Table 2 Quality of Included Studies

Acarturk, 2004[13]

not random; all patients

admitted for diagnostic

coronary angiography

chart review, angiography results

NR

NR

TG value in the blood was used

as a continuous number

(variable). OR expresses

increased risk per unit of serum

TG level

Bansal, 2007[14]

derived from women health study, previously completed randomized controlled trial of aspirin and vitamin E

chart review, events adjudicated by

an end point committee

adjusted for treatment

assignment to ASA,

vitamin E, beta

carotene, age, BP,

smoking status, and use

of hormone therapy,

levels of total

cholesterol and HDL-C,

history of DM, BMI,

high-sensitivity C-

reactive protein

0

TG value in the blood was used

as a continuous number

(variable). OR expresses increased risk per unit of serum TG level

Barrett-Connor, 1987[15]

random sample

chart review, ICD or billing codes,

death certificates

adjust by TG level, age,

BP, BMI, smoking

habit, DM, family

history of heart attack

0.5%

Compared normal to "borderline HTG", defined as TG between

240-500 mg/dL (2.7-5.65

mmol/L)

Bass, 1993[16]

subset of female participants

in the Lipid Research Clinics' Follow-up Study

chart review, annual checkups

Adjusted for age, HTN,

DM, smoking, history

of heart disease and

estrogen use

NR

Compared TG < 200 mg/dL

(< 2.25 mmol/L) to elevated 200

to 399 mg/dL (2.25 to 4.49

mmol/L) and high > 400 mg/dL (> 4.50 mmol/L)

Bonaventure, 2010[17]

not random and not consecutive: recruited from electoral rolls

Death certificates and autopsy

reports, ascertained the same way

in cases and controls

medical history of MI,

stroke, or peripheral

arterial disease, as well

as smoking and alcohol consumption status

(never, former, current),

excess weight, elevated

BP, DM, apolipoprotein

E (APOE) genotype,

low-dose aspirin intake, and lipid-lowering treatment

NR

They compared tertiles or

quintiles: TG < 83.4 mg/dL

(< 0.94), 84.2-117.8 mg/L (0.95-1.33), and > 118.7 mg/dL (1.34 mmol/L)

Carlson, 1988[18]

consecutive sample (all patients presenting with HTG)

chart review, ascertained the same

way in cases and controls, done

without knowledge of patients' TG

level

NR

13.4%

3 groups according to TG levels. Low = TG < 132.9 mg/dL (1.5 mmol/L), intermediate = TG

132.9-177.2 mg/dL (1.5-2.0 mmol/L), high = TG > 177.2 (2.0 mmol/L).

Chan, 2005[19]

not random;

consecutive patients with type 2 DM, not HPTG

chart review, death registry

Adjusted for sex

and age. stepwise linear regression with BMI,

WC, HbA1c, FPG and

HOMA as independent

variables and lipid

profile as dependent

variable

0

Unclear

Chester, 1995[20]

consecutive sample (all men presenting with HTG and are awaiting routine angioplasty)

chart review,

done without knowledge of

patients' TG level

The potential predictor

variables-that is, risk

factors assessed at

baseline angiography,

for adverse events were

analyzed using the

multiple logistic

regression models.

2

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: mmol/L

Czernichow, 2007[21]

consecutive sample

self report, chart review, ICD or

billing codes, ascertained

differently: self report in all

patients, however if a CVD event

was reported -- chart review and

ICD billing codes were reviewed

for those individuals only

Age

NR

Age-adjusted relative risk

correlate to one standard

deviation increase in TG levels

Drexel, 2005[22]

consecutive sample

follow up investigation after 2.3

years, Time and causes of death

were obtained from national

surveys, hospital records

age, sex, and use of

lipid-lowering

medication

0

Unclear

Eberly, 2003[23]

not random; likely consecutive sample: 2863 men with both nonfasting and fasting TG levels measured at screens 1 and 2

self report, chart review, ICD or

billing codes, death certificates

age, lipids subfractions,

glucose level, BP,

cigarettes smoked per

day, alcohol use, BMI

and race

0

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: mg/dL

Egger, 1999[24]

not random; likely consecutive sample: Participants of the Caerphilly Heart Disease Study

self report, chart review, ICD or

billing codes

age, all three lipid

factors, laboratory error

and within person

variation, blood

glucose and diastolic

BP, BMI, smoking and

markers for pre-existent

disease

12.5

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: mmol/L

Ellingsen, 2003[25]

not random; likely

consecutive sample: 1232

healthy men with elevated cholesterol or coronary risk

score included in the study

from a pool of 16202

screened men

chart review, ICD or billing codes, ascertained the same way in cases

and controls

adjusted for age, BMI,

cigarette smoking, total cholesterol,

triacylglycerol, glucose,

BP, dietary score,

alcohol intake, and

activity level

0

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: high TG > or = 178.1 mg/dL (2.01 mmol/L)

