Three studies have investigated the ability of blood test analytes to predict biological age. First, when considering the top 20 variables that were associated with biological age in aging.ai, albumin contributed most to this prediction, almost 2x more than circulating levels of glucose (Mamoshina et al. 2018):
Second, albumin was one of the 9 blood test variables that were best able to predict biological age when using the Phenotypic Age calculator. However, as shown below, it didn’t come in first place, but fifth. Interestingly, the analyte that contributed most to biological age prediction was the red cell distribution width (RDW%), with glucose again in second place (Levine et al. 2018):
Third, Earls et al. (2019) used the Klemera-Doubal algorithm (Klemera and Doubal, 2006) in conjunction with blood test data to predict biological age. Regardless if the blood was analyzed by Labcorp or Quest, higher levels of albumin (the left side of both images below) were associated with the greatest reduction in biological age, up to 5 years! In contrast, HbA1c was associated with a higher biological age when measured by Labcorp (top image, right side), and second to lead in the Quest analysis (bottom image, right side). Interestingly, glucose came in third and fifth in the Labcorb and Quest data sets, respectively, in terms of its positive association with biological age.
Glucose would’ve been an obvious choice, but would you have guessed that albumin may be just as important, and potentially more important for predicting biological age?
If you’re interested, please have a look at my book!
References
Earls JC, Rappaport N, Heath L, Wilmanski T, Magis AT, Schork NJ, Omenn GS, Lovejoy J, Hood L, Price ND. Multi-OmicBiologicalAgeEstimation and Its Correlation With Wellness and DiseasePhenotypes: A LongitudinalStudy of 3,558Individuals. J Gerontol A Biol Sci Med Sci. 2019 Nov 13;74(Supplement_1):S52-S60. doi: 10.1093/gerona/glz220.
Klemera P, Doubal S. A new approach to the concept and computation of biological age. Mech Ageing Dev. 2006;127:240–248. doi:10.1016/j. mad.2005.10.004
Levine ME, Lu AT, Quach A, Chen BH, Assimes TL, Bandinelli S, Hou L, Baccarelli AA, Stewart JD, Li Y, Whitsel EA, Wilson JG, Reiner AP, Aviv A, Lohman K, Liu Y, Ferrucci L, Horvath S. An epigeneticbiomarker of aging for lifespan and healthspan. Aging (Albany NY). 2018 Apr 18;10(4):573-591. doi: 10.18632/aging.101414.
Mamoshina P, Kochetov K, Putin E, Cortese F, Aliper A, Lee WS, Ahn SM, Uhn L, Skjodt N, Kovalchuk O, Scheibye-Knudsen M, Zhavoronkov A. Population specific biomarkers of human aging: a big data study using South Korean, Canadian and Eastern European patient populations. J Gerontol A Biol Sci Med Sci. 2018 Jan 11.
Creatinine is one of the 9 blood test variables included on the biological age calculator, Phenotypic Age (https://michaellustgarten.wordpress.com/2019/09/09/quantifying-biological-age/). The reference range for women and men is 0.5 – 1.1, and 0.6 – 1.2 mg/dL respectively, but within that range, what’s optimal for health and longevity?
To answer that question, it’s important to know how circulating levels of creatinine change during aging, and its association with risk of death for all causes. Creatinine increases during aging, as reported in studies of 9,389 adults (age range, 30 – 75y; Levine 2013), and in 377,686 subjects (age range, 18 – 85y; Wang et al. 2017). However, the absolute values for these changes, i.e. from 0.8 to 1.0 mg/dL, for ex., was not reported in either study.
In terms of all-cause mortality risk, creatinine levels of 0.8 mg/dL (blue line; 95% confidence interval (CI), red dotted line) were associated with the lowest risk of death for all causes, with risk being significantly reduced for creatinine values between 0.6 – 1.1 mg/dL in the 30,760 older adults (median age, 69y) of Solinger and Rothman (2013):
Note the U-shaped mortality curve for creatinine: all-cause mortality increased when it was both less than or greater than 0.8 mg/dL. More specifically, risk of death for all causes was significantly increased when serum levels of creatinine were less than ~0.55 and greater than 1.5 mg/dL.
