YKL-40 and alcoholic liver and pancreas damage and disease in 86258 individuals from the general population: cohort and Mendelian randomization studies.
We tested the hypothesis that observationally and genetically increased plasma YKL-40 concentrations are associated with alcoholic liver and pancreas damage and disease. For this purpose, we performed cohort and mendelian randomization studies in 86258 individuals from the Danish general population, with measured plasma YKL-40 concentrations (n = 21646) and CHI3L1 rs4950928 genotype (n = 84738). First, in the cohort study, we examined association between increased plasma YKL-40 concentrations and alcoholic liver and pancreas damage and disease. Second, in the mendelian randomization study, we examined whether CHI3L1 rs4950928 genotype is associated with plasma YKL-40 concentrations and with alcoholic liver and pancreas damage and disease. Third, in cohort and mendelian randomization studies combined, we compared the association of observationally and genetically increased YKL-40 concentrations with risk of alcoholic liver and pancreas damage and disease. Finally, we estimated the absolute 10-year risk of alcoholic liver cirrhosis as a function of combined increased plasma YKL-40 concentrations and excessive alcohol intake. Among our participants, only about 9% reported not drinking alcohol (a combination of people who stopped drinking and lifetime abstainers) (see Supplemental Table 1, which accompanies the online version of this article at http://www.clinchem.org/content/ vol60/issue11), and we therefore included all participants in all analyses.
We used 2 independent prospective studies with individuals selected to reflect the adult Danish population aged 20 years and older, the Copenhagen City Heart Study (CCHS)  (16) and the Copenhagen General Population Study (CGPS) (17-19). All participants were whites of Danish descent, and there was no overlap between the 2 studies. Before examination in both studies, participants filled out a self-administered questionnaire which was completed together with an investigator on the day of attendance, where a physical examination was performed and blood was drawn for immediate biochemical analyses, as well as for storage at -80 [degrees]C for later DNA and biochemical analyses.
THE CCHS STUDY
We used the 1991-1994 examination of the CCHS. Of 9601 participants, 8773 were genotyped for rs4950928. Plasma YKL-40 concentrations were measured in samples frozen for 12-15 years at -80 [degrees]C in 8444 of these participants (20).
THE CGPS STUDY
CGPS was initiated in 2003 with ongoing enrollment. Of the first 76 657 participants, 75 965 were genotyped for rs4950928. Plasma YKL-40 concentrations were measured in fresh samples in 3211 participants and in samples frozen for 7 years at -80[degrees]C in 9991 participants.
Diagnoses of diseases listed in online Supplemental Table 2 were obtained from the national Danish Patient Registry and the national Danish Causes of Death Registry, and were defined according to the 8th and 10th revisions of the World Health Organization International Classification of Diseases (ICD). Alcoholic liver cirrhosis was ICD8: 571.09; ICD10: K70.3. Any alcoholic liver disease was ICD8: 155.09-155.89, 456.00-456.09, 570.00-570.99, 571.09, 571.10-571.99, 573.00-573.09, 785.19-785.39; ICD10: C22, I85, R18, K70. Alcoholic pancreatitis was ICD8: 577.00-577.02; ICD10: K86. Any pancreatitis was ICD8:577.0-577.9; ICD10: K85, K86. We had suchinformation on all participants.
The study was approved by Herlev Hospital and a Danish scientific ethics committee (nos. 100.2039/91 and 01-144/01) and was conducted according to the Declaration of Helsinki. Participants gave written informed consent.
PLASMA YKL-40 CONCENTRATIONS
Plasma YKL-40 concentrations were determined in duplicate by use of a 2-site, sandwich-type ELISA (Quidel Corp.), with streptavidin-coated microplate wells, a biotinylated-Fab monoclonal capture antibody, and an alkaline phosphatase-labeled polyclonal detection antibody. The recovery of the ELISA was 102%, and the detection limit was 10 [micro]g/L. The intraassay CVs were 5% at 40 [micro]g/L and 4% at 104 [micro]g/L and 155 [micro]g/L. The interassay CV was <6%.
We stratified plasma YKL-40 concentrations into 5 percentile categories to evaluate both tertiles in the lower range and extreme phenotypes in the upper range; the latter was a priori presumed to infer increased risk and was identical to previous reports (21, 22): 0%-33%, 34%-66%, 67%-90%, 91%-95%, and 96%-100%. Because plasma YKL-40 increases with increasing age, YKL-40 percentile categories al lowed for age differences (see online Supplemental Fig. 1). For the absolute 10-year risk of alcoholic liver cirrhosis, the YKL-40 percentile was calculated as a function of age in years and plasma YKL-40 in [micro]g/L with the following equation (23):
YKL-40 percentile = 100/(1 + YKL [40.sup.-3] x [1.062.sup.age] x 5000)
This equation allows age adjustment ofplasma YKL-40 concentrations for clinical use.
MARKERS OF POSSIBLE LIVER AND PANCREAS DAMAGE
We used alanine aminotransferase; alkaline phosphatase; albumin; erythrocyte mean corpuscular volume; bilirubin; [gamma]-glutamyl transpeptidase; coagulation factors II, VII, and X; pancreatic amylase; C-reactive protein (CRP); and fibrinogen as markers of possible liver and pancreas damage. These were measured by use of standard hospital assays subjected to a daily internal quality control assessing assay precision and a monthly external quality control assessing assay accuracy. Erythrocyte mean corpuscular volume and coagulation factors II, VII, and X were measured only in the participants from the CGPS. For the CCHS, samples frozen for 12-15 years were used for analyses, except for albumin, which was measured on fresh samples. For the CGPS, fresh blood samples were used for analyses.
