905 were invited in 1998 to participate in a survey similar to the LSADT and carried out by the same lay interviewers. Population-based registers were used to evaluate whether the sample was representative. Participants and nonparticipants were highly comparable with regard to marital status, institutionalization, and hospitalization patterns, although men and rural area residents were more likely to participate than women and urban residents. Despite the known difficulties of conducting surveys among the very old, the study showed that it was possible to conduct a nationwide survey, including more than 2000 fairly representative nonagenarians, using lay interviewers.20,21 A total of 129 full blood samples were obtained from participants from one county (Funen). Sampling date and telomere length were determined for a total of 118 of these individuals age 92 to 93 years.

The Longitudinal Danish Centenarian Study

This study is a nationwide survey of all persons living in Denmark who celebrated their 100th birthday during the period 1 April 1995 to 31 May 1996. The residence of all centenarians in the study population was identified through the Civil Registration System. Approximately 2 weeks after their 100th birthday, all centenarians received a letter explaining the study and asking their permission for a geriatrician and a geriatric nurse to visit them for an interview and physical examination, including phlebotomy. A total of 207 centenarians participated in the interview.22–27 One and a half years later, the surviving participants were revisited, and a total of 62 blood samples were obtained. Telomere length was determined for a total of 42 of these individuals age 101 years. For the remaining 20 samples, the amount or the quality of the DNA was insufficient.

In all 3 surveys, informed consent was obtained from the participants after explaining the nature and possible consequences of this study. This study was approved by the Danish ethics committee.

A total of 812 blood samples contained enough DNA of sufficient quality to generate reliable data. The age span of the examined individuals ranged from 73 years to 101 years, and the sex distribution was 260 males (mean age at the time of blood sampling 81 years) and 552 females (mean age at the time of blood sampling 82 years).

Red blood cells were removed by osmotic lysis, and DNA was extracted from peripheral blood using standard procedures. In brief, leukocytes were pelleted by standard techniques as previously described.28,29 Using this method, DNA was prepared from a population of peripheral blood leukocytes consisting of lymphocytes, monocytes, neutrophils, and eosinophils.

Telomere Length Analysis

The methods for measurement of mean terminal restriction fragment length have been described elsewhere30 and is also described in an electronic appendix, available with the online version of this article.

Follow Up and Statistical Analysis

The participants were followed from blood drawing in 1997 or 1998 through 7 January 2005, in the Danish Civil Registration system, which registers date of death or emigration of all Danish persons. Of the 812 participants, 400 were alive at the end of follow up and 412 had died.

The sex-specific survival pattern was analyzed using Cox regression models. Different models were used, ie, telomere length as the independent variable, age as the independent variable, and telomere length and age as independent variables. Furthermore, we treated both telomere length and age as continuous variables as well as categorical variables with 4 categories. For telomere length, we used sex-specific quartiles, and for age, the following groups: 73–79, 80–89, 90–99, and 100+ years. Because any correlation between twins in a pair could lead to too narrow confidence intervals, the analyses were performed using Stata version 8 (Stata Corp., College Station, TX), including the option cluster, with twin pair number being the cluster variable. Tied data were treated using the “Breslow” option in Stata. For the individuals from the 1905 cohort and for the centenarians, who were not twins, each person was given a unique pseudo pair number.

For the twin population, we also performed a matched analysis to test whether the cotwin with the shortest telomere tended to die first.


In the cohort of 812 elderly and oldest old (mean age 81 years, median age 79 years), the mean telomere length observed was 7.67 kb (standard error of mean [SEM] 0.05). The cohort comprised 260 males (mean age 81 years, median age 78 years, mean telomere length 7.54 kb [SEM 0.08]) and 552 females (mean age 82 years, median age 79 years, mean telomere length 7.73 kb [SEM 0.06]). Age range was 73 to 101 years for both sexes. For the 400 individuals who were dead at the time of the follow up, the mean remaining lifespan (from the date of the blood sample to the date of death) was 5.5 years with a range of 0.04 to 8.1 years. The survival analysis results are presented in Table 1.

Telomere Length, Age, and Survival

Comparison of the Cox regressions, with telomere length treated as both a continuous and as a categorical variable, demonstrated that for men as well as for women, a linear association between mortality and telomere length could be assumed, although the pattern was less regular for females. The following Cox regression analyses have therefore assumed linear association between mortality and telomere length for both sexes. The hazard ratios (HRs) are 0.89 (95% confidence interval = 0.76–1.04) for men and 0.79 (95% confidence interval = 0.72–0.88) for women (Table 1, model 1). When telomere length is regarded as a continuous variable, these figures suggest that increasing the telomere length by 1 kb decreases the risk of dying by 11% for men and by 21% for women.