Few studies have investigated the association between serum (or plasma) levels of creatinine with all-cause mortality risk, as most studies use creatinine in conjunction with chronological age, gender, and ethnicity to estimate kidney function (eGFR). For example, the MDRD equation (Levey et al. 2006) is commonly used to calculate eGFR, and if you’re interested in converting your creatinine levels into eGFR, here’s a link to calculate it (https://www.mdcalc.com/mdrd-gfr-equation). As creatinine goes up, eGFR goes down, and is indicative of worse kidney function. Based on that, we should expect to see an age-related decrease in kidney function, as measured by eGFR. Is this true?
eGFR decreases during aging, from values ~125 in 20 year old women and men to ~50 mL/min/1.73m^2 in adults older than 90y in the 385,918 subjects (age range, 18 – 100y) of Wang et al. (2017):
Similarly, eGFR decreased from ~90 (thick black line; 95% CI, dashed lines: 75 – 130 mL/min/1.73m^2) in young men (18-24y) to less than 70 (thick black line; 95% CI, dashed lines: 45 – 90 mL/min/1.73m^2) in men older than 75y (Baba et al. 2015):
In women, eGFR decreased from values ~100 (thick black line; 95% CI, dashed lines: 70 – 135 mL/min/1.73m^2) in youth to ~70 (thick black line; 95% CI, dashed lines: 50 – 95 mL/min/1.73m^2) in women older than 75y (Baba et al. 2015):
Similar findings have been reported for the age-related decline in eGFR in other studies, including Wetzels et al. (2007). When comparing young adults (18-24 year olds) with older adults (> 85y), median eGFR values declined from ~95 to ~65 mL/min/1.73m^2 in men, and from ~90 to ~60 mL/min/1.73m^2 in women.
What’s the effect of reduced kidney function (i.e. increased creatinine, decreased eGFR) on risk of death for all causes? In a meta-analysis of 46 studies that included 2,051,158 subjects, risk of death for all causes was significantly increased when eGFR was less than 52 in women (red, below), and less than 44 in men (blue), when compared with eGFR values between 90 – 104 mL/min/1.73m^2 (95 was used as the reference; Nitsch et al. 2013):
In sum, creatinine increases during aging, which is associated with an increased all-cause mortality risk. Similarly, eGFR, which includes circulating values for creatinine, decreases during aging, and is also associated with an increased all-cause mortality risk. Therefore, if you’re tracking your creatinine levels with the goal of optimizing your biological age, it’s important to try to keep creatinine levels relatively low (i.e. around 0.8 mg/dL), and to avoid its age-related increase!
If you’re interested, please have a look at my book!
References
Baba M, Shimbo T, Horio M, Ando M, Yasuda Y, Komatsu Y, Masuda K, Matsuo S, Maruyama S. Longitudinal Study of the Decline in Renal Function in Healthy Subjects. PLoS One. 2015 Jun 10;10(6):e0129036.
Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S, Kusek JW, Van Lente F; Chronic Kidney Disease Epidemiology Collaboration. Using standardizedserumcreatininevalues in the modification of diet in renal diseasestudyequation for estimatingglomerular filtration rate. Ann Intern Med. 2006 Aug 15;145(4):247-54.
Levine ME. Modeling the rate of senescence: can estimated biological age predict mortality more accurately than chronological age? J Gerontol A Biol Sci Med Sci. 2013 Jun;68(6):667-74. doi: 10.1093/gerona/gls233.
Nitsch D, Grams M, Sang Y, Black C, Cirillo M, Djurdjev O, Iseki K, Jassal SK, Kimm H, Kronenberg F, Oien CM, Levey AS, Levin A, Woodward M, Hemmelgarn BR; Chronic Kidney Disease Prognosis Consortium. Associations of estimated glomerular filtration rate and albuminuria with mortality and renal failure by sex: a meta-analysis. BMJ. 2013 Jan 29;346:f324. doi: 10.1136/bmj.f324.
Solinger AB, Rothman SI. Risks of mortality associated with common laboratory tests: a novel, simple and meaningful way to set decision limits from data available in the Electronic Medical Record. Clin Chem Lab Med. 2013 Sep;51(9):1803-13. doi: 10.1515/cclm-2013-0167.
Wang Z, Li L, Glicksberg BS, Israel A, Dudley JT, Ma’ayan A. Predictingage by miningelectronicmedicalrecords with deeplearningcharacterizesdifferences between chronological and physiologicalage. J Biomed Inform. 2017 Dec;76:59-68. doi: 10.1016/j.jbi.2017.11.003.