Participants reported on present and past smoking habits and weekly alcohol consumption in the questionnaire. We subdivided participants into never-and ever-smokers, the latter category including both previous and current smokers. For ever-smokers, a pack-year was defined as 20 g tobacco smoked per day for 1 year. School education was reported as number of years of basic schooling and divided into three groups: up to 7 years, 8-10 years, and 11 or more years. Body mass index was calculated as measured weight in kilograms divided by measured height in meters squared.
We genotyped rs4950928 in 84738 participants with DNA available for genotyping (>99.9% were genotyped due to reruns). Genotyping was performed by use of TaqMan (Applied Biosystems by Life Technologies Corp.) assays. We used a forward (AGTTCCCA TAAAAGGGCTGGTTT) and a reverse (CCCAGGC CCTGTACTTCCTTTATAT) primer for the PCR amplification and a common (CTCCCCCACGCGGC) and a variant (ACTCCCCGACGCGGC) probe to determine genotype.
We used STATA version 12. Trend tests were performed by use of Cuzick nonparametric test for trend. General linear models calculated [r.sup.2] values. Hardy-Weinberg equilibrium was investigated by use of a [chi square] test. We performed no corrections for multiple comparisons.
Risk estimates and corresponding 95% confidence intervals for alcoholic liver and pancreas disease were computed as hazard ratios (HRs) by use of Cox regression analysis. The multifactorially adjusted model included adjustments for age, sex, ever smoking (yes/no), cumulative smoking (packyears), alcohol consumption (number of drinks per week, 1 drink [approximately equal to] 12 g alcohol), body mass index (kg/
[m.sup.2]), education ([less than or equal to] 7, 8-10, [greater than or equal to]11 years of schooling), and study population (CCHS or CGPS). We assessed the assumption of proportional hazards graphically by plotting -ln (survival probability) against ln (analysis time) as well as by using Schoenfeld residuals, and did not detect any major violations of the proportional hazard assumption. Participants with events before blood sampling were excluded from analyses of plasma YKL-40 concentrations, i.e., we used left truncation (delayed entry), meaning that participants only entered into the Cox model at their age at examination in 1991-1994 or 2003-2010, when blood for plasma YKL-40 measurement was drawn. Recurrent events were not considered.
In the mendelian randomization study for analyses of rs4950928 genotype, we did not exclude participants, because genotype is constant throughout life, and participants were considered at risk from the day of birth. Otherwise, analyses were like those described above for the cohort study.
For cohort and mendelian randomization studies combined, observational HRs and genetic odds ratios (ORs) for a doubling in YKL-40 concentrations on a continuous scale were assessed by calculating these ratios for an increase of 1 unit of base 2 logarithm of plasma YKL-40; in individuals without a YKL-40 measurement and for genetically increased concentrations, the well-known association between genotype and YKL-40 concentrations in other participants were applied, taking age, sex, and study population into account. We chose a doubling in YKL-40 concentration as an outcome, as we have done previously (20), because this allows for a simple interpretation of results. Furthermore, YKL-40 increased exponentially with increasing age, and CHI3L1 rs4950928 single nucleotide polymorphism was responsible for a genetically doubling or tripling in plasma YKL-40 concentrations.
In the instrumental variable analyses, we used a 2-stage least squares regression, where rs4950928 genotype examines the effect of a genetically doubling in plasma YKL-40 on the concentration of markers of possible liver and pancreas damage. The first stage of the 2-stage least squares estimator is a linear regression of rs4950928 genotype on plasma YKL-40, which generates predicted values for plasma YKL40. The second stage is a linear regression of outcome on predicted plasma YKL-40 concentrations. To estimate the effect of a genetic doubling in plasma YKL-40 on alcoholic liver and pancreas disease, we used a control function estimator with a logistic regression at second stage. The control function estimator follows the same principle as the 2-stage estimator, but additionally includes the estimated residuals from the first stage regression. These residuals may be correlated to unmeasured confounders, in which case they will help to control the effect of unmeasured confounders on outcome. The strength of rs4950928 as instrument was assessed from the first-stage regression, where an F statistic >10 usually indicates sufficient statistical strength (8). An [r.sup.2] value expressed as a percentage was used as a measure of the contribution of rs4950928 genotype to the variation in plasma YKL-40 concentrations.
Absolute risks of alcoholic liver cirrhosis by plasma YKL-40 percentile groups; smoking (never-smokers/ ever-smokers); or sex (woman/man), age (<50, 50-70, >70 years), and alcohol consumption ([less than or equal to] 7, 7-20, >20 drinks per week) were estimated by using the regression coefficients from a Poisson regression; we examined alcoholic liver cirrhosis, as it was the endpoint associated most strongly with plasma YKL-40 concentrations. Absolute risks are presented as estimated incidence rates (events per 10 years) in percentages.
The combined studies included 86258 individuals from the Danish general population, of whom 84738 were genotyped and 21 646 had plasma YKL-40 concentrations measured; we only show combined results, as the results for the 2 studies separately were similar. From 1976 through 2011, with information on all participants, 226 individuals developed alcoholic liver cirrhosis, 836 any alcoholic liver disease, 378 alcoholic pancreatitis, and 730 any pancreatitis (see online Supplemental Table 2). CHI3L1 rs4950928 genotypes were in Hardy-Weinberg equilibrium (P = 0.37), with a minor allele frequency of 21%. Baseline characteristics as a function of YKL-40 percentile categories, alcoholic liver and pancreas disease, and rs4950928 genotype are shown in Table 1 and online Supplemental Tables 2-4. Collectively, these data illustrate that YKL-40 concentrations and alcoholic liver and pancreas disease are strongly associated with many covariates and potential confounders, whereas genotype is associated with none, and therefore can be used as a largely unconfounded instrument to study the importance of lifelong increased YKL-40 concentrations not prone to reverse causation.
