Joint analysis of individual participants’ data from 17 studies on the association of the IL6 variant -174G>C with circulating glucose levels, interleukin-6 levels, and body mass index

Figures & data

Table

BMI	body mass index
CHD	Coronary heart disease
HWE	Hardy-Weinberg equilibrium
IL-6	Interleukin-6
IL6	Interleukin-6 gene
OR	odds ratio
T2DM	type 2 diabetes mellitus
2-h	2-hour
95% CI	95% confidence interval

Table I.  Characteristics of included studies.

Study^a	Full study name	Country^b	n T2DM/non-diabetic subjects^c
BOTNIA	Botnia Study	SF	731/557
CAPPP	Captopril Prevention Project	S	42/424
DANISH	Danish Study	DK	1212/4399
EDSC	Ealing Diabetes Study of Coagulation	UK	299/0
EPIC-POTSDAM	European Prospective Investigation into Cancer and Nutrition Potsdam (EPIC-Potsdam)	D	0/348
FUSION 1	The Finland-United States Investigation of NIDDM Genetics, 1st sampling wave	SF	508/367
FUSION 2	The Finland-United States Investigation of NIDDM Genetics, 2nd sampling wave	SF	437/201
GIRONA	Girona Genetics of Diabetes Study	E	42/123
KORA-MIFAM	KORA MI Family Study	D	95/881
KORA-S4	KORA Survey S4	D	225/1190
KORA-T2DMFAM	KORA T2DM Family Study	D	776/513
MONICA/KORA-BASE	MONICA/KORA Case Cohort Study S123 (MONICA/KORA-S123)	D	101/1744
MONICA-S3	MONICA/KORA Survey S3	D	151/3551
NPHS II	Second Northwick Park Heart Study	UK	0/2652
RMIFAM	Regensburg Ml Family Study	D	662/2614
TGN	Tarraco Study	E	166/64
UDACS	University College Diabetes and Cardiovascular Study	UK	560/0

^aAbbreviated study name used in the present publication.

^bCountry of recruitment: D=Germany, DK=Denmark, E=Spain, SF=Finland, S=Sweden, UK=United Kingdom

^cNumber of type 2 diabetic/non-diabetic subjects included in analyses of the outcome BMI

MI=myocardial infarction; NIDDM=non-insulin-dependent diabetes mellitus.

Figure 1.  Forest plot, illustrating the study-specific β-coefficients with 95% CIs for the association between IL6 -174G>C, dominant model for the C-allele, and fasting glucose in (A) the full data analysis (combined group of non-diabetic and prevalent type 2 diabetes mellitus (T2DM) subjects), (B) only in non-diabetic subjects, and (C) only in prevalent T2DM subjects, adjusted for age, sex, and body mass index (BMI). Additionally, the summary fixed-effect β-coefficient is shown. I² measures the impact of inconsistency across studies and can range between 0% and 100%.

Table II.  Summary results of association between IL6 -174G>C and fasting glucose, 2-h glucose, body mass index (BMI), or interleukin-6 (IL-6) levels. The fixed-effect results are presented for the full data analysis^a and for analyses stratified for type 2 diabetes status (T2DM).

	Full data analysis^a					Non-diabetic subjects					Type 2 diabetic subjects
	n		β-coefficient (P-value)^b			n		β-coefficient (P-value)^b			n		β-coefficient (P-value)^b
Outcome	Studies	Subjects	CC vs GG	GC vs GG	Dominant^c	Studies	Subjects	CC vs GG	GC vs GG	Dominant^b	Studies	Subjects	CC vs GG	GC vs GG	Dominant^b
Fasting glucose (mmol/L)	7	9440	−0.081 (0.108)	−0.099 (0.011)	−0.091 (0.014)	9	7420	0.007 (0.626)	0.007 (0.523)	0.007 (0.501)	6	2412	−0.322 (0.076)	−0.354 (0.013)	−0.317 (0.020)
2-hour glucose (mmol/L)	6	7398	−0.127 (0.127)	−0.053 (0.441)	−0.075 (0.243)	6	6731	−0.022 (0.592)	−0.003 (0.920)	−0.009 (0.775)	4	618	−0.043 (0.927)	−0.114 (0.760)	−0.135 (0.703)
BMI (kg/m^2)	15	24117	0.097 (0.197)	0.086 (0.151)	0.088 (0.120)	13	19007	0.057 (0.478)	0.080 (0.213)	0.071 (0.242)	11	5026	0.216 (0.255)	0.154 (0.321)	0.176 (0.226)
Circulating IL-6^d	7	5659	0.034 (0.292)	0.018 (0.467)	0.024 (0.304)	6	4621	0.001 (0.969)	0.022 (0.441)	0.018 (0.504)	4	966	0.159 (0.044)	0.004 (0.938)	0.042 (0.415)

^aAnalysis of non-diabetic and prevalent T2DM subjects together. The sum of the number of non-diabetic and T2DM subjects in stratified analyses does not yield the number of subjects in the full data analysis, because stratified analyses were performed in individual studies only if there were at least 50 non-diabetic or T2DM subjects available.