Wetzels JF, Kiemeney LA, Swinkels DW, Willems HL, den Heijer M. Age– and gender-specificreferencevalues of estimatedGFR in Caucasians: the NijmegenBiomedicalStudy. Kidney Int. 2007 Sep;72(5):632-7.
The percentage of lymphocytes is one of the 9 blood test variables included in the biological age calculator, Phenotypic Age (https://michaellustgarten.wordpress.com/2019/09/09/quantifying-biological-age/). The reference range for lymphocyte % is 20 – 40% of the total amount of white blood cells (WBCs), but are higher or lower values optimal for health and longevity?
To answer that question, it’s important to know how levels of lymphocytes change during aging, and its association with risk of death for all causes. In one of the earliest studies to examine how the percentage of lymphocytes changes with age, Levine (2013) reported that lymphocyte % significantly decreased during aging in 9,389 adults (age range, 30 – 75y). However, the absolute values for these changes, i.e. from 40% to 30%, for ex., was not reported.
Similarly, lymphocyte % decreased during aging in a much larger study (377,686 subjects; age range, 18 – 85y; Wang et al. 2017):
Interestingly, for women, lymphocyte % decreased from 27% to 21% from 20 – 35y, increased from 21% to 26% from 35 – 55y, then again decreased from 26% to 20% from 55y to 85y. In contrast, lymphocyte % more steadily decreased for men, from 28% to 17% from 20 – 85y.
Based on the aging data, higher values for lymphocyte % are are associated with biologic youth, whereas lower values are found in older adults. Although there are few studies that have investigated associations between lymphocyte % with aging or disease risk, in contrast, more studies have been published for absolute levels of lymphocytes.
In a small study of 106 older adults (> 85y) that were healthy (i.e. free of disease) at baseline, lymphocytes less than 1.14*10^9 cells/L (equivalent to 1140*10^6 cells/L) was associated with an increased risk of death for all causes, when compared with 1850*10^6 cells/L (Izaks et al. 2003):
In a larger study (624 subjects), lymphocytes less than 1540*10^6 cells/L was associated with a significantly shorter average lifespan (~5y; 0.5 proportion remaining below), when compared with 1540 – 2040*10^6 cells/L . Also note that survival for the group that had 1540 – 2040*10^9 lymphocytes/L was not significantly different from the group that had more than 2040*10^9 lymphocytes/L (Leng et al. 2005):
In agreement with the smaller studies, lymphocytes < 1300 and < 1200*10^6 cells/L in women and men (red and blue, far left), respectively was associated with an increased all-cause mortality risk, when compared with average lymphocyte values ~1900*10^6 cells/L (decile 5) in a larger study that included 262,394 non-smokers (age range, 37 – 73y; Welsh et al. 2018):
Collectively, these data suggest that higher values for lymphocyte % and for the absolute amount of lymphocytes may be optimal for minimizing disease risk and for maximizing longevity. If both are low, can they be raised? Circulating levels of lymphocytes are reduced during zinc deficiency (Fraker and King, 2001), so monitoring zinc intake, then increasing it to at least the RDA may be a first step towards increasing lymphocyte levels and %.
If you’re interested, please have a look at my book!
References
Fraker PJ, King LE. A distinct role for apoptosis in the changes in lymphopoiesis and myelopoiesis created by deficiencies in zinc. FASEB J. 2001 Dec;15(14):2572-8.
Izaks GJ, Remarque EJ, Becker SV, Westendorp RG. Lymphocyte count and mortality risk in older persons. The Leiden 85-Plus Study. J Am Geriatr Soc. 2003 Oct;51(10):1461-5.
Leng SX, Xue QL, Huang Y, Ferrucci L, Fried LP, Walston JD. Baselinetotal and specificdifferentialwhitebloodcellcounts and 5-yearall-causemortality in community-dwellingolderwomen. Exp Gerontol. 2005 Dec;40(12):982-7.
Levine ME. Modeling the rate of senescence: can estimated biological age predict mortality more accurately than chronological age? J Gerontol A Biol Sci Med Sci. 2013 Jun;68(6):667-74. doi: 10.1093/gerona/gls233.
Wang Z, Li L, Glicksberg BS, Israel A, Dudley JT, Ma’ayan A. Predictingage by miningelectronicmedicalrecords with deeplearningcharacterizesdifferences between chronological and physiologicalage. J Biomed Inform. 2017 Dec;76:59-68. doi: 10.1016/j.jbi.2017.11.003.
