- Email: [email protected]
Serum Magnesium Levels and In-Hospital Mortality in Acute Myocardial Infarction
JACC VOL. 69, NO. 22, 2017
Letters
JUNE 6, 2017:2769–76
were time stamped. As this dataset represents real-
Serum Magnesium Levels and In-Hospital Mortality in Acute Myocardial Infarction
world care, local providers dictated all laboratory draws, electrolyte supplementation, and other aspects of clinical care. Hierarchical Cox regression evaluated the association between the most recent Mg level (as a timevarying covariate) and in-hospital mortality. As a secondary analysis, logistic regression was used to
Based primarily on cellular data (1) and community-
examine the association between admission Mg levels
based
an
and in-hospital malignant arrhythmia (ventricular
association between low magnesium (Mg) levels and
tachycardia or fibrillation or cardiac arrest). As
malignant
guidelines
occurrences of arrhythmia were not time-stamped
recommend Mg supplementation to maintain levels
(only noted on discharge diagnoses), we could not
>2.0
infarction
analyze Mg levels as a time-varying covariate. Both
(AMI) (3). These recommendations are despite the
models were adjusted for age, sex, race, smoking,
lack of supporting data from >26 randomized trials
alcohol abuse, diabetes, heart failure, cerebrovascular
of routine Mg supplementation (4). We used the
disease, kidney disease, liver disease, and admission
Cerner Health Facts database, a national database of
concentrations of potassium, creatinine, and hemat-
patients hospitalized with AMI, to determine the
ocrit. Hospital site was included as a random effect to
association of serum Mg levels with in-hospital
account for clustering of patients within sites.
epidemiological arrhythmia
mg/dl
during
studies and
acute
(2)
showing
mortality, myocardial
The University of Missouri Kansas City institu-
mortality and malignant cardiac arrhythmia. Eligible patients were hospitalized between 2005
tional review board determined this was non-human
and 2015 with a clinically relevant AMI, defined as a
subjects research, and a waiver of informed consent
principal discharge diagnosis of AMI and cardiac
was granted. Analyses were conducted using SAS
biomarkers >3 upper limit of normal ULN within
version 9.4 software (SAS Institute, Cary, North
48 h of admission, or a secondary discharge diagnosis
Carolina) and R version 3.2.1 software (R Core Team,
of AMI and cardiac biomarkers >10 ULN within 48 h
Vienna, Austria).
of admission. All data were extracted and deidenti-
Among 10,806 patients with a clinically relevant
fied from electronic health records. Laboratory data
AMI hospitalized at 77 U.S. hospitals, median age was
F I G U R E 1 Mg Level and In-Hospital Mortality
4.0
2.0
p < 0.001
1.0
0.5 1.2
1.6
2.0
2.4
Most Recent Mg Level
2.8
Odds Ratio for In-Hospital Malignant Arrhythmia
B Relative Hazard for In-Hospital Mortality
A
4.0
2.0
p < 0.001
1.0
0.5 1.2
1.6
2.0
2.4
2.8
Initial Mg Level in First 24h
(A) Adjusted association between most recent serum magnesium (Mg) level and in-hospital mortality. Model accounts for demographic and clinical factors, Mg checks during hospitalization, and Mg as a time-varying covariate. (B) Adjusted association between admission Mg levels and in-hospital malignant arrhythmia.
2771
2772
JACC VOL. 69, NO. 22, 2017
Letters
JUNE 6, 2017:2769–76
66 years of age, 61% were male, and 81% were white. Median length of stay was 3.1 (interquartile range [IQR]: 2.1 to 5.8) days. Sixty-three percent of the subjects had $1 Mg level checked, and the median Mg concentration at admission was 1.9 (IQR: 1.8 to 2.1) mg/dl. Patients who presented with low Mg concentrations were more likely to be female and have diabetes and heart failure. Patients with
Please note: Drs. Shafiq and Qintar are supported by T32 grant T32HL110837 from the National Heart, Lung, and Blood Institute. Dr. Kosiborod has received research grants from AstraZeneca, Gilead Sciences, Genentech, Sanofi, Amgen, ZS Pharma; consulting honoraria from AstraZeneca, Sanofi, Eli Lilly, GlaxoSmithKline, Boehringer Ingelheim, Takeda, Merck (Diabetes), Novo Nordisk, Glytec, ZS Pharma; and is a member of the speakers bureau for Amgen. Dr. Goyal has received advisory board honoraria from ZS Pharma. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
