Table 1. Characteristics of the prospective studies included in the meta-analysis of published studies on magnesium intake and the risk of CVD events.
| Source | Location/Follow-up, y | Study Population (n) | Sex/Age, y | Recruitment Time | Outcome(s) | Outcomes Assessment | Magnesium intake Assessment |
| Ascherio (1998) | United States, 8 | 43,738 | Male, 40–75 | 1986 | Stroke | Ascertained by self-report; subclassified accordingto the criteria of the National Survey of Stroke. | Validated FFQ |
| Liao (1998) | United States, 7 | 13,922 | Female/Male, 45–64 | 1987–1989 | CHD | Ascertained by self-report and medical records, deathcertificates; reviewed by members of the ARIC Morbidityand Mortality Classification Committee | FFQ |
| Iso (1999) | United States, 14 | 85,764 | Female, 34–59 | 1976 | Stroke | Ascertained by self-report and medical records;confirmed according to the criteria of theNational Survey of Stroke. | Validated FFQ |
| Abbott (2003) | United States, 30 | 7172 | Male, 45–68 | 1965–1968 | CHD | Confirmation by the Honolulu Heart ProgramMorbidity and Mortality Review Committee. | 24-hour dietary recall |
| Al-Delaimy (2004) | United States, 12 | 39,633 | Male, 40–75 | 1986 | CHD | Ascertained by self-report and medical records. | Validated FFQ |
| Song (2005) | United States, 10 | 39,876 | Female, 39–89 | 1991 | CVD | Confirmed through medical records, autopsy reports,and death certificates. | Validated FFQ |
| Larsson (2008) | Finnish,13.6 | 26,556 | Male, 50–69 | 1985–1988 | Stroke | Identified by National Hospital Discharge Register andthe National Register of Causes of Death and classifiedaccording to ICD-8,9,10 (ICD-8 codes 430–434,436; ICD-9codes 430, 431, 433, 434, 436; ICD-10 codes I60, I61, I63,and I64) | Validated FFQ |
| Weng (2008) | China, 10.6 | 1772 | Female/Male, ≥40 | 1990–1993 | Stroke | Ascertained by self-reported; classified according toICD-9-CM (codes 430 to 438) | FFQ |
| Ohira (2009) | United States, 15 | 14,221 | Female/Male, 45–64 | 1987–1989 | Stroke | Ascertained by self-report, local hospitals, state vitalstatistics offices; and classified according to ICD-9(codes 430–438) | FFQ |
| Kaluza (2010) | Swedish, 10 | 23,366 | Male, 45–79 | 1997–1998 | CVD death | Review of Swedish Death and Population Registersand classified according to ICD-10 (codes I00–I79) | Validated FFQ |
| Chiuve (2011) | United States, 10 | 88,375 | Female, 30–55 | 1976 | CVD death | Documented by medical records | FFQ |
| Larsson (2011) | Swedish,10.4 | 34,670 | Female, 49–83 | 1987–1990 | Stroke | Ascertained by Swedish Hospital Discharge Registry andidentified according to ICD-10 (codes I60, I61, I63, I64) | Validated FFQ |
| Zhang (2012) | Japan, 14.7 | 58,615 | Female/Male, 40–79 | 1988–1990 | CVD death | A systematic review of death certificates; and classifiedaccording to ICD-9 (codes 390–459) and ICD-10 (codes101–199) | Validated FFQ |
CHD = coronary heart disease; CVD = cardiovascular disease; FFQ = food frequency questionnaire; ICD = International Classification of Diseases.