^bSummary estimates from generalized linear models adjusted for age and sex (for BMI and IL-6 levels: additionally adjusted for T2DM status; for glucose and IL-6 levels: additionally adjusted for BMI). Main findings are printed in bold.

^cDominant model comparing CC- and GC-subjects versus GG.

^dIL-6 β-coefficients were computed and are displayed on the logarithm of the original scale (pg/mL).

Figure 2.  Forest plot, illustrating the study-specific β-coefficients with 95% CIs for the association between IL6 -174G>C, dominant model for the C-allele, and body mass index (BMI) of (A) the full data analysis, (B) only non-diabetic subjects, and (C) only prevalent type 2 diabetes mellitus (T2DM) subjects, adjusted for age, sex, and T2DM status (A). Please refer to Figure 1 legend for more details.

Figure 3.  Forest plot, illustrating the study-specific β-coefficients with 95% CIs for the association between IL6 -174 CC versus GG and natural logarithm of circulating interleukin-6 (IL-6) levels of (A) the full data analysis, (B) only non-diabetic subjects, and (C) only prevalent type 2 diabetes mellitus (T2DM) subjects, adjusted for age, sex, body mass index (BMI), and T2DM status (A). Please refer to Figure 1 legend for more details.

Figure A-1.  (online appendix). Funnel plot for the association between IL6 -174G>C and body mass index (BMI). For each study, the β-coefficient of the dominant model for the C-allele, adjusted for age, sex, and type 2 diabetes status, is plotted against its standard error as a measure of study precision. All studies with IL6 -174G>C genotypes of non-diabetic individuals in Hardy-Weinberg equilibrium (HWE) are included. Black circles represent BMI published studies, white circles represent BMI unpublished studies. The vertical line marks the pooled β-coefficient (0.09 kg/m²).

Table A -I. (online appendix). Methods used for quantifying circulating interleukin (IL)-6 levels and genotyping IL6 -174G>C (rs1800795 according to http://www.ncbi.nlm.nih.gov).

Study	IL-6 level quantification	Genotyping method	Call rate (%)^a
Botnia Study	n.a.	Allelic discrimination assay-by-design on ABI 7900 (Applied Biosystems)	100
Captopril Prevention Project	Sandwich ELISA, R&D Systems, Abingdon, UK	Dynamic allele specific hybridization (DASH)	100
Danish Study	n.a.	Chip-based MALDI-TOF MS (MassArray, Sequenom)	96
Ealing Diabetes Study of Coagulation	n.a.	Nla III RFLP, MADGE	97
European Prospective Investigation into Cancer and Nutrition Potsdam	Sandwich ELISA, R&D Systems, Abingdon, UK	SNuPE, MegaBACE 1000	100
The Finland-United States Investigation of NIDDM Genetics	n.a.	Illumina GoldenGate	98
Girona Genetics of Diabetes Study	Solid-phase, enzyme-labeled, chemiluminescent sequential immunometric assay, DPC DIPESA S.A., Madrid, Spain	SfaNI RFLP	99
KORA MI Family Study	Sandwich ELISA, R&D Systems, Abingdon, UK	Hsp92II RFLP, PAGE	97
KORA Survey S4	Sandwich ELISA, CLB, Amsterdam, Netherlands	Chip-based MALDI-TOF MS (MassArray, Sequenom)	97
KORA T2DM Family Study	n.a.	Chip-based MALDI-TOF MS (MassArray, Sequenom)	98
MONICA/KORA Case Cohort Study S123	Sandwich ELISA, CLB, Amsterdam, Netherlands	Chip-based MALDI-TOF MS (MassArray, Sequenom)	99
MONICA/KORA Survey S3	n.a.	Chip-based MALDI-TOF MS (MassArray, Sequenom)	99
Regensburg MI Family Study	n.a.	Hsp92II RFLP, PAGE	97
Second Northwick Park Heart Study	n.a.	PCR by MADGE, NIaIII RFLP	99
Tarraco Study	n.a.	SfaNI RFLP	99
University College Diabetes and Cardiovascular Study	Sandwich ELISA, R&D Systems, Abingdon, UK	PCR by MADGE, NIaIII RFLP	99

^aSuccessfully genotyped individuals in percent of all subjects intended for genotyping.

n.a.=not applicable; NIDDM=non-insulin-dependent diabetes mellitus; MI=myocardial infarction.