REFERENCES
admission Mg concentrations of <1.8, 1.8, 1.9 to 2.0,
1. Agus ZS. Hypomagnesemia. J Am Soc Nephrol 1999;10:1616–22.
and >2 mg/dl had in-hospital mortality rates of 7.4%
2. Ford ES. Serum magnesium and ischaemic heart disease: findings from a national sample of US adults. Int J Epidemiol 1999;28:645–51.
versus 4.1% versus 4.7% versus 9.7%, respectively (unadjusted p < 0.001). After we adjusted for patient factors, there continued to be a U-shaped relationship between
most
recent
Mg
level
and
mortality
(Figure 1A), with the lowest hazard of death at most recent Mg concentration of w1.8 mg/dl. A similar relationship was evident between admission Mg
3. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 1999 Guidelines for the Management of Patients With Acute Myocardial Infarction). J Am Coll Cardiol 2004;44:671–719. 4. Li J, Zhang Q, Zhang M, Egger M. Intravenous magnesium for acute myocardial infarction. Cochrane Database Syst Rev 2007:CD002755.
levels and odds of malignant arrhythmia (Figure 1B). Previously
conducted
large
multicenter
trials
found no benefit of routine intravenous Mg supplementation in AMI (1); however, those studies did not specifically assess the relationship between Mg concentrations
and
in-hospital
outcomes.
In
a
Concerns About the Stability of hsTnI Assay After 20 Years of Storage
retrospective multicenter study of >10,000 patients with AMI, we found that serum Mg concentrations were
associated
with
in-hospital
mortality and
malignant arrhythmias in a U-shaped manner, with the lowest risk of adverse events at serum Mg levels of w1.7 to 1.9 mg/dl. Although our results confirm risk associated with hypomagnesemia, they also demonstrate increased risk with higher Mg levels and suggest that the optimal range of Mg in patients with AMI may be lower than that currently recommended by AMI guidelines. Importantly, these results represent associations and therefore cannot explicitly define a specific safe Mg level; however, they do suggest that aggressive supplementation may
not
be
necessary
(or
safe)
for
Mg
It was with great interest that we read the paper by Ford et al. (1) on the potential utility of highsensitivity troponin I (hsTnI) for risk stratification in primary prevention and for monitoring the response of statin therapy on hsTnI concentrations. There
is accumulating evidence
that patient
selection by troponin level and monitoring of response to therapy in acute cardiac care and in primary prevention will become a powerful tool in the future. The present article, however, has severe limitations, and we feel committed to dampen exaggerated expectations for a new indication for hsTnI. Our first concerns relate to the stability of the
concentrations slightly <2.0 mg/dl.
troponin molecule after 25 years (last patient in 1991)
*Ali Shafiq, MD, MSc Abhinav Goyal, MD, MHS Philip G. Jones, MS Suman Sahil, BS, BEng Mark Hoffman, PhD Mohammed Qintar, MD Donna M. Buchanan, PhD Mikhail Kosiborod, MD Suzanne V. Arnold, MD, MHA
Coronary Prevention Study) trial. Information about
of sample storage in the WOSCOPS (West of Scotland sample stability after such a long freezing period are not available for cTnI assays. In particular, the package insert of the Abbott Architect STAT hsTnI (Abbott Diagnostics, Abbott Park, Illinois) provides data for sample stability during freezing only for a maximum of 30 days at temperatures around 30 (2).