PLASMA YKL-40 CONCENTRATION AND ALCOHOLIC LIVER AND PANCREAS DAMAGE AND DISEASE: COHORT STUDY
The higher percentiles of plasma YKL-40 were associated with markers of possible liver and pancreas damage, that is, with increased concentrations of alanine aminotransferase, bilirubin, alkaline phosphatase, [gamma]-glutamyl transferase, erythrocyte mean corpuscular volume, CRP, and fibrinogen and with decreased concentrations of albumin; coagulation factors II, VII, and X; and pancreatic amylase (Fig. 1 and online Supplemental Fig. 2). For some markers, this was most pronounced in the upper 10 percentiles of plasma YKL-40 concentrations.
Risk of all disease endpoints increased with increased plasma YKL-40 percentile category, and was particularly high for individuals in the 96%-100% plasma YKL-40 percentile category (Fig. 2). The multifactorially adjusted HR for alcoholic liver cirrhosis was 41 (95% CI, 14-118) for the 96%-100% vs 0%33% YKL-40 percentile category. Corresponding HRs were 7.9 (5.1-12) for any alcoholic liver disease, 4.1 (1.7-10) for alcoholic pancreatitis, and 3.4 (1.9-6.1) for any pancreatitis.
CHI3L1 RS4950928 GENOTYPE, PLASMA YKL-40 CONCENTRATION, AND ALCOHOLIC LIVER AND PANCREAS DAMAGE AND DISEASE: MENDELIAN RANDOMIZATION STUDY
There was a doubling (99% increase) in geometric mean of plasma YKL-40 concentrations from CC (4.4% of all individuals) to GC (32.9%) genotype and a tripling (196% increase) in concentrations from CC to GG (62.7%) genotype (see online Supplemental Fig. 3; P-trend <1 x [10.sup.-300]); genotype explained 14% of the variation in plasma YKL-40 concentrations and the F statistic was 1690, indicating that rs4950928 is a very strong instrument to examine the effect of lifelong increased plasma YKL-40 concentrations. Geometric mean for plasma YKL-40 was 24 ng/mL for individuals with the CC genotype, 48 ng/mL for the GC genotype, and 71 ng/mL for the GG genotype. CHI3L1 rs4950928 genotype was not associated with markers of possible liver and pancreas damage (see online Supplemental Fig. 4), or with risk of alcoholic liver cirrhosis, any alcoholic liver disease, alcoholic pancreatitis, and any pancreatitis (Fig. 3).
OBSERVATIONALLY VS GENETICALLY INCREASED PLASMA YKL-40 CONCENTRATIONS AND ALCOHOLIC LIVER AND PANCREAS DAMAGE AND DISEASE: COHORT AND MENDELIAN RANDOMIZATION STUDIES
A doubling in plasma YKL-40 concentrations was associated with multifactorially adjusted increases for alanine aminotransferase of 4.2%, bilirubin of 2.5%, alkaline phosphatase of 5.0%, [gamma]-glutamyl transferase of 16%, erythrocyte mean corpuscular volume of 0.5%, CRP of 19%, and fibrinogen of 0.9% and with decreases for albumin of 0.7%, coagulation factors II, VII, and X of 0.9%, and pancreatic amylase of 1.4% (see online Supplemental Fig. 5). In contrast, a doubling in genetically increased plasma YKL-40 was associated with none of these markers of possible liver and pancreas damage.
A doubling in YKL-40 concentration was associated with a multifactorially adjusted observational HR of 2.8 (2.4-3.3) for alcoholic liver cirrhosis and a corresponding OR for a genetic doubling in YKL-40 concentrations of 1.1 (0.7-1.5). Corresponding risk estimates were 2.0 (1.8-2.2) observationally and 1.0 (0.8-1.1) genetically for any alcoholic liver disease, 1.4 (1.1-1.9) observationally and 1.1 (0.8-1.5) genetically for alcoholic pancreatitis, and 1.3 (1.1-1.6) observationally and 1.0 (0.8-1.3) genetically for any pancreatitis (Fig. 4).
ABSOLUTE 10-YEAR RISK OF ALCOHOLIC LIVER CIRRHOSIS
Although genetically increased plasma YKL-40 concentrations were not associated with risk of alcoholic liver cirrhosis, YKL-40 might still be a useful marker of future risk of this disease. Therefore, we calculated the absolute 10-year risk of alcoholic liver cirrhosis as a function of age, smoking (or sex), alcohol consumption, and plasma YKL-40 percentile categories (Fig. 5; online Supplemental Fig. 6). The absolute 10-year risk of alcoholic liver cirrhosis was increased in ever-smokers vs never-smokers, excessive drinkers (>20 drinks/week) vs modest drinkers (<7 drinks/week), and 51- to 70-year-olds vs <50-year-olds, and the risk increased stepwise with increasing plasma YKL-40 categories (all P < 6 X [10.sup.-8]). The risk was lower in individuals >70 years old than in individuals <50 (P = 2 X [10.sup.-5]), possibly explained by the early age of onset and low life expectancy of this disease. There was no difference (P = 0.20) in risk of alcoholic liver cirrhosis in moderate (7-20 drinks/week) vs modest (<7 drinks/ week) drinkers.
