Increased levels of C-reactive protein in serum from blood donors before the onset of rheumatoid arthritis

Abstract

Objective

We previously reported that approximately half of the patients with rheumatoid arthritis (RA) have specific serologic abnormalities (elevated serum concentrations of IgM rheumatoid factor and/or anti–cyclic citrullinated peptide antibodies) starting several years before the onset of symptoms. In this study, the presence of serologic signs of inflammation in patients with preclinical RA was investigated with serial measurements of C-reactive protein (CRP).

Methods

Seventy-nine patients (61% female; mean age at onset of symptoms 51 years) who had been blood donors before the onset of RA were identified. Frozen serum samples from each donor were retrieved, together with 1 sample from a control donor matched for age, sex, and date of donation. CRP was measured using a highly sensitive latex-enhanced assay. The dates of donation were categorized into 15 1-year periods preceding the onset of RA symptoms. For each period, the median CRP levels in the patient and control groups were compared using the Mann-Whitney U test. The course of CRP concentrations over time in the patient group was estimated with random coefficient analysis.

Results

A median of 13 samples (range 1–51) per patient were available; the earliest donation was made a median of 7.5 years (range 0.4–14.5 years) before the onset of symptoms. A total of 1,078 patient samples and 1,071 control samples were tested. For all 1-year periods, the median CRP concentration was increased in the patient group compared with the control group, but this difference was statistically significant only for the periods 0–1 year, 1–2 years, and 4–5 years before the onset of symptoms. The CRP concentration increased significantly over time in patients with preclinical RA; levels were slightly higher in the group of patients who had serologic abnormalities before the onset of symptoms than in those without such serologic abnormalities.

Conclusion

After observing specific serologic abnormalities 5 years before the onset of RA symptoms, we now report increased levels of CRP in blood donors in whom RA later developed; these increases were most common within the 2 years before the onset of symptoms. The preclinical increase in CRP levels was observed both in donors with and in those without serologic abnormalities.

The pathogenesis of rheumatoid arthritis (RA) is poorly understood, but there is evidence for an immunologic disturbance preceding clinical RA. This process is associated with the occurrence of antibodies against immunoglobulins (rheumatoid factor [RF]) (1–7) and citrullinated peptides (mostly measured as anti–cyclic citrullinated peptide [anti-CCP]) (6, 7). Recently, the presence of 1 or both of these antibodies was demonstrated in half of the serum samples obtained from a group of 79 blood donors in whom RA later developed. Seroconversion to anti-CCP and IgM-RF positivity, respectively, occurred a median of 5 years and 2 years before symptom onset (6). In a comparable study of blood donors in whom RA later developed, Rantapää-Dahlqvist et al reported sensitivities of 20% and 34%, respectively, for detecting IgM-RF and anti-CCP more than 1 year before the onset of symptoms (7).

Because the immunologic derangement preceding overt RA is often present for several years, the existence of subclinical inflammation during this period can be assumed. There have been conflicting reports on this issue (8, 9). In Finland, Aho et al found that the prerheumatic immunologic process was not associated with inflammation, as measured by the C-reactive protein (CRP) concentration in serum samples obtained from 124 RA patients up to 20 years before the onset of disease (8). In contrast, Masi et al reported a higher frequency of increased CRP concentrations in serum samples obtained from 18 male RA patients 3–20 years before the onset of symptoms compared with the concentrations in samples from 72 matched controls (9). However, these studies used single serum samples and therefore provide no information about the increase in CRP concentrations during the period before the onset of RA symptoms.

In this study, the presence of inflammation in the preclinical phase of RA was investigated with serial measurements of CRP in serum samples obtained from the blood of donors before RA developed and in samples from control blood donors.

PATIENTS AND METHODS

Patients.

Since 1984, the Sanquin Blood Bank Northwest Region in Amsterdam, The Netherlands, has stored (at −30°C) 1-ml aliquots of serum from donated blood. The majority of the donors, who are age ≤70 years, donate 2–4 times per year over long periods of time. The bank serves an area with approximately 2 million inhabitants. Within this area, most RA patients are registered at the Jan van Breemen Institute, a regional network of outpatient clinics. Seventy-nine patients with RA according to the 1987 American College of Rheumatology (formerly, the American Rheumatism Association) criteria (10) were identified as having donated blood before the onset of symptoms, as described previously (6). The following data were collected from the medical records: time of the start of symptoms and smoking status at the first visit. The study was approved by the local Institutional Review Board.

Procedures.

Patient sera obtained during the period from 1984 to 1999 were collected as previously described (6). Briefly, for each RA sample, 1 masked control sample from the same container was selected, matched for sex, age (±3 years), and time of blood donation. This method of control selection was chosen to ensure identical storage conditions for the samples from patients and controls.