However,
time-dependent
changes
in
cTnI
immune reactivity have been published previously
*Aurora Cardiovascular Services
for freezing periods as short as 48 to 72 h. Recovery
950 North 12th Street, 2nd Floorj
of cTnT shows a more stable time course in
Milwaukee, Wisconsin 53233
refrigerated and frozen heparin plasma pools at
E-mail: alishafi[email protected]
temperatures
http://dx.doi.org/10.1016/j.jacc.2017.03.579
because capture and detection antibodies react in
between
4 C
and
84 C,
likely
Letters
JUNE 6, 2017:2769–76
were time stamped. As this dataset represents real-
Serum Magnesium Levels and In-Hospital Mortality in Acute Myocardial Infarction
world care, local providers dictated all laboratory draws, electrolyte supplementation, and other aspects of clinical care. Hierarchical Cox regression evaluated the association between the most recent Mg level (as a timevarying covariate) and in-hospital mortality. As a secondary analysis, logistic regression was used to
Based primarily on cellular data (1) and community-
examine the association between admission Mg levels
based
an
and in-hospital malignant arrhythmia (ventricular
association between low magnesium (Mg) levels and
tachycardia or fibrillation or cardiac arrest). As
malignant
guidelines
occurrences of arrhythmia were not time-stamped
recommend Mg supplementation to maintain levels
(only noted on discharge diagnoses), we could not
>2.0
infarction
analyze Mg levels as a time-varying covariate. Both
(AMI) (3). These recommendations are despite the
models were adjusted for age, sex, race, smoking,
lack of supporting data from >26 randomized trials
alcohol abuse, diabetes, heart failure, cerebrovascular
of routine Mg supplementation (4). We used the
disease, kidney disease, liver disease, and admission
Cerner Health Facts database, a national database of
concentrations of potassium, creatinine, and hemat-
patients hospitalized with AMI, to determine the
ocrit. Hospital site was included as a random effect to
association of serum Mg levels with in-hospital
account for clustering of patients within sites.
epidemiological arrhythmia
mg/dl
during
studies and
acute
(2)
showing
mortality, myocardial
The University of Missouri Kansas City institu-
mortality and malignant cardiac arrhythmia. Eligible patients were hospitalized between 2005
tional review board determined this was non-human
and 2015 with a clinically relevant AMI, defined as a
subjects research, and a waiver of informed consent
principal discharge diagnosis of AMI and cardiac
was granted. Analyses were conducted using SAS
biomarkers >3 upper limit of normal ULN within
version 9.4 software (SAS Institute, Cary, North
48 h of admission, or a secondary discharge diagnosis
Carolina) and R version 3.2.1 software (R Core Team,
of AMI and cardiac biomarkers >10 ULN within 48 h
Vienna, Austria).
of admission. All data were extracted and deidenti-
Among 10,806 patients with a clinically relevant
fied from electronic health records. Laboratory data
AMI hospitalized at 77 U.S. hospitals, median age was
F I G U R E 1 Mg Level and In-Hospital Mortality
4.0
2.0
p < 0.001
1.0
0.5 1.2
1.6
2.0
2.4
Most Recent Mg Level
2.8
Odds Ratio for In-Hospital Malignant Arrhythmia
B Relative Hazard for In-Hospital Mortality
A
4.0
2.0
p < 0.001
1.0
0.5 1.2
1.6
2.0
2.4
2.8
Initial Mg Level in First 24h
(A) Adjusted association between most recent serum magnesium (Mg) level and in-hospital mortality. Model accounts for demographic and clinical factors, Mg checks during hospitalization, and Mg as a time-varying covariate. (B) Adjusted association between admission Mg levels and in-hospital malignant arrhythmia.
2771
2772
JACC VOL. 69, NO. 22, 2017
Letters
JUNE 6, 2017:2769–76
66 years of age, 61% were male, and 81% were white. Median length of stay was 3.1 (interquartile range [IQR]: 2.1 to 5.8) days. Sixty-three percent of the subjects had $1 Mg level checked, and the median Mg concentration at admission was 1.9 (IQR: 1.8 to 2.1) mg/dl. Patients who presented with low Mg concentrations were more likely to be female and have diabetes and heart failure. Patients with
Please note: Drs. Shafiq and Qintar are supported by T32 grant T32HL110837 from the National Heart, Lung, and Blood Institute. Dr. Kosiborod has received research grants from AstraZeneca, Gilead Sciences, Genentech, Sanofi, Amgen, ZS Pharma; consulting honoraria from AstraZeneca, Sanofi, Eli Lilly, GlaxoSmithKline, Boehringer Ingelheim, Takeda, Merck (Diabetes), Novo Nordisk, Glytec, ZS Pharma; and is a member of the speakers bureau for Amgen. Dr. Goyal has received advisory board honoraria from ZS Pharma. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