Studying 86 258 individuals from the general population, we found that increased plasma YKL-40 concentrations were associated with alcoholic liver and pancreas damage and disease, whereas genetically increased concentrations were not. These are novel findings.
Our results that rs4950928 genotype is not associated with risk of alcoholic liver disease are in conflict with an earlier report on the development of liver fibrosis in 440 patients with hepatitis C (13), but are supported by a later report from the larger HALT-C study of patients with chronic hepatitis C infection (24). One explanation for this discrepancy could be that the rs4950928 genotype is more important in early rather than late and more advanced liver fibrosis. Thus, more studies are needed to explore the role of this polymorphism in patients with chronic hepatitis C infection.
Although our study does not support the hypothesis that lifelong increased plasma YKL-40 is a cause of alcoholic liver and pancreas disease, it is still possible that YKL-40 concentrations are associated with degree of liver fibrosis and prognosis in patients with alcoholic liver disease (4, 6). Thus, YKL-40 could be a prognostic marker that plays a role in alcoholic liver disease progression after diagnosis, and that, potentially, inhibition of YKL-40 activity might be used in treatment of alcoholic liver disease. However, the design of the study was not adequate to investigate noninvasive fibrosis markers, and it is also possible that the higher YKL-40 concentrations observed among individuals who subsequently developed alcoholic liver cirrhosis were indicative of established fibrosis in a prediagnostic stage. Even in this case, measurement of plasma YKL-40 in alcoholic patients might be a potential alternative to invasive liver biopsy in identifying individuals at very high risk of alcoholic liver cirrhosis, thereby possibly improving treatment and compliance to alcohol reduction in these patients. Although such a potential use of plasma YKL-40 measurement needs to be tested directly in other studies, it has the potential to lower costs, patient morbidity, and liver sampling errors.
Alcoholic liver cirrhosis is by definition caused by excessive alcohol consumption, regardless of plasma YKL-40 concentrations. Indeed, there was no difference in risk of alcoholic liver cirrhosis between participants drinking <7 vs 7-20 drinks/week. However, the risk of alcoholic liver cirrhosis was increased among participants with plasma YKL-40 concentrations within the top 5% who drank <7 drinks/week. Possibly this can be explained by the fact that the absolute risk was estimated by using the regression coefficients from Poisson regression, which assumes that risk increases multiplicatively with each increase in plasma YKL-40 percentile category regardless of smoking, sex, age, and alcohol consumption.
The threshold of alcohol consumption needed to develop alcoholic liver disease varies: although some moderate drinkers do develop alcoholic liver disease, not all heavy drinkers do. Alcohol consumption in this study was self reported at baseline, we did not have information on lifetime abstainers, and some heavy drinkers might have underreported their alcohol consumption. Indeed, half of the participants who developed any alcoholic liver disease after the initiation of the study had reported a weekly alcohol consumption of 10 drinks or less at baseline; thus we included all the participants in the analyses regardless of alcohol consumption.
It may seem surprising that increased YKL-40 concentrations were inversely associated with pancreatic amylase and directly associated with risk of pancreatitis. Although acute pancreatitis is associated with increased concentrations of pancreatic amylase, chronic pancreatitis is not, and it may even be associated with decreased concentrations of pancreatic amylase. This might explain the fact that increased plasma YKL-40 was associated with decreasing concentrations of pancreatic amylase as well as with increased risk of pancreatitis.
Limitations of this study include survival bias and selective nonresponse to the invitation owing to individuals being very sick with liver and/or pancreatic disease. Both of these concerns would likely lead to under-representation of these individuals in our sample of the general population and thereby bias the results toward the null hypothesis. Therefore, these limitations cannot explain our findings for alcoholic liver disease and pancreatitis. Second, the diagnoses were ascertained from the hospital system, and not from a dedicated clinic with the specific intent of studying these diseases. It is possible that some of the healthy controls might have had the diagnosis, had they been examined more thoroughly, which might have significantly influenced the results. Taken together, considering the presumed survival bias, selective noncompliance, and false-negative controls, our significant risk estimates of liver and pancreatic diseases are probably rather conservative and the true risks might be even higher. Our negative results regarding rs4950928 are likely accurate, since the distribution of genotypes does not deviate from Hardy-Weinberg equilibrium. Third, our study was based on fresh samples as well as samples frozen for 7-15 years at -80[degrees]C. Because plasma YKL-40 remains stable regardless of repetitive thawing and freezing (25), and because internal controls in our laboratory have demonstrated that plasma YKL-40 is stable for at least 16 years when stored at--80[degrees]C, we do not believe that YKL-40 measurement in frozen samples has distorted our analyses to a large degree. Finally, we only studied white individuals of Danish descent, and our results are therefore not necessarily directly applicable to other ethnic groups; however, we are not aware of data to suggest that the present results should not be applicable to most ethnic groups.
In conclusion, in this large study of individuals from the general population, we found that a plasma YKL-40 concentration within the top 5% was a marker of alcoholic liver cirrhosis, but found no evidence to support a causal relationship.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, oranalysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts ofInterest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts ofinterest:
Employment or Leadership: None declared.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: A.D. Kjaergaard, University of Copenhagen; B.G. Nordestgaard, the Research Council at Herlev Hospital, the Danish Heart Foundation, the Danish Medical Research Council; J.S. Johansen, Chief Physician Johan Boserup and Lise Boserups Foundation, Vera and Carl Johan Michaelsen's Foundation, ToyotaFonden Denmark.