Table A -II. (online appendix). Description of included studies.

Study name (official abbreviation)	Study description ^a
Botnia Study	The Botnia Study began in 1990 as a family-based study aiming to identify genes increasing susceptibility to T2DM. Type 2 diabetic subjects from the area of five health care centers in the Botnia region of Western Finland were invited to participate together with their family members. For the purpose of this joint analysis, unrelated individuals were genotyped for IL6 -174G>C. According to a priori defined criteria, one type 2 diabetic individual per family was selected. Non-diabetic subjects comprise cases’ spouses and unrelated individuals, all being 35 years or older.
Captopril Prevention Project (CAPPP)	CAPPP is a prospective randomized clinical trial conducted in Sweden and Finland during the 1990s. Patients aged 25–66 years, with a measured diastolic blood pressure of 100 mmHg or more on two occasions, were recruited at health centers and randomly assigned to captopril or conventional antihypertensive treatment. Exclusion criteria were secondary hypertension, serum creatinine concentration of more than 150 µmol/L, and disorders that required treatment with β-blockers. Cases had T2DM at base-line or were diagnosed during the follow-up. This joint analysis includes a substudy of the Swedish part of CAPPP, which has been genotyped for IL6 -174G>C. This substudy comprises all patients that got myocardial infarction (MI), plus two control subjects without MI per patient, matched with respect to gender, age, and smoking. Further details: 4.
Danish Study	The DANISH case-control study of T2DM involves all 4568 subjects with normal glucose tolerance (NGT) from the Inter99 cohort and 1389 unrelated type 2 diabetic patients recruited from the outpatient clinic at Steno Diabetes Center, Copenhagen and the Research Center for Prevention and Health through the Inter99 study. The Inter99 cohort is a population-based randomized non-pharmacological intervention study for prevention of cardiovascular disease done at the Research Center for Prevention and Health involving 6514 Caucasian subjects (6164 with data from an oral glucose tolerance test). Further details: 5.
Ealing Diabetes Study of Coagulation (EDSC)	The type 2 diabetic individuals of the EDSC study were recruited consecutively from the Ealing Hospital diabetes clinic in London, UK. Patients completed a questionnaire with details of age, ethnicity, smoking habit, fasting status, duration of diabetes, and other clinical details. Blood was collected for plasma and DNA analysis. Several further parameters, such as BMI, were measured. Type 2 diabetic individuals (n=927) comprised primarily three ethnic groups: Indian Asian, n=503; UK white, n=331; black Afro-Caribbean, n=93. Further details: 6. To ensure comparability with the other Caucasian studies, only the white subjects were included in this joint analysis.
European Prospective Investigation into Cancer and Nutrition Potsdam (EPIC-Potsdam)	A nested case-control study was designed within the European Prospective Investigation into Cancer and Nutrition Potsdam cohort (EPIC-POTSDAM_nCC-T2DM), which is part of the European multicenter, population-based EPIC-study including 27,548 subjects from the area around Potsdam, Germany (women aged 35–65 years and men aged 40–65 years). Base-line examination and blood sampling were conducted between 1994 and 1998. Data presented in this joint analysis are based on the first follow-up questionnaires sent to the study participants on average 2.3 years after base-line examination. Further details: 7. To ensure comparability with the other cross-sectional studies, only the non-diabetic control subjects were included in this joint analysis. Cases were free of T2DM at base-line and developed their incident T2DM during the follow-up. Analyses of their data are presented separately.
The Finland-United States Investigation of NIDDM Genetics (FUSION)	The index probands in the FUSION study were identified primarily from the National Hospital Discharge Registry (NHDR), which includes records since 1970 of all hospitalized patients with diabetes, and from previous studies carried out by the National Public Health Institute in Finland. From the NHDR, all patients who were hospitalized with a diagnosis of T2DM in Finland during 1987–1993 were identified in the first wave of sampling (FUSION 1). In the second wave of sampling (FUSION 2), patients hospitalized with T2DM during 1994–1995 were identified. Potential families for FUSION 2 also included some identified during FUSION 1 but not invited to participate at that time due to distance from the study clinics. An index proband with his family was eligible for participation in the FUSION study if 1) the proband or another affected sibling was diagnosed with T2DM between 35 and 60 years of age, 2) there was no history of type 1 diabetes in first-degree relatives, 3) the proband had one or more living full siblings diagnosed with T2DM at any age, and 4) at least one parent was apparently non-diabetic, with preference given to families with living parents or parents who had lived a long life without known diabetes. Further details on FUSION 1: 8; on FUSION 2: 9. Participants of FUSION 1 and FUSION 2 were analyzed separately. ‘FUSION 1’ in this joint analysis comprises one type 2 diabetic individual from each FUSION 1 family, and non-diabetic spouses of type 2 diabetic FUSION participants, as well as elderly subjects that were all born in 1925 and were normal glucose tolerant by oral glucose tolerance tests (OGTTs) at both ages 65 and 70. ‘FUSION 2’ comprises the sibling generation of the FUSION 2 sampling wave.
Girona Genetics of Diabetes Study	The type 2 diabetic patients of the Girona Genetics of Diabetes Study were consecutively recruited subjects from the diabetes clinics at the Hospital of Girona, Spain. The non-diabetic subjects are unrelated healthy Caucasian middle-aged subjects recruited from the general population. Further details: 2.