REFERENCES
admission Mg concentrations of <1.8, 1.8, 1.9 to 2.0,
1. Agus ZS. Hypomagnesemia. J Am Soc Nephrol 1999;10:1616–22.
and >2 mg/dl had in-hospital mortality rates of 7.4%
2. Ford ES. Serum magnesium and ischaemic heart disease: findings from a national sample of US adults. Int J Epidemiol 1999;28:645–51.
versus 4.1% versus 4.7% versus 9.7%, respectively (unadjusted p < 0.001). After we adjusted for patient factors, there continued to be a U-shaped relationship between
most
recent
Mg
level
and
mortality
(Figure 1A), with the lowest hazard of death at most recent Mg concentration of w1.8 mg/dl. A similar relationship was evident between admission Mg
3. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 1999 Guidelines for the Management of Patients With Acute Myocardial Infarction). J Am Coll Cardiol 2004;44:671–719. 4. Li J, Zhang Q, Zhang M, Egger M. Intravenous magnesium for acute myocardial infarction. Cochrane Database Syst Rev 2007:CD002755.
levels and odds of malignant arrhythmia (Figure 1B). Previously
conducted
large
multicenter
trials
found no benefit of routine intravenous Mg supplementation in AMI (1); however, those studies did not specifically assess the relationship between Mg concentrations
and
in-hospital
outcomes.
In
a
Concerns About the Stability of hsTnI Assay After 20 Years of Storage
retrospective multicenter study of >10,000 patients with AMI, we found that serum Mg concentrations were
associated
with
in-hospital
mortality and
malignant arrhythmias in a U-shaped manner, with the lowest risk of adverse events at serum Mg levels of w1.7 to 1.9 mg/dl. Although our results confirm risk associated with hypomagnesemia, they also demonstrate increased risk with higher Mg levels and suggest that the optimal range of Mg in patients with AMI may be lower than that currently recommended by AMI guidelines. Importantly, these results represent associations and therefore cannot explicitly define a specific safe Mg level; however, they do suggest that aggressive supplementation may
not
be
necessary
(or
safe)
for
Mg
It was with great interest that we read the paper by Ford et al. (1) on the potential utility of highsensitivity troponin I (hsTnI) for risk stratification in primary prevention and for monitoring the response of statin therapy on hsTnI concentrations. There
is accumulating evidence
that patient
selection by troponin level and monitoring of response to therapy in acute cardiac care and in primary prevention will become a powerful tool in the future. The present article, however, has severe limitations, and we feel committed to dampen exaggerated expectations for a new indication for hsTnI. Our first concerns relate to the stability of the
concentrations slightly <2.0 mg/dl.
troponin molecule after 25 years (last patient in 1991)
*Ali Shafiq, MD, MSc Abhinav Goyal, MD, MHS Philip G. Jones, MS Suman Sahil, BS, BEng Mark Hoffman, PhD Mohammed Qintar, MD Donna M. Buchanan, PhD Mikhail Kosiborod, MD Suzanne V. Arnold, MD, MHA
Coronary Prevention Study) trial. Information about
of sample storage in the WOSCOPS (West of Scotland sample stability after such a long freezing period are not available for cTnI assays. In particular, the package insert of the Abbott Architect STAT hsTnI (Abbott Diagnostics, Abbott Park, Illinois) provides data for sample stability during freezing only for a maximum of 30 days at temperatures around 30 (2).
However,
time-dependent
changes
in
cTnI
immune reactivity have been published previously
*Aurora Cardiovascular Services
for freezing periods as short as 48 to 72 h. Recovery
950 North 12th Street, 2nd Floorj
of cTnT shows a more stable time course in
Milwaukee, Wisconsin 53233
refrigerated and frozen heparin plasma pools at
E-mail: alishafi[email protected]
temperatures
http://dx.doi.org/10.1016/j.jacc.2017.03.579
because capture and detection antibodies react in
between
4 C
and
84 C,
likely