Expert Testimony: None declared.
Patents: None declared.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.
Acknowledgments: We thank Birgit Hertz, Nina Dahl Kjersgaard, Hanne Damm, Tonni Love Hansen, AnjaJochumsen, Ulla Kjaerullf-Hansen, Dorthe Mogensen, Marianne Sorensen, Rikke Mirsbach Henkel, Annette Lundgren-Beck, and Debbie Nadelmann, all Herlev Hospital, for excellent technical assistance. The participants and staffs of the Copenhagen City Heart Study and the Copenhagen General Population Study are thanked for their willingness to participate.
(1.) Lee CG, Da Silva CA, Dela Cruz CS, Ahangari F, Ma B, Kang MJ et al. Role of chitin and chitinase/ chitinase-like proteins in inflammation, tissue remodeling, and injury. Annu Rev Physiol 2011;73: 479-501.
(2.) Shao R. YKL-40 acts as an angiogenic factor to promote tumor angiogenesis. Front Physiol 2013; 4:122.
(3.) Shultz NA, Johansen JS. YKL-40- a protein in the field of translational medicine: a role as a biomarker in cancer patients? Cancers 2010;2:145391.
(4.) Johansen JS, Moller S, Price PA, Bendtsen F, Junge J, Garbarsch C, Henriksen JH. Plasma YKL40: a new potential marker of fibrosis in patients with alcoholic cirrhosis? Scand J Gastroenterol 1997;32:582-90.
(5.) Johansen JS, Christoffersen P, Moller S, Price PA, Henriksen JH, Garbarsch C, Bendtsen F. Serum YKL-40 is increased in patients with hepatic fibrosis. J Hepatol 2000;32:911-20.
(6.) Nojgaard C, Johansen JS, Christensen E, Skovgaard LT, Price PA, Becker U. Serum levels of YKL-40 and PIIINP as prognostic markers in patients with alcoholic liver disease. J Hepatol 2003; 39:179-86.
(7.) Hansen M, Nielsen AR, Vilsboll T, Lund A, Krarup T, Knop FK, Vestergaard H. Increased levels of YKL-40 and interleukin 6 in patients with chronic pancreatitis and secondary diabetes. Pancreas 2012;41:1316-8.
(8.) Lawlor DA, Harbord RM, Sterne JA, Timpson N, Davey SG. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat Med 2008;27:1133-163.
(9.) Smith GD, Ebrahim S. Mendelian randomization: prospects, potentials, and limitations. Int J Epidemiol 2004;33:30-42.
(10.) Ober C, Tan Z, Sun Y, Possick JD, Pan L, Nicolae R et al. Effect of variation in CHI3L1 on serum YKL-40 level, risk of asthma, and lung function. N Engl J Med 2008;358:1682-91.
(11.) Zhao X, Tang R, Gao B, Shi Y, Zhou J, Guo S et al. Functional variants in the promoter region of Chitinase 3-like 1 (CHI3L1) and susceptibility to schizophrenia. Am J Hum Genet 2007;80:12-8.
(12.) Sohn MH, Lee JH, Kim KW, Kim SW, Lee SH, Kim K-E et al. Genetic variation in the promoter region of chitinase 3-like 1 is associated with atopy. Am J Respir Crit Care Med 2009;179:449-56.
(13.) Berres ML, Papen S, Pauels K, Schmitz P, Zaldivar MM, Hellerbrand C et al. A functional variation in CHI3L1 is associated with severity of liver fibrosis and YKL-40 serum levels in chronic hepatitis C infection. J Hepatol 2009;50:370-6.
(14.) Nielsen KR, Steffensen R, Boegsted M, Baech J, Lundbye-Christensen S, Hetland ML et al. Promoter polymorphisms in the chitinase 3-like 1 gene influence the serum concentration of YKL-40 in Danish patients with rheumatoid arthritis and in healthy subjects. Arthritis Res Ther 2011;13:R109.
(15.) Yamamori H, Hashimoto R, Ohi K, Yasuda Y, Fukumoto M, Kasahara E et al. A promoter variant in the chitinase 3-like 1 gene is associated with serum YKL-40 level and personality trait. Neurosci Lett 2012;513:204-8.
(16.) Nordestgaard BG, Benn M, Schnohr P, Tybjaerg-Hansen A. Nonfasting triglycerides and risk of myocardial infarction, ischemic heart disease, and death in men and women. JAMA 2007;298:299308.
(17.) Frikke-Schmidt R, Nordestgaard BG, Stene MC. Association of loss-of-function mutations in the ABCA1 gene with high-density lipoprotein cholesterol levels and risk of ischemic heart disease. JAMA 2008;299:2524-32.
(18.) Kamstrup PR, Tybjaerg-Hansen A, Steffensen R, Nordestgaard BG. Genetically elevated lipoprotein (a) and increased risk of myocardial infarction. JAMA 2009;301:2331-9.
(19.) Zacho J, Tybjaerg-Hansen A, Jensen JS, Grande P, Sillesen H, Nordestgaard BG. Genetically elevated C-reactive protein and ischemic vascular disease. N Engl J Med 2008;359:1897-908.
(20.) Kjaergaard AD, Bojesen SE, Johansen JS, Nordestgaard BG. Elevated plasma YKL-40 levels and ischemic stroke in the general population. Ann Neurol 2010;68:672-80.