KORA Studies in chronological order	KORA (Cooperative Health Research in the Region Augsburg) is a regional research platform in the German city of Augsburg and the two adjacent counties, for population-based studies, subsequent follow-up studies, and family studies in the fields of epidemiology, health economics, and health care research. KORA was established in 1996 to expand the WHO (World Health Organization) MONICA (Monitoring of Trends and Determinants in Cardiovascular Disease) project in Augsburg. In the framework of MONICA, three independent cross-sectional population-representative surveys were conducted in 1984/85 (S1), 1989/90 (S2), and 1994/95 (S3), and a population-based acute myocardial infarction registry was set up. The study subjects of all Augsburg MONICA and KORA surveys and the family studies on myocardial infarction and T2DM are of German nationality and were studied by physical examination, blood testing, and a standardized interview in KORA study centers. All tests were carried out by specially trained personnel. Further details: 48–50. Some individuals were originally recruited for two or more studies, but were assigned to one of the included KORA studies for the purpose of this joint analysis according to a priori defined criteria.
MONICA/KORA Survey S3	The MONICA/KORA Survey S3 originally investigated 4856 individuals. Study participants that are included in the MONICA/KORA Case Cohort Study S123 were eliminated from subjects of the MONICA/KORA Survey S3 for this joint analysis.
KORA MI Family Study	Patients with MI prior to the age of 60 years and their siblings were identified through the acute myocardial infarction registry. The diagnosis of MI was established according to the MONICA diagnostic criteria. Of 1254 patients contacted, 609 agreed to participate in the study (532 men, aged 56.1±0.3 years). Moreover, 540 siblings without MI (251 men, aged 54.6±0.4 years from 325 families) were recruited and examined by the same protocol.
KORA Survey S4 (KORA S4)	The KORA S4 studied a population-representative sample of 4261 subjects, 25–74 years old, during the years 1999–2001. The sample design followed the guidelines of the three previous MONICA Augsburg surveys. In the age-range of 55–74 years, 1653 persons participated in an OGTT. These participants were genotyped for IL6 -174G>C and included in this joint analysis. Further details: 13.
KORA T2DM Family Study (KORA T2DMFAM)	In 2001/2002, 605 nuclear families were enrolled in the KORA T2DM Family Study. Families were ascertained through an index proband with known T2DM, who had at least one full sib or both parents willing to participate in the study. All available members of the index probands’ nuclear families, i.e. full sibs and parents, were included. Index probands were all from the city or region of Augsburg. They were recruited from T2DM patients of the Central Hospital of Augsburg, from earlier MONICA and KORA studies, from the acute myocardial infarction registry, or via public relations. All participants were living in Germany, and all were of European origin. Most subjects were extensively phenotyped in the KORA study center, some were examined by their family doctor, who decided whether or not the subject had T2DM and took blood samples for DNA analyses. Data of the sibling generation was included in this joint analysis.
MONICA/KORA Case Cohort Study S123	All participants of at least one of the three MONICA Augsburg surveys were prospectively followed for the MONICA/KORA Case Cohort Study S123. The study was restricted to participants aged 35–74 years at base-line, since the incidence of T2DM is low in younger subjects. A stratified random sample of the source population, containing 1885 subjects, was selected. A total of 555 incident cases of T2DM were observed between participants’ study start dates and 31st December 2002. Further details: 14. For the purposes of this joint analysis on quantitative phenotypes, the base-line data of the MONICA/KORA Case Cohort Study S123 (prevalent T2DM and non-diabetic subjects) without the participants who developed their incident T2DM during the follow-up were used to ensure comparability with the other cross-sectional studies (MONICA/KORA-BASE). Analyses of the subjects with incident T2DM are presented separately.
Regensburg MI Family Study	The kindreds of the Regensburg MI Family Study were ascertained through MI index patients, who were identified by screening 93,500 patient charts in seven cardiac in-hospital rehabilitation centers distributed throughout Germany. Index patients had all suffered from MI before 60 years. If at least one sibling had suffered from MI or had severe coronary artery disease or by-pass surgery, the index patient with all available parents and siblings were contacted and invited to participate in the study. All participating individuals filled out a standardized questionnaire that focused on cardiovascular risk factors, medical diagnoses, life-style and medication. Further details: 15. Data of the sibling generation were included in this joint analysis.
Second Northwick Park Heart Study (NPHS II)	For the NPHS II Study, 3012 unrelated healthy Caucasian middle-aged male subjects were recruited from nine general medical practices scattered throughout the UK and prospectively followed from 1989. Sixty-eight subjects with diabetes at base-line were excluded from analysis. Further details: 16.
Tarraco Study	For the Tarraco study, 211 unrelated type 2 diabetic subjects were recruited from the outpatient clinic at Hospital Universitari de Tarragona ‘Joan XXIII’ during the years 2000–2004. Simultaneously, 118 healthy subjects were recruited from the same hospital. Further details: 2.
University College Diabetes and Cardiovascular Study (UDACS)	The UDACS Study comprises 1011 consecutive subjects recruited from the diabetes clinic at University College London Hospitals NHS Trust (UCLH) between the years 2001 and 2002. Patients completed a questionnaire with details of age, ethnicity, smoking habit, fasting status, duration of diabetes, and other clinical details. Blood was collected for plasma and DNA analysis. Several further parameters, such as BMI, were measured. No subjects requiring renal dialysis were recruited. Further details: 16.