CRP was measured using a highly sensitive latex-enhanced assay supplied by Roche Diagnostics (Almere, The Netherlands) on a Hitachi 911 analyzer (Roche Diagnostics), according to the manufacturer's instructions. This test measures the CRP concentration very sensitively, in the range of 0.1–20 mg/liter.

Statistical analysis.

In a first analysis, the time axis (dates of blood donations) was divided into 15 1-year periods preceding the onset of symptoms. For each period, the median CRP concentrations in the patient and control groups were compared using the Mann-Whitney U test.

In a second analysis, all of the patient and control samples collected 0–10 years before the onset of symptoms were used to estimate the increase in CRP concentrations in the patient group over time, compared with that in the control group. Samples obtained 10–15 years before the onset of symptoms were not used in this analysis, because of their small number. The increase in CRP levels over time was estimated with random coefficient analysis (11). Basically, with this longitudinal regression technique, both the intercept and the increase over time are allowed to differ between patients. The increase over time was modeled either linearly or (if necessary) quadratically. In the patient group, the time course of the increase in CRP levels was corrected for age, sex, and smoking status at the time of the diagnosis of RA, because smoking is associated with elevated CRP concentrations (12). Because the distribution of CRP concentrations was skewed to the right, the natural log of CRP concentrations was used in all analyses.

To study whether the degree of inflammation was different between the group of RA patients who had serologic abnormalities and those who did not have such abnormalities before the onset of symptoms, the increase in CRP levels over time was compared in a third analysis. Patients who were positive for IgM-RF or anti-CCP at least once before the onset of symptoms were counted as having serologic abnormalities (6).

In addition, all patient samples were used in a time-lag analysis to determine whether increased CRP levels preceded the development of antibodies (IgM-RF/anti-CCP) in the preclinical phase of RA or vice versa. Concentrations of IgM-RF and anti-CCP were associated with CRP levels measured at the same time point and 1, 2, and 3 years before, and 1, 2, and 3 years after. By comparing the magnitude of the different regression coefficients with each other, one can determine whether or not a time lag is present in the relationship under investigation (11). The magnitudes of the regression coefficients were investigated with random coefficient analysis, with correction for age and sex. In all analyses, the natural log of IgM-RF, anti-CCP, and CRP concentrations was used because of the non-normal distribution of these variables. Random coefficient analyses were performed with the statistical program MLwiN (Multilevel Models Project, Institute of Education, University of London, London, UK).

RESULTS

Patient characteristics.

Seventy-nine patients with RA (61% women; mean age at symptom onset 51 years) were included. A median of 13 serum samples (range 1–51) per patient were available, the earliest of which had been collected a median of 7.5 years (range 0.4–14.5 years) before the onset of symptoms. In total, 1,078 patient sera and 1,071 matched control sera were tested (for 7 patient sera, no control serum matching the criteria was available).

Test results.

For all 1-year periods, the median CRP concentration was higher in the patient group compared with the control group, but this difference was statistically significant only for the periods 0–1 year, 1–2 years, and 4–5 years before the onset of symptoms (Table 1). Due to multiple testing, the calculated P values must be interpreted cautiously.

Table 1. C-reactive protein levels before the onset of symptoms in patients with rheumatoid arthritis and controls*

Time period before onset of symptoms	No. of patient samples	C-reactive protein level, mg/liter
		Patients	Controls	P
* Values are the median (interquartile range).
0–1 year	130	2.2 (1.0–3.9)	1.0 (0.5–2.8)	0.001
1–2 years	120	1.9 (1.0–3.2)	1.5 (0.7–2.6)	0.035
2–3 years	120	1.7 (0.7–2.7)	1.4 (0.5–2.6)	0.492
3–4 years	99	1.6 (0.8–3.2)	1.3 (0.6–3.2)	0.125
4–5 years	115	1.7 (0.6–2.8)	1.0 (0.6–2.0)	0.012

The CRP levels before the onset of symptoms in patients with RA and controls are shown in Figure 1. Whereas the mean CRP level in the control group remained stable over time (the slight increase was not statistically significant), the mean CRP level in the patient group showed a statistically significant increase over time, with the highest level observed at the time of symptom onset. The progression of CRP concentrations over time within the patient group was not confounded by smoking status (data not shown).

Figure 1. C-reactive protein levels before the onset of symptoms in patients with rheumatoid arthritis (RA) and controls.