(21.) Johansen JS, Bojesen SE, Mylin AK, Frikke-Schmidt R, Price PA, Nordestgaard BG. Elevated plasma YKL-40 predicts increased risk of gastrointestinal cancer and decreased survival after any cancer diagnosis in the general population. J Clin Oncol 2009;27:572-8.
(22.) Johansen JS, Bojesen SE, Tybjaerg-Hansen A, Mylin AK, Price PA, Nordestgaard BG. Plasma YKL-40 and total and disease-specific mortality in the general population. Clin Chem 2010;56: 1580-91.
(23.) Bojesen SE, Johansen JS, Nordestgaard BG. Plasma YKL-40 levels in healthy subjects from the general population. Clin Chim Acta 2011;412: 709-12.
(24.) Fontana RJ, Litman HJ, Dienstag JL, Bonkovsky HL, Su G, Sterling RK, Lok AS. YKL-40 genetic polymorphisms and the risk of liver disease progression in patients with advanced fibrosis due to chronic hepatitis C. Liver Int 2012;32:665-74.
(25.) Hogdall EV, Johansen JS, Kjaer SK, Price PA, Blaakjaer J, Hogdall CK. Stability of YKL-40 concentration in blood samples. Scand J Clin Lab Invest 2000;60:247-51.
Alisa D. Kjaergaard, [1,2,3] Stig E. Bojesen, [1,2,3,4] Borge G. Nordestgaard, [1,2,3,4] and Julia S. Johansen [2,5,6] *
 Department of Clinical Biochemistry, Herlev Hospital,  Faculty of Health and Medical Sciences,  The Copenhagen General Population Study, Herlev University Hospital,  The Copenhagen City Heart Study, Frederiksberg Hospital, Copenhagen University Hospital,  Department of Medicine, Herlev Hospital, and  Department of Oncology, Herlev Hospital, Copenhagen University Hospital, University of Copenhagen, Denmark.
* Address correspondence to this author at: Copenhagen University Hospital, Herlev Ringvej 75, DK-2730 Herlev, Denmark. E-mail email@example.com. Received June 17, 2014; accepted August 20, 2014.
Previously published online at DOI: 10.1373/clinchem.2014.229096
 Nonstandard abbreviations: CCHS, Copenhagen City Heart Study; CGPS, Copenhagen General Population Study; ICD, International Classification of Diseases; CRP, C-reactive protein; HR, hazard ratio; OR, odds ratio.
Table 1. Characteristics of participants according to plasma YKL-40 percentile categories (a) Plasma YKL-40 percentile categories (b) Characteristic 0%-33% 34%-66% 67%-90% Number of 7179 7137 5174 participants, N Women, % 56 56 56 Age, years 58 (48-69) 58 (48-69) 58 (48-69) Alcohol consumption, 6(2-12) 7(3-14) 8(3-17) weekly drinks Ever-smoker, % 66 67 69 Cumulative smoking, 8 (0-26) 9 (0-28) 11 (0-30) pack years Body mass index, 25 (23-28) 25 (23-28) 26 (23-29) kg/[m.sup.2] Education 20/42/38 21/41/38 21/41/38 ([less than or equal to] 7/8-10/ [greater than or equal to] 11 years), % Plasma YKL-40 percentile categories (b) Characteristic 91%-95% 96%-100% P Number of 1078 1078 21 646 participants, N Women, % 56 56 0.99 Age, years 58 (48-69) 58 (48-69) 0.73 Alcohol consumption, 11 (4-21) 14(5-28) 5 x [10.sup.-137] weekly drinks Ever-smoker, % 76 77 1 x [10.sup.-19] Cumulative smoking, 17(0-38) 20(1-40) 6 x [10.sup.-49] pack years Body mass index, 26 (23-29) 26 (23-29) 1 x [10.sup.-16] kg/[m.sup.2] Education 23/44/33 26/43/31 4 x [10.sup.-6] ([less than or equal to] 7/8-10/ [greater than or equal to] 11 years), % P value for disease (c) Characteristic ALC (e) ALD Number of 85 528 85 528 participants, N Women, % 4 x [10.sup.-11] 8 x [10.sup.-13] Age, years 4 x [10.sup.-4] 8 x [10.sup.-23] Alcohol consumption, 5 x [10.sup.-5] 0.008 weekly drinks Ever-smoker, % 2 x [10.sup.-20] 6 x [10.sup.-28] Cumulative smoking, 1 x [10.sup.-49] 5 x [10.sup.-66] pack years Body mass index, 0.01 1 x [10.sup.-4] kg/[m.sup.2] Education 2 x [10.sup.-12] 3 x [10.sup.-23] ([less than or equal to] 7/8-10/ [greater than or equal to] 11 years), % P value for disease (c) Characteristic AP P Number of 85 528 85 528 participants, N Women, % 2 x [10.sup.-5] 0.004 Age, years 6 x [10.sup.-12] 2 x [10.sup.-28] Alcohol consumption, 0.17 0.82 weekly drinks Ever-smoker, % 5 x [10.sup.-9] 4 x [10.sup.-9] Cumulative smoking, 7 x [10.sup.-23] 1 x [10.sup.