^aThe numbers of participants presented for the original studies do not always match the numbers used in the analyses of this joint analysis. The reason is that some subjects of the original studies were not included in the joint analysis because of study overlap, missing genotype data, or missing phenotype data.

Table A -III. (online appendix). References of included studies and overview on analyzed outcomes.

Study	No. T2DM/Non-diab^a	Outcome^b Fasting glucose	2-h glucose	BMI	IL-6 Level	Reference^c
BOTNIA	731/557	3	3	3	4	n.a.
CAPPP	42/424	2	4	1	2	4
DANISH	1212/4399	1	3	1	4	5
EDSC	299/0	3	4	3	4	n.a.
EPIC-POTSDAM	0/348	4	4	2	2	7
FUSION 1	508/367	3	3	3	4	n.a.
FUSION 2	437/201	3	3	3	4	n.a.
GIRONA	42/123	3	3	3	3	2
KORA-MIFAM	95/881	4	4	2	2	17
KORA-S4	225/1190	3	3	1	1	18
KORA-T2DMFAM	776/513	3	3	3	4	n.a.
MONICA/KORA-BASE	101/1744	4	4	3	3	n.a.
MONICA-S3	151/3551	4	4	2	4	17
NPHS II	0/2652	4	4	1	4	16
RMIFAM	662/2614	4	4	2	4	17
TGN	166/64	4	4	3	4	2
UDACS	560/0	4	4	1	3	16

^aNumber of type 2 diabetic (T2DM)/non-diabetic (non-diab) subjects included in analyses of the outcome BMI.

^bPublication of analyses on association between IL6 -174G>C and outcomes fasting glucose, 2-h glucose, body mass index (BMI), or interleukin-6 (IL-6) levels 1=as main outcome, 2=as additional outcome; 3=completely unpublished before study recruitment for joint analysis. 4=No or not enough outcome data available for analysis.

^cReferences of association studies between IL6 -174G>C and circulating glucose or IL-6 levels, BMI, or T2DM.

n.a.=not applicable

Table A -IV. (online appendix). Overview of studies and numbers of participants included in the analyses of the quantitative traits fasting glucose, 2-h glucose, body mass index, and circulating interleukin (IL)-6 levels.

