Methods

The method was guided by Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA). The inclusion criteria were prospective studies published in adult patients with acute or chronic heart failure or AMI or patients presenting with acute dyspnea or acute chest pain or stable or unstable angina. In addition, they must have had adrenomedullin measured at baseline and related to major adverse cardiovascular events (MACE) or mortality. Excluded were review articles, pediatric studies, and animal studies.

A systematic search of studies was carried out on original published studies using PubMed (MEDLINE) and the Cochrane Collaboration databases using key words and MeSH terms. For patients with heart failure, the MeSH terms or key words were (“Adrenomedullin” OR “ADM” OR “MRproADM”) AND “Heart Failure.” For patients with myocardial infarction, these were (“Myocardial Infarction” OR “Angina Pectoris” OR “Angina, Stable” OR “Angina, Unstable”) AND (“Adrenomedullin” OR “ADM” OR “MRproADM”). Searches of references of published reports obtained from the aforementioned systematic search were also undertaken. The search was restricted to reports published in English from January 1, 1993, to June 30, 2014, in humans. Figure 1 shows the flow chart for selection for this systematic review.

Flowchart of search of studies and identification of relevant English-language studies on of adrenomedullin and death in patients with heart failure and adrenomedullin and MACE and death in patients with myocardial infarction (January 1, 1993, to June 30, 2014).

The main outcome measure of this systematic review for patients with heart failure was death. The outcome measure for patients with myocardial infarction was MACE as primary end point measure, with death as secondary end point. The end point, MACE, was defined variably across studies as a composite of myocardial infarction, stroke, hospitalization for heart failure, and all-cause mortality, although in some studies not all these components were present.

Two authors independently perused all the relevant titles and abstracted data from selected reports with any disagreements settled by a consensus or opinion from other co-authors. Some studies had very small sample sizes, but we included given the small number of studies.

Because of the high level of heterogeneity across studies, no quantitative pooling of results from various studies through a meta-analysis was undertaken. This heterogeneity was mainly due to the discrepancy in type of adrenomedullin measured (mature ADM vs its surrogate MRproADM), discrepancy in units of ADM used by different studies, the varying use of ADM as a continuous and a categorical variable across studies, the different cutoffs of ADM when used as a categorical variable, and differing logarithmic transformations or unit change of ADM when used as a continuous variable in various statistical models.

Results

Table 1 lists studies which looked at the relation between ADM and mortality in patients with heart failure. We identified 14 studies in total, and these were mainly cohort studies. Baseline plasma levels of MRproADM were significantly higher in patients who died during follow-up compared with the levels of those who survived. The area under the curve (AUC) for adrenomedullin predicting mortality was only reported for 12 studies using MRproADM in patients with heart failure, and this ranged from 0.68 to 0.81 (95% confidence interval [CI] 0.63 to 0.91) across studies. Overall, the earlier studies using mature adrenomedullin demonstrated a statistically significant association from hazard ratios (HRs) and 95% CIs with all-cause mortality. Subsequent studies using MRproADM showed similar results. When taken as a continuous variable, a 1 nmol/l increase in MRproADM was associated with HRs ranging from 1.77 to 2.79 (95% CI 1.29 to 5.95) for death in patients with heart failure. When taken as a log10 of MRproADM measured in nanomoles per liter, the HRs ranged from 1.82 to 10.26 for death (95% CI 1.26 to 80.50). Various cutoffs were used for categorical adrenomedullin as presented in Table 1 , and these included upper tertile, receiver operating characteristic optimal cutoff, and above the median value. Only one study showed a negative result of MRproADM on mortality in heart failure patients.

Table 2 depicts studies which looked at the relation between adrenomedullin and MACE in patients with acute coronary syndrome. We identified 9 cohort studies in total. Baseline plasma levels of MRproADM were significantly higher in patients who experienced a MACE during follow-up compared with the levels of those who did not. The AUC for MRproADM predicting MACE ranged from 0.64 to 0.80 (95% CI 0.51 to 0.87) across studies in patients with AMI. When taken as a continuous variable, 1 nmol/L increase in MRproADM was associated with HRs for MACE ranging from 1.78 to 4.10 (95% CI 1.20 to 10.12), whereas log10 of MRproADM had HRs of 3.63 to 9.75 (95% CI 1.48 to 26.16) for MACE across studies in patients with AMI. Only 1 study showed a negative result of MRproADM on MACE in patients with AMI. This review also identified studies which have shown that MRproADM significantly enhances the integrated discrimination improvement (IDI) and the net reclassification improvement (NRI) when added to other risk markers such as brain natriuretic peptide and the GRACE score in patients with AMI. For example, 1 study showed an NRI of 34.1% (range 18.6% to 49.5%) for 30-day mortality when MRproADM was added to the GRACE score.

Table 3 depicts studies which looked at the relation between adrenomedullin and death in patients with AMI. We identified 8 cohort studies in total. The AUC for MRproADM predicting death ranged from 0.79 to 0.84 (95% CI 0.73 to 0.90) across studies in patients with AMI. When taken as a log10 of MRproADM measured in nanomoles per liter, death was associated with HRs ranging from 4.86 to 16.68 (95% CI 4.56 to 60.99) across studies in patients with AMI. One study showed a negative result of MRproADM on mortality in patients with AMI.

Table 3

Studies showing the predictive effect of adrenomedullin on all-cause mortality in acute myocardial infarction patients

Author	Year	Sample size	Number of Deaths	HR Death (95% CI or P value)	AUC (95% CI)	Duration follow-up	Measure of Adrenomedullin	Primary endpoint	Modelled adrenomedullin
Richards	1998	121	27	1.21 (not significant)	NA	2 4 month	Mature ADM	Death, HF	Continuous, 1 pmol/l
Richards	2001	297	35	3.92 (1.76-8.70)	NA	18 months	Mature ADM	Death, HF	Above median, pmol/l
Nagaya	1999	113	16	1.04 (1.01-1.08)	NA	25 months	Mature ADM	Cardiovascular death	Continuous, 1 pmol/l
Katayama	2005	124	9	10.63 (1.06-107.06)	NA	90 days	Mature ADM	Death	Continuous 1, fmol/l
Khan	2007	983	101	4.86 (p=0.001)	NA	2.1 years	MRproADM	Death, HF	Log Base10 MRproADM, nmol/l
Dhillon	2010	745	120	16.68 (4.56-60.99)	0.79 (0.74-0.84)	2 years	MRproADM	Death, HF, AMI	Log Base10 MRproADM, nmol/l
Klip	2011	214	31	3.51 (1.38-5.49)	0.81 (0.73-0.90)	2.5 years	MRproADM	Death, AMI, Stroke, resuscitated CA	Doubling of MRproADM, nmol/l
Haaf	2013	1179 Acute chest pain	79	2.79 (1.81-4.33)	0.84 (0.81-0.86)	2 years	MRproADM	Death	Continuous, 1 nmol/l

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Evaluation of a Structured Training Program for Arterial Femoral Sheath Removal After Percutaneous Arterial Catheter Procedures by Assistant Personnel
2. Comparison of Left Ventricular Torsion and Strain With Biventricular Pacing in Patients With Underlying Right Bundle Branch Block Versus Those With Left Bundle Branch Block
3. Correlation of Precordial Voltages to Left Ventricular Mass on Echocardiogram in Adolescent Patients With Hypertrophic Cardiomyopathy Compared With that in Adolescent Athletes
4. Measuring the Base of the Triangle is a Useful Tool to Differentiate Between Electrocardiographic Patterns of Healthy Athletes and Brugada Syndrome

Stay updated, free articles. Join our Telegram channel

Join