High red cell distribution width as a prognostic indicator in heart failure

Table 1 Baseline characteristics

Characteristic	Combined studies (n = 11460)	Low quartile (n = 6562)	High quartile (n = 5098)
Demographic characteristics
Female sex (%)	41.5	40.4	41.5
Age (years) (mean)	78.9	71.3	78.8
Body mass index (kg/m²) (mean)	29.0	30.1	28.9
Systolic blood pressure (mmHg) (mean)	142.0	138.0	142.0
Diastolic blood pressure (mmHg) (mean)	83.3	83.0	83.3
Heart rate (bpm) (mean)	78.0	80.6	78.0
NYHA I-II (%)	64.6	49.4	64.6
NYHA III-IV (%)	35.2	50.5	35.2
Left ventricular ejection fraction (%) (mean)	42.7	41.5	42.7
Previous myocardial infarction (%)	41.5	37.0	41.5
Atrial fibrillation/flutter (%)	21.2	30.0	21.2
Hypertension (%)	62.0	57.0	62.0
Diabetes (%)	37.9	39.9	37.9
Angiotensin-converting-enzyme inhibitors/angiotensin II receptor blocker (%)	67.9	66.2	67.9
Beta-blocker (%)	62.4	61.8	62.4
Mineralocorticoid receptor antagonist (%)	17.9	21.0	17.9
Calcium channel blocker (%)	10.4	11.7	10.4
Statin (%)	51.2	44.8	51.2
Loop diuretics (%)	68.8	79.4	68.8
Baseline laboratory studies
White blood cell (cells/μL) (mean)	7.7	7.6	7.7
Hemoglobin (g/dL) (mean)	12.0	11.5	12.0
Mean corpuscular volume (fL) (mean)	89.3	86.1	89.3
Estimated glomerular filtration rate (mL/minute/1.73 m²) (mean)	66.1	56.6	66.1
Creatinine (mg/dL) (mean)	1.2	1.5	1.2
Blood urea nitrogen (mg/dL) (mean)	36.8	41.5	36.8
Total cholesterol (mg/dL) (mean)	165.6	152.5	165.6

NYHA: New York Heart Association.

Table 2 Quality assessment of each study included in meta-analysis

Ref.	Year	Population (n)	RDW quartiles	Primary outcome	Follow-up duration	Newcastle-Ottawa Scale score (selection/comparability/outcome)	Notes on quality
Zalawadiya et al[7]	2011	635	Q1: < 14%, Q4: > 16.5%	All-cause mortality	1.5 years	3/1/2 (6/9)	Moderate quality. Clear cohort selection, but limited adjustment for confounders (e.g., only age and sex). Older publication may affect generalizability
Bonaque et al[8]	2012	698	Q1: Not specified, Q4: Not specified	Mortality	2.5 years	3/2/3 (8/9)	Good quality. Robust outcome assessment and follow-up, but RDW quartile definitions are unclear, limiting comparability
Makhoul et al[9]	2013	614	Q1: < 15.2%, Q4: > 15.3%	Clinical outcomes (heart failure hospitalization, mortality)	1 year	3/1/2 (6/9)	Moderate quality. Adequate selection and outcome reporting, but small sample size and limited confounder adjustment reduce robustness
Sotiropoulos et al[10]	2016	402	Q1: 12.2%-14.2%, Q4: 16.7%-32.1%	Mortality	1 year	3/1/2 (6/9)	Moderate quality. Small sample size limits statistical power. Clear RDW quartiles, but limited adjustment for confounders
Cheng et al[11]	2016	978	Q1: ≤ 14.3%, Q4: > 14.3%	Mortality	30 months	3/2/3 (8/9)	Good quality. Strong outcome assessment and longer follow-up, but focus on cardiorenal anemia syndrome may introduce confounding
Wasilewski et al[12]	2018	1734	Q1: ≤ 13.4%, Q4: > 14.6%	Mortality	660 days (approximately 1.8 years)	4/2/3 (9/9)	High quality. Large cohort, robust selection, and comprehensive confounder adjustment. Clear outcome reporting
Kim et al[5]	2023	6399	Q1: 12.7% ± 0.5%, Q4: 15.0% ± 1.9%	All-cause mortality	3 months	4/2/2 (8/9)	Good quality. Large sample size and clear selection, but short follow-up (3 months) limits long-term prognostic insights

The Newcastle-Ottawa Scale assesses selection (0-4), comparability (0-2), and outcome (0-3), with a maximum score of 9. Studies with scores ≥ 7 are considered high quality, 5-6 moderate quality, and < 5 Low quality. Quality limitations include small sample sizes (e.g., Sotiropoulos et al[10], Makhoul et al[9]), short follow-up periods (e.g., Kim et al[5]), and limited confounder adjustment in older studies (e.g., Zalawadiya et al[7]). Strengths include large cohorts (e.g., Kim et al[5], Wasilewski et al[12]) and clear outcome reporting in most studies. RDW: Red cell distribution width.