Gaziano, 1997[26]

not random, likely

consecutive sample:

Men/women < 76 yrs. age

with no prior history of CAD discharged from one of 6

Boston area hospitals with the diagnosis of confirmed MI

chart review, medical exam/lab

analysis,

ascertained differently: cases were interviewed 8 weeks after MI

Adjusted for age, sex,

history of HTN, history

of DM, body mass

index, type A

personality, family

history of previous MI,

alcohol consumption,

physical activity,

smoking, caloric intake

12

they compared quintiles, highest compared to lowest

Goldberg, 2009[27]

consecutive sample (all

patients presenting with HTG)

chart review, telephone calls,

ascertained the same way in cases

and controls

A time-to-event

regression model was

performed to establish

the role of baseline lipid subfractions, other

metabolic risk factors,

lifestyle variables, and demographic

characteristics in

relation to the development

of CAD.

3.8

high triglyceride level > = 248.1 mg/dL (2.8 mmol/L)

Habib, 2006[28]

Data from the United States Renal Data System database collected during the

prospective Dialysis

Morbidity and Mortality

Study Wave 2 study

chart review

age, gender, race,

weight, height, primary

cause of ESRD,

hemoglobin, serum

albumin, serum calcium phosphate product,

serum bicarbonate,

residual kidney

creatinine clearance, PD parameters (dialysate

effluent volume,

dialysis creatinine

clearance, D/P creatinine ratio after a 4 h dwell), use of lipid-modifying medications and comorbidity characteristics

0

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: HR is using a

reference of TG levels 200-300

mg/dl (2.2-3.4 mmol/L)

Haim, 1999[29]

not random; likely

consecutive sample

chart review, ICD or billing codes

age, previous MI, DM,

NYHA class, HTN,

LDL cholesterol,

glucose, chronic

obstructive pulmonary

disease, peripheral

vascular disease, stroke,

angina pectoris,

smoking, and lipids

0.37

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: mg/dL

Hoogeveen, 2001[30]

Random sample

chart review, clinical exam and investigations,

ascertained the same way in cases

and controls

Logistic regression

applied but no specific adjustments are

mentioned

12

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: 10 mg/dL (0.11 mmol/L)

Jonsdottir, 2002[31]

not random; likely

consecutive: subjects of the Reykjavik Study

self report, chart review, ICD or billing codes

age, high-density lipoprotein cholesterol, total/low-density lipoprotein cholesterol, smoking, body mass index and BP

0.6

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: mmol/L

Lamarche, 1995[32]

random sample

chart review, Examination/EKG/death certificate

Adjusted for age,

systolic BP, DM,

alcohol consumption,

and tobacco use

27

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: TG > 203.8 mg/dL (2.3 mmol/L)

Lloret Linares, 2008[33]

not random and not

consecutive: Patients referred

by their general practitioner

or general hospital for very

high TG levels to

Endocrinology Dept. between 2000 and 2005

self report, chart review

Adjusted for age at hospitalization

NR

TG: lowest 95.1-180 mg/dL

(1.1-2.0 mmol/L) vs. highest

360-1505 gm/dL (4.1-17

mmol/L).

Lu, 2003[34]

not random; likely

consecutive: cohort chosen

from the strong heart study to include only DM, no baseline CVD

through death certificates and tribal

and Indian Health Service hospital records and by direct contact of

study personnel with the study participants and their families

Adjusted for age, BMI,

smoking status, study

center, systolic BP,

HbA1c, fibrinogen,

insulin, and ratio of

albumin to creatinine

0

They compared tertiles or

quintiles: TG: lower < 111; 111-

175; higher > 175 mg/dL

(lower < 1.2; 1.2-2.0; higher > 2.0 mmol/L)

Malone, 2009[35]

Not random; likely

consecutive: Retrospective

data from 3 integrated health-care systems that

systematically collect

and store detailed patient-level data.