-25] pack years Body mass index, 0.12 4 x [10.sup.4] kg/[m.sup.2] Education 4 x [10.sup.-12] 4 x [10.sup.-25] ([less than or equal to] 7/8-10/ [greater than or equal to] 11 years), % P value for genotype Characteristic rs4950928 (d) Number of 84 738 participants, N Women, % 0.83 Age, years 0.69 Alcohol consumption, 0.97 weekly drinks Ever-smoker, % 0.54 Cumulative smoking, 0.43 pack years Body mass index, 0.71 kg/[m.sup.2] Education 0.65 ([less than or equal to] 7/8-10/ [greater than or equal to] 11 years), % (a) Baseline values are expressed as n, %, or median (interquartile range). (b,c,d) p values are from Cuzick nonparametric test for trend and represent comparisons (b) across YKL-40 percentile categories,c between participants with events vs controls (see online Supplemental Table 1), and d across genotype coded as 0, 1, and 2 for number of G-alleles (see online Supplemental Table 2). (e) ALC, alcoholic liver cirrhosis; ALP, any alcoholic liver disease; AP, alcoholic pancreatitis, P, any pancreatitis. Fig. 2. HRs of alcoholic liver and pancreas disease according to age-corrected plasma YKL-40 percentile category. Individuals with event before blood sampling were excluded from analyses of the total of 21646 participants with measured plasma YKL-40 concentrations; therefore, the numbers of participants varies between studies of the different disease endpoints. Multifactorially adjusted model included adjustments for age, sex, alcohol consumption, smoking, body mass index, education, and study population. YKL-40 No. of Age- and sex- percentiles events/total adjust Alcoholic liver cirrhosis 96%-100% 34/1060 70 (25-196) 91%-95% 8/1075 16 (4.7-52) 67%-90% 10/5163 4.1 (1.3-13) 34%-66% 7/7136 2.0 (0.6-6.7) 0%-33% 4/7172 Trend: P = 2 x [10.sup.-29] Any alcoholic liver disease 96%-100% 55/1043 11 (7.6-17) 91%-95% 19/1064 3.7 (2.2-6.4) 67%-90% 42/5128 1.7 (1.1-2.6) 34%-66% 35/7108 0.9 (0.6-1.5) 0%-33% 42/7143 Trend: P = 9 x [10.sup.-34] Alcoholic pancreatitis 96%-100% 10/1066 6.2 (2.7-14) 91% 95% 1/1068 0.6 (0.1-4.6) 67%-90% 14/5148 1.7 (0.8-3.7) 34%-66% 12/7111 1.0 (0.5-2.2) 0%-33% 13/7165 Trend: P = 2 x [10.sup.-4] Any pancreatitis 96%-100% 20/1065 4.4 (2.6-7.6) 91%-95% 7/1065 1.5 (0.7-3.4) 67%-90% 27/5139 1.2 (0.7-1.9) 34%-66% 39/7101 1.1 (0.7-1.8) 0%-33% 38/7155 Trend: P = 2 x [10.sup.-5] YKL-40 No. of Age-, sex-, and percentiles events/total alcohol-adjustment Alcoholic liver cirrhosis 96%-100% 34/1060 49 (17-142) 91%-95% 8/1075 13 (3.9-43) 67%-90% 10/5163 3.8 (1.2-12) 34%-66% 7/7136 1.9 (0.6-6.5) 0%-33% 4/7172 Trend: P = 8 x [10-.sup.23] Any alcoholic liver disease 96%-100% 55/1043 8.8 (5.8-14) 91%-95% 19/1064 3.2 (1.9-5.6) 67%-90% 42/5128 1.6 (1.0-2.4) 34%-66% 35/7108 0.9(0.6-1.4) 0%-33% 42/7143 Trend: P = 2 x [10.sup.-24] Alcoholic pancreatitis 96%-100% 10/1066 4.4 (1.8-11) 91% 95% 1/1068 0.5 (0.1-3.8) 67%-90% 14/5148 1.6 (0.7-3.4) 34%-66% 12/7111 1.0 (0.4-2.2) 0%-33% 13/7165 Trend: P = 0.006 Any pancreatitis 96%-100% 20/1065 3.7 (2.1-6.6) 91%-95% 7/1065 1.3 (0.6-3.0) 67%-90% 27/5139 1.1 (0.7-1.8) 34%-66% 39/7101 1.1 (0.7-1.8) 0%-33% 38/7155 Trend: P = 6 x [10.sup.-4] YKL-40 No. of Multifactorially percentiles events/total adjusted Alcoholic liver cirrhosis 96%-100% 34/1060 41 (14-118) 91%-95% 8/1075 10 (3.1-35) 67%-90% 10/5163 3.4 (1.1-11) 34%-66% 7/7136 1.8 (0.5-6.2) 0%-33% 4/7172 Trend: P = 2 x [10.sup.-20] Any alcoholic liver disease 96%-100% 55/1043 7.9 (5.1-12) 91%-95% 19/1064 2.9 (1.7-5.0) 67%-90% 42/5128 1.5 (1.0-2.3) 34%-66% 35/7108 0.9 (0.6-1.4) 0%-33% 42/7143 Trend: P = 1 x [10.sup.