Next, we included age as an independent variable in the Cox regression. As the HRs demonstrate, the association between mortality and age is nonlinear, with mortality accelerating with age (Table 1, model 3); therefore, age is used as a categorical variable. The age group 73–79 years is used as reference group. Using both telomere length (continuous) and age (categorical) as independent variables, the HR for telomere length changes to 0.97 (0.83–1.14) for males and 0.93 (0.85–1.03) for females (Table 1, model 4). Finally, we stratified by sex and age decade and found a correlation between telomere length and survival in only one of the 8 strata, namely 80- to 89-year-old women.

The group of centenarians in this study comprised 42 individuals who had a blood sample drawn and telomere length measured when they were 101.5 years old. At the end of follow up, all were dead, and the number of days they remained alive was recorded. The relationship between telomere length and remaining survival for this well-defined group is shown graphically in Figure 1. Not even in this well-defined group do we find a clear relationship between telomere length and survival.

Relationship between survival and telomere length in centenarians. In this scatterplot, only the 101-year-olds are included, eliminating any effect due to age. At the end of follow up, all centenarians were dead. Spearman's rank correlation: −0.13 (P = 0.42).

Intrapair comparison showed that among 175 twin pairs in which at least one died during follow up, it was the twin with the shorter telomere length who died first in 97 (55%) of the pairs (95% confidence interval = 48–63%). If considering only monozygotic twins, the corresponding proportion was 51% (39–63%).

From our previous work31 and from the work by others,32 we know there is a gel-to-gel variation despite the standardization procedure for the terminal restriction fragment assay analyses. The overall coefficient of variation of the terminal restriction fragment assay was found to be 12%, of which approximately half was gel-to-gel variation. Due to the sampling procedure, two thirds of the twin pairs had the 2 cotwins’ samples run on different gels, whereas the remaining third (93 pairs) had the cotwins’ samples run on the same gel. Restricting the survival analysis to the latter 93 pairs gave the same result as in the overall sample: in 47 of the 93 pairs (51%; 42–69%), the twin with the shortest telomere length died first.

Valdes et al33 recently showed a negative association between telomere length and both obesity and cigarette smoking in women, which makes the control for these factors pertinent. We therefore included an analysis of the correlation between obesity and telomere length as well as between cigarette smoking and telomere length for the 2 sexes separately. We could not confirm the negative association between telomere length and obesity or cigarette smoking using the same categories as Valdes et al.


In this article, we have measured the telomere length in leukocytes from 812 individuals age 73 to 101 years. We used the terminal restriction fragment assay for the telomere length measurements instead of one of the more recently developed assay systems based on quantitative hybridization or polymerase chain reaction.34 Those assays are in general fast and easy to use, but apart from the FISH-based assays, they have the disadvantage of measuring telomere amounts rather than telomere length. This makes it very important to pipette DNA precisely, which is notoriously difficult. In contrast, precise pipetting is not mandatory in the terminal restriction fragment assay, because telomere length is directly measured. However, in the terminal restriction fragment assay, a block of subtelomeric DNA of approximately 2 kb is included in the telomere length measure. This might be a source of error if this subtelomeric block is of variable length between different individuals. To test the extent of possible error introduced by subtelomeric block, we initially examined whether our dataset could verify the findings of the 2 other research groups that women have longer telomeres than men, at least up to the age of 75 years.17,35 In our data, women up to the age of 75 had longer telomeres than men of the same age. This indicates that the terminal restriction fragment method is able to demonstrate even minor differences in telomere length.

We measured telomere length with the purpose of investigating whether this parameter can be used as a predictor of remaining lifespan. The reason for this assumption is that telomeres shorten as a consequence of DNA loss at each round of replication, suggesting that telomeres shorten during organismal aging. Therefore, it would be expected that a survival analysis considering only telomere length as a factor would show telomere length as a predictor of survival. Indeed, the Cox regression survival analysis shows that long telomeres predict survival (Table 1, model 1). However, when age is included as a covariate in the Cox regression analysis, it becomes evident that age is a much stronger predictor of survival than telomere length, and adjustment for age weakens an already weak association between telomere length and survival. Hence, the predictor value of telomere length is so weak compared with age that it is not possible to use this molecular information as a predictor of mortality in a study of this sample size.

