In-6. [29] Because of the plethora of variables besides infections, overt inflammation, and atherosclerosis that may affect the biomarkers of inflammation such as weight change, drugs, level of physical activity, changes in diet, smoking status, depression and trauma, and 301-00-8 site likely other indeterminate factors, it is not surprising that CRP values may exhibit swings that are often unpredictable.In a recent individual participant meta-analysis that examined CRP and vascular risk, Kaptoge et al. [27] calculated regression dilution ratios and found in 22,124 subjects with 2 CRP measurements a mean of 5 years apart that CRP (log transformed) exhibited year-to-year intra-individual consistency similar to KS 176 supplier cholesterol (not log transformed) and systolic blood pressure. The design and 23977191 methods of the study and its focus on outcomes do not address the problematic of variability encountered when CRP is used in daily clinical practice for 
risk stratification and individualized management based on a threshold risk value. In contrast to these studies, others have suggested that CRP exhibits considerable variability with intra-individual coefficients of variation for CRP that are 4? fold larger than for cholesterol. [16,17] Similar or greater variability of CRP has been found in other studies. [14,15,20,22].Potential LimitationsOur study group was relatively small in comparison to some of the previous studies examining intra-individual CRP variability. However, previous studies have neither been as systematic in their design and analysis nor as intense in terms of numbers of measurements per subject and time-points as the current study. We used a total of 1500 observations to estimate variability of CRP over time. While that size is small compared to other studies that focused on outcomes, it clearly was an adequate size for estimating variance, as evidenced by our reasonably small interval estimates. The study group was highly selected; most had CAD and were on statin therapy. On the other hand, neither clinical group nor use of lipid-lowering therapy was retained in the hierarchical model of CRP variability. However, even if CAD status and statin use blunted CRP variability, this would suggest that CRP exhibits even more variability in the general population than was found in this study. Finally, our design and methods did not allow us to characterize intra-individual variability of CRP based on sex, age bracket, or ethnic group.Previous StudiesPrevious studies that have examined CRP variability have produced conflicting results. This inconsistency and the possible reasons to account for it have not been well addressed or debated. For example, in contrast to our findings, Ockene et al. [19] studied 113 healthy adults with 5 measurements of CRP and total cholesterol over one year and concluded that CRP had measurement stability similar to cholesterol. This conclusion is surprising because the authors reported a within-subject standard deviation of cholesterol of 0.447 mmol/L, which would be less than 10 a cut-off risk value of 5.2 mmol/L for cholesterol, while the within-subject standard deviation of CRP was 1.2 mg/L, which would be 60 of the cut-off risk value of 2 mg/L. Three other studies have suggested that CRP is relatively stable within individuals on serial measurement. The `stable’ values were measured over a period of 5?2 years and often excluded elevated CRP values without any clinical justification. [18,21] More recently, Glynn et al. [23] st.In-6. [29] Because of the plethora of variables besides infections, overt inflammation, and atherosclerosis 
that may affect the biomarkers of inflammation such as weight change, drugs, level of physical activity, changes in diet, smoking status, depression and trauma, and likely other indeterminate factors, it is not surprising that CRP values may exhibit swings that are often unpredictable.In a recent individual participant meta-analysis that examined CRP and vascular risk, Kaptoge et al. [27] calculated regression dilution ratios and found in 22,124 subjects with 2 CRP measurements a mean of 5 years apart that CRP (log transformed) exhibited year-to-year intra-individual consistency similar to cholesterol (not log transformed) and systolic blood pressure. The design and 23977191 methods of the study and its focus on outcomes do not address the problematic of variability encountered when CRP is used in daily clinical practice for risk stratification and individualized management based on a threshold risk value. In contrast to these studies, others have suggested that CRP exhibits considerable variability with intra-individual coefficients of variation for CRP that are 4? fold larger than for cholesterol. [16,17] Similar or greater variability of CRP has been found in other studies. [14,15,20,22].Potential LimitationsOur study group was relatively small in comparison to some of the previous studies examining intra-individual CRP variability. However, previous studies have neither been as systematic in their design and analysis nor as intense in terms of numbers of measurements per subject and time-points as the current study. We used a total of 1500 observations to estimate variability of CRP over time. While that size is small compared to other studies that focused on outcomes, it clearly was an adequate size for estimating variance, as evidenced by our reasonably small interval estimates. The study group was highly selected; most had CAD and were on statin therapy. On the other hand, neither clinical group nor use of lipid-lowering therapy was retained in the hierarchical model of CRP variability. However, even if CAD status and statin use blunted CRP variability, this would suggest that CRP exhibits even more variability in the general population than was found in this study. Finally, our design and methods did not allow us to characterize intra-individual variability of CRP based on sex, age bracket, or ethnic group.Previous StudiesPrevious studies that have examined CRP variability have produced conflicting results. This inconsistency and the possible reasons to account for it have not been well addressed or debated. For example, in contrast to our findings, Ockene et al. [19] studied 113 healthy adults with 5 measurements of CRP and total cholesterol over one year and concluded that CRP had measurement stability similar to cholesterol. This conclusion is surprising because the authors reported a within-subject standard deviation of cholesterol of 0.447 mmol/L, which would be less than 10 a cut-off risk value of 5.2 mmol/L for cholesterol, while the within-subject standard deviation of CRP was 1.2 mg/L, which would be 60 of the cut-off risk value of 2 mg/L. Three other studies have suggested that CRP is relatively stable within individuals on serial measurement. The `stable’ values were measured over a period of 5?2 years and often excluded elevated CRP values without any clinical justification. [18,21] More recently, Glynn et al. [23] st.