-21] Alcoholic pancreatitis 96%-100% 10/1066 4.1 (1.7-10) 91% 95% 1/1068 0.4 (0.1-3.4) 67%-90% 14/5148 1.5 (0.7-3.2) 34%-66% 12/7111 1.0 (0.4-2.1) 0%-33% 13/7165 Trend: P = 0.01 Any pancreatitis 96%-100% 20/1065 3.4 (1.9-6.1) 91%-95% 7/1065 1.2 (0.5-2.8) 67%-90% 27/5139 1.1 (0.6-1.8) 34%-66% 39/7101 1.1 (0.7-1.7) 0%-33% 38/7155 Trend: P = 0.002 HR ratio (95% CI) Fig. 3. HRs of alcoholic liver and pancreas disease according to CHI3L1 rs4950928 genotype. The rare homozygote (CC) is the reference group. Individuals with event before blood sampling were not excluded from analyses of the total of 84738 genotyped participants. The multifactorially adjusted model included adjustments for age, sex, alcohol consumption, smoking, body mass index, education, and study population. [DELTA]YKL-40 indicates the genotype-associated difference in YKL-40 concentration from that of the CC genotype. rs4950928 [DELTA]YKL-40 Age-and sex- Age-, sex-, and events/total adjusted alcohol-adjusted Alcoholic liver cirrhosis GG 133/53150 +196% 1.3 (0.6-2.8) 1.3 (0.6-2.7) GC 71/27 867 +99% 1.3 (0.6-2.9) 1.3 (0.6-2.8) CC 7/3721 0% Any alcoholic liver disease GG 499/53150 +196% 0.9 (0.6-1.2) 0.9 (0.6-1.2) GC 265/27867 +99% 0.9 (0.7-1.3) 0.9 (0.6-1.2) CC 39/3721 0% Alcoholic pancreatitis GG 239/53150 +196% 0.9 (0.6-1.5) 1.0 (0.7-1.4) GC 108/27867 +99% 0.8 (0.5-1.3) 0.9 (0.6-1.3) Any pancreatitis GC 225/27867 +99% 1.0 (0.7-1.4) 0.9 (0.6-1.5) 0.9 (0.6-1.3) 0.8 (0.5-1.3) CC 33/3721 0% rs4950928 [DELTA]YKL-40 Multifactorial events/total adjusted Alcoholic liver cirrhosis GG 133/53150 +196% 1.2 (0.5-2.5) GC 71/27 867 +99% 1.2 (0.5-2.6) CC 7/3721 0% Any alcoholic liver disease GG 499/53150 +196% 0.9 (0.6-1.2) GC 265/27867 +99% 0.9 (0.6-1.2) CC 39/3721 0% Alcoholic pancreatitis GG 239/53150 +196% 0.9 (0.7-1.3) GC 108/27867 +99% 0.9 (0.6-1.3) Any pancreatitis GC 225/27867 +99% 0.8 (0.5-1.3) 0.8 (0.5-1.3) CC 33/3721 0% HR ratio (95% CI) Fig. 4. Observational HRs and genetic ORs of alcoholic liver and pancreas disease for a doubling in plasma YKL-40 concentrations. The multifactorially adjusted model included adjustment for age, sex, alcohol consumption, smoking, body mass index, education, and study population. Observational HRs and genetic Ors by a doubling in plasma YKL-40 Alcoholic liver cirrhosis HR/OR (95% CI) Observational Age- and sex-adjusted 3.1 (2.7-3.6) Age-, sex-, and alcohol-adjusted 2.8 (2.4-3.3) Multifactorially adjusted 2.8 (2.4-3.3) Genetic CHI3L1 rsd4950928 1.1 (0.7-1.5) Any alcoholic liver disease Observational Age- and sex-adjusted 2.2 (2.0-2.4) Age-, sex-, and alcohol-adjusted 2.1 (1.8-2.3) Multifactorially adjusted 2.0 (1.8-2.2) Genetic CHI3L1 rs4950928 1.0 (0.8-1.1) Alcoholic pancreatitis Observational Age- and sex-adjusted 1.6 (1.3-2.0) Age-, sex-, and alcohol-adjusted 1.5 (1.2-1.9) Multifactorially adjusted 1.4 (1.1-1.9) Genetic CHI3L1 rs4950928 1.1 (0.8-1.5) Any pancreatitis Observational Age- and sex-adjusted 1.4 (1.2-1.7) Age-, sex-, and alcohol-adjusted 1.4 (1.2-1.6) Multifactorially adjusted 1.3 (1.1-1.6) Genetic CHI3L1 rs4950928 1.0 (0.8-1.3) Alcoholic liver cirrhosis P value Observational Age- and sex-adjusted 9 x [10.sup.-55] Age-, sex-, and alcohol-adjusted 3 x [10.sup.-41] Multifactorially adjusted 1 x [10.sup.-33] Genetic CHI3L1 rsd4950928 0.75 Any alcoholic liver disease Observational Age- and sex-adjusted 1 x [10.sup.-51] Age-, sex-, and alcohol-adjusted 3 x [10.sup.-38] Multifactorially adjusted 3 x [10.sup.-33] Genetic CHI3L1 rs4950928 0.58 Alcoholic pancreatitis Observational Age- and sex-adjusted 5 x [10.sup.-5] Age-, sex-, and alcohol-adjusted 0.002 Multifactorially adjusted 0.005 Genetic CHI3L1 rs4950928 0.46 Any pancreatitis Observational Age- and sex-adjusted 5 x [10.sup.-6] Age-, sex-, and alcohol-adjusted 2 x [10.sup.-4] Multifactorially adjusted 0.001 Genetic CHI3L1 rs4950928 0.78
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Proteomics and Protein Markers|
|Author:||Kjaergaard, Alisa D.; Bojesen, Stig E.; Nordestgaard, Borge G.; Johansen, Julia S.|
|Date:||Nov 1, 2014|
|Previous Article:||Total and cause-specific mortality by moderately and markedly increased ferritin concentrations: general population study and metaanalysis.|
|Next Article:||Prognostic value of midregional pro-A-type natriuretic peptide and N-terminal pro-B-type natriuretic peptide in patients with stable coronary heart...|