	Fasting glucose			2-h glucose			Body mass index			Circulating IL-6 levels
Study	Non-diab^a	T2DM^a	Full^a	Non-diab	T2DM	Full	Non-diab	T2DM	Full	Non-diab	T2DM	Full
CAPPP	286	(37)^b	323				424	(42)^b	466	420	(42)^b	462
DANISH	4397	1139	5536	4396	366	4762	4399	1212	5611
EDSC		106						299	299
EPIC-POTSDAM	65						348		348	346		346
FUSION 1	358	486	844	359	(37)^b	396	367	508	875
FUSION 2	194	397	591	192	78	270	201	437	638
GIRONA	123	(41)^b	164	89	(12)^b	101	123	(42)^b	165	84	(30)^b	114
KORA-MIFAM							881	95	976	872	94	966
KORA-S4	1186	132	1318	1190	118	1308	1190	225	1415	1183	222	1405
KORA-T2DMFAM	512	152	664	505	56	561	513	776	1289
MONICA/KORA-BASE							1744	101	1845	1716	101	1817
MONICA-S3	299						3551	151	3702
NPHS II							2652		2652
RMIFAM							2614	662	3276
UDACS								560	560		549	549
Sum main analyses	7420	2412	9440	6731	618	7398	19007	5026	24117	4621	966	5659
Studies with Hardy-Weinberg equilibrium (HWE) violation (included in sensitivity analyses):
BOTNIA	557	728	1285	539	462	1001	557	731	1288
TGN		52					64	166	230
Sum all studies	7977	3192	10725	7270	1080	8399	19628	5923	25635	4621	966	5659

^aNumber of ‘Non-diab’=non-diabetic subjects, ‘T2DM’=type 2 diabetic subjects, ‘Full’=non-diabetic and type 2 diabetic subjects combined (used for full data analysis).

^bThese type 2 diabetic subjects were not included in the T2DM status-specific analysis because there were less than 50 type 2 diabetic subjects in this study with data on the respective outcome.

Table A -V. (online appendix). Characteristics of study participants in the joint analysis.

					Mean±standard error		IL6 -174G>C
Study	Status^a	T2DM diagnosis^b	No.^c	Male (%)	Age (years)	BMI (kg/m^2)	% GC	% CC	P HWE^d
BOTNIA	1	4, 10	731	53	60±10	29±5	53	25	0.18
	2	2, 3, 4, 6	557	48	53±12	26±4	55	24	0.03
CAPPP	1	5, 7, 10	42	81	58±5	30±4	45	26	0.55
	2	4	424	72	57±7	27±4	45	24	0.11
DANISH	1	3, 4, 5, 6, 10	1212	59	57±11	30±5	48	23	0.28
	2	5, 7	4399	47	45±8	26±4	48	23	0.06
EDSC	1	3, 4, 10	299	63	63±14	30±6	45	22	0.13
	2	n.a.
EPIC-POTSDAM	1	n.a.
	2	1, 2	348	59	55±7	27±4	55	18	0.07
FUSION 1	1	3, 4, 5, 6, 10	508	55	63±7	30±5	47	30	0.15
	2	2, 3, 4, 7	367	41	66±6	27±4	50	30	1.00
FUSION 2	1	3, 4, 5, 6, 10	437	55	64±9	30±5	49	29	0.89
	2	2, 3, 4, 6	201	38	58±9	27±4	52	28	0.30
GIRONA	1	5, 6, 10	42	76	53±10	31±5	50	10	0.73
	2	4, 6	123	82	47±13	26±4	50	24	1.00
KORA-MIFAM	1	1, 3, 8	95	85	58±6	30±5	50	15	0.66
	2	1, 2, 8	881	67	55±8	28±4	47	17	0.71
KORA-S4	1	2, 3, 5, 6, 10	225	59	65±5	31±5	54	16	0.17
	2	1, 2, 4, 6	1190	51	64±5	28±4	52	17	0.06
KORA-T2DMFAM	1	2, 3, 5, 6, 10	776	58	61±10	31±5	52	16	0.02
	2	1, 2, 4, 6	513	43	58±11	28±5	52	16	0.17
MONICA/KORA-BASE	1	2, 10	101	66	62±7	30±4	49	21	0.85
	2	1	1744	53	52±10	27±4	49	17	0.55
MONICA-S3	1	2, 3, 8, 10	151	58	63±8	30±4	46	23	0.33
	2	1, 2, 8	3551	51	48±14	27±4	50	19	0.26
NPHSII	1	n.a.
	2	1	2652	100	56±3	26±3	50	18	0.47
RMIFAM	1	1, 9	662	70	61±7	28±4	51	18	0.26
	2	1, 9	2614	68	58±9	27±3	48	19	0.52
TGN	1	1, 4	166	34	60±10	29±5	50	08	0.16
	2	1, 3, 4, 6	64	41	49±14	30±5	61	08	0.04
UDACS	1	4, 10	560	59	66±11	30±6	45	13	0.85
	2	n.a.