Welsh C, Welsh P, Mark PB, Celis-Morales CA, Lewsey J, Gray SR, Lyall DM, Iliodromiti S, Gill JMR, Pell J, Jhund PS, Sattar N. Association of Total and Differential Leukocyte Counts With Cardiovascular Disease and Mortality in the UK Biobank. Arterioscler Thromb Vasc Biol. 2018 Jun;38(6):1415-1423. doi: 10.1161/ATVBAHA.118.310945.
Shown below is my average daily dietary intake from January 1 through November 7th, 2019 (n=306 days). I weigh all of my food with a food scale, so these aren’t estimated amounts:
In terms of weight (or volume), green tea is atop the list, as I drink 20 oz/day. Carrots come in second place (for why, see https://michaellustgarten.wordpress.com/2018/07/06/serum-albumin-and-acm/), followed by strawberries, red bell peppers, bananas, watermelon (for the lycopene), cauliflower, blueberries, blackberries, and raspberries. Note that I mix the bananas and berries in my green smoothies, which I drink 3-4x/week, which includes spinach (#11) and parsley (#23).
What does my average daily macro- and micro-nutrient data look like for 2019?
In terms of supplements, I use 1000 IU of vitamin D from November – May, and I take a methylfolate-methylB12-B6 supplement, to help keep my homocysteine levels low.
If you’re interested, please have a look at my book!
While that value is 23% younger than my chronological age (46y), I knew that I could do better! So I tried again on September 17, 2019. Basically, the same biological age, 35.58y:
An 23% younger biological age on 2 separate dates, months apart might be good for most, but not for me. So, I tried again on October 29th, 2019, and voila, a biological age of 31.3y, which is 32% younger than my chronological age! How did I do it?
From my last blood test until my most recent blood test, I attempted a mild caloric restriction. To maintain my body weight, I require about 2800 calories per day, an amount which is based on daily body weight weighing in conjunction with daily dietary tracking. For the period of time that elapsed between my last 2 blood tests, I averaged 2657 calories/day, which is 3.2% less than the 2745 calories/day that I averaged for the dietary period that corresponded to my September blood test. That I was also in a very mild caloric restriction is confirmed by a reduction in my average body weight, which was (purposefully) down 0.7 lbs from September 17 to October 29th, when compared with the dietary period that corresponded to my September blood test (August 20 – September 17).
This is a superficial analysis of how I further reduced my biological age, but in future posts I’ll report the average dietary intake that corresponded to my relatively youthful biologic age!
If you’re interested, please have a look at my book!
High sensitivity C-Reactive Protein (CRP) is one of the 10 variables included in the biological age calculator, PhenoAge (https://michaellustgarten.wordpress.com/2019/09/09/quantifying-biological-age/). The reference range for CRP is 0 – 3 mg/L, but within that range, what’s optimal? To answer that question, it’s important to know how CRP changes during aging, and what levels are associated with an increased risk of death for all causes.
First, CRP increases 3-5 fold during aging in women and men, respectively (Ferrucci et al. 2005):
Investigating further, CRP continues to increase from relatively low levels of 0.7 mg/L (equivalent to 0.07 mg/dL) in 85-99 year olds to 2.5 mg/L (equivalent to 0.25 mg/dL) in adults older than 110y (Arai et al. 2015):
In support of this finding, the average CRP level in 98 centenarians (average age, 101y) was 5.4 mg/L, when compared with 3.2 mg/L in 70 year olds (Montoliu et al. 2014).
Based on the data for how CRP changes during aging, lower values would be expected to better in terms of risk of death for all causes. How low is optimal for CRP?
Several studies have investigated this issue. Risk of death for all causes was significantly reduced when CRP was < 3 mg/L, when compared with > 3 mg/L in the 11,193 subjects (average age 63y) of Oluleye et al. (2013). In terms of CRP values less than 3 mg/L, CRP < 1.0 mg/L was associated with significantly reduced risk of death for all causes in the 5,248 subjects (average age, 54y) of Hamer et al. (2010), in the 3,620 subjects (average age, 58y) of Koenig et al. (2008), in the 2,240 older adults (average age, 69y) of Elkind et al. (2009), and in the 1,519 subjects (average age 72y) of Kuoppamäki et al. (2015).