Chart review, ICD or billing codes

Adjusted for age, sex,

smoking status and site

N/A

lower/normal TG - 80.0 mg/dl

(0.9 mmol/L); higher TG - TG = 217.4 mg/dl (2.4 mmol/L)

Mazza, 2005[36]

random sample

chart review, ICD or billing codes, through the Register Office, general practitioners

Gender, age, DM,

obesity, lipids

subfractions, serum uric

acid, BP, smoking,

alcohol and proteinuria

0

They compared tertiles or

quintiles: TG: First (low) < 97.5 mg/dL (1.01 mmol/L); Fifth (high) > = 156.8 mg/dL (1.77 mmol/L)

Mora, 2008[37]

Random sample enrolled in

the Women's Health Study

Follow-up questionnaires every 6-

12 months

Adjusted for age,

randomized treatment assignment, smoking

status, menopausal

status, postmenopausal

hormone use, BP, DM,

and BMI

NR

They compared tertiles or

quintiles: TG: First (low) < 89.5 mg/dL (1.01 mmol/L); Fifth (high) > = 180.7 mg/dL (2.04 mmol/L)

Noda, 2010[38]

not random and not

consecutive: death related to a MI defined a case, then 2 controls were selected

randomly matched by age

Death registration from 1997-2000,

done without knowledge of

patients' TG level,

ascertained the same way in cases

and controls

Adjusted for age and 6

risk factors for MI

NR

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum TG level: High TG > = 150 mg/dL (1.7 mmoml/L)

Rubins, 1999[39]

not random and not

consecutive: to obtain

population with appropriate

lipid levels, a multi stage screening method that

included two lipid profiles obtained one week apart

chart review, clinical and radiologic

data, ascertained the same way in

cases and controls

Adjustment for baseline

variables in the Cox

models had a trivial

effect on the estimates

of the hazard ratios

2.3%

Two groups: TG < 150 mg/dl (1.7 mmol/L) and TG > 150 mg/dl (1.7 mmol/L)

Samuelsson, 1994[40]

random sample

chart review

traditional risk factors,

end-organ damage

status

NR

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: RR reported for every

88.6 mg/dL (1 mmol/L) increase

in TG level

Schupf, 2005[41]

random sample

self report, chart review,

interviewing relatives

Adjusted for age, sex,

ethnicity, and level of

education, for BMI or

APOE; a history of

HTN, DM, heart

disease, stroke, or

cancer; or current

smoking

0

They compared tertiles or

quintiles: Lowest - < = 98.9 mg/dl

(1.1 mmol/L), highest

- > 191.2 mg/dl (2.1 mmol/L); RR compared the lowest quartile to

the highest quartile.

Sprecher, 2000[42]

not random; likely

consecutive: diabetic patients undergoing primary isolated CABG between 1982 and

1992 at Cleveland Clinic

chart review, clinical exam, labs

and CVIR (Cardiovascular

Information Registry)

age, sex, left ventricular

function, coronary

anatomy, history of

HTN, BMI, and total

cholesterol

highest quartile compared to

lower three quartiles (normal)

Tanko, 2005[43]

not random and not

consecutive:

recruited via a questionnaire surveys

self report, chart review: Central

Registry of the Danish Ministry of

Health

Adjusted for age,

smoking, and LDL-C),

waist circumference

NR

TG value in the blood was used

as a continuous number

(variable). Here OR expresses increased risk per unit of serum

TG level: presented as 2 cutoffs - > 128.5 mg/dL (1.45 mmol/L) - > 149.7 mg/dL (1.69 mmol/L)

Tsai, 2008[44]

not random; likely

consecutive:

civil servants and teachers

who took the annual physical examination at the Taipei Outpatient Center

chart review, annual exam, national

death files

Adjusted for age,

gender, fasting glucose,

BP, BMI, smoking

NR

They compared tertiles or

quintiles: TG

normal < 150 mg/dL (1.7

mmol/L, abnormal 150 mg/dL

(1.7 mmol/L)-199 mg/dL (2.2 mmol/L), and high abnormal

> 200 mg/dL (2.25 mmol/L)

Upmeier, 2009[45]

not random and not

consecutive:

mailed invitation to

participate to all residents of Turku born in 1920

Self report, chart review,

ICD or billing codes

Adjusted for gender,

body mass index,

smoking and any history

of angina pectoris,

stroke, DM, and HTN

NR

They compared TG level

quartiles, highest to lowest

Valdivielso, 2009[46]

not random; likely consecutive

Chart review and self report

age, sex, smoking,

HTN, DM, and lipids

fractions

NR

Categorized as normal when TG

was < 150 mg/dL (< 1.69

mmol/L); the remainder were considered to be HTG

Wier, 2003[47]

not random; likely

consecutive

chart review,

done without knowledge of

patients' TG level

age, time of resolution

of symptoms, smoking,

BP, presence of atrial

fibrillation and

hyperglycemia

0

They compared tertiles or

quintiles: TG, mmol/l: < = 0.9;

1.0-1.3; 1.4-1.8; > = 1.9.

Mg/dL: < = 79.7; 88.6-115.2;

124.0-159.5; > = 168.3