C-Reactive Protein, Fibrinogen, and Incident Heart Failure in the Strong Heart Study Population

Authors


We thank Dr Kawada for his suggestions. In response to his letter, we have performed additional analyses to examine the relationship between incident heart failure (HF) and 4 categories of combined C-reactive protein (CRP) and fibrinogen, defined according to the 75th percentile cutoff values for fibrinogen and CRP. Considering the group of patients with both fibrinogen and CRP values < their 75th percentile as a reference, hazard ratios and corresponding 95% confidence intervals of each of the other 3 groups were estimated using multivariate Cox proportional hazards models adjusted for age and sex (multivariate model 1), and for conventional cardiovascular risk factors including age, sex, low-density lipoprotein cholesterol, current smoking status, estimated glomerular filtration rate, hypertension, and diabetes (multivariate model 2).

As shown in Table 1, compared with the category with both fibrinogen and CRP values < the 75th percentile, the category with fibrinogen values > the 75th percentile, but CRP values below this threshold, was significantly predictive of incident heart failure in both multivariate analyses (hazard ratios [HRs]; 95% confidence intervals [CIs], 2.09 [1.45–3.01] and 1.61 [1.09–2.38], respectively). In contrast, the category with CRP values > the 75th percentile, but fibrinogen levels below this threshold, did not show significantly increased risk of incident heart failure (HRs [95% CIs], 1.11 [0.69–1.78] and 0.96 [0.57–1.60], respectively). The category with both fibrinogen and CRP values > their 75th percentile had the highest risk of incident heart failure (HRs [95% CIs], 2.38 [1.67–3.39] and 1.76 [1.20–2.57], respectively). Hence, these additional analyses confirm our published observations and add to the evidence suggesting that elevated fibrinogen, but not CRP, levels are associated with increased risk of incident heart failure among American Indians, a population with a high prevalence of diabetes and obesity.

Table 1. Hazard Ratios and 95% Confidence Intervals for Incident Heart Failure According to the Combined Categories of CRP and Fibrinogen
 CRP and Fibrinogen <75th Percentile (n=1891)CRP <75th Percentile and Fibrinogen >75th Percentile (n=376)CRP >75th Percentile and Fibrinogen <75th Percentile (n=365)CRP and Fibrinogen >75th Percentile (n=380)
  1. Cox proportional hazards models were used to analyze the associations between C-reactive protein (CRP)/fibrinogen and incident HF. aAdjusted for age (continuous variable) and sex (men vs women). bAdjusted for age (continuous), sex (men vs women), low-density lipoprotein cholesterol (continuous), current smoker (yes vs no), estimated glomerular filtration rate (continuous), hypertension (yes vs no), and diabetes (yes vs no).

Univariate model1 (referent)2.23 (1.55–3.20)1.07 (0.67–1.72)2.46 (1.73–3.49)
Multivariate model 1a1 (referent)2.09 (1.45–3.01)1.11 (0.69–1.78)2.38 (1.67–3.39)
Multivariate model 2b1 (referent)1.61 (1.09–2.38)0.96 (0.57–1.60)1.76 (1.20–2.57)

Ancillary