A drawback of the present investigation is that the telomere length measure represents the mean of all telomeres of all the leukocytes in the blood sample. Our study does not correlate minimum telomere length with survival, which would be of interest. Researchers have tried to calculate a value for the shortest telomere length using terminal restriction fragment assay smears.36 However, the method is difficult because the identification of the lower edge of the smear is very imprecise. Another drawback of the present method is that full blood is used for DNA isolation. This means that the mean telomeric length determined by the terminal restriction fragment assay can be influenced by the relative distribution of lymphocytes and granulocytes in the blood samples (lymphocytes lose telomeres slightly faster than granulocytes).37

Our results are in some disagreement with the results of the Cawthon et al17 study of 143 individuals age 60 to 97 years. The drawbacks listed here apply to the data produced by Cawthon et al as well. One possible reason is that Cawthon et al use a different method for mean telomere length determination, but there could be other methodological explanations as well. A strength of the Cawthon data is that they have followed their cohort for a longer time period than in our study. A consequence is that we have a somewhat lower total mortality than Cawthon et al (51% vs 71%), although our number of subjects is more than 5-fold larger and hence should have sufficient power to confirm the previous finding. The Cawthon et al paper also adjusted for age in their Cox regression analysis in their first analysis by using age categories and telomere length, but also using age-adjusted mortality rates and telomere length as a continuous variable. The age range in the Cawthon study of 143 individuals was 60 to 97 years, whereas in the present study, it is 73 to 101 years. Considering that most of the 42 surviving participants in the Cawthon study probably are among the youngest after the more than 15 years of follow up, it seems likely that the age distribution is very similar in the 2 studies.

Finally, we had a unique opportunity to do intrapair comparisons among twins, in which the effect of the genetic factor is controlled for (100% for monozygotic and 50% for dizygotic pairs). This approach did not reveal evidence for an association between telomere length and survival among the elderly, supporting our other analyses.


