The report from Aushon contained valid assay data for all samples we sent (n = 1279). There were no insufficient quantities or values below the detectectable range.

I combined CRP data from wave 1 (hsCRP) with CRP data from wave 3 (Quest's low sensitivity CRP) with these wave 3 hsCRP assay data.

      CRP.1    CRP.3    CRPLS   
CRP.1  1.00     0.28***  0.27***
CRP.3  0.28***  1.00     0.93***
CRPLS  0.27***  0.93***  1.00   

n
      CRP.1 CRP.3 CRPLS
CRP.1  1047  1047  1040
CRP.3  1047  1278  1269
CRPLS  1040  1269  1269

hsCRP at wave 1 (CRP.1) is correlated significantly (though weakly) with hsCRP at wave 3 (CRP.3) [r = 0.28] and low sensitivity CRP (CRPLS) is correlated strongly with hsCRP at wave 3 [r = 0.93]:

Looking at change in hsCRP over time by age, sex, race, and poverty status, there are significant differences associated with these measures but no evidence for interactions among them or with time. In particular, this suggests that there are no differences among these groups in rate of hsCRP change. All groups are changing at the same rate despite overall mean differences in levels of hsCRP per the figure below.

     Simultaneous Tests for General Linear Hypotheses

Fit: lmer(formula = CRP ~ (Age + Sex + Race + PovStat) + (Age | HNDid), 
    data = longCRP, control = lmerControl(check.nobs.vs.nRE = "ignore"), 
    na.action = na.omit)

Linear Hypotheses:
                     Estimate  Std. Error  z value         Pr(>|z|)
(Intercept) == 0   5.62805218  0.32192065 17.48273       < 2.22e-16
Age == 0           0.06464369  0.01778786  3.63415        0.0002789
SexMen == 0       -2.10564142  0.34058052 -6.18251 0.00000000063092
RaceAfrAm == 0     0.99940085  0.34649754  2.88429        0.0039229
PovStatBelow == 0  0.65124491  0.34703051  1.87662        0.0605700
(Univariate p values reported)