Correction to: Arthritis Res Ther

https://doi.org/10.1186/s13075-018-1597-9

Following publication of the original article [1], it has come to our attention that we did not appropriately convert the units for CRP from the National Health and Nutrition Examination Survey (NHANES) from mg/dL to mg/L. In this study, NHANES was considered the general population cohort for this study; interestingly, the correction resulted in changes to a broader range of CRP values in NHANES from 0.01 to 18.0 mg/dL to 0.10 to 180.0 mg/L.

The overall results did not change the message for this study. We continued to observe a general positive association between CRP and blood pressure (BP) in all three groups, RA, non-RA and NHANES. For RA and non-RA this relationship held until a CRP level of approximately 7 mg/L when an inverse association was observed between CRP and systolic BP. In NHANES, we also observed a similar association of higher CRP associated with SBP until approximately 7 mg/L when the there is no association between CRP and SBP. In NHANES, 15% have a CRP above 7 mg/L compared to 32% in the RA and 21% non-RA cohorts. Thus, with the smaller percentage of subjects with a CRP above 7 mg/L in NHANES, there was potentially less power to detect an association.

Line-by-line corrections of our results are given below:

  1. 1.

    Results, page 3, last sentence in the right column should read: “…in the non-RA outpatient population, and 1.80 mg/L (range 0.10 to 180.10 mg/L).”

  2. 2.

    Results, page 4, 2nd paragraph, first sentence. The range of CRP was “0.10-180.10 mg/L in NHANES.”

  3. 3.

    Results, page 4, 3rd paragraph. “In comparison, within NHANES, we generally observed a positive association between CRP and SBP also until 7mg/dL when there is no association between CRP and SBP.”

  4. 4.

    Results, page 4, 2nd column, last sentence of the second paragraph, “Findings from the sensitivity analysis showed a general linear association between CRP and SBP in NHANES after trimming of the CRP outliers; this finding is consistent with previous studies in the general population. For RA and non-RA where higher levels are typically observed, the relationships were similar to the main analyses.”

  5. 5.

    In Table 1, the median (IQR) of CRP in NHANES should be 1.80 (0.70, 4.30) mg/L.

  6. 6.

    Figure 1 The relationship between C-reactive protein levels (CRP) and systolic blood pressure with 95% confidence intervals, in the rheumatoid arthritis (RA) outpatient population and the general population (National Health and Nutrition Examination Survey (NHANES)). RA outpatient population CRP range 0.20–126.90 mg/L; NHANES CRP range 0.10–180.10 mg/L (attached)

  7. 7.

    Figure 2 The relationship between C-reactive protein (CRP) levels and systolic blood pressure with 95% confidence intervals, in the non-rheumatoid arthritis (RA) outpatient population and the general population (National Health and Nutrition Examination Survey (NHANES)). Non-RA outpatient population CRP range 0.10–416.20 mg/L; NHANES 0.10–180.10 mg/L (attached)

  8. 8.

    We have also updated the supplementary figures to reflect the updated data:

    Fig. 1
    figure 1

    The relationship between C-reactive protein levels (CRP) and systolic blood pressure with 95% confidence intervals, in the rheumatoid arthritis (RA) outpatient population and the general population (National Health and Nutrition Examination Survey (NHANES)). RA outpatient population CRP range 0.20–126.90 mg/L; NHANES CRP range 0.10–180.10 mg/L

    Full size image
    Fig. 2
    figure 2

    The relationship between C-reactive protein (CRP) levels and systolic blood pressure with 95% confidence intervals, in the non-rheumatoid arthritis (non-RA) outpatient population and the general population (National Health and Nutrition Examination Survey (NHANES)). Non-RA outpatient population CRP range 0.10–416.20mg/L; NHANES 0.10–180.10 mg/L

    Full size image