1. Griffith JD, Comeau L, Rosenfield S, et al. Mammalian telomeres end in a large duplex loop. Cell. 1999;97:503–514.
2. Olovnikov A. A theory of marginotomy. The incomplete copying of template margin in enzymic synthesis of polynucleotides and biological significance of the phenomenon. J Theor Biol. 1973;41:181–190.
3. Allsopp RC, Chang E, Kashefi-Aazam M, et al. Telomere shortening is associated with cell division in vitro and in vivo. Exp Cell Res. 1995;220:194–200.
4. Allsopp RC, Harley CB. Evidence for a critical telomere length in senescent human fibroblasts. Exp Cell Res. 1995;219:130–136.
5. Harley CB. Telomere loss: mitotic clock or genetic time bomb? Mutat Res. 1991;256:271–282.
6. Campisi J. The biology of replicative senescence. Eur J Cancer. 1997;33:703–709.
7. Martens UM, Zijlmans JM, Poon SS, et al. Short telomeres on human chromosome 17p. Nat Genet. 1998;18:76–80.
8. Cherif H, Tarry JL, Ozanne SE, et al. Ageing and telomeres: a study into organ- and gender-specific telomere shortening. Nucleic Acids Res. 2003;31:1576–1583.
9. Blasco MA, Lee HW, Hande MP, et al. Telomere shortening and tumor formation by mouse cells lacking telomerase RNA. Cell. 1997;91:25–34.
10. Herrera E, Samper E, Martin-Caballero J, et al. Disease states associated with telomerase deficiency appear earlier in mice with short telomeres. Embo J. 1999;18:2950–2960.
11. Lee HW, Blasco MA, Gottlieb GJ, et al. Essential role of mouse telomerase in highly proliferative organs. Nature. 1998;392:569–574.
12. Bessler M, Wilson DB, Mason PJ. Dyskeratosis congenita and telomerase. Curr Opin Pediatr. 2004;16:23–28.
13. Brouilette S, Singh RK, Thompson JR, et al. White cell telomere length and risk of premature myocardial infarction. Arterioscler Thromb Vasc Biol. 2003;23:842–846.
14. Benetos A, Gardner JP, Zureik M, et al. Short telemores are associated with increased carotid atherosclerosis in hypertensive subjects. Hypertension. 2004;43:182–185.
15. von Zglinicki T, Serra V, Lorenz M, et al. Short telomeres in patients with vascular dementia: an indicator of low antioxidative capacity and a possible risk factor? Lab Invest. 2000;80:1739–1747.
16. Panossian LA, Porter VR, Valenzuela HF, et al. Telomere shortening in T cells correlates with Alzheimer's disease status. Neurobiol Aging. 2003;24:77–84.
17. Cawthon RM, Smith KR, O'Brien E, et al. Association between telomere length in blood and mortality in people aged 60 years or older. Lancet. 2003;361:393–395.
18. Kyvik KO, Christensen K, Skytthe A, et al. The Danish Twin Register. Dan Med Bull. 1996;43:467–470.
19. Christensen K, Kristiansen M, Hagen-Larsen H, et al. X-linked genetic factors regulate hematopoietic stem-cell kinetics in females. Blood. 2000;95:2449–2451.
20. Bathum L, Petersen HC, Rosholm JU, et al. Evidence for a substantial genetic influence on biochemical liver function tests: results from a population-based Danish twin study. Clin Chem. 2001;47:81–87.
21. Nybo H, Gaist D, Jeune B, et al. The Danish 1905 Cohort. A genetic–epidemiological nationwide survey. Journal of Aging and Health. 2001;13:32–46.
22. Andersen-Ranberg K, Jeune B, Hoier-Madsen M, et al. Thyroid function, morphology and prevalence of thyroid disease in a population-based study of Danish centenarians. J Am Geriatr Soc. 1999;47:1238–1243.
23. Bathum L, Andersen-Ranberg K, Boldsen J, et al. Genotypes for the cytochrome P450 enzymes CYP2D6 and CYP2C19 in human longevity. Role of CYP2D6 and CYP2C19 in longevity. Eur J Clin Pharmacol. 1998;54:427–430.
24. Bladbjerg EM, Andersen-Ranberg K, de Maat MP, et al. Longevity is independent of common variations in genes associated with cardiovascular risk. Thromb Haemost. 1998;82:1100–1105.
25. Bruunsgaard H, Andersen-Ranberg K, Jeune B, et al. A high plasma concentration of TNF-alpha is associated with dementia in centenarians. J Gerontol A Biol Sci Med Sci. 1999;54:M357–364.
26. Gerdes LU, Jeune B, Ranberg KA, et al. Estimation of apolipoprotein E genotype-specific relative mortality risks from the distribution of genotypes in centenarians and middle- aged men: apolipoprotein E gene is a ‘frailty gene,’ not a ‘longevity gene’. Genet Epidemiol. 2000;19:202–210.
27. Nybo H, Gaist D, Jeune B, et al. Functional status and self-rated health in 2,262 nonagenarians: the Danish 1905 Cohort Survey. J Am Geriatr Soc. 2001;49:601–609.
28. Shannon KM, Turhan AG, Chang SS, et al. Familial bone marrow monosomy 7. Evidence that the predisposing locus is not on the long arm of chromosome 7. J Clin Invest. 1989;84:984–989.
29. Shannon KM, Turhan AG, Rogers PC, et al. Evidence implicating heterozygous deletion of chromosome 7 in the pathogenesis of familial leukemia associated with monosomy 7. Genomics. 1992;14:121–125.
30. Harley CB, Futcher AB, Greider CW. Telomeres shorten during ageing of human fibroblasts. Nature. 1990;345:458–460.
31. Bischoff C, Graakjaer J, Petersen HC, et al. The heritability of telomere length among the elderly and oldest-old. Twin Res Hum Genet. 2005;5:433–439.
32. Slagboom PE, Droog S, Boomsma DI. Genetic determination of telomere size in humans: a twin study of three age groups. Am J Hum Genet. 1994;55:866–869.
33. Valdes AM, Andrew T, Gardner JP, et al. Obesity, cigarette smoking, and telomere length in women. Lancet. 2005;366:662–664.
34. Nakagawa S, Gemmell NJ, Burke T. Measuring vertebrate telomeres: applications and limitations. Mol Ecol. 2004;13:2523–2533.
35. Benetos A, Okuda K, Lajemi M, et al. Telomere length as an indicator of biological aging: the gender effect and relation with pulse pressure and pulse wave velocity. Hypertension. 2001;37:381–385.
36. Butler MG, Tilburt J, DeVries A, et al. Comparison of chromosome telomere integrity in multiple tissues from subjects at different ages. Cancer Genet Cytogenet. 1998;105:138–144.
37. Rufer N, Brummendorf TH, Kølvraa S, et al. Telomere fluorescence measurements in granulocytes and T lymphocyte subsets point to a high turnover of hematopoietic stem cells and memory T cells in early childhood. J Exp Med. 1999;190:157–167.

Supplemental Digital Content

© 2006 Lippincott Williams & Wilkins, Inc.