The CRP levels before the onset of RA symptoms in patients with and those without serologic abnormalities are shown in Figure 2. Both groups of patients had a statistically significant increase in the CRP concentration over time, with the highest levels observed at the time of symptom onset. The group of patients who had serologic abnormalities before the onset of symptoms had slightly higher mean CRP levels at all time points, compared with the patients who did not have serologic abnormalities before symptom onset (P < 0.05).

Figure 2. C-reactive protein levels in patients who were positive for IgM rheumatoid factor or anti–cyclic citrullinated protein at least once before the onset of rheumatoid arthritis (RA) symptoms (serologic abnormalities) and in those without serologic abnormalities before the onset of symptoms.

Time-lag analyses were used to determine whether increased CRP levels preceded the development of antibodies (IgM-RF/anti-CCP) or vice versa. No trend was apparent (data not shown). Therefore, these analyses could not clarify the sequence of increased CRP levels and the development of antibodies in the preclinical phase of RA.

DISCUSSION

In the group of healthy blood donors in whom RA later developed, the median serum CRP level was higher than that in the control group for all 15 1-year periods; however, the difference was statistically significant only for the periods 0–1 year, 1–2 years, and 4–5 years before the onset of symptoms. The CRP concentration increased significantly over time in the subjects with preclinical RA, with slightly higher levels in the group of patients who had serologic abnormalities before symptom onset than in the group of patients without serologic abnormalities.

The median CRP levels in the patient and control groups were compared within the 15 1-year periods. In each of these periods, samples from a different set of patients and sometimes more than 1 sample from the same patient were included. Because multiple high or low CRP levels in some patients might influence the median CRP concentration in the whole group, multiple linear regression analysis was also performed, with correction for the intercept and development over time for each patient. In addition, the relationship between the CRP level and time was corrected for age and sex. The results may have been confounded by differences in smoking habits between patients and controls (12), because we have only limited information on smoking habits in the patient group. Current smoking status at the time of the diagnosis of RA did not appear to be related to CRP levels before the onset of symptoms.

The present results are consistent with those in the study by Masi et al (9), who reported that 4 of 18 male RA patients had increased (>8 mg/liter) CRP levels before the onset of symptoms. In contrast, Aho et al did not observe increased CRP levels in serum obtained from RA patients up to 20 years before the onset of RA compared with healthy controls (8). Possible explanations for the discrepancies could be the small number of measured sera and the long period of time between collection of blood samples and the onset of symptoms in these earlier studies. Another argument in favor of the existence of preclinical inflammation is the finding of synovitis in clinically uninvolved knee joints of patients with early RA (13).

Patients who had serologic abnormalities before the onset of symptoms had slightly higher CRP concentrations compared with patients without serologic abnormalities, which indicates an intensified inflammatory process in patients with IgM-RF and/or anti-CCP positivity before the onset of symptoms. However, both groups showed a very similar pattern of increase in the CRP concentration over time, which does not indicate a pivotal role for autoantibodies in the development of inflammation during the preclinical phase of RA. Rather, the somewhat higher CRP concentrations observed in patients with serologic abnormalities suggest that the production of autoantibodies is a phenomenon secondary to an increased level of inflammation in these patients. Even the appearance of IgM-RF or anti-CCP years earlier may have been a consequence of an inflammatory stimulus that was either too brief or too weak to be detected with the presently used technique and frequency of samples. Unfortunately, time-lag analyses could not clarify whether increased CRP levels preceded the development of antibodies in the preclinical phase of RA, or vice versa.

Although the differences between patients and controls were statistically significant in the 2 years before the onset of RA symptoms, it is uncertain whether these differences are also clinically relevant. An example of the clinical relevance of even small changes in the CRP concentration, albeit in a different disease, is given in a report of the influence of CRP levels on the risk of cardiovascular diseases in a healthy North American population (14). In that study, it was shown that a difference of <1 mg/liter in CRP levels could lead to an increased risk of cardiovascular disease. Therefore, the small difference in CRP concentrations observed in the present study may well be a significant factor in the development of later symptomatic inflammation. It must be noted that the findings of this study can be used only in a population of patients and are not suitable for decision-making in individual patient care.

In conclusion, after previously observing specific serologic abnormalities in RA patients 5 years before the onset of symptoms, we now report increased levels of CRP in blood donors in whom RA later developed; these increased levels were most common within the 2 years before the onset of symptoms. This increase in CRP levels was demonstrated both in subjects with and in those without serologic abnormalities.

Acknowledgements

Margret de Koning is greatly appreciated for her help in locating sera from the patients and matched controls. Nora Schram wrote the necessary computer programs and set up the databases. We thank Maarten Boers for helpful discussions concerning the time sequence of the increase in antibody and CRP levels.