Similarly, CRP levels close to 1 mg/dL have also been associated with a significantly reduced risk of death for all causes, including < 0.86 mg/L in the 11,409 adults (average age, 59y) of Shen et al. (2019), and < 0.83 mg/L in the 1,476 men (average age, 53y) of Laaksonen et al. (2005).
In terms of the association between CRP with risk of death for all causes, can we go lower than ~0.8 mg/L? CRP values < 0.5 mg/L were associated with reduced all-cause mortality risk, whereas values > 3 mg/L were associated with increased risk in the 16,850 non-smokers and non-users of hormone replacement therapy (average age, 58y) of Ahmadi-Abhari et al. (2013). Similarly, CRP between 0.5 – 1 mg/L, the area on the chart (see below) where the black line and the shaded 95% confidence interval have a hazard ratio < 1, was associated with a significantly reduced risk of death for all causes, whereas CRP > 5 mg/L was associated with an increased all-cause mortality risk in the 7,015 subjects of Zuo et al. (2016):
Going even lower, CRP < 0.33 mg/L (Tertile 1, white circle) was associated with a maximally reduced all-cause mortality risk when compared with values > 0.86 mg/L (Tertile 3, black circle) in the 1,034 older adults (average age, ~72y) of Shinkai et al. (2008):
Similarly, CRP values between 0.03 – 0.33 mg/L in men and between 0.03 – 0.25 mg/L in women (blue lines for both men and women, Tertile 1) was associated with a significantly reduced risk of death for all causes in the 11,080 subjects (average age, 62y) of Nisa et al. (2016). In contrast, CRP > 0.85 mg/L in men and > 0.62 mg/L in women (close to significance in women, p=0.06; green lines for both, Tertile 3) was associated with an increased all-cause mortality risk:
Lowest risk of death for all causes was also identified when CRP was between 0.1 – 0.3 mg/L (1st; Tertile 1), when compared with CRP > 0.8 mg/L (3rd; Tertile 3) in the 7,740 older adults (average age, 64y) of Makita et al. (2009):
Can we go lower than 0.3 mg/L for CRP and all-cause mortality risk? Yes! CRP < 0.21 mg/L was associated with a maximally reduced risk of death for all causes, and mortality risk significantly increased above 0.44 mg/L in the 2,589 subjects (average age, 59y) of Arima et al. (2008):
Collectively, based on these data, with the goal of optimal health and lifespan, I’d suggest that CRP levels should be as low as possible, and avoiding the age-related CRP increase. What are my CRP values? I’ve only measured it 6x, including once in 2009 (0.2 mg/L), once in 2018 (0.67 mg/L), and 4x in 2019 (0.41, 0.34, 0.47, 0.29 mg/L). My average value for 2019 is 0.38 mg/L, and I’d like to cut that in half. I have a blood test scheduled for next week, so stay tuned for that data!
Here’s the short version of this post in video format!
If you’re interested, please have a look at my book!
References
Ahmadi-Abhari S, Luben RN, Wareham NJ, Khaw KT. Seventeenyearrisk of all-cause and cause-specificmortalityassociated with C-reactiveprotein, fibrinogen and leukocytecount in men and women: the EPIC-Norfolkstudy. Eur J Epidemiol. 2013 Jul;28(7):541-50. doi: 10.1007/s10654-013-9819-6.
Arai Y, Martin-Ruiz CM, Takayama M, Abe Y, Takebayashi T, Koyasu S, Suematsu M, Hirose N, von Zglinicki T. Inflammation, But Not Telomere Length, Predicts Successful Ageing at Extreme Old Age: A Longitudinal Study of Semi-supercentenarians. EBioMedicine. 2015 Jul 29;2(10):1549-58. doi: 10.1016/j.ebiom.2015.07.029.
Arima H, Kubo M, Yonemoto K, Doi Y, Ninomiya T, Tanizaki Y, Hata J, Matsumura K, Iida M, Kiyohara Y. High–sensitivityC-reactiveprotein and coronaryheartdisease in a generalpopulation of Japanese: the Hisayamastudy. Arterioscler Thromb Vasc Biol. 2008 Jul;28(7):1385-91. doi: 10.1161/ATVBAHA.107.157164.
Elkind MS, Luna JM, Moon YP, Liu KM, Spitalnik SL, Paik MC, Sacco RL. High-sensitivityC-reactive proteinpredictsmortality but not stroke: the Northern Manhattan Study. Neurology. 2009 Oct 20;73(16):1300-7. doi: 10.1212/WNL.0b013e3181bd10bc.