^a1=type 2 diabetic subjects, 2=non-diabetic subjects.

^bType of type 2 diabetes mellitus (T2DM)-diagnosis for type 2 diabetic subjects: 1=interview question; 2=interview question with diabetes confirmation by doctor; 3=diabetes medication; 4=doctor diagnosis; 5=fasting glucose ≥7.0 mmol/L; 6=oral glucose tolerance test with 2-hour glucose ≥11.1 mmol/L (WHO-OGTT); 7=WHO-OGTT in some participants to confirm diagnosis; 8=random glucose ≥11.1 mmol/L; 9=glycated hemoglobin (HbA1c) ≥6.2%; 10=exclusion of subjects with type 1 diabetes.

Type of ‘No T2DM’-diagnosis for non-diabetic subjects: 1=interview question; 2=no diabetes medication; 3=doctor diagnosis; 4=fasting glucose<7.0 mmol/L; 5=fasting glucose <6.1 mmol/l; 6=oral glucose tolerance test (OGTT) with 2-hour glucose<11.1 mmol/L; 7=OGTT with 2-hour glucose <7.8 mmol/L; 8=random glucose <11.1 mmol/L; 9=HbA1c <6.2%.

^cNumber of individuals included in analyses for association between IL6 -174G>C and body mass index BMI (kg/m²).

^dP-value of exact test for Hardy-Weinberg equilibrium (HWE), using 10,000 Monte Carlo permutations. In the four family-studies FUSION 2, KORA-MIFAM, KORA-T2DMFAM, and RMIFAM, HWE was calculated, including only one randomly drawn type 2 diabetic, or non-diabetic subject, respectively, per family.

n.a.=not applicable.

Table VI.  (online appendix). Sensitivity analyses, excluding published studies.^a

Outcome	Egger's test^b	Excluded studies	Individual n	Individual β-estimates (95% CI)	Summary n^c	Summary β-estimate (95% CI)^c
Full data analyses: C-allele dominant model
Fasting glucose (mmol/L)^d	0.12	DANISH	5536	−0.08 ((−0.20)–(0.03))	3904	−0.09 ((−0.19)–(−0.002))
2-h glucose (mmol/L)^d	0.23	n.a.	n.a.	n.a.	7398	−0.08 ((−0.20)–(0.05))
BMI ^e	0.06	CAPPP	466	1.03 (0.26–1.81)	13,413	0.01 ((−0.14)–(0.16))
		DANISH	5611	0.04 ((−0.22)–(0.29))
		KORA-S4	1415	0.13 ((−0.35)–(0.60))
		NPHS II	2652	0.22 ((−0.06)–(0.50))
		UDACS	560	0.86 ((−0.05)–(1.77))
Analysis in type 2 diabetic subjects: CC- versus GG-genotype
Ln IL-6 (pg/mL)^d	0.62	KORA-S4	222	0.15 ((−0.20)–(0.51))	744	0.14 ((−0.02)–(0.30))

^aPublished for the respective outcome as a main result at the time of study recruitment.

^bP-value of Egger's regression test for publication bias including all studies of main analysis.

^cOf unpublished studies.

^dAdjusted for age, sex, and body mass index BMI (kg/m²).

^eAdjusted for age, sex, and type 2 diabetes mellitus (T2DM).

n.a.=not applicable (because all studies were unpublished).

Table VII.  (online appendix). Full data sensitivity analyses, additionally including studies showing Hardy-Weinberg equilibrium (HWE) violation in non-diabetic subjects (C-allele dominant model).