Kuoppamäki M, Salminen M, Vahlberg T, Irjala K, Kivelä SL, Räihä I. HighsensitiveC-reactive protein (hsCRP), cardiovascularevents and mortality in the aged: a prospective9-yearfollow-up study. Arch Gerontol Geriatr. 2015 Jan-Feb;60(1):112-7. doi: 10.1016/j.archger.2014.10.002.
Ferrucci L, Corsi A, Lauretani F, Bandinelli S, Bartali B, Taub DD, Guralnik JM, Longo DL. The origins of age-related proinflammatory state. Blood. 2005 Mar 15;105(6):2294-9. Epub 2004 Nov 30.
Hamer M, Chida Y, Stamatakis E. Association of very highlyelevatedC-reactive proteinconcentration with cardiovascular events and all-cause mortality. Clin Chem. 2010 Jan;56(1):132-5. doi: 10.1373/clinchem.2009.130740.
Koenig W, Khuseyinova N, Baumert J, Meisinger C. Prospectivestudy of high-sensitivityC-reactiveprotein as a determinant of mortality: results from the MONICA/KORAAugsburgCohortStudy, 1984-1998. Clin Chem. 2008 Feb;54(2):335-42.
Laaksonen DE, Niskanen L, Nyyssönen K, Punnonen K, Tuomainen TP, Salonen JT. C-reactive protein in the prediction of cardiovascular and overall mortality in middle-aged men: a population-based cohort study. Eur Heart J. 2005 Sep;26(17):1783-9.
Montoliu I, Scherer M, Beguelin F, DaSilva L, Mari D, Salvioli S, Martin FP, Capri M, Bucci L, Ostan R, Garagnani P, Monti D, Biagi E, Brigidi P, Kussmann M, Rezzi S, Franceschi C, Collino S. Serum profiling of healthy aging identifies phospho- and sphingolipid species as markers of human longevity. Aging (Albany NY). 2014 Jan;6(1):9-25.
Makita S, Nakamura M, Satoh K, Tanaka F, Onoda T, Kawamura K, Ohsawa M, Tanno K, Itai K, Sakata K, Okayama A, Terayama Y, Yoshida Y, Ogawa A. SerumC-reactiveproteinlevels can be used to predictfutureischemicstroke and mortality in Japanesemen from the generalpopulation. Atherosclerosis. 2009 May;204(1):234-8. doi: 10.1016/j.atherosclerosis.2008.07.040.
Oluleye OW, Folsom AR, Nambi V, Lutsey PL, Ballantyne CM; ARIC Study Investigators. Troponin T, B-type natriuretic peptide, C-reactive protein, and cause-specific mortality. Ann Epidemiol. 2013 Feb;23(2):66-73. doi: 10.1016/j.annepidem.2012.11.004.
Nisa H, Hirata A, Kohno M, Kiyohara C, Ohnaka K. High-Sensitivity C-Reactive Protein and Risks of All-Cause and Cause-Specific Mortality in a Japanese Population. Asian Pac J Cancer Prev. 2016;17(5):2643-8.
Shen Y, Zhang Y, Xiong S, Zhu X, Ke C. High-sensitivity C-reactive protein and cystatin C independently and jointly predict all-cause mortality among the middle-aged and elderly Chinese population. Clin Biochem. 2019 Mar;65:7-14. doi:10.1016/j.clinbiochem.2018.12.012.
Shinkai S, Chaves PH, Fujiwara Y, Watanabe S, Shibata H, Yoshida H, Suzuki T. Beta2-microglobulin for riskstratification of totalmortality in the elderlypopulation: comparison with cystatin C and C-reactive protein. Arch Intern Med. 2008 Jan 28;168(2):200-6. doi: 10.1001/archinternmed.2007.64.
Zuo H, Ueland PM, Ulvik A, Eussen SJ, Vollset SE, Nygård O, Midttun Ø, Theofylaktopoulou D, Meyer K, Tell GS. PlasmaBiomarkers of Inflammation, the KynureninePathway, and Risks of All-Cause, Cancer, and CardiovascularDiseaseMortality: The Hordaland Health Study. Am J Epidemiol. 2016 Feb 15;183(4):249-58. doi: 10.1093/aje/kwv242.
Michael Lustgarten 