Outcome	Additionally included studies	Individual n	Individual β-estimates (95% CI)	Summary n	Summary β-estimate (95% CI)
Fasting glucose (mmol/L)^a	BOTNIA	1285	0.01 ((−0.38)–(0.39))	10,725	−0.09 ((−0.16)–(−0.02))
2-h glucose (mmol/L)^a	BOTNIA	1001	0.44 ((−0.37)–(1.25))	8399	−0.06 ((−0.19)–(0.06))
BMI ^b	BOTNIA	1288	−0.02 ((−0.58)–(0.55))	25,635	0.08 ((−0.03)–(0.19))
	TGN	230	−0.27 ((−1.52)–(0.99))

^aAdjusted for age, sex, and body mass index (BMI).

^bAdjusted for age, sex, and type 2 diabetes mellitus (T2DM).

Wernstedt I, Eriksson AL, Berndtsson A, Hoffstedt J, Skrtic S, Hedner T, et al. A common polymorphism in the interleukin-6 gene promoter is associated with overweight. Int J Obes Relat Metab Disord. 2004; 28: 1272–9

 PubMed Web of Science ®Google Scholar

Hamid YH, Rose CS, Urhammer SA, Glumer C, Nolsoe R, Kristiansen OP, et al. Variations of the interleukin-6 promoter are associated with features of the metabolic syndrome in Caucasian Danes. Diabetologia. 2005; 48: 251–60

 Google Scholar

Flavell DM, Ireland H, Stephens JW, Hawe E, Acharya J, Mather H, et al. Peroxisome proliferator-activated receptor alpha gene variation influences age of onset and progression of type 2 diabetes. Diabetes. 2005; 54: 582–6

 PubMed Web of Science ®Google Scholar

Mohlig M, Boeing H, Spranger J, Osterhoff M, Kroke A, Fisher E, et al. Body mass index and C-174G interleukin-6 promoter polymorphism interact in predicting type 2 diabetes. J Clin Endocrinol Metab. 2004; 89: 1885–90

 Google Scholar

Valle T, Tuomilehto J, Bergman RN, Ghosh S, Hauser ER, Eriksson J, et al. Mapping genes for NIDDM. Design of the Finland-United States Investigation of NIDDM Genetics (FUSION) Study. Diabetes Care. 1998; 21: 949–58

 PubMed Web of Science ®Google Scholar

Silander K, Mohlke KL, Scott LJ, Peck EC, Hollstein P, Skol AD, et al. Genetic variation near the hepatocyte nuclear factor-4 alpha gene predicts susceptibility to type 2 diabetes. Diabetes. 2004; 53: 1141–9

 PubMed Web of Science ®Google Scholar

Vozarova B, Fernandez-Real JM, Knowler WC, Gallart L, Hanson RL, Gruber JD, et al. The interleukin-6 ((174) G/C promoter polymorphism is associated with type-2 diabetes mellitus in Native Americans and Caucasians. Hum Genet. 2003; 112: 409–13

 PubMed Web of Science ®Google Scholar

Rathmann W, Haastert B, Icks A, Löwel H, Meisinger C, Holle R, et al. High prevalence of undiagnosed diabetes mellitus in Southern Germany: Target populations for efficient screening. The KORA survey 2000. Diabetologia. 2003; 46: 182–9

 Google Scholar

Thorand B, Baumert J, Kolb H, Meisinger C, Chambless L, Koenig W, et al. Sex Differences in the prediction of type 2 diabetes by inflammatory markers: results from the MONICA/KORA Augsburg case-cohort study, 1984–2002. Diabetes Care. 2007; 30: 854–60

 Google Scholar

Broeckel U, Hengstenberg C, Mayer B, Holmer S, Martin LJ, Comuzzie AG, et al. A comprehensive linkage analysis for myocardial infarction and its related risk factors. Nat Genet. 2002; 30: 210–4

 PubMed Web of Science ®Google Scholar

Stephens JW, Hurel SJ, Cooper JA, Acharya J, Miller GJ, Humphries SE. A common functional variant in the interleukin-6 gene is associated with increased body mass index in subjects with type 2 diabetes mellitus. Mol Genet Metab. 2004; 82: 180–6

 Google Scholar

Lieb W, Pavlik R, Erdmann J, Mayer B, Holmer SR, Fischer M, et al. No association of interleukin-6 gene polymorphism ((174 G/C) with myocardial infarction or traditional cardiovascular risk factors. Int J Cardiol. 2004; 97: 205–12

 Google Scholar

Illig T, Bongardt F, Schopfer A, Muller-Scholze S, Rathmann W, Koenig W, et al. Significant association of the interleukin-6 gene polymorphisms C-174G and A-598G with type 2 diabetes. J Clin Endocrinol Metab. 2004; 89: 5053–8

 PubMed Web